Merge branch 'dev'

CHANGELOG.md

@@ -1,5 +1,15 @@
+# v2.14.0
+- **[#969](https://github.com/xmrig/xmrig/pull/969) Added new algorithm `cryptonight/rwz`, short alias `cn/rwz` (also known as CryptoNight ReverseWaltz), for upcoming [Graft](https://www.graft.network/) fork.**
+- **[#931](https://github.com/xmrig/xmrig/issues/931) Added new algorithm `cryptonight/zls`, short alias `cn/zls` for [Zelerius Network](https://zelerius.org) fork.**
+- **[#940](https://github.com/xmrig/xmrig/issues/940) Added new algorithm `cryptonight/double`, short alias `cn/double` (also known as CryptoNight HeavyX), for [X-CASH](https://x-cash.org/).**
+- [#951](https://github.com/xmrig/xmrig/issues/951#issuecomment-469581529) Fixed crash if AVX was disabled on OS level.
+- [#952](https://github.com/xmrig/xmrig/issues/952) Fixed compile error on some Linux.
+- [#957](https://github.com/xmrig/xmrig/issues/957#issuecomment-468890667) Added support for embedded config.
+- [#958](https://github.com/xmrig/xmrig/pull/958) Fixed incorrect user agent on ARM platforms.
+- [#968](https://github.com/xmrig/xmrig/pull/968) Optimized `cn/r` algorithm performance.
+
 # v2.13.1
-[#946](https://github.com/xmrig/xmrig/pull/946) Optimized software AES implementations for CPUs without hardware AES support. `cn/r`, `cn/wow` up to 2.6 times faster, 4-9% improvements for other algorithms.
+- [#946](https://github.com/xmrig/xmrig/pull/946) Optimized software AES implementations for CPUs without hardware AES support. `cn/r`, `cn/wow` up to 2.6 times faster, 4-9% improvements for other algorithms.
 
 # v2.13.0
 - **[#938](https://github.com/xmrig/xmrig/issues/938) Added support for new algorithm `cryptonight/r`, short alias `cn/r` (also known as CryptoNightR or CryptoNight variant 4), for upcoming [Monero](https://www.getmonero.org/) fork on March 9, thanks [@SChernykh](https://github.com/SChernykh).**

CMakeLists.txt

@@ -12,6 +12,7 @@ option(WITH_TLS       "Enable OpenSSL support" ON)
 option(WITH_ASM             "Enable ASM PoW implementations" ON)
 option(BUILD_STATIC         "Build static binary" OFF)
 option(ARM_TARGET           "Force use specific ARM target 8 or 7" 0)
+option(WITH_EMBEDDED_CONFIG "Enable internal embedded JSON config" OFF)
 
 include (CheckIncludeFile)
 include (cmake/cpu.cmake)
@@ -65,6 +66,7 @@ set(HEADERS
     src/common/utils/mm_malloc.h
     src/common/xmrig.h
     src/core/ConfigLoader_platform.h
+    src/core/ConfigLoader_default.h
     src/core/Controller.h
     src/interfaces/IJobResultListener.h
     src/interfaces/IThread.h
@@ -154,7 +156,6 @@ set(SOURCES_CRYPTO
     src/crypto/c_blake256.c
     src/crypto/c_jh.c
     src/crypto/c_skein.c
-    src/crypto/CryptonightR_gen.cpp
    )
 
 if (WIN32)
@@ -249,6 +250,10 @@ if (NOT WITH_CN_PICO)
     add_definitions(/DXMRIG_NO_CN_PICO)
 endif()
 
+if (WITH_EMBEDDED_CONFIG)
+    add_definitions(/DXMRIG_FEATURE_EMBEDDED_CONFIG)
+endif()
+
 if (WITH_HTTPD)
     find_package(MHD)
 

cmake/asm.cmake

@@ -23,7 +23,7 @@ if (WITH_ASM AND NOT XMRIG_ARM AND CMAKE_SIZEOF_VOID_P EQUAL 8)
         if (WIN32 AND CMAKE_C_COMPILER_ID MATCHES GNU)
             set(XMRIG_ASM_FILES
                 "src/crypto/asm/win64/cn_main_loop.S"
-                "src/crypto/asm/win64/CryptonightR_template.S"
+                "src/crypto/asm/CryptonightR_template.S"
             )
         else()
             set(XMRIG_ASM_FILES
@@ -36,7 +36,7 @@ if (WITH_ASM AND NOT XMRIG_ARM AND CMAKE_SIZEOF_VOID_P EQUAL 8)
     endif()
 
     add_library(${XMRIG_ASM_LIBRARY} STATIC ${XMRIG_ASM_FILES})
-    set(XMRIG_ASM_SOURCES src/crypto/Asm.h src/crypto/Asm.cpp)
+    set(XMRIG_ASM_SOURCES src/crypto/Asm.h src/crypto/Asm.cpp src/crypto/CryptonightR_gen.cpp)
     set_property(TARGET ${XMRIG_ASM_LIBRARY} PROPERTY LINKER_LANGUAGE C)
 else()
     set(XMRIG_ASM_SOURCES "")

src/base/io/Json.cpp

@@ -47,3 +47,47 @@ const char *xmrig::Json::getString(const rapidjson::Value &obj, const char *key,
 
     return defaultValue;
 }
+
+
+int xmrig::Json::getInt(const rapidjson::Value &obj, const char *key, int defaultValue)
+{
+    auto i = obj.FindMember(key);
+    if (i != obj.MemberEnd() && i->value.IsInt()) {
+        return i->value.GetInt();
+    }
+
+    return defaultValue;
+}
+
+
+int64_t xmrig::Json::getInt64(const rapidjson::Value &obj, const char *key, int64_t defaultValue)
+{
+    auto i = obj.FindMember(key);
+    if (i != obj.MemberEnd() && i->value.IsInt64()) {
+        return i->value.GetInt64();
+    }
+
+    return defaultValue;
+}
+
+
+uint64_t xmrig::Json::getUint64(const rapidjson::Value &obj, const char *key, uint64_t defaultValue)
+{
+    auto i = obj.FindMember(key);
+    if (i != obj.MemberEnd() && i->value.IsUint64()) {
+        return i->value.GetUint64();
+    }
+
+    return defaultValue;
+}
+
+
+unsigned xmrig::Json::getUint(const rapidjson::Value &obj, const char *key, unsigned defaultValue)
+{
+    auto i = obj.FindMember(key);
+    if (i != obj.MemberEnd() && i->value.IsUint()) {
+        return i->value.GetUint();
+    }
+
+    return defaultValue;
+}

src/base/io/Json.h

@@ -37,6 +37,10 @@ class Json
 public:
     static bool getBool(const rapidjson::Value &obj, const char *key, bool defaultValue = false);
     static const char *getString(const rapidjson::Value &obj, const char *key, const char *defaultValue = nullptr);
+    static int getInt(const rapidjson::Value &obj, const char *key, int defaultValue = 0);
+    static int64_t getInt64(const rapidjson::Value &obj, const char *key, int64_t defaultValue = 0);
+    static uint64_t getUint64(const rapidjson::Value &obj, const char *key, uint64_t defaultValue = 0);
+    static unsigned getUint(const rapidjson::Value &obj, const char *key, unsigned defaultValue = 0);
 
     static bool get(const char *fileName, rapidjson::Document &doc);
     static bool save(const char *fileName, const rapidjson::Document &doc);

src/base/kernel/Entry.cpp

@@ -23,6 +23,7 @@
  */
 
 
+#include <stdio.h>
 #include <uv.h>
 
 

src/base/net/Pool.cpp

@@ -45,6 +45,8 @@
 #endif
 
 
+namespace xmrig {
+
 static const char *kEnabled     = "enabled";
 static const char *kFingerprint = "tls-fingerprint";
 static const char *kKeepalive   = "keepalive";
@@ -56,6 +58,8 @@ static const char *kUrl         = "url";
 static const char *kUser        = "user";
 static const char *kVariant     = "variant";
 
+}
+
 
 xmrig::Pool::Pool() :
     m_enabled(true),
@@ -128,6 +132,7 @@ xmrig::Pool::Pool(const rapidjson::Value &object) :
 
 
 xmrig::Pool::Pool(const char *host, uint16_t port, const char *user, const char *password, int keepAlive, bool nicehash, bool tls) :
+    m_enabled(true),
     m_nicehash(nicehash),
     m_tls(tls),
     m_keepAlive(keepAlive),
@@ -492,6 +497,9 @@ void xmrig::Pool::rebuild()
     addVariant(VARIANT_XAO);
     addVariant(VARIANT_RTO);
     addVariant(VARIANT_GPU);
+    addVariant(VARIANT_RWZ);
+    addVariant(VARIANT_ZLS);
+    addVariant(VARIANT_DOUBLE);
     addVariant(VARIANT_AUTO);
 #   endif
 }

src/common/config/ConfigLoader.cpp

@@ -53,6 +53,11 @@
 #include "rapidjson/fwd.h"
 
 
+#ifdef XMRIG_FEATURE_EMBEDDED_CONFIG
+#   include "core/ConfigLoader_default.h"
+#endif
+
+
 xmrig::ConfigWatcher *xmrig::ConfigLoader::m_watcher     = nullptr;
 xmrig::IConfigCreator *xmrig::ConfigLoader::m_creator    = nullptr;
 xmrig::IConfigListener *xmrig::ConfigLoader::m_listener  = nullptr;
@@ -180,6 +185,15 @@ xmrig::IConfig *xmrig::ConfigLoader::load(Process *process, IConfigCreator *crea
         loadFromFile(config, process->location(Process::ExeLocation, "config.json"));
     }
 
+#   ifdef XMRIG_FEATURE_EMBEDDED_CONFIG
+    if (!config->finalize()) {
+        delete config;
+
+        config = m_creator->create();
+        loadFromJSON(config, default_config);
+    }
+#   endif
+
     if (!config->finalize()) {
         if (!config->algorithm().isValid()) {
             fprintf(stderr, "No valid algorithm specified. Exiting.\n");

src/common/cpu/BasicCpuInfo.cpp

@@ -36,6 +36,10 @@
 #   define bit_AES (1 << 25)
 #endif
 
+#ifndef bit_OSXSAVE
+#   define bit_OSXSAVE (1 << 27)
+#endif
+
 #ifndef bit_AVX2
 #   define bit_AVX2 (1 << 5)
 #endif
@@ -107,10 +111,19 @@ static inline bool has_avx2()
 }
 
 
+static inline bool has_ossave()
+{
+    int32_t cpu_info[4] = { 0 };
+    cpuid(PROCESSOR_INFO, cpu_info);
+
+    return (cpu_info[ECX_Reg] & bit_OSXSAVE) != 0;
+}
+
+
 xmrig::BasicCpuInfo::BasicCpuInfo() :
     m_assembly(ASM_NONE),
     m_aes(has_aes_ni()),
-    m_avx2(has_avx2()),
+    m_avx2(has_avx2() && has_ossave()),
     m_brand(),
     m_threads(std::thread::hardware_concurrency())
 {

src/common/crypto/Algorithm.cpp

@@ -66,6 +66,9 @@ static AlgoData const algorithms[] = {
     { "cryptonight/xtlv9",     "cn/xtlv9",     xmrig::CRYPTONIGHT,       xmrig::VARIANT_HALF   },
     { "cryptonight/wow",       "cn/wow",       xmrig::CRYPTONIGHT,       xmrig::VARIANT_WOW    },
     { "cryptonight/r",         "cn/r",         xmrig::CRYPTONIGHT,       xmrig::VARIANT_4      },
+    { "cryptonight/rwz",       "cn/rwz",       xmrig::CRYPTONIGHT,       xmrig::VARIANT_RWZ    },
+    { "cryptonight/zls",       "cn/zls",       xmrig::CRYPTONIGHT,       xmrig::VARIANT_ZLS    },
+    { "cryptonight/double",    "cn/double",    xmrig::CRYPTONIGHT,       xmrig::VARIANT_DOUBLE },
 
 #   ifndef XMRIG_NO_AEON
     { "cryptonight-lite",      "cn-lite",      xmrig::CRYPTONIGHT_LITE,  xmrig::VARIANT_AUTO },
@@ -133,6 +136,9 @@ static const char *variants[] = {
     "gpu",
     "wow",
     "r",
+    "rwz",
+    "zls",
+    "double"
 };
 
 

src/common/net/Job.cpp

@@ -137,6 +137,12 @@ bool xmrig::Job::setBlob(const char *blob)
         else if (m_algorithm.variant() == VARIANT_WOW && m_blob[0] < 11) {
             m_algorithm.setVariant(VARIANT_2);
         }
+        else if (m_algorithm.variant() == VARIANT_RWZ && m_blob[0] < 12) {
+            m_algorithm.setVariant(VARIANT_2);
+        }
+        else if (m_algorithm.variant() == VARIANT_ZLS && m_blob[0] < 8) {
+            m_algorithm.setVariant(VARIANT_2);
+        }
     }
 
 #   ifdef XMRIG_PROXY_PROJECT

src/common/xmrig.h

@@ -76,6 +76,9 @@ enum Variant {
     VARIANT_GPU    = 11, // CryptoNight-GPU (Ryo)
     VARIANT_WOW    = 12, // CryptoNightR (Wownero)
     VARIANT_4      = 13, // CryptoNightR (Monero's variant 4)
+    VARIANT_RWZ    = 14, // CryptoNight variant 2 with 3/4 iterations and reversed shuffle operation (Graft)
+    VARIANT_ZLS    = 15, // CryptoNight variant 2 with 3/4 iterations (Zelerius)
+    VARIANT_DOUBLE = 16, // CryptoNight variant 2 with double iterations (X-CASH)
     VARIANT_MAX
 };
 

src/config.json

@@ -39,5 +39,5 @@
     "safe": false,
     "threads": null,
     "user-agent": null,
-    "watch": false
+    "watch": true
 }

src/core/ConfigLoader_default.h

@@ -0,0 +1,84 @@
+/* XMRig
+ * Copyright 2010      Jeff Garzik <jgarzik@pobox.com>
+ * Copyright 2012-2014 pooler      <pooler@litecoinpool.org>
+ * Copyright 2014      Lucas Jones <https://github.com/lucasjones>
+ * Copyright 2014-2016 Wolf9466    <https://github.com/OhGodAPet>
+ * Copyright 2016      Jay D Dee   <jayddee246@gmail.com>
+ * Copyright 2017-2018 XMR-Stak    <https://github.com/fireice-uk>, <https://github.com/psychocrypt>
+ * Copyright 2018-2019 SChernykh   <https://github.com/SChernykh>
+ * Copyright 2016-2019 XMRig       <https://github.com/xmrig>, <support@xmrig.com>
+ *
+ *   This program is free software: you can redistribute it and/or modify
+ *   it under the terms of the GNU General Public License as published by
+ *   the Free Software Foundation, either version 3 of the License, or
+ *   (at your option) any later version.
+ *
+ *   This program is distributed in the hope that it will be useful,
+ *   but WITHOUT ANY WARRANTY; without even the implied warranty of
+ *   MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+ *   GNU General Public License for more details.
+ *
+ *   You should have received a copy of the GNU General Public License
+ *   along with this program. If not, see <http://www.gnu.org/licenses/>.
+ */
+
+#ifndef XMRIG_CONFIGLOADER_DEFAULT_H
+#define XMRIG_CONFIGLOADER_DEFAULT_H
+
+
+namespace xmrig {
+
+
+#ifdef XMRIG_FEATURE_EMBEDDED_CONFIG
+const static char *default_config =
+R"===(
+{
+    "algo": "cryptonight",
+    "api": {
+        "port": 0,
+        "access-token": null,
+        "id": null,
+        "worker-id": null,
+        "ipv6": false,
+        "restricted": true
+    },
+    "asm": true,
+    "autosave": true,
+    "av": 0,
+    "background": false,
+    "colors": true,
+    "cpu-affinity": null,
+    "cpu-priority": null,
+    "donate-level": 5,
+    "huge-pages": true,
+    "hw-aes": null,
+    "log-file": null,
+    "max-cpu-usage": 100,
+    "pools": [
+        {
+            "url": "donate.v2.xmrig.com:3333",
+            "user": "YOUR_WALLET_ADDRESS",
+            "pass": "x",
+            "rig-id": null,
+            "nicehash": false,
+            "keepalive": false,
+            "variant": -1,
+            "tls": false,
+            "tls-fingerprint": null
+        }
+    ],
+    "print-time": 60,
+    "retries": 5,
+    "retry-pause": 5,
+    "safe": false,
+    "threads": null,
+    "user-agent": null,
+    "watch": false
+}
+)===";
+#endif
+
+
+} /* namespace xmrig */
+
+#endif /* XMRIG_CONFIGLOADER_DEFAULT_H */

src/core/cpu/AdvancedCpuInfo.cpp

@@ -85,7 +85,7 @@ xmrig::AdvancedCpuInfo::AdvancedCpuInfo() :
         }
     }
 
-    m_avx2 = data.flags[CPU_FEATURE_AVX2];
+    m_avx2 = data.flags[CPU_FEATURE_AVX2] && data.flags[CPU_FEATURE_OSXSAVE];
 }
 
 

src/crypto/CryptoNight_arm.h

@@ -436,7 +436,7 @@ static inline void cryptonight_monero_tweak(const uint8_t* l, uint64_t idx, __m1
     uint64_t* mem_out = (uint64_t*)&l[idx];
 
     if (BASE == xmrig::VARIANT_2) {
-        VARIANT2_SHUFFLE(l, idx, ax0, bx0, bx1, cx);
+        VARIANT2_SHUFFLE(l, idx, ax0, bx0, bx1, cx, (VARIANT == xmrig::VARIANT_RWZ ? 1 : 0));
         _mm_store_si128((__m128i *)mem_out, _mm_xor_si128(bx0, cx));
     } else {
         __m128i tmp = _mm_xor_si128(bx0, cx);
@@ -530,9 +530,9 @@ inline void cryptonight_single_hash(const uint8_t *__restrict__ input, size_t si
 
         if (BASE == xmrig::VARIANT_2) {
             if (VARIANT == xmrig::VARIANT_4) {
-                VARIANT2_SHUFFLE(l0, idx0 & MASK, ax0, bx0, bx1, cx);
+                VARIANT2_SHUFFLE(l0, idx0 & MASK, ax0, bx0, bx1, cx, 0);
             } else {
-                VARIANT2_SHUFFLE2(l0, idx0 & MASK, ax0, bx0, bx1, hi, lo);
+                VARIANT2_SHUFFLE2(l0, idx0 & MASK, ax0, bx0, bx1, hi, lo, (VARIANT == xmrig::VARIANT_RWZ ? 1 : 0));
             }
         }
 
@@ -709,9 +709,9 @@ inline void cryptonight_double_hash(const uint8_t *__restrict__ input, size_t si
 
         if (BASE == xmrig::VARIANT_2) {
             if (VARIANT == xmrig::VARIANT_4) {
-                VARIANT2_SHUFFLE(l0, idx0 & MASK, ax0, bx00, bx01, cx0);
+                VARIANT2_SHUFFLE(l0, idx0 & MASK, ax0, bx00, bx01, cx0, 0);
             } else {
-                VARIANT2_SHUFFLE2(l0, idx0 & MASK, ax0, bx00, bx01, hi, lo);
+                VARIANT2_SHUFFLE2(l0, idx0 & MASK, ax0, bx00, bx01, hi, lo, (VARIANT == xmrig::VARIANT_RWZ ? 1 : 0));
             }
         }
 
@@ -767,9 +767,9 @@ inline void cryptonight_double_hash(const uint8_t *__restrict__ input, size_t si
 
         if (BASE == xmrig::VARIANT_2) {
             if (VARIANT == xmrig::VARIANT_4) {
-                VARIANT2_SHUFFLE(l1, idx1 & MASK, ax1, bx10, bx11, cx1);
+                VARIANT2_SHUFFLE(l1, idx1 & MASK, ax1, bx10, bx11, cx1, 0);
             } else {
-                VARIANT2_SHUFFLE2(l1, idx1 & MASK, ax1, bx10, bx11, hi, lo);
+                VARIANT2_SHUFFLE2(l1, idx1 & MASK, ax1, bx10, bx11, hi, lo, (VARIANT == xmrig::VARIANT_RWZ ? 1 : 0));
             }
         }
 

src/crypto/CryptoNight_constants.h

@@ -42,6 +42,9 @@ constexpr const uint32_t CRYPTONIGHT_MASK         = 0x1FFFF0;
 constexpr const uint32_t CRYPTONIGHT_ITER         = 0x80000;
 constexpr const uint32_t CRYPTONIGHT_HALF_ITER    = 0x40000;
 constexpr const uint32_t CRYPTONIGHT_XAO_ITER     = 0x100000;
+constexpr const uint32_t CRYPTONIGHT_DOUBLE_ITER  = 0x100000;
+constexpr const uint32_t CRYPTONIGHT_WALTZ_ITER   = 0x60000;
+constexpr const uint32_t CRYPTONIGHT_ZLS_ITER     = 0x60000;
 
 constexpr const uint32_t CRYPTONIGHT_GPU_ITER     = 0xC000;
 constexpr const uint32_t CRYPTONIGHT_GPU_MASK     = 0x1FFFC0;
@@ -134,6 +137,9 @@ template<> inline constexpr uint32_t cn_select_iter<CRYPTONIGHT, VARIANT_MSR>()
 template<> inline constexpr uint32_t cn_select_iter<CRYPTONIGHT, VARIANT_XAO>()        { return CRYPTONIGHT_XAO_ITER; }
 template<> inline constexpr uint32_t cn_select_iter<CRYPTONIGHT, VARIANT_RTO>()        { return CRYPTONIGHT_ITER; }
 template<> inline constexpr uint32_t cn_select_iter<CRYPTONIGHT, VARIANT_GPU>()        { return CRYPTONIGHT_GPU_ITER; }
+template<> inline constexpr uint32_t cn_select_iter<CRYPTONIGHT, VARIANT_RWZ>()        { return CRYPTONIGHT_WALTZ_ITER; }
+template<> inline constexpr uint32_t cn_select_iter<CRYPTONIGHT, VARIANT_ZLS>()        { return CRYPTONIGHT_ZLS_ITER; }
+template<> inline constexpr uint32_t cn_select_iter<CRYPTONIGHT, VARIANT_DOUBLE>()     { return CRYPTONIGHT_DOUBLE_ITER; }
 template<> inline constexpr uint32_t cn_select_iter<CRYPTONIGHT_LITE, VARIANT_0>()     { return CRYPTONIGHT_LITE_ITER; }
 template<> inline constexpr uint32_t cn_select_iter<CRYPTONIGHT_LITE, VARIANT_1>()     { return CRYPTONIGHT_LITE_ITER; }
 template<> inline constexpr uint32_t cn_select_iter<CRYPTONIGHT_HEAVY, VARIANT_0>()    { return CRYPTONIGHT_HEAVY_ITER; }
@@ -153,11 +159,16 @@ inline uint32_t cn_select_iter(Algo algorithm, Variant variant)
         return CRYPTONIGHT_GPU_ITER;
 
     case VARIANT_RTO:
+    case VARIANT_DOUBLE:
         return CRYPTONIGHT_XAO_ITER;
 
     case VARIANT_TRTL:
         return CRYPTONIGHT_TRTL_ITER;
 
+    case VARIANT_RWZ:
+    case VARIANT_ZLS:
+        return CRYPTONIGHT_WALTZ_ITER;
+
     default:
         break;
     }
@@ -199,6 +210,9 @@ template<> inline constexpr Variant cn_base_variant<VARIANT_TRTL>()  { return VA
 template<> inline constexpr Variant cn_base_variant<VARIANT_GPU>()    { return VARIANT_GPU; }
 template<> inline constexpr Variant cn_base_variant<VARIANT_WOW>()    { return VARIANT_2; }
 template<> inline constexpr Variant cn_base_variant<VARIANT_4>()      { return VARIANT_2; }
+template<> inline constexpr Variant cn_base_variant<VARIANT_RWZ>()    { return VARIANT_2; }
+template<> inline constexpr Variant cn_base_variant<VARIANT_ZLS>()    { return VARIANT_2; }
+template<> inline constexpr Variant cn_base_variant<VARIANT_DOUBLE>() { return VARIANT_2; }
 
 
 template<Variant variant> inline constexpr bool cn_is_cryptonight_r() { return false; }

src/crypto/CryptoNight_monero.h

@@ -83,11 +83,11 @@
         sqrt_result_xmm_##part = int_sqrt_v2(cx_0 + division_result); \
     } while (0)
 
-#   define VARIANT2_SHUFFLE(base_ptr, offset, _a, _b, _b1, _c) \
+#   define VARIANT2_SHUFFLE(base_ptr, offset, _a, _b, _b1, _c, reverse) \
     do { \
-        const __m128i chunk1 = _mm_load_si128((__m128i *)((base_ptr) + ((offset) ^ 0x10))); \
+        const __m128i chunk1 = _mm_load_si128((__m128i *)((base_ptr) + ((offset) ^ (reverse ? 0x30 : 0x10)))); \
         const __m128i chunk2 = _mm_load_si128((__m128i *)((base_ptr) + ((offset) ^ 0x20))); \
-        const __m128i chunk3 = _mm_load_si128((__m128i *)((base_ptr) + ((offset) ^ 0x30))); \
+        const __m128i chunk3 = _mm_load_si128((__m128i *)((base_ptr) + ((offset) ^ (reverse ? 0x10 : 0x30)))); \
         _mm_store_si128((__m128i *)((base_ptr) + ((offset) ^ 0x10)), _mm_add_epi64(chunk3, _b1)); \
         _mm_store_si128((__m128i *)((base_ptr) + ((offset) ^ 0x20)), _mm_add_epi64(chunk1, _b)); \
         _mm_store_si128((__m128i *)((base_ptr) + ((offset) ^ 0x30)), _mm_add_epi64(chunk2, _a)); \
@@ -96,15 +96,20 @@
         } \
     } while (0)
 
-#   define VARIANT2_SHUFFLE2(base_ptr, offset, _a, _b, _b1, hi, lo) \
+#   define VARIANT2_SHUFFLE2(base_ptr, offset, _a, _b, _b1, hi, lo, reverse) \
     do { \
         const __m128i chunk1 = _mm_xor_si128(_mm_load_si128((__m128i *)((base_ptr) + ((offset) ^ 0x10))), _mm_set_epi64x(lo, hi)); \
         const __m128i chunk2 = _mm_load_si128((__m128i *)((base_ptr) + ((offset) ^ 0x20))); \
         hi ^= ((uint64_t*)((base_ptr) + ((offset) ^ 0x20)))[0]; \
         lo ^= ((uint64_t*)((base_ptr) + ((offset) ^ 0x20)))[1]; \
         const __m128i chunk3 = _mm_load_si128((__m128i *)((base_ptr) + ((offset) ^ 0x30))); \
+        if (reverse) { \
+            _mm_store_si128((__m128i *)((base_ptr) + ((offset) ^ 0x10)), _mm_add_epi64(chunk1, _b1)); \
+            _mm_store_si128((__m128i *)((base_ptr) + ((offset) ^ 0x20)), _mm_add_epi64(chunk3, _b)); \
+        } else { \
             _mm_store_si128((__m128i *)((base_ptr) + ((offset) ^ 0x10)), _mm_add_epi64(chunk3, _b1)); \
             _mm_store_si128((__m128i *)((base_ptr) + ((offset) ^ 0x20)), _mm_add_epi64(chunk1, _b)); \
+        } \
         _mm_store_si128((__m128i *)((base_ptr) + ((offset) ^ 0x30)), _mm_add_epi64(chunk2, _a)); \
     } while (0)
 
@@ -128,11 +133,11 @@
         sqrt_result_##part += ((r2 + b > sqrt_input) ? -1 : 0) + ((r2 + (1ULL << 32) < sqrt_input - s) ? 1 : 0); \
     } while (0)
 
-#   define VARIANT2_SHUFFLE(base_ptr, offset, _a, _b, _b1, _c) \
+#   define VARIANT2_SHUFFLE(base_ptr, offset, _a, _b, _b1, _c, reverse) \
     do { \
-        const uint64x2_t chunk1 = vld1q_u64((uint64_t*)((base_ptr) + ((offset) ^ 0x10))); \
+        const uint64x2_t chunk1 = vld1q_u64((uint64_t*)((base_ptr) + ((offset) ^ (reverse ? 0x30 : 0x10)))); \
         const uint64x2_t chunk2 = vld1q_u64((uint64_t*)((base_ptr) + ((offset) ^ 0x20))); \
-        const uint64x2_t chunk3 = vld1q_u64((uint64_t*)((base_ptr) + ((offset) ^ 0x30))); \
+        const uint64x2_t chunk3 = vld1q_u64((uint64_t*)((base_ptr) + ((offset) ^ (reverse ? 0x10 : 0x30)))); \
         vst1q_u64((uint64_t*)((base_ptr) + ((offset) ^ 0x10)), vaddq_u64(chunk3, vreinterpretq_u64_u8(_b1))); \
         vst1q_u64((uint64_t*)((base_ptr) + ((offset) ^ 0x20)), vaddq_u64(chunk1, vreinterpretq_u64_u8(_b))); \
         vst1q_u64((uint64_t*)((base_ptr) + ((offset) ^ 0x30)), vaddq_u64(chunk2, vreinterpretq_u64_u8(_a))); \
@@ -141,15 +146,20 @@
         } \
     } while (0)
 
-#   define VARIANT2_SHUFFLE2(base_ptr, offset, _a, _b, _b1, hi, lo) \
+#   define VARIANT2_SHUFFLE2(base_ptr, offset, _a, _b, _b1, hi, lo, reverse) \
     do { \
         const uint64x2_t chunk1 = veorq_u64(vld1q_u64((uint64_t*)((base_ptr) + ((offset) ^ 0x10))), vcombine_u64(vcreate_u64(hi), vcreate_u64(lo))); \
         const uint64x2_t chunk2 = vld1q_u64((uint64_t*)((base_ptr) + ((offset) ^ 0x20))); \
         hi ^= ((uint64_t*)((base_ptr) + ((offset) ^ 0x20)))[0]; \
         lo ^= ((uint64_t*)((base_ptr) + ((offset) ^ 0x20)))[1]; \
         const uint64x2_t chunk3 = vld1q_u64((uint64_t*)((base_ptr) + ((offset) ^ 0x30))); \
+        if (reverse) { \
+            vst1q_u64((uint64_t*)((base_ptr) + ((offset) ^ 0x10)), vaddq_u64(chunk1, vreinterpretq_u64_u8(_b1))); \
+            vst1q_u64((uint64_t*)((base_ptr) + ((offset) ^ 0x20)), vaddq_u64(chunk3, vreinterpretq_u64_u8(_b))); \
+        } else { \
             vst1q_u64((uint64_t*)((base_ptr) + ((offset) ^ 0x10)), vaddq_u64(chunk3, vreinterpretq_u64_u8(_b1))); \
             vst1q_u64((uint64_t*)((base_ptr) + ((offset) ^ 0x20)), vaddq_u64(chunk1, vreinterpretq_u64_u8(_b))); \
+        } \
         vst1q_u64((uint64_t*)((base_ptr) + ((offset) ^ 0x30)), vaddq_u64(chunk2, vreinterpretq_u64_u8(_a))); \
     } while (0)
 #endif

src/crypto/CryptoNight_test.h

@@ -58,27 +58,58 @@ const static uint8_t test_input[380] = {
     0xCF, 0x50, 0x29, 0x6A, 0x07, 0x0B, 0x93, 0x8F, 0x8F, 0xA8, 0x10, 0x04
 };
 
-const static char* test_input_WOW = R"===(9d47bf4c41b7e8e727e681715acb47fa1677cdba9ca7bcb05ad8cc8abd5daa66 5468697320697320612074657374205468697320697320612074657374205468697320697320612074657374 1806260
-0d4a495cb844a3ca8ba4edb8e6bcf829ef1c06d9cdea2b62ca46c2a21b8b0a79 4c6f72656d20697073756d20646f6c6f722073697420616d65742c20636f6e73656374657475722061646970697363696e67 1806261
-a1d6d848b5c5915fccd2f64cf216c6b1a02cf7c77bc80d8d4e51b419e88ff0dd 656c69742c2073656420646f20656975736d6f642074656d706f7220696e6369646964756e74207574206c61626f7265 1806262
-af3a8544a0221a148c2ac90484b19861e3afca33fe17021efb8ad6496b567915 657420646f6c6f7265206d61676e6120616c697175612e20557420656e696d206164206d696e696d2076656e69616d2c 1806263
-313399e0963ae8a99dab8af66d343e097dae0c0feb08dbc43ccdafef5515f413 71756973206e6f737472756420657865726369746174696f6e20756c6c616d636f206c61626f726973206e697369 1806264
-6021c6ef90bff9ae94a7506d623d3a7a86c1756d655f50dd558f716d64622a34 757420616c697175697020657820656120636f6d6d6f646f20636f6e7365717561742e20447569732061757465 1806265
-2b13000535f3db5f9b9b84a65c4351f386cd2cdedebb8c3ad2eab086e6a3fee5 697275726520646f6c6f7220696e20726570726568656e646572697420696e20766f6c7570746174652076656c6974 1806266
-fc0e1dad8e895749dc90eb690bc1ba059a1cd772afaaf65a106bf9e5e6b80503 657373652063696c6c756d20646f6c6f726520657520667567696174206e756c6c612070617269617475722e 1806267
-b60b0afe144deff7d903ed2d5545e77ebe66a3c51fee7016eeb8fee9eb630c0f 4578636570746575722073696e74206f6363616563617420637570696461746174206e6f6e2070726f6964656e742c 1806268
-64774b27e7d5fec862fc4c0c13ac6bf09123b6f05bb0e4b75c97f379a2b3a679 73756e7420696e2063756c706120717569206f666669636961206465736572756e74206d6f6c6c697420616e696d20696420657374206c61626f72756d2e 1806269)===";
 
-const static char* test_input_R = R"===(f759588ad57e758467295443a9bd71490abff8e9dad1b95b6bf2f5d0d78387bc 5468697320697320612074657374205468697320697320612074657374205468697320697320612074657374 1806260
+struct cn_r_test_input_data
-5bb833deca2bdd7252a9ccd7b4ce0b6a4854515794b56c207262f7a5b9bdb566 4c6f72656d20697073756d20646f6c6f722073697420616d65742c20636f6e73656374657475722061646970697363696e67 1806261
+{
-1ee6728da60fbd8d7d55b2b1ade487a3cf52a2c3ac6f520db12c27d8921f6cab 656c69742c2073656420646f20656975736d6f642074656d706f7220696e6369646964756e74207574206c61626f7265 1806262
+    uint64_t height;
-6969fe2ddfb758438d48049f302fc2108a4fcc93e37669170e6db4b0b9b4c4cb 657420646f6c6f7265206d61676e6120616c697175612e20557420656e696d206164206d696e696d2076656e69616d2c 1806263
+    size_t size;
-7f3048b4e90d0cbe7a57c0394f37338a01fae3adfdc0e5126d863a895eb04e02 71756973206e6f737472756420657865726369746174696f6e20756c6c616d636f206c61626f726973206e697369 1806264
+    uint8_t data[64];
-1d290443a4b542af04a82f6b2494a6ee7f20f2754c58e0849032483a56e8e2ef 757420616c697175697020657820656120636f6d6d6f646f20636f6e7365717561742e20447569732061757465 1806265
+};
-c43cc6567436a86afbd6aa9eaa7c276e9806830334b614b2bee23cc76634f6fd 697275726520646f6c6f7220696e20726570726568656e646572697420696e20766f6c7570746174652076656c6974 1806266
+
-87be2479c0c4e8edfdfaa5603e93f4265b3f8224c1c5946feb424819d18990a4 657373652063696c6c756d20646f6c6f726520657520667567696174206e756c6c612070617269617475722e 1806267
+
-dd9d6a6d8e47465cceac0877ef889b93e7eba979557e3935d7f86dce11b070f3 4578636570746575722073696e74206f6363616563617420637570696461746174206e6f6e2070726f6964656e742c 1806268
+const static cn_r_test_input_data cn_r_test_input[] = {
-75c6f2ae49a20521de97285b431e717125847fb8935ed84a61e7f8d36a2c3d8e 73756e7420696e2063756c706120717569206f666669636961206465736572756e74206d6f6c6c697420616e696d20696420657374206c61626f72756d2e 1806269)===";
+    { 1806260, 44, { 0x54, 0x68, 0x69, 0x73, 0x20, 0x69, 0x73, 0x20, 0x61, 0x20, 0x74, 0x65, 0x73, 0x74, 0x20, 0x54, 0x68, 0x69, 0x73, 0x20, 0x69, 0x73, 0x20, 0x61, 0x20, 0x74, 0x65, 0x73, 0x74, 0x20, 0x54, 0x68, 0x69, 0x73, 0x20, 0x69, 0x73, 0x20, 0x61, 0x20, 0x74, 0x65, 0x73, 0x74 } },
+    { 1806261, 50, { 0x4c, 0x6f, 0x72, 0x65, 0x6d, 0x20, 0x69, 0x70, 0x73, 0x75, 0x6d, 0x20, 0x64, 0x6f, 0x6c, 0x6f, 0x72, 0x20, 0x73, 0x69, 0x74, 0x20, 0x61, 0x6d, 0x65, 0x74, 0x2c, 0x20, 0x63, 0x6f, 0x6e, 0x73, 0x65, 0x63, 0x74, 0x65, 0x74, 0x75, 0x72, 0x20, 0x61, 0x64, 0x69, 0x70, 0x69, 0x73, 0x63, 0x69, 0x6e, 0x67 } },
+    { 1806262, 48, { 0x65, 0x6c, 0x69, 0x74, 0x2c, 0x20, 0x73, 0x65, 0x64, 0x20, 0x64, 0x6f, 0x20, 0x65, 0x69, 0x75, 0x73, 0x6d, 0x6f, 0x64, 0x20, 0x74, 0x65, 0x6d, 0x70, 0x6f, 0x72, 0x20, 0x69, 0x6e, 0x63, 0x69, 0x64, 0x69, 0x64, 0x75, 0x6e, 0x74, 0x20, 0x75, 0x74, 0x20, 0x6c, 0x61, 0x62, 0x6f, 0x72, 0x65 } },
+    { 1806263, 48, { 0x65, 0x74, 0x20, 0x64, 0x6f, 0x6c, 0x6f, 0x72, 0x65, 0x20, 0x6d, 0x61, 0x67, 0x6e, 0x61, 0x20, 0x61, 0x6c, 0x69, 0x71, 0x75, 0x61, 0x2e, 0x20, 0x55, 0x74, 0x20, 0x65, 0x6e, 0x69, 0x6d, 0x20, 0x61, 0x64, 0x20, 0x6d, 0x69, 0x6e, 0x69, 0x6d, 0x20, 0x76, 0x65, 0x6e, 0x69, 0x61, 0x6d, 0x2c } },
+    { 1806264, 46, { 0x71, 0x75, 0x69, 0x73, 0x20, 0x6e, 0x6f, 0x73, 0x74, 0x72, 0x75, 0x64, 0x20, 0x65, 0x78, 0x65, 0x72, 0x63, 0x69, 0x74, 0x61, 0x74, 0x69, 0x6f, 0x6e, 0x20, 0x75, 0x6c, 0x6c, 0x61, 0x6d, 0x63, 0x6f, 0x20, 0x6c, 0x61, 0x62, 0x6f, 0x72, 0x69, 0x73, 0x20, 0x6e, 0x69, 0x73, 0x69 } },
+    { 1806265, 45, { 0x75, 0x74, 0x20, 0x61, 0x6c, 0x69, 0x71, 0x75, 0x69, 0x70, 0x20, 0x65, 0x78, 0x20, 0x65, 0x61, 0x20, 0x63, 0x6f, 0x6d, 0x6d, 0x6f, 0x64, 0x6f, 0x20, 0x63, 0x6f, 0x6e, 0x73, 0x65, 0x71, 0x75, 0x61, 0x74, 0x2e, 0x20, 0x44, 0x75, 0x69, 0x73, 0x20, 0x61, 0x75, 0x74, 0x65 } },
+    { 1806266, 47, { 0x69, 0x72, 0x75, 0x72, 0x65, 0x20, 0x64, 0x6f, 0x6c, 0x6f, 0x72, 0x20, 0x69, 0x6e, 0x20, 0x72, 0x65, 0x70, 0x72, 0x65, 0x68, 0x65, 0x6e, 0x64, 0x65, 0x72, 0x69, 0x74, 0x20, 0x69, 0x6e, 0x20, 0x76, 0x6f, 0x6c, 0x75, 0x70, 0x74, 0x61, 0x74, 0x65, 0x20, 0x76, 0x65, 0x6c, 0x69, 0x74 } },
+    { 1806267, 44, { 0x65, 0x73, 0x73, 0x65, 0x20, 0x63, 0x69, 0x6c, 0x6c, 0x75, 0x6d, 0x20, 0x64, 0x6f, 0x6c, 0x6f, 0x72, 0x65, 0x20, 0x65, 0x75, 0x20, 0x66, 0x75, 0x67, 0x69, 0x61, 0x74, 0x20, 0x6e, 0x75, 0x6c, 0x6c, 0x61, 0x20, 0x70, 0x61, 0x72, 0x69, 0x61, 0x74, 0x75, 0x72, 0x2e } },
+    { 1806268, 47, { 0x45, 0x78, 0x63, 0x65, 0x70, 0x74, 0x65, 0x75, 0x72, 0x20, 0x73, 0x69, 0x6e, 0x74, 0x20, 0x6f, 0x63, 0x63, 0x61, 0x65, 0x63, 0x61, 0x74, 0x20, 0x63, 0x75, 0x70, 0x69, 0x64, 0x61, 0x74, 0x61, 0x74, 0x20, 0x6e, 0x6f, 0x6e, 0x20, 0x70, 0x72, 0x6f, 0x69, 0x64, 0x65, 0x6e, 0x74, 0x2c } },
+    { 1806269, 62, { 0x73, 0x75, 0x6e, 0x74, 0x20, 0x69, 0x6e, 0x20, 0x63, 0x75, 0x6c, 0x70, 0x61, 0x20, 0x71, 0x75, 0x69, 0x20, 0x6f, 0x66, 0x66, 0x69, 0x63, 0x69, 0x61, 0x20, 0x64, 0x65, 0x73, 0x65, 0x72, 0x75, 0x6e, 0x74, 0x20, 0x6d, 0x6f, 0x6c, 0x6c, 0x69, 0x74, 0x20, 0x61, 0x6e, 0x69, 0x6d, 0x20, 0x69, 0x64, 0x20, 0x65, 0x73, 0x74, 0x20, 0x6c, 0x61, 0x62, 0x6f, 0x72, 0x75, 0x6d, 0x2e } },
+};
+
+
+// "cn/wow"
+const static uint8_t test_output_wow[] = {
+    0x9d, 0x47, 0xbf, 0x4c, 0x41, 0xb7, 0xe8, 0xe7, 0x27, 0xe6, 0x81, 0x71, 0x5a, 0xcb, 0x47, 0xfa, 0x16, 0x77, 0xcd, 0xba, 0x9c, 0xa7, 0xbc, 0xb0, 0x5a, 0xd8, 0xcc, 0x8a, 0xbd, 0x5d, 0xaa, 0x66,
+    0x0d, 0x4a, 0x49, 0x5c, 0xb8, 0x44, 0xa3, 0xca, 0x8b, 0xa4, 0xed, 0xb8, 0xe6, 0xbc, 0xf8, 0x29, 0xef, 0x1c, 0x06, 0xd9, 0xcd, 0xea, 0x2b, 0x62, 0xca, 0x46, 0xc2, 0xa2, 0x1b, 0x8b, 0x0a, 0x79,
+    0xa1, 0xd6, 0xd8, 0x48, 0xb5, 0xc5, 0x91, 0x5f, 0xcc, 0xd2, 0xf6, 0x4c, 0xf2, 0x16, 0xc6, 0xb1, 0xa0, 0x2c, 0xf7, 0xc7, 0x7b, 0xc8, 0x0d, 0x8d, 0x4e, 0x51, 0xb4, 0x19, 0xe8, 0x8f, 0xf0, 0xdd,
+    0xaf, 0x3a, 0x85, 0x44, 0xa0, 0x22, 0x1a, 0x14, 0x8c, 0x2a, 0xc9, 0x04, 0x84, 0xb1, 0x98, 0x61, 0xe3, 0xaf, 0xca, 0x33, 0xfe, 0x17, 0x02, 0x1e, 0xfb, 0x8a, 0xd6, 0x49, 0x6b, 0x56, 0x79, 0x15,
+    0x31, 0x33, 0x99, 0xe0, 0x96, 0x3a, 0xe8, 0xa9, 0x9d, 0xab, 0x8a, 0xf6, 0x6d, 0x34, 0x3e, 0x09, 0x7d, 0xae, 0x0c, 0x0f, 0xeb, 0x08, 0xdb, 0xc4, 0x3c, 0xcd, 0xaf, 0xef, 0x55, 0x15, 0xf4, 0x13,
+    0x60, 0x21, 0xc6, 0xef, 0x90, 0xbf, 0xf9, 0xae, 0x94, 0xa7, 0x50, 0x6d, 0x62, 0x3d, 0x3a, 0x7a, 0x86, 0xc1, 0x75, 0x6d, 0x65, 0x5f, 0x50, 0xdd, 0x55, 0x8f, 0x71, 0x6d, 0x64, 0x62, 0x2a, 0x34,
+    0x2b, 0x13, 0x00, 0x05, 0x35, 0xf3, 0xdb, 0x5f, 0x9b, 0x9b, 0x84, 0xa6, 0x5c, 0x43, 0x51, 0xf3, 0x86, 0xcd, 0x2c, 0xde, 0xde, 0xbb, 0x8c, 0x3a, 0xd2, 0xea, 0xb0, 0x86, 0xe6, 0xa3, 0xfe, 0xe5,
+    0xfc, 0x0e, 0x1d, 0xad, 0x8e, 0x89, 0x57, 0x49, 0xdc, 0x90, 0xeb, 0x69, 0x0b, 0xc1, 0xba, 0x05, 0x9a, 0x1c, 0xd7, 0x72, 0xaf, 0xaa, 0xf6, 0x5a, 0x10, 0x6b, 0xf9, 0xe5, 0xe6, 0xb8, 0x05, 0x03,
+    0xb6, 0x0b, 0x0a, 0xfe, 0x14, 0x4d, 0xef, 0xf7, 0xd9, 0x03, 0xed, 0x2d, 0x55, 0x45, 0xe7, 0x7e, 0xbe, 0x66, 0xa3, 0xc5, 0x1f, 0xee, 0x70, 0x16, 0xee, 0xb8, 0xfe, 0xe9, 0xeb, 0x63, 0x0c, 0x0f,
+    0x64, 0x77, 0x4b, 0x27, 0xe7, 0xd5, 0xfe, 0xc8, 0x62, 0xfc, 0x4c, 0x0c, 0x13, 0xac, 0x6b, 0xf0, 0x91, 0x23, 0xb6, 0xf0, 0x5b, 0xb0, 0xe4, 0xb7, 0x5c, 0x97, 0xf3, 0x79, 0xa2, 0xb3, 0xa6, 0x79,
+};
+
+
+// "cn/r"
+const static uint8_t test_output_r[] = {
+    0xf7, 0x59, 0x58, 0x8a, 0xd5, 0x7e, 0x75, 0x84, 0x67, 0x29, 0x54, 0x43, 0xa9, 0xbd, 0x71, 0x49, 0x0a, 0xbf, 0xf8, 0xe9, 0xda, 0xd1, 0xb9, 0x5b, 0x6b, 0xf2, 0xf5, 0xd0, 0xd7, 0x83, 0x87, 0xbc,
+    0x5b, 0xb8, 0x33, 0xde, 0xca, 0x2b, 0xdd, 0x72, 0x52, 0xa9, 0xcc, 0xd7, 0xb4, 0xce, 0x0b, 0x6a, 0x48, 0x54, 0x51, 0x57, 0x94, 0xb5, 0x6c, 0x20, 0x72, 0x62, 0xf7, 0xa5, 0xb9, 0xbd, 0xb5, 0x66,
+    0x1e, 0xe6, 0x72, 0x8d, 0xa6, 0x0f, 0xbd, 0x8d, 0x7d, 0x55, 0xb2, 0xb1, 0xad, 0xe4, 0x87, 0xa3, 0xcf, 0x52, 0xa2, 0xc3, 0xac, 0x6f, 0x52, 0x0d, 0xb1, 0x2c, 0x27, 0xd8, 0x92, 0x1f, 0x6c, 0xab,
+    0x69, 0x69, 0xfe, 0x2d, 0xdf, 0xb7, 0x58, 0x43, 0x8d, 0x48, 0x04, 0x9f, 0x30, 0x2f, 0xc2, 0x10, 0x8a, 0x4f, 0xcc, 0x93, 0xe3, 0x76, 0x69, 0x17, 0x0e, 0x6d, 0xb4, 0xb0, 0xb9, 0xb4, 0xc4, 0xcb,
+    0x7f, 0x30, 0x48, 0xb4, 0xe9, 0x0d, 0x0c, 0xbe, 0x7a, 0x57, 0xc0, 0x39, 0x4f, 0x37, 0x33, 0x8a, 0x01, 0xfa, 0xe3, 0xad, 0xfd, 0xc0, 0xe5, 0x12, 0x6d, 0x86, 0x3a, 0x89, 0x5e, 0xb0, 0x4e, 0x02,
+    0x1d, 0x29, 0x04, 0x43, 0xa4, 0xb5, 0x42, 0xaf, 0x04, 0xa8, 0x2f, 0x6b, 0x24, 0x94, 0xa6, 0xee, 0x7f, 0x20, 0xf2, 0x75, 0x4c, 0x58, 0xe0, 0x84, 0x90, 0x32, 0x48, 0x3a, 0x56, 0xe8, 0xe2, 0xef,
+    0xc4, 0x3c, 0xc6, 0x56, 0x74, 0x36, 0xa8, 0x6a, 0xfb, 0xd6, 0xaa, 0x9e, 0xaa, 0x7c, 0x27, 0x6e, 0x98, 0x06, 0x83, 0x03, 0x34, 0xb6, 0x14, 0xb2, 0xbe, 0xe2, 0x3c, 0xc7, 0x66, 0x34, 0xf6, 0xfd,
+    0x87, 0xbe, 0x24, 0x79, 0xc0, 0xc4, 0xe8, 0xed, 0xfd, 0xfa, 0xa5, 0x60, 0x3e, 0x93, 0xf4, 0x26, 0x5b, 0x3f, 0x82, 0x24, 0xc1, 0xc5, 0x94, 0x6f, 0xeb, 0x42, 0x48, 0x19, 0xd1, 0x89, 0x90, 0xa4,
+    0xdd, 0x9d, 0x6a, 0x6d, 0x8e, 0x47, 0x46, 0x5c, 0xce, 0xac, 0x08, 0x77, 0xef, 0x88, 0x9b, 0x93, 0xe7, 0xeb, 0xa9, 0x79, 0x55, 0x7e, 0x39, 0x35, 0xd7, 0xf8, 0x6d, 0xce, 0x11, 0xb0, 0x70, 0xf3,
+    0x75, 0xc6, 0xf2, 0xae, 0x49, 0xa2, 0x05, 0x21, 0xde, 0x97, 0x28, 0x5b, 0x43, 0x1e, 0x71, 0x71, 0x25, 0x84, 0x7f, 0xb8, 0x93, 0x5e, 0xd8, 0x4a, 0x61, 0xe7, 0xf8, 0xd3, 0x6a, 0x2c, 0x3d, 0x8e,
+};
+
 
 // "cn/0"
 const static uint8_t test_output_v0[160] = {
@@ -199,6 +230,47 @@ const static uint8_t test_output_rto[160] = {
     0xE7, 0x81, 0x4E, 0x2A, 0xBD, 0x62, 0xC1, 0x1B, 0x7C, 0xB9, 0x33, 0x7B, 0xEE, 0x95, 0x80, 0xB3
 };
 
+// "cn/rwz"
+const static uint8_t test_output_rwz[160] = {
+    0x5f, 0x56, 0xc6, 0xb0, 0x99, 0x6b, 0xa2, 0x3e, 0x0b, 0xba, 0x07, 0x29, 0xc9, 0x90, 0x74, 0x85,
+    0x5a, 0x10, 0xe3, 0x08, 0x7f, 0xdb, 0xfe, 0x94, 0x75, 0x33, 0x54, 0x73, 0x76, 0xf0, 0x75, 0xb8,
+    0x8b, 0x70, 0x43, 0x9a, 0xfc, 0xf5, 0xeb, 0x15, 0xbb, 0xf9, 0xad, 0x9d, 0x2a, 0xbd, 0x72, 0x52,
+    0x49, 0x54, 0x0b, 0x91, 0xea, 0x61, 0x7f, 0x98, 0x7d, 0x39, 0x17, 0xb7, 0xd7, 0x65, 0xff, 0x75,
+    0x13, 0x21, 0x1d, 0xce, 0x61, 0x5a, 0xdc, 0x5f, 0x8c, 0xcb, 0x1f, 0x6f, 0xbb, 0x92, 0x88, 0xc3,
+    0xe3, 0xe2, 0xfc, 0x4f, 0x62, 0xfb, 0xf0, 0x48, 0x02, 0x01, 0xd3, 0xbe, 0x77, 0x6a, 0x40, 0xca,
+    0x9a, 0xe9, 0xba, 0x0c, 0xc0, 0x2b, 0x11, 0xf6, 0x9b, 0xee, 0x24, 0x3a, 0xd8, 0x86, 0x18, 0xd0,
+    0xe8, 0xeb, 0xcb, 0x38, 0x2c, 0xf5, 0x99, 0x83, 0x14, 0x7b, 0x0c, 0x20, 0xbe, 0x50, 0xf4, 0x87,
+    0x83, 0x41, 0x75, 0xd8, 0xd1, 0xdd, 0x4b, 0x73, 0xb3, 0x92, 0x8f, 0xe6, 0x1c, 0x72, 0x70, 0xf5,
+    0x7c, 0xf6, 0x23, 0x3a, 0xb4, 0x5f, 0xdf, 0xde, 0xa6, 0x5a, 0x58, 0xec, 0x13, 0x5a, 0x23, 0x2f
+};
+
+// "cn/zls"
+const static uint8_t test_output_zls[160] = {
+    0x51, 0x6E, 0x33, 0xC6, 0xE4, 0x46, 0xAB, 0xBC, 0xCD, 0xAD, 0x18, 0xC0, 0x4C, 0xD9, 0xA2, 0x5E,
+    0x64, 0x10, 0x28, 0x53, 0xB2, 0x0A, 0x42, 0xDF, 0xDE, 0xAA, 0x8B, 0x59, 0x9E, 0xCF, 0x40, 0xE2,
+    0x0D, 0x62, 0x5B, 0x42, 0x18, 0xE2, 0x76, 0xAD, 0xD0, 0x74, 0x90, 0x60, 0x8D, 0xC4, 0xC7, 0x80,
+    0x17, 0xB5, 0x1B, 0x25, 0x31, 0x39, 0x87, 0xD2, 0x2D, 0x6A, 0x9D, 0x1C, 0x74, 0xF4, 0x43, 0x22,
+    0x4B, 0x97, 0x1F, 0x6A, 0xD0, 0xBE, 0x00, 0x74, 0xEC, 0xC5, 0xD8, 0x3B, 0xE6, 0xF4, 0x03, 0x8A,
+    0x7B, 0xBA, 0x80, 0xCC, 0x9F, 0x00, 0xCB, 0xC2, 0x14, 0x8F, 0xF3, 0xD8, 0x92, 0x73, 0xBF, 0x17,
+    0x3D, 0x9B, 0x22, 0xA3, 0x61, 0x94, 0x41, 0x9E, 0xF9, 0x68, 0x1D, 0x42, 0x48, 0x3B, 0x39, 0x45,
+    0xE2, 0xE6, 0x16, 0x84, 0xFC, 0x21, 0xE6, 0xDA, 0x38, 0x7F, 0x17, 0xAB, 0xD3, 0xF2, 0xCE, 0x1A,
+    0x2F, 0x35, 0xD5, 0x74, 0xFA, 0x45, 0x3B, 0x06, 0xD1, 0x4E, 0x84, 0x3A, 0x5D, 0xE3, 0x0E, 0xA5,
+    0x00, 0x08, 0x64, 0xF0, 0xA6, 0xC8, 0x94, 0x45, 0x08, 0xED, 0x03, 0x95, 0x52, 0xE9, 0xBC, 0x5F
+};
+
+// "cn/double"
+const static uint8_t test_output_double[160] = {
+    0xAE, 0xFB, 0xB3, 0xF0, 0xCC, 0x88, 0x04, 0x6D, 0x11, 0x9F, 0x6C, 0x54, 0xB9, 0x6D, 0x90, 0xC9,
+    0xE8, 0x84, 0xEA, 0x3B, 0x59, 0x83, 0xA6, 0x0D, 0x50, 0xA4, 0x2D, 0x7D, 0x3E, 0xBE, 0x48, 0x21,
+    0x49, 0xCE, 0x8E, 0xF3, 0xBC, 0x8A, 0x36, 0xBF, 0x86, 0x37, 0x89, 0x55, 0x09, 0xBA, 0x22, 0xF8,
+    0xEB, 0x3A, 0xE1, 0xDC, 0x91, 0xF7, 0x62, 0x4B, 0x9F, 0x48, 0xE6, 0x92, 0xBD, 0xE4, 0x5D, 0xC1,
+    0xF1, 0x3C, 0x63, 0x1D, 0xEB, 0x0B, 0x04, 0xA3, 0x30, 0xD5, 0x11, 0x15, 0x4C, 0xCE, 0xEF, 0x4F,
+    0xDF, 0x69, 0xE3, 0x9E, 0xD2, 0x68, 0xFC, 0x1B, 0x6F, 0xE8, 0x08, 0x9C, 0xBB, 0xA5, 0x2B, 0x60,
+    0x52, 0x0F, 0xE5, 0xD2, 0xF3, 0x8A, 0xB3, 0xE1, 0x76, 0x7F, 0x44, 0x25, 0x76, 0xEC, 0xFF, 0xA2,
+    0x0C, 0x64, 0xD0, 0x0E, 0x32, 0x33, 0x28, 0x20, 0x73, 0xE0, 0x31, 0x66, 0x4E, 0x54, 0x83, 0x49,
+    0x51, 0x55, 0x4D, 0x2E, 0x22, 0xB7, 0x51, 0x09, 0x73, 0x61, 0x7E, 0x6A, 0x57, 0x0B, 0x28, 0x3C,
+    0x5E, 0x2E, 0xC1, 0x80, 0x89, 0x39, 0xB3, 0x54, 0x39, 0x52, 0x0E, 0x69, 0x3D, 0xF6, 0xC5, 0x4A
+};
 
 #ifndef XMRIG_NO_AEON
 // "cn-lite/0"
@@ -295,8 +367,6 @@ const static uint8_t test_output_pico_trtl[160] = {
 };
 #endif
 
-unsigned char hf_hex2bin(char c, bool &err);
-char hf_bin2hex(unsigned char c);
 
 #ifndef XMRIG_NO_CN_GPU
 // "cn/gpu"

src/crypto/CryptoNight_x86.h

@@ -531,7 +531,7 @@ template<xmrig::Variant VARIANT, xmrig::Variant BASE>
 static inline void cryptonight_monero_tweak(uint64_t* mem_out, const uint8_t* l, uint64_t idx, __m128i ax0, __m128i bx0, __m128i bx1, __m128i& cx)
 {
     if (BASE == xmrig::VARIANT_2) {
-        VARIANT2_SHUFFLE(l, idx, ax0, bx0, bx1, cx);
+        VARIANT2_SHUFFLE(l, idx, ax0, bx0, bx1, cx, (VARIANT == xmrig::VARIANT_RWZ ? 1 : 0));
         _mm_store_si128((__m128i *)mem_out, _mm_xor_si128(bx0, cx));
     } else {
         __m128i tmp = _mm_xor_si128(bx0, cx);
@@ -653,9 +653,9 @@ inline void cryptonight_single_hash(const uint8_t *__restrict__ input, size_t si
 
         if (BASE == xmrig::VARIANT_2) {
             if (VARIANT == xmrig::VARIANT_4) {
-                VARIANT2_SHUFFLE(l0, idx0 & MASK, ax0, bx0, bx1, cx);
+                VARIANT2_SHUFFLE(l0, idx0 & MASK, ax0, bx0, bx1, cx, 0);
             } else {
-                VARIANT2_SHUFFLE2(l0, idx0 & MASK, ax0, bx0, bx1, hi, lo);
+                VARIANT2_SHUFFLE2(l0, idx0 & MASK, ax0, bx0, bx1, hi, lo, (VARIANT == xmrig::VARIANT_RWZ ? 1 : 0));
             }
         }
 
@@ -754,6 +754,8 @@ extern "C" void cnv2_mainloop_ivybridge_asm(cryptonight_ctx *ctx);
 extern "C" void cnv2_mainloop_ryzen_asm(cryptonight_ctx *ctx);
 extern "C" void cnv2_mainloop_bulldozer_asm(cryptonight_ctx *ctx);
 extern "C" void cnv2_double_mainloop_sandybridge_asm(cryptonight_ctx* ctx0, cryptonight_ctx* ctx1);
+extern "C" void cnv2_rwz_mainloop_asm(cryptonight_ctx *ctx);
+extern "C" void cnv2_rwz_double_mainloop_asm(cryptonight_ctx* ctx0, cryptonight_ctx* ctx1);
 
 extern xmrig::CpuThread::cn_mainloop_fun        cn_half_mainloop_ivybridge_asm;
 extern xmrig::CpuThread::cn_mainloop_fun        cn_half_mainloop_ryzen_asm;
@@ -765,6 +767,16 @@ extern xmrig::CpuThread::cn_mainloop_fun        cn_trtl_mainloop_ryzen_asm;
 extern xmrig::CpuThread::cn_mainloop_fun        cn_trtl_mainloop_bulldozer_asm;
 extern xmrig::CpuThread::cn_mainloop_double_fun cn_trtl_double_mainloop_sandybridge_asm;
 
+extern xmrig::CpuThread::cn_mainloop_fun        cn_zls_mainloop_ivybridge_asm;
+extern xmrig::CpuThread::cn_mainloop_fun        cn_zls_mainloop_ryzen_asm;
+extern xmrig::CpuThread::cn_mainloop_fun        cn_zls_mainloop_bulldozer_asm;
+extern xmrig::CpuThread::cn_mainloop_double_fun cn_zls_double_mainloop_sandybridge_asm;
+
+extern xmrig::CpuThread::cn_mainloop_fun        cn_double_mainloop_ivybridge_asm;
+extern xmrig::CpuThread::cn_mainloop_fun        cn_double_mainloop_ryzen_asm;
+extern xmrig::CpuThread::cn_mainloop_fun        cn_double_mainloop_bulldozer_asm;
+extern xmrig::CpuThread::cn_mainloop_double_fun cn_double_double_mainloop_sandybridge_asm;
+
 void wow_compile_code(const V4_Instruction* code, int code_size, void* machine_code, xmrig::Assembly ASM);
 void v4_compile_code(const V4_Instruction* code, int code_size, void* machine_code, xmrig::Assembly ASM);
 void wow_compile_code_double(const V4_Instruction* code, int code_size, void* machine_code, xmrig::Assembly ASM);
@@ -843,6 +855,31 @@ inline void cryptonight_single_hash_asm(const uint8_t *__restrict__ input, size_
             cn_trtl_mainloop_bulldozer_asm(ctx[0]);
         }
     }
+    else if (VARIANT == xmrig::VARIANT_RWZ) {
+        cnv2_rwz_mainloop_asm(ctx[0]);
+    }
+    else if (VARIANT == xmrig::VARIANT_ZLS) {
+        if (ASM == xmrig::ASM_INTEL) {
+            cn_zls_mainloop_ivybridge_asm(ctx[0]);
+        }
+        else if (ASM == xmrig::ASM_RYZEN) {
+            cn_zls_mainloop_ryzen_asm(ctx[0]);
+        }
+        else {
+            cn_zls_mainloop_bulldozer_asm(ctx[0]);
+        }
+    }
+    else if (VARIANT == xmrig::VARIANT_DOUBLE) {
+        if (ASM == xmrig::ASM_INTEL) {
+            cn_double_mainloop_ivybridge_asm(ctx[0]);
+        }
+        else if (ASM == xmrig::ASM_RYZEN) {
+            cn_double_mainloop_ryzen_asm(ctx[0]);
+        }
+        else {
+            cn_double_mainloop_bulldozer_asm(ctx[0]);
+        }
+    }
     else if (xmrig::cn_is_cryptonight_r<VARIANT>()) {
         ctx[0]->generated_code(ctx[0]);
     }
@@ -881,6 +918,15 @@ inline void cryptonight_double_hash_asm(const uint8_t *__restrict__ input, size_
     else if (VARIANT == xmrig::VARIANT_TRTL) {
         cn_trtl_double_mainloop_sandybridge_asm(ctx[0], ctx[1]);
     }
+    else if (VARIANT == xmrig::VARIANT_RWZ) {
+        cnv2_rwz_double_mainloop_asm(ctx[0], ctx[1]);
+    }
+    else if (VARIANT == xmrig::VARIANT_ZLS) {
+        cn_zls_double_mainloop_sandybridge_asm(ctx[0], ctx[1]);
+    }
+    else if (VARIANT == xmrig::VARIANT_DOUBLE) {
+        cn_double_double_mainloop_sandybridge_asm(ctx[0], ctx[1]);
+    }
     else if (xmrig::cn_is_cryptonight_r<VARIANT>()) {
         ctx[0]->generated_code_double(ctx[0], ctx[1]);
     }
@@ -995,9 +1041,9 @@ inline void cryptonight_double_hash(const uint8_t *__restrict__ input, size_t si
 
         if (BASE == xmrig::VARIANT_2) {
             if (VARIANT == xmrig::VARIANT_4) {
-                VARIANT2_SHUFFLE(l0, idx0 & MASK, ax0, bx00, bx01, cx0);
+                VARIANT2_SHUFFLE(l0, idx0 & MASK, ax0, bx00, bx01, cx0, 0);
             } else {
-                VARIANT2_SHUFFLE2(l0, idx0 & MASK, ax0, bx00, bx01, hi, lo);
+                VARIANT2_SHUFFLE2(l0, idx0 & MASK, ax0, bx00, bx01, hi, lo, (VARIANT == xmrig::VARIANT_RWZ ? 1 : 0));
             }
         }
 
@@ -1051,9 +1097,9 @@ inline void cryptonight_double_hash(const uint8_t *__restrict__ input, size_t si
 
         if (BASE == xmrig::VARIANT_2) {
             if (VARIANT == xmrig::VARIANT_4) {
-                VARIANT2_SHUFFLE(l1, idx1 & MASK, ax1, bx10, bx11, cx1);
+                VARIANT2_SHUFFLE(l1, idx1 & MASK, ax1, bx10, bx11, cx1, 0);
             } else {
-                VARIANT2_SHUFFLE2(l1, idx1 & MASK, ax1, bx10, bx11, hi, lo);
+                VARIANT2_SHUFFLE2(l1, idx1 & MASK, ax1, bx10, bx11, hi, lo, (VARIANT == xmrig::VARIANT_RWZ ? 1 : 0));
             }
         }
 
@@ -1155,9 +1201,9 @@ inline void cryptonight_double_hash(const uint8_t *__restrict__ input, size_t si
     lo = __umul128(idx, cl##part, &hi);                 \
     if (BASE == xmrig::VARIANT_2) {                     \
         if (VARIANT == xmrig::VARIANT_4) { \
-            VARIANT2_SHUFFLE(l, idx & MASK, a, b0, b1, c); \
+            VARIANT2_SHUFFLE(l, idx & MASK, a, b0, b1, c, 0); \
         } else {                                        \
-            VARIANT2_SHUFFLE2(l, idx & MASK, a, b0, b1, hi, lo); \
+            VARIANT2_SHUFFLE2(l, idx & MASK, a, b0, b1, hi, lo, (VARIANT == xmrig::VARIANT_RWZ ? 1 : 0)); \
         } \
     }                                                   \
     if (VARIANT == xmrig::VARIANT_4) { \

src/crypto/CryptonightR_gen.cpp

@@ -31,7 +31,6 @@ typedef void(*void_func)();
 #include "crypto/asm/CryptonightR_template.h"
 #include "Mem.h"
 
-#if !defined XMRIG_ARM && !defined XMRIG_NO_ASM
 
 static inline void add_code(uint8_t* &p, void (*p1)(), void (*p2)())
 {
@@ -186,5 +185,3 @@ void v4_soft_aes_compile_code(const V4_Instruction* code, int code_size, void* m
 
     Mem::flushInstructionCache(machine_code, p - p0);
 }
-
-#endif

src/crypto/asm/CryptonightR_template.inc

@@ -70,29 +70,30 @@ FN_PREFIX(CryptonightR_template_mainloop):
 
 	aesenc	xmm5, xmm4
 
-	mov	r12d, r9d
+	mov	r13d, r9d
 	mov	eax, r9d
 	xor	r9d, 48
-	xor	r12d, 16
+	xor	r13d, 16
 	xor	eax, 32
 	movdqu	xmm0, XMMWORD PTR [r9+r11]
 	movaps xmm3, xmm0
-	movdqu	xmm2, XMMWORD PTR [r12+r11]
+	movdqu	xmm2, XMMWORD PTR [r13+r11]
 	movdqu	xmm1, XMMWORD PTR [rax+r11]
 	pxor xmm0, xmm2
 	pxor xmm5, xmm1
 	pxor xmm5, xmm0
-	paddq	xmm3, xmm7
-	paddq	xmm2, xmm6
-	paddq	xmm1, xmm4
-	movdqu	XMMWORD PTR [r12+r11], xmm3
-	movdqu	XMMWORD PTR [rax+r11], xmm2
-	movdqu	XMMWORD PTR [r9+r11], xmm1
 
 	movq	r12, xmm5
 	movd	r10d, xmm5
 	and	r10d, 2097136
 
+	paddq	xmm3, xmm7
+	paddq	xmm2, xmm6
+	paddq	xmm1, xmm4
+	movdqu	XMMWORD PTR [r13+r11], xmm3
+	movdqu	XMMWORD PTR [rax+r11], xmm2
+	movdqu	XMMWORD PTR [r9+r11], xmm1
+
 	movdqa	xmm0, xmm5
 	pxor	xmm0, xmm6
 	movdqu	XMMWORD PTR [rdx], xmm0
@@ -102,14 +103,16 @@ FN_PREFIX(CryptonightR_template_mainloop):
 	shl rdx, 32
 	or	r13, rdx
 
-	xor	r13, QWORD PTR [r10+r11]
-	mov	r14, QWORD PTR [r10+r11+8]
-
 	movd eax, xmm6
 	movd edx, xmm7
 	pextrd r9d, xmm7, 2
 
+	xor	r13, QWORD PTR [r10+r11]
+	mov	r14, QWORD PTR [r10+r11+8]
+
 FN_PREFIX(CryptonightR_template_part2):
+	lea	rcx, [r10+r11]
+
 	mov eax, edi
 	mov edx, ebp
 	shl rdx, 32
@@ -124,6 +127,8 @@ FN_PREFIX(CryptonightR_template_part2):
 
 	mov	rax, r13
 	mul	r12
+	add	r15, rax
+	add	rsp, rdx
 
 	mov	r9d, r10d
 	mov	r12d, r10d
@@ -145,13 +150,10 @@ FN_PREFIX(CryptonightR_template_part2):
 	movdqu	XMMWORD PTR [r10+r11], xmm3
 
 	movdqa	xmm7, xmm6
-	add	r15, rax
+	mov	QWORD PTR [rcx], rsp
-	add	rsp, rdx
-	xor	r10, 48
-	mov	QWORD PTR [r10+r11], rsp
 	xor	rsp, r13
 	mov	r9d, esp
-	mov	QWORD PTR [r10+r11+8], r15
+	mov	QWORD PTR [rcx+8], r15
 	and	r9d, 2097136
 	xor	r15, r14
 	movdqa	xmm6, xmm5

src/crypto/asm/CryptonightR_template_win.inc

@@ -70,29 +70,30 @@ CryptonightR_template_mainloop:
 
 	aesenc	xmm5, xmm4
 
-	mov	r12d, r9d
+	mov	r13d, r9d
 	mov	eax, r9d
 	xor	r9d, 48
-	xor	r12d, 16
+	xor	r13d, 16
 	xor	eax, 32
 	movdqu	xmm0, XMMWORD PTR [r9+r11]
 	movaps xmm3, xmm0
-	movdqu	xmm2, XMMWORD PTR [r12+r11]
+	movdqu	xmm2, XMMWORD PTR [r13+r11]
 	movdqu	xmm1, XMMWORD PTR [rax+r11]
 	pxor xmm0, xmm2
 	pxor xmm5, xmm1
 	pxor xmm5, xmm0
-	paddq	xmm3, xmm7
-	paddq	xmm2, xmm6
-	paddq	xmm1, xmm4
-	movdqu	XMMWORD PTR [r12+r11], xmm3
-	movdqu	XMMWORD PTR [rax+r11], xmm2
-	movdqu	XMMWORD PTR [r9+r11], xmm1
 
 	movq	r12, xmm5
 	movd	r10d, xmm5
 	and	r10d, 2097136
 
+	paddq	xmm3, xmm7
+	paddq	xmm2, xmm6
+	paddq	xmm1, xmm4
+	movdqu	XMMWORD PTR [r13+r11], xmm3
+	movdqu	XMMWORD PTR [rax+r11], xmm2
+	movdqu	XMMWORD PTR [r9+r11], xmm1
+
 	movdqa	xmm0, xmm5
 	pxor	xmm0, xmm6
 	movdqu	XMMWORD PTR [rdx], xmm0
@@ -102,14 +103,16 @@ CryptonightR_template_mainloop:
 	shl rdx, 32
 	or	r13, rdx
 
-	xor	r13, QWORD PTR [r10+r11]
-	mov	r14, QWORD PTR [r10+r11+8]
-
 	movd eax, xmm6
 	movd edx, xmm7
 	pextrd r9d, xmm7, 2
 
+	xor	r13, QWORD PTR [r10+r11]
+	mov	r14, QWORD PTR [r10+r11+8]
+
 CryptonightR_template_part2:
+	lea	rcx, [r10+r11]
+
 	mov eax, edi
 	mov edx, ebp
 	shl rdx, 32
@@ -124,6 +127,8 @@ CryptonightR_template_part2:
 
 	mov	rax, r13
 	mul	r12
+	add	r15, rax
+	add	rsp, rdx
 
 	mov	r9d, r10d
 	mov	r12d, r10d
@@ -145,13 +150,10 @@ CryptonightR_template_part2:
 	movdqu	XMMWORD PTR [r10+r11], xmm3
 
 	movdqa	xmm7, xmm6
-	add	r15, rax
+	mov	QWORD PTR [rcx], rsp
-	add	rsp, rdx
-	xor	r10, 48
-	mov	QWORD PTR [r10+r11], rsp
 	xor	rsp, r13
 	mov	r9d, esp
-	mov	QWORD PTR [r10+r11+8], r15
+	mov	QWORD PTR [rcx+8], r15
 	and	r9d, 2097136
 	xor	r15, r14
 	movdqa	xmm6, xmm5

src/crypto/asm/cn2/cnv2_rwz_double_main_loop.inc

@@ -0,0 +1,410 @@
+	mov	rax, rsp
+	push	rbx
+	push	rbp
+	push	rsi
+	push	rdi
+	push	r12
+	push	r13
+	push	r14
+	push	r15
+	sub	rsp, 184
+
+	stmxcsr DWORD PTR [rsp+272]
+	mov DWORD PTR [rsp+276], 24448
+	ldmxcsr DWORD PTR [rsp+276]
+
+	mov	r13, QWORD PTR [rcx+224]
+	mov	r9, rdx
+	mov	r10, QWORD PTR [rcx+32]
+	mov	r8, rcx
+	xor	r10, QWORD PTR [rcx]
+	mov	r14d, 393216
+	mov	r11, QWORD PTR [rcx+40]
+	xor	r11, QWORD PTR [rcx+8]
+	mov	rsi, QWORD PTR [rdx+224]
+	mov	rdx, QWORD PTR [rcx+56]
+	xor	rdx, QWORD PTR [rcx+24]
+	mov	rdi, QWORD PTR [r9+32]
+	xor	rdi, QWORD PTR [r9]
+	mov	rbp, QWORD PTR [r9+40]
+	xor	rbp, QWORD PTR [r9+8]
+	movq	xmm0, rdx
+	movaps	XMMWORD PTR [rax-88], xmm6
+	movaps	XMMWORD PTR [rax-104], xmm7
+	movaps	XMMWORD PTR [rax-120], xmm8
+	movaps	XMMWORD PTR [rsp+112], xmm9
+	movaps	XMMWORD PTR [rsp+96], xmm10
+	movaps	XMMWORD PTR [rsp+80], xmm11
+	movaps	XMMWORD PTR [rsp+64], xmm12
+	movaps	XMMWORD PTR [rsp+48], xmm13
+	movaps	XMMWORD PTR [rsp+32], xmm14
+	movaps	XMMWORD PTR [rsp+16], xmm15
+	mov	rdx, r10
+	movq	xmm4, QWORD PTR [r8+96]
+	and	edx, 2097136
+	mov	rax, QWORD PTR [rcx+48]
+	xorps	xmm13, xmm13
+	xor	rax, QWORD PTR [rcx+16]
+	mov	rcx, QWORD PTR [rcx+88]
+	xor	rcx, QWORD PTR [r8+72]
+	movq	xmm5, QWORD PTR [r8+104]
+	movq	xmm7, rax
+
+	mov eax, 1
+	shl rax, 52
+	movq xmm14, rax
+	punpcklqdq xmm14, xmm14
+
+	mov eax, 1023
+	shl rax, 52
+	movq xmm12, rax
+	punpcklqdq xmm12, xmm12
+
+	mov	rax, QWORD PTR [r8+80]
+	xor	rax, QWORD PTR [r8+64]
+	punpcklqdq xmm7, xmm0
+	movq	xmm0, rcx
+	mov	rcx, QWORD PTR [r9+56]
+	xor	rcx, QWORD PTR [r9+24]
+	movq	xmm3, rax
+	mov	rax, QWORD PTR [r9+48]
+	xor	rax, QWORD PTR [r9+16]
+	punpcklqdq xmm3, xmm0
+	movq	xmm0, rcx
+	mov	QWORD PTR [rsp], r13
+	mov	rcx, QWORD PTR [r9+88]
+	xor	rcx, QWORD PTR [r9+72]
+	movq	xmm6, rax
+	mov	rax, QWORD PTR [r9+80]
+	xor	rax, QWORD PTR [r9+64]
+	punpcklqdq xmm6, xmm0
+	movq	xmm0, rcx
+	mov	QWORD PTR [rsp+256], r10
+	mov	rcx, rdi
+	mov	QWORD PTR [rsp+264], r11
+	movq	xmm8, rax
+	and	ecx, 2097136
+	punpcklqdq xmm8, xmm0
+	movq	xmm0, QWORD PTR [r9+96]
+	punpcklqdq xmm4, xmm0
+	movq	xmm0, QWORD PTR [r9+104]
+	lea	r8, QWORD PTR [rcx+rsi]
+	movdqu	xmm11, XMMWORD PTR [r8]
+	punpcklqdq xmm5, xmm0
+	lea	r9, QWORD PTR [rdx+r13]
+	movdqu	xmm15, XMMWORD PTR [r9]
+
+	ALIGN(64)
+rwz_main_loop_double:
+	movdqu	xmm9, xmm15
+	mov eax, edx
+	mov ebx, edx
+	xor eax, 16
+	xor ebx, 32
+	xor edx, 48
+
+	movq	xmm0, r11
+	movq	xmm2, r10
+	punpcklqdq xmm2, xmm0
+	aesenc	xmm9, xmm2
+
+	movdqu	xmm0, XMMWORD PTR [rdx+r13]
+	movdqu	xmm1, XMMWORD PTR [rbx+r13]
+	paddq	xmm0, xmm7
+	paddq	xmm1, xmm2
+	movdqu	XMMWORD PTR [rbx+r13], xmm0
+	movdqu	xmm0, XMMWORD PTR [rax+r13]
+	movdqu	XMMWORD PTR [rdx+r13], xmm1
+	paddq	xmm0, xmm3
+	movdqu	XMMWORD PTR [rax+r13], xmm0
+
+	movq	r11, xmm9
+	mov	edx, r11d
+	and	edx, 2097136
+	movdqa	xmm0, xmm9
+	pxor	xmm0, xmm7
+	movdqu	XMMWORD PTR [r9], xmm0
+
+	lea	rbx, QWORD PTR [rdx+r13]
+	mov	r10, QWORD PTR [rdx+r13]
+
+	movdqu	xmm10, xmm11
+	movq	xmm0, rbp
+	movq	xmm11, rdi
+	punpcklqdq xmm11, xmm0
+	aesenc	xmm10, xmm11
+
+	mov eax, ecx
+	mov r12d, ecx
+	xor eax, 16
+	xor r12d, 32
+	xor ecx, 48
+
+	movdqu	xmm0, XMMWORD PTR [rcx+rsi]
+	paddq	xmm0, xmm6
+	movdqu	xmm1, XMMWORD PTR [r12+rsi]
+	movdqu	XMMWORD PTR [r12+rsi], xmm0
+	paddq	xmm1, xmm11
+	movdqu	xmm0, XMMWORD PTR [rax+rsi]
+	movdqu	XMMWORD PTR [rcx+rsi], xmm1
+	paddq	xmm0, xmm8
+	movdqu	XMMWORD PTR [rax+rsi], xmm0
+
+	movq	rcx, xmm10
+	and	ecx, 2097136
+
+	movdqa	xmm0, xmm10
+	pxor	xmm0, xmm6
+	movdqu	XMMWORD PTR [r8], xmm0
+	mov r12, QWORD PTR [rcx+rsi]
+
+	mov	r9, QWORD PTR [rbx+8]
+
+	xor edx, 16
+	mov r8d, edx
+	mov r15d, edx
+
+	movq	rdx, xmm5
+	shl	rdx, 32
+	movq	rax, xmm4
+	xor	rdx, rax
+	xor	r10, rdx
+	mov	rax, r10
+	mul	r11
+	mov r11d, r8d
+	xor r11d, 48
+	movq xmm0, rdx
+	xor rdx, [r11+r13]
+	movq xmm1, rax
+	xor rax, [r11+r13+8]
+	punpcklqdq xmm0, xmm1
+
+	pxor xmm0, XMMWORD PTR [r8+r13]
+	movdqu	xmm1, XMMWORD PTR [r11+r13]
+	paddq	xmm0, xmm3
+	paddq	xmm1, xmm2
+	movdqu	XMMWORD PTR [r8+r13], xmm0
+	xor	r8d, 32
+	movdqu	xmm0, XMMWORD PTR [r8+r13]
+	movdqu	XMMWORD PTR [r8+r13], xmm1
+	paddq	xmm0, xmm7
+	movdqu	XMMWORD PTR [r11+r13], xmm0
+
+	mov	r11, QWORD PTR [rsp+256]
+	add	r11, rdx
+	mov	rdx, QWORD PTR [rsp+264]
+	add	rdx, rax
+	mov	QWORD PTR [rbx], r11
+	xor	r11, r10
+	mov	QWORD PTR [rbx+8], rdx
+	xor	rdx, r9
+	mov	QWORD PTR [rsp+256], r11
+	and	r11d, 2097136
+	mov	QWORD PTR [rsp+264], rdx
+	mov	QWORD PTR [rsp+8], r11
+	lea	r15, QWORD PTR [r11+r13]
+	movdqu xmm15, XMMWORD PTR [r11+r13]
+	lea	r13, QWORD PTR [rsi+rcx]
+	movdqa	xmm0, xmm5
+	psrldq	xmm0, 8
+	movaps	xmm2, xmm13
+	movq	r10, xmm0
+	psllq	xmm5, 1
+	shl	r10, 32
+	movdqa	xmm0, xmm9
+	psrldq	xmm0, 8
+	movdqa	xmm1, xmm10
+	movq	r11, xmm0
+	psrldq	xmm1, 8
+	movq	r8, xmm1
+	psrldq	xmm4, 8
+	movaps	xmm0, xmm13
+	movq	rax, xmm4
+	xor	r10, rax
+	movaps	xmm1, xmm13
+	xor	r10, r12
+	lea	rax, QWORD PTR [r11+1]
+	shr	rax, 1
+	movdqa	xmm3, xmm9
+	punpcklqdq xmm3, xmm10
+	paddq	xmm5, xmm3
+	movq	rdx, xmm5
+	psrldq	xmm5, 8
+	cvtsi2sd xmm2, rax
+	or	edx, -2147483647
+	lea	rax, QWORD PTR [r8+1]
+	shr	rax, 1
+	movq	r9, xmm5
+	cvtsi2sd xmm0, rax
+	or	r9d, -2147483647
+	cvtsi2sd xmm1, rdx
+	unpcklpd xmm2, xmm0
+	movaps	xmm0, xmm13
+	cvtsi2sd xmm0, r9
+	unpcklpd xmm1, xmm0
+	divpd	xmm2, xmm1
+	paddq	xmm2, xmm14
+	cvttsd2si rax, xmm2
+	psrldq	xmm2, 8
+	mov	rbx, rax
+	imul	rax, rdx
+	sub	r11, rax
+	js	rwz_div_fix_1
+rwz_div_fix_1_ret:
+
+	cvttsd2si rdx, xmm2
+	mov	rax, rdx
+	imul	rax, r9
+	movd	xmm2, r11d
+	movd	xmm4, ebx
+	sub	r8, rax
+	js	rwz_div_fix_2
+rwz_div_fix_2_ret:
+
+	movd	xmm1, r8d
+	movd	xmm0, edx
+	punpckldq xmm2, xmm1
+	punpckldq xmm4, xmm0
+	punpckldq xmm4, xmm2
+	paddq	xmm3, xmm4
+	movdqa	xmm0, xmm3
+	psrlq	xmm0, 12
+	paddq	xmm0, xmm12
+	sqrtpd	xmm1, xmm0
+	movq	r9, xmm1
+	movdqa xmm5, xmm1
+	psrlq xmm5, 19
+	test	r9, 524287
+	je	rwz_sqrt_fix_1
+rwz_sqrt_fix_1_ret:
+
+	movq r9, xmm10
+	psrldq	xmm1, 8
+	movq	r8, xmm1
+	test	r8, 524287
+	je	rwz_sqrt_fix_2
+rwz_sqrt_fix_2_ret:
+
+	mov r12d, ecx
+	mov r8d, ecx
+	xor r12d, 16
+	xor r8d, 32
+	xor ecx, 48
+	mov	rax, r10
+	mul	r9
+	movq xmm0, rax
+	movq xmm3, rdx
+	punpcklqdq xmm3, xmm0
+
+	movdqu	xmm0, XMMWORD PTR [r12+rsi]
+	pxor xmm0, xmm3
+	movdqu	xmm1, XMMWORD PTR [r8+rsi]
+	xor rdx, [r8+rsi]
+	xor rax, [r8+rsi+8]
+	movdqu	xmm3, XMMWORD PTR [rcx+rsi]
+	paddq	xmm3, xmm6
+	paddq	xmm1, xmm11
+	paddq	xmm0, xmm8
+	movdqu	XMMWORD PTR [r8+rsi], xmm3
+	movdqu	XMMWORD PTR [rcx+rsi], xmm1
+	movdqu	XMMWORD PTR [r12+rsi], xmm0
+
+	add	rdi, rdx
+	mov	QWORD PTR [r13], rdi
+	xor	rdi, r10
+	mov	ecx, edi
+	and	ecx, 2097136
+	lea	r8, QWORD PTR [rcx+rsi]
+
+	mov rdx, QWORD PTR [r13+8]	
+	add	rbp, rax
+	mov	QWORD PTR [r13+8], rbp
+	movdqu xmm11, XMMWORD PTR [rcx+rsi]
+	xor	rbp, rdx
+	mov	r13, QWORD PTR [rsp]
+	movdqa	xmm3, xmm7
+	mov	rdx, QWORD PTR [rsp+8]
+	movdqa	xmm8, xmm6
+	mov	r10, QWORD PTR [rsp+256]
+	movdqa	xmm7, xmm9
+	mov	r11, QWORD PTR [rsp+264]
+	movdqa	xmm6, xmm10
+	mov	r9, r15
+	dec r14d
+	jne	rwz_main_loop_double
+
+	ldmxcsr DWORD PTR [rsp+272]
+	movaps	xmm13, XMMWORD PTR [rsp+48]
+	lea	r11, QWORD PTR [rsp+184]
+	movaps	xmm6, XMMWORD PTR [r11-24]
+	movaps	xmm7, XMMWORD PTR [r11-40]
+	movaps	xmm8, XMMWORD PTR [r11-56]
+	movaps	xmm9, XMMWORD PTR [r11-72]
+	movaps	xmm10, XMMWORD PTR [r11-88]
+	movaps	xmm11, XMMWORD PTR [r11-104]
+	movaps	xmm12, XMMWORD PTR [r11-120]
+	movaps	xmm14, XMMWORD PTR [rsp+32]
+	movaps	xmm15, XMMWORD PTR [rsp+16]
+	mov	rsp, r11
+	pop	r15
+	pop	r14
+	pop	r13
+	pop	r12
+	pop	rdi
+	pop	rsi
+	pop	rbp
+	pop	rbx
+	jmp rwz_cnv2_double_mainloop_asm_endp
+
+rwz_div_fix_1:
+	dec	rbx
+	add	r11, rdx
+	jmp	rwz_div_fix_1_ret
+
+rwz_div_fix_2:
+	dec	rdx
+	add	r8, r9
+	jmp	rwz_div_fix_2_ret
+
+rwz_sqrt_fix_1:
+	movq	r8, xmm3
+	movdqa xmm0, xmm5
+	psrldq xmm0, 8
+	dec	r9
+	mov r11d, -1022
+	shl r11, 32
+	mov	rax, r9
+	shr	r9, 19
+	shr	rax, 20
+	mov	rdx, r9
+	sub	rdx, rax
+	lea	rdx, [rdx+r11+1]
+	add	rax, r11
+	imul	rdx, rax
+	sub	rdx, r8
+	adc	r9, 0
+	movq xmm5, r9
+	punpcklqdq xmm5, xmm0
+	jmp	rwz_sqrt_fix_1_ret
+
+rwz_sqrt_fix_2:
+	psrldq	xmm3, 8
+	movq	r11, xmm3
+	dec	r8
+	mov ebx, -1022
+	shl rbx, 32
+	mov	rax, r8
+	shr	r8, 19
+	shr	rax, 20
+	mov	rdx, r8
+	sub	rdx, rax
+	lea	rdx, [rdx+rbx+1]
+	add	rax, rbx
+	imul	rdx, rax
+	sub	rdx, r11
+	adc	r8, 0
+	movq xmm0, r8
+	punpcklqdq xmm5, xmm0
+	jmp	rwz_sqrt_fix_2_ret
+
+rwz_cnv2_double_mainloop_asm_endp:

src/crypto/asm/cn2/cnv2_rwz_main_loop.inc

@@ -0,0 +1,186 @@
+	mov	 QWORD PTR [rsp+24], rbx
+	push	 rbp
+	push	 rsi
+	push	 rdi
+	push	 r12
+	push	 r13
+	push	 r14
+	push	 r15
+	sub	 rsp, 80
+
+	stmxcsr DWORD PTR [rsp]
+	mov DWORD PTR [rsp+4], 24448
+	ldmxcsr DWORD PTR [rsp+4]
+
+	mov	 rax, QWORD PTR [rcx+48]
+	mov	 r9, rcx
+	xor	 rax, QWORD PTR [rcx+16]
+	mov	 esi, 393216
+	mov	 r8, QWORD PTR [rcx+32]
+	mov	 r13d, -2147483647
+	xor	 r8, QWORD PTR [rcx]
+	mov	 r11, QWORD PTR [rcx+40]
+	mov	 r10, r8
+	mov	 rdx, QWORD PTR [rcx+56]
+	movq	 xmm4, rax
+	xor	 rdx, QWORD PTR [rcx+24]
+	xor	 r11, QWORD PTR [rcx+8]
+	mov	 rbx, QWORD PTR [rcx+224]
+	mov	 rax, QWORD PTR [r9+80]
+	xor	 rax, QWORD PTR [r9+64]
+	movq	 xmm0, rdx
+	mov	 rcx, QWORD PTR [rcx+88]
+	xor	 rcx, QWORD PTR [r9+72]
+	movq	 xmm3, QWORD PTR [r9+104]
+	movaps	 XMMWORD PTR [rsp+64], xmm6
+	movaps	 XMMWORD PTR [rsp+48], xmm7
+	movaps	 XMMWORD PTR [rsp+32], xmm8
+	and	 r10d, 2097136
+	movq	 xmm5, rax
+
+	xor eax, eax
+	mov QWORD PTR [rsp+16], rax
+
+	mov ax, 1023
+	shl rax, 52
+	movq xmm8, rax
+	mov r15, QWORD PTR [r9+96]
+	punpcklqdq xmm4, xmm0
+	movq	 xmm0, rcx
+	punpcklqdq xmm5, xmm0
+	movdqu	 xmm6, XMMWORD PTR [r10+rbx]
+
+	ALIGN(64)
+rwz_main_loop:
+	lea	 rdx, QWORD PTR [r10+rbx]
+	mov	 ecx, r10d
+	mov	 eax, r10d
+	mov rdi, r15
+	xor	 ecx, 16
+	xor	 eax, 32
+	xor	 r10d, 48
+	movq	 xmm0, r11
+	movq	 xmm7, r8
+	punpcklqdq xmm7, xmm0
+	aesenc	 xmm6, xmm7
+	movq	 rbp, xmm6
+	mov	 r9, rbp
+	and	 r9d, 2097136
+	movdqu	 xmm0, XMMWORD PTR [rcx+rbx]
+	movdqu	 xmm1, XMMWORD PTR [rax+rbx]
+	movdqu	 xmm2, XMMWORD PTR [r10+rbx]
+	paddq	 xmm0, xmm5
+	paddq	 xmm1, xmm7
+	paddq	 xmm2, xmm4
+	movdqu	 XMMWORD PTR [rcx+rbx], xmm0
+	movdqu	 XMMWORD PTR [rax+rbx], xmm2
+	movdqu	 XMMWORD PTR [r10+rbx], xmm1
+	mov r10, r9
+	xor r10d, 32
+	movq	 rcx, xmm3
+	mov	 rax, rcx
+	shl	 rax, 32
+	xor	 rdi, rax
+	movdqa	 xmm0, xmm6
+	pxor	 xmm0, xmm4
+	movdqu	 XMMWORD PTR [rdx], xmm0
+	xor	 rdi, QWORD PTR [r9+rbx]
+	lea	 r14, QWORD PTR [r9+rbx]
+	mov	 r12, QWORD PTR [r14+8]
+	xor	 edx, edx
+	lea	 r9d, DWORD PTR [ecx+ecx]
+	add	 r9d, ebp
+	movdqa	 xmm0, xmm6
+	psrldq	 xmm0, 8
+	or	 r9d, r13d
+	movq	 rax, xmm0
+	div	 r9
+	xorps xmm3, xmm3
+	mov	 eax, eax
+	shl	 rdx, 32
+	add	 rdx, rax
+	lea	 r9, QWORD PTR [rdx+rbp]
+	mov r15, rdx
+	mov	 rax, r9
+	shr	 rax, 12
+	movq	 xmm0, rax
+	paddq	 xmm0, xmm8
+	sqrtsd	 xmm3, xmm0
+	psubq	 xmm3, XMMWORD PTR [rsp+16]
+	movq	 rdx, xmm3
+	test	 edx, 524287
+	je	 rwz_sqrt_fixup
+	psrlq	 xmm3, 19
+rwz_sqrt_fixup_ret:
+
+	mov	 ecx, r10d
+	mov	 rax, rdi
+	mul	 rbp
+	movq xmm2, rdx
+	xor rdx, [rcx+rbx]
+	add	 r8, rdx
+	mov	 QWORD PTR [r14], r8
+	xor	 r8, rdi
+	mov edi, r8d
+	and edi, 2097136
+	movq xmm0, rax
+	xor rax, [rcx+rbx+8]
+	add	 r11, rax
+	mov	 QWORD PTR [r14+8], r11
+	punpcklqdq xmm2, xmm0
+
+	mov	 r9d, r10d
+	xor	 r9d, 48
+	xor	 r10d, 16
+	pxor	 xmm2, XMMWORD PTR [r9+rbx]
+	movdqu	 xmm0, XMMWORD PTR [r10+rbx]
+	paddq	 xmm0, xmm4
+	movdqu	 xmm1, XMMWORD PTR [rcx+rbx]
+	paddq	 xmm2, xmm5
+	paddq	 xmm1, xmm7
+	movdqa	 xmm5, xmm4
+	movdqu	 XMMWORD PTR [r9+rbx], xmm2
+	movdqa	 xmm4, xmm6
+	movdqu	 XMMWORD PTR [rcx+rbx], xmm0
+	movdqu	 XMMWORD PTR [r10+rbx], xmm1
+	movdqu xmm6, [rdi+rbx]
+	mov	 r10d, edi
+	xor	 r11, r12
+	dec rsi
+	jne	 rwz_main_loop
+
+	ldmxcsr DWORD PTR [rsp]
+	mov	 rbx, QWORD PTR [rsp+160]
+	movaps	 xmm6, XMMWORD PTR [rsp+64]
+	movaps	 xmm7, XMMWORD PTR [rsp+48]
+	movaps	 xmm8, XMMWORD PTR [rsp+32]
+	add	 rsp, 80
+	pop	 r15
+	pop	 r14
+	pop	 r13
+	pop	 r12
+	pop	 rdi
+	pop	 rsi
+	pop	 rbp
+	jmp cnv2_rwz_main_loop_endp
+
+rwz_sqrt_fixup:
+	dec	 rdx
+	mov r13d, -1022
+	shl r13, 32
+	mov	 rax, rdx
+	shr	 rdx, 19
+	shr	 rax, 20
+	mov	 rcx, rdx
+	sub	 rcx, rax
+	add	 rax, r13
+	not r13
+	sub	 rcx, r13
+	mov	 r13d, -2147483647
+	imul	 rcx, rax
+	sub	 rcx, r9
+	adc	 rdx, 0
+	movq	 xmm3, rdx
+	jmp	 rwz_sqrt_fixup_ret
+
+cnv2_rwz_main_loop_endp:

src/crypto/asm/cn_main_loop.S

@@ -15,6 +15,8 @@
 .global FN_PREFIX(cnv2_mainloop_ryzen_asm)
 .global FN_PREFIX(cnv2_mainloop_bulldozer_asm)
 .global FN_PREFIX(cnv2_double_mainloop_sandybridge_asm)
+.global FN_PREFIX(cnv2_rwz_mainloop_asm)
+.global FN_PREFIX(cnv2_rwz_double_mainloop_asm)
 
 ALIGN(64)
 FN_PREFIX(cnv2_mainloop_ivybridge_asm):
@@ -52,3 +54,21 @@ FN_PREFIX(cnv2_double_mainloop_sandybridge_asm):
 	add rsp, 48
 	ret 0
 	mov eax, 3735929054
+
+ALIGN(64)
+FN_PREFIX(cnv2_rwz_mainloop_asm):
+	sub rsp, 48
+	mov rcx, rdi
+	#include "cn2/cnv2_rwz_main_loop.inc"
+	add rsp, 48
+	ret 0
+	mov eax, 3735929054
+
+ALIGN(64)
+FN_PREFIX(cnv2_rwz_double_mainloop_asm):
+	sub rsp, 48
+	mov rcx, rdi
+	#include "cn2/cnv2_rwz_double_main_loop.inc"
+	add rsp, 48
+	ret 0
+	mov eax, 3735929054

src/crypto/asm/cn_main_loop.asm

@@ -3,6 +3,8 @@ PUBLIC cnv2_mainloop_ivybridge_asm
 PUBLIC cnv2_mainloop_ryzen_asm
 PUBLIC cnv2_mainloop_bulldozer_asm
 PUBLIC cnv2_double_mainloop_sandybridge_asm
+PUBLIC cnv2_rwz_mainloop_asm
+PUBLIC cnv2_rwz_double_mainloop_asm
 
 ALIGN(64)
 cnv2_mainloop_ivybridge_asm PROC
@@ -32,5 +34,19 @@ cnv2_double_mainloop_sandybridge_asm PROC
 	mov eax, 3735929054
 cnv2_double_mainloop_sandybridge_asm ENDP
 
+ALIGN(64)
+cnv2_rwz_mainloop_asm PROC
+	INCLUDE cn2/cnv2_rwz_main_loop.inc
+	ret 0
+	mov eax, 3735929054
+cnv2_rwz_mainloop_asm ENDP
+
+ALIGN(64)
+cnv2_rwz_double_mainloop_asm PROC
+	INCLUDE cn2/cnv2_rwz_double_main_loop.inc
+	ret 0
+	mov eax, 3735929054
+cnv2_rwz_double_mainloop_asm ENDP
+
 _TEXT_CNV2_MAINLOOP ENDS
 END

src/crypto/asm/win64/CryptonightR_soft_aes_template.inc

@@ -1,279 +0,0 @@
-PUBLIC FN_PREFIX(CryptonightR_soft_aes_template_part1)
-PUBLIC FN_PREFIX(CryptonightR_soft_aes_template_mainloop)
-PUBLIC FN_PREFIX(CryptonightR_soft_aes_template_part2)
-PUBLIC FN_PREFIX(CryptonightR_soft_aes_template_part3)
-PUBLIC FN_PREFIX(CryptonightR_soft_aes_template_end)
-
-ALIGN(64)
-FN_PREFIX(CryptonightR_soft_aes_template_part1):
-	mov	QWORD PTR [rsp+8], rcx
-	push	rbx
-	push	rbp
-	push	rsi
-	push	rdi
-	push	r12
-	push	r13
-	push	r14
-	push	r15
-	sub	rsp, 232
-
-	mov	eax, [rcx+96]
-	mov	ebx, [rcx+100]
-	mov	esi, [rcx+104]
-	mov	edx, [rcx+108]
-	mov [rsp+144], eax
-	mov [rsp+148], ebx
-	mov [rsp+152], esi
-	mov [rsp+156], edx
-
-	mov	rax, QWORD PTR [rcx+48]
-	mov	r10, rcx
-	xor	rax, QWORD PTR [rcx+16]
-	mov	r8, QWORD PTR [rcx+32]
-	xor	r8, QWORD PTR [rcx]
-	mov	r9, QWORD PTR [rcx+40]
-	xor	r9, QWORD PTR [rcx+8]
-	movd	xmm4, rax
-	mov	rdx, QWORD PTR [rcx+56]
-	xor	rdx, QWORD PTR [rcx+24]
-	mov	r11, QWORD PTR [rcx+224]
-	mov	rcx, QWORD PTR [rcx+88]
-	xor	rcx, QWORD PTR [r10+72]
-	mov	rax, QWORD PTR [r10+80]
-	movd	xmm0, rdx
-	xor	rax, QWORD PTR [r10+64]
-
-	movaps	XMMWORD PTR [rsp+16], xmm6
-	movaps	XMMWORD PTR [rsp+32], xmm7
-	movaps	XMMWORD PTR [rsp+48], xmm8
-	movaps	XMMWORD PTR [rsp+64], xmm9
-	movaps	XMMWORD PTR [rsp+80], xmm10
-	movaps	XMMWORD PTR [rsp+96], xmm11
-	movaps	XMMWORD PTR [rsp+112], xmm12
-	movaps	XMMWORD PTR [rsp+128], xmm13
-
-	movd	xmm5, rax
-
-	mov	rax, r8
-	punpcklqdq xmm4, xmm0
-	and	eax, 2097136
-	movd	xmm10, QWORD PTR [r10+96]
-	movd	xmm0, rcx
-	mov	rcx, QWORD PTR [r10+104]
-	xorps	xmm9, xmm9
-	mov	QWORD PTR [rsp+328], rax
-	movd	xmm12, r11
-	mov	QWORD PTR [rsp+320], r9
-	punpcklqdq xmm5, xmm0
-	movd xmm13, rcx
-	mov r12d, 524288
-
-	ALIGN(64)
-FN_PREFIX(CryptonightR_soft_aes_template_mainloop):
-	movd xmm11, r12d
-	mov	r12, QWORD PTR [r10+272]
-	lea	r13, QWORD PTR [rax+r11]
-	mov	esi, DWORD PTR [r13]
-	movd	xmm0, r9
-	mov	r10d, DWORD PTR [r13+4]
-	movd	xmm7, r8
-	mov	ebp, DWORD PTR [r13+12]
-	mov	r14d, DWORD PTR [r13+8]
-	mov	rdx, QWORD PTR [rsp+328]
-	movzx	ecx, sil
-	shr	esi, 8
-	punpcklqdq xmm7, xmm0
-	mov	r15d, DWORD PTR [r12+rcx*4]
-	movzx	ecx, r10b
-	shr	r10d, 8
-	mov	edi, DWORD PTR [r12+rcx*4]
-	movzx	ecx, r14b
-	shr	r14d, 8
-	mov	ebx, DWORD PTR [r12+rcx*4]
-	movzx	ecx, bpl
-	shr	ebp, 8
-	mov	r9d, DWORD PTR [r12+rcx*4]
-	movzx	ecx, r10b
-	shr	r10d, 8
-	xor	r15d, DWORD PTR [r12+rcx*4+1024]
-	movzx	ecx, r14b
-	shr	r14d, 8
-	mov	eax, r14d
-	shr	eax, 8
-	xor	edi, DWORD PTR [r12+rcx*4+1024]
-	add	eax, 256
-	movzx	ecx, bpl
-	shr	ebp, 8
-	xor	ebx, DWORD PTR [r12+rcx*4+1024]
-	movzx	ecx, sil
-	shr	esi, 8
-	xor	r9d, DWORD PTR [r12+rcx*4+1024]
-	add	r12, 2048
-	movzx	ecx, r10b
-	shr	r10d, 8
-	add	r10d, 256
-	mov	r11d, DWORD PTR [r12+rax*4]
-	xor	r11d, DWORD PTR [r12+rcx*4]
-	xor	r11d, r9d
-	movzx	ecx, sil
-	mov	r10d, DWORD PTR [r12+r10*4]
-	shr	esi, 8
-	add	esi, 256
-	xor	r10d, DWORD PTR [r12+rcx*4]
-	movzx	ecx, bpl
-	xor	r10d, ebx
-	shr	ebp, 8
-	movd	xmm1, r11d
-	add	ebp, 256
-	movd	r11, xmm12
-	mov	r9d, DWORD PTR [r12+rcx*4]
-	xor	r9d, DWORD PTR [r12+rsi*4]
-	mov	eax, DWORD PTR [r12+rbp*4]
-	xor	r9d, edi
-	movzx	ecx, r14b
-	movd	xmm0, r10d
-	movd	xmm2, r9d
-	xor	eax, DWORD PTR [r12+rcx*4]
-	mov	rcx, rdx
-	xor	eax, r15d
-	punpckldq xmm2, xmm1
-	xor	rcx, 16
-	movd	xmm6, eax
-	mov	rax, rdx
-	punpckldq xmm6, xmm0
-	xor	rax, 32
-	punpckldq xmm6, xmm2
-	xor	rdx, 48
-	movdqu	xmm2, XMMWORD PTR [rcx+r11]
-	pxor xmm6, xmm2
-	pxor	xmm6, xmm7
-	paddq	xmm2, xmm4
-	movdqu	xmm1, XMMWORD PTR [rax+r11]
-	movdqu	xmm0, XMMWORD PTR [rdx+r11]
-	pxor xmm6, xmm1
-	pxor xmm6, xmm0
-	paddq	xmm0, xmm5
-	movdqu	XMMWORD PTR [rcx+r11], xmm0
-	movdqu	XMMWORD PTR [rax+r11], xmm2
-	movd rcx, xmm13
-	paddq	xmm1, xmm7
-	movdqu	XMMWORD PTR [rdx+r11], xmm1
-	movd	rdi, xmm6
-	mov	r10, rdi
-	and	r10d, 2097136
-	movdqa	xmm0, xmm6
-	pxor	xmm0, xmm4
-	movdqu	XMMWORD PTR [r13], xmm0
-
-	mov ebx, [rsp+144]
-	mov ebp, [rsp+152]
-	add ebx, [rsp+148]
-	add ebp, [rsp+156]
-	shl rbp, 32
-	or rbx, rbp
-
-	xor rbx, QWORD PTR [r10+r11]
-	lea	r14, QWORD PTR [r10+r11]
-	mov	rbp, QWORD PTR [r14+8]
-
-	mov [rsp+160], rbx
-	mov [rsp+168], rdi
-	mov [rsp+176], rbp
-	mov [rsp+184], r10
-	mov r10, rsp
-
-	mov ebx, [rsp+144]
-	mov esi, [rsp+148]
-	mov edi, [rsp+152]
-	mov ebp, [rsp+156]
-
-	movd esp, xmm7
-	movaps xmm0, xmm7
-	psrldq xmm0, 8
-	movd r15d, xmm0
-	movd eax, xmm4
-	movd edx, xmm5
-	movaps xmm0, xmm5
-	psrldq xmm0, 8
-	movd r9d, xmm0
-
-FN_PREFIX(CryptonightR_soft_aes_template_part2):
-	mov rsp, r10
-	mov [rsp+144], ebx
-	mov [rsp+148], esi
-	mov [rsp+152], edi
-	mov [rsp+156], ebp
-
-	mov edi, edi
-	shl rbp, 32
-	or rbp, rdi
-	xor r8, rbp
-
-	mov ebx, ebx
-	shl rsi, 32
-	or rsi, rbx
-	xor QWORD PTR [rsp+320], rsi
-
-	mov rbx, [rsp+160]
-	mov rdi, [rsp+168]
-	mov rbp, [rsp+176]
-	mov r10, [rsp+184]
-
-	mov	r9, r10
-	xor	r9, 16
-	mov	rcx, r10
-	xor	rcx, 32
-	xor	r10, 48
-	mov	rax, rbx
-	mul	rdi
-	movdqu	xmm2, XMMWORD PTR [r9+r11]
-	movdqu	xmm1, XMMWORD PTR [rcx+r11]
-	pxor xmm6, xmm2
-	pxor xmm6, xmm1
-	paddq	xmm1, xmm7
-	add	r8, rdx
-	movdqu	xmm0, XMMWORD PTR [r10+r11]
-	pxor xmm6, xmm0
-	paddq	xmm0, xmm5
-	paddq	xmm2, xmm4
-	movdqu	XMMWORD PTR [r9+r11], xmm0
-	movdqa	xmm5, xmm4
-	mov	r9, QWORD PTR [rsp+320]
-	movdqa	xmm4, xmm6
-	add	r9, rax
-	movdqu	XMMWORD PTR [rcx+r11], xmm2
-	movdqu	XMMWORD PTR [r10+r11], xmm1
-	mov	r10, QWORD PTR [rsp+304]
-	movd r12d, xmm11
-	mov	QWORD PTR [r14], r8
-	xor	r8, rbx
-	mov	rax, r8
-	mov	QWORD PTR [r14+8], r9
-	and	eax, 2097136
-	xor	r9, rbp
-	mov	QWORD PTR [rsp+320], r9
-	mov	QWORD PTR [rsp+328], rax
-	sub	r12d, 1
-	jne	FN_PREFIX(CryptonightR_soft_aes_template_mainloop)
-
-FN_PREFIX(CryptonightR_soft_aes_template_part3):
-	movaps	xmm6, XMMWORD PTR [rsp+16]
-	movaps	xmm7, XMMWORD PTR [rsp+32]
-	movaps	xmm8, XMMWORD PTR [rsp+48]
-	movaps	xmm9, XMMWORD PTR [rsp+64]
-	movaps	xmm10, XMMWORD PTR [rsp+80]
-	movaps	xmm11, XMMWORD PTR [rsp+96]
-	movaps	xmm12, XMMWORD PTR [rsp+112]
-	movaps	xmm13, XMMWORD PTR [rsp+128]
-
-	add	rsp, 232
-	pop	r15
-	pop	r14
-	pop	r13
-	pop	r12
-	pop	rdi
-	pop	rsi
-	pop	rbp
-	pop	rbx
-	ret
-FN_PREFIX(CryptonightR_soft_aes_template_end):

src/crypto/asm/win64/CryptonightR_template.inc

@@ -1,529 +0,0 @@
-PUBLIC FN_PREFIX(CryptonightR_template_part1)
-PUBLIC FN_PREFIX(CryptonightR_template_mainloop)
-PUBLIC FN_PREFIX(CryptonightR_template_part2)
-PUBLIC FN_PREFIX(CryptonightR_template_part3)
-PUBLIC FN_PREFIX(CryptonightR_template_end)
-PUBLIC FN_PREFIX(CryptonightR_template_double_part1)
-PUBLIC FN_PREFIX(CryptonightR_template_double_mainloop)
-PUBLIC FN_PREFIX(CryptonightR_template_double_part2)
-PUBLIC FN_PREFIX(CryptonightR_template_double_part3)
-PUBLIC FN_PREFIX(CryptonightR_template_double_part4)
-PUBLIC FN_PREFIX(CryptonightR_template_double_end)
-
-ALIGN(64)
-FN_PREFIX(CryptonightR_template_part1):
-	mov	QWORD PTR [rsp+16], rbx
-	mov	QWORD PTR [rsp+24], rbp
-	mov	QWORD PTR [rsp+32], rsi
-	push	r10
-	push	r11
-	push	r12
-	push	r13
-	push	r14
-	push	r15
-	push	rdi
-	sub	rsp, 64
-	mov	r12, rcx
-	mov	r8, QWORD PTR [r12+32]
-	mov	rdx, r12
-	xor	r8, QWORD PTR [r12]
-	mov	r15, QWORD PTR [r12+40]
-	mov	r9, r8
-	xor	r15, QWORD PTR [r12+8]
-	mov	r11, QWORD PTR [r12+224]
-	mov	r12, QWORD PTR [r12+56]
-	xor	r12, QWORD PTR [rdx+24]
-	mov	rax, QWORD PTR [rdx+48]
-	xor	rax, QWORD PTR [rdx+16]
-	movaps	XMMWORD PTR [rsp+48], xmm6
-	movd	xmm0, r12
-	movaps	XMMWORD PTR [rsp+32], xmm7
-	movaps	XMMWORD PTR [rsp+16], xmm8
-	movaps	XMMWORD PTR [rsp], xmm9
-	mov	r12, QWORD PTR [rdx+88]
-	xor	r12, QWORD PTR [rdx+72]
-	movd	xmm6, rax
-	mov	rax, QWORD PTR [rdx+80]
-	xor	rax, QWORD PTR [rdx+64]
-	punpcklqdq xmm6, xmm0
-	and	r9d, 2097136
-	movd	xmm0, r12
-	movd	xmm7, rax
-	punpcklqdq xmm7, xmm0
-	mov r10d, r9d
-	movd	xmm9, rsp
-	mov rsp, r8
-	mov	r8d, 524288
-
-	mov	ebx, [rdx+96]
-	mov	esi, [rdx+100]
-	mov	edi, [rdx+104]
-	mov	ebp, [rdx+108]
-
-	ALIGN(64)
-FN_PREFIX(CryptonightR_template_mainloop):
-	movdqa	xmm5, XMMWORD PTR [r9+r11]
-	movd	xmm0, r15
-	movd	xmm4, rsp
-	punpcklqdq xmm4, xmm0
-	lea	rdx, QWORD PTR [r9+r11]
-
-	aesenc	xmm5, xmm4
-
-	mov	r12d, r9d
-	mov	eax, r9d
-	xor	r9d, 48
-	xor	r12d, 16
-	xor	eax, 32
-	movdqu	xmm0, XMMWORD PTR [r9+r11]
-	movaps xmm3, xmm0
-	movdqu	xmm2, XMMWORD PTR [r12+r11]
-	movdqu	xmm1, XMMWORD PTR [rax+r11]
-	pxor xmm0, xmm2
-	pxor xmm5, xmm1
-	pxor xmm5, xmm0
-	paddq	xmm3, xmm7
-	paddq	xmm2, xmm6
-	paddq	xmm1, xmm4
-	movdqu	XMMWORD PTR [r12+r11], xmm3
-	movdqu	XMMWORD PTR [rax+r11], xmm2
-	movdqu	XMMWORD PTR [r9+r11], xmm1
-
-	movd	r12, xmm5
-	movd	r10d, xmm5
-	and	r10d, 2097136
-
-	movdqa	xmm0, xmm5
-	pxor	xmm0, xmm6
-	movdqu	XMMWORD PTR [rdx], xmm0
-
-	lea	r13d, [ebx+esi]
-	lea	edx, [edi+ebp]
-	shl rdx, 32
-	or	r13, rdx
-
-	xor	r13, QWORD PTR [r10+r11]
-	mov	r14, QWORD PTR [r10+r11+8]
-
-	movd eax, xmm6
-	movd edx, xmm7
-	pextrd r9d, xmm7, 2
-
-FN_PREFIX(CryptonightR_template_part2):
-	mov eax, edi
-	mov edx, ebp
-	shl rdx, 32
-	or rax, rdx
-	xor rsp, rax
-
-	mov eax, ebx
-	mov edx, esi
-	shl rdx, 32
-	or rax, rdx
-	xor r15, rax
-
-	mov	rax, r13
-	mul	r12
-
-	mov	r9d, r10d
-	mov	r12d, r10d
-	xor	r9d, 16
-	xor	r12d, 32
-	xor	r10d, 48
-	movdqa	xmm1, XMMWORD PTR [r12+r11]
-	movaps xmm3, xmm1
-	movdqa	xmm2, XMMWORD PTR [r9+r11]
-	movdqa	xmm0, XMMWORD PTR [r10+r11]
-	pxor xmm1, xmm2
-	pxor xmm5, xmm0
-	pxor xmm5, xmm1
-	paddq	xmm3, xmm4
-	paddq	xmm2, xmm6
-	paddq	xmm0, xmm7
-	movdqu	XMMWORD PTR [r9+r11], xmm0
-	movdqu	XMMWORD PTR [r12+r11], xmm2
-	movdqu	XMMWORD PTR [r10+r11], xmm3
-
-	movdqa	xmm7, xmm6
-	add	r15, rax
-	add	rsp, rdx
-	xor	r10, 48
-	mov	QWORD PTR [r10+r11], rsp
-	xor	rsp, r13
-	mov	r9d, esp
-	mov	QWORD PTR [r10+r11+8], r15
-	and	r9d, 2097136
-	xor	r15, r14
-	movdqa	xmm6, xmm5
-	dec	r8d
-	jnz	FN_PREFIX(CryptonightR_template_mainloop)
-
-FN_PREFIX(CryptonightR_template_part3):
-	movd	rsp, xmm9
-
-	mov	rbx, QWORD PTR [rsp+136]
-	mov	rbp, QWORD PTR [rsp+144]
-	mov	rsi, QWORD PTR [rsp+152]
-	movaps	xmm6, XMMWORD PTR [rsp+48]
-	movaps	xmm7, XMMWORD PTR [rsp+32]
-	movaps	xmm8, XMMWORD PTR [rsp+16]
-	movaps	xmm9, XMMWORD PTR [rsp]
-	add	rsp, 64
-	pop	rdi
-	pop	r15
-	pop	r14
-	pop	r13
-	pop	r12
-	pop	r11
-	pop	r10
-	ret	0
-FN_PREFIX(CryptonightR_template_end):
-
-ALIGN(64)
-FN_PREFIX(CryptonightR_template_double_part1):
-	mov	QWORD PTR [rsp+24], rbx
-	push	rbp
-	push	rsi
-	push	rdi
-	push	r12
-	push	r13
-	push	r14
-	push	r15
-	sub	rsp, 320
-	mov	r14, QWORD PTR [rcx+32]
-	mov	r8, rcx
-	xor	r14, QWORD PTR [rcx]
-	mov	r12, QWORD PTR [rcx+40]
-	mov	ebx, r14d
-	mov	rsi, QWORD PTR [rcx+224]
-	and	ebx, 2097136
-	xor	r12, QWORD PTR [rcx+8]
-	mov	rcx, QWORD PTR [rcx+56]
-	xor	rcx, QWORD PTR [r8+24]
-	mov	rax, QWORD PTR [r8+48]
-	xor	rax, QWORD PTR [r8+16]
-	mov	r15, QWORD PTR [rdx+32]
-	xor	r15, QWORD PTR [rdx]
-	movd	xmm0, rcx
-	mov	rcx, QWORD PTR [r8+88]
-	xor	rcx, QWORD PTR [r8+72]
-	mov	r13, QWORD PTR [rdx+40]
-	mov	rdi, QWORD PTR [rdx+224]
-	xor	r13, QWORD PTR [rdx+8]
-	movaps	XMMWORD PTR [rsp+160], xmm6
-	movaps	XMMWORD PTR [rsp+176], xmm7
-	movaps	XMMWORD PTR [rsp+192], xmm8
-	movaps	XMMWORD PTR [rsp+208], xmm9
-	movaps	XMMWORD PTR [rsp+224], xmm10
-	movaps	XMMWORD PTR [rsp+240], xmm11
-	movaps	XMMWORD PTR [rsp+256], xmm12
-	movaps	XMMWORD PTR [rsp+272], xmm13
-	movaps	XMMWORD PTR [rsp+288], xmm14
-	movaps	XMMWORD PTR [rsp+304], xmm15
-	movd	xmm7, rax
-	mov	rax, QWORD PTR [r8+80]
-	xor	rax, QWORD PTR [r8+64]
-
-	movaps xmm1, XMMWORD PTR [rdx+96]
-	movaps xmm2, XMMWORD PTR [r8+96]
-	movaps XMMWORD PTR [rsp], xmm1
-	movaps XMMWORD PTR [rsp+16], xmm2
-
-	mov	r8d, r15d
-	punpcklqdq xmm7, xmm0
-	movd	xmm0, rcx
-	mov	rcx, QWORD PTR [rdx+56]
-	xor	rcx, QWORD PTR [rdx+24]
-	movd	xmm9, rax
-	mov	QWORD PTR [rsp+128], rsi
-	mov	rax, QWORD PTR [rdx+48]
-	xor	rax, QWORD PTR [rdx+16]
-	punpcklqdq xmm9, xmm0
-	movd	xmm0, rcx
-	mov	rcx, QWORD PTR [rdx+88]
-	xor	rcx, QWORD PTR [rdx+72]
-	movd	xmm8, rax
-	mov	QWORD PTR [rsp+136], rdi
-	mov	rax, QWORD PTR [rdx+80]
-	xor	rax, QWORD PTR [rdx+64]
-	punpcklqdq xmm8, xmm0
-	and	r8d, 2097136
-	movd	xmm0, rcx
-	mov	r11d, 524288
-	movd	xmm10, rax
-	punpcklqdq xmm10, xmm0
-	
-	movd xmm14, QWORD PTR [rsp+128]
-	movd xmm15, QWORD PTR [rsp+136]
-
-	ALIGN(64)
-FN_PREFIX(CryptonightR_template_double_mainloop):
-	movdqu	xmm6, XMMWORD PTR [rbx+rsi]
-	movd	xmm0, r12
-	mov	ecx, ebx
-	movd	xmm3, r14
-	punpcklqdq xmm3, xmm0
-	xor	ebx, 16
-	aesenc	xmm6, xmm3
-	movd	xmm4, r15
-	movdqu	xmm0, XMMWORD PTR [rbx+rsi]
-	pxor	xmm6, xmm0
-	xor	ebx, 48
-	paddq	xmm0, xmm7
-	movdqu	xmm1, XMMWORD PTR [rbx+rsi]
-	pxor	xmm6, xmm1
-	movdqu	XMMWORD PTR [rbx+rsi], xmm0
-	paddq	xmm1, xmm3
-	xor	ebx, 16
-	mov	eax, ebx
-	xor	rax, 32
-	movdqu	xmm0, XMMWORD PTR [rbx+rsi]
-	pxor	xmm6, xmm0
-	movd	rdx, xmm6
-	movdqu	XMMWORD PTR [rbx+rsi], xmm1
-	paddq	xmm0, xmm9
-	movdqu	XMMWORD PTR [rax+rsi], xmm0
-	movdqa	xmm0, xmm6
-	pxor	xmm0, xmm7
-	movdqu	XMMWORD PTR [rcx+rsi], xmm0
-	mov	esi, edx
-	movdqu	xmm5, XMMWORD PTR [r8+rdi]
-	and	esi, 2097136
-	mov	ecx, r8d
-	movd	xmm0, r13
-	punpcklqdq xmm4, xmm0
-	xor	r8d, 16
-	aesenc	xmm5, xmm4
-	movdqu	xmm0, XMMWORD PTR [r8+rdi]
-	pxor	xmm5, xmm0
-	xor	r8d, 48
-	paddq	xmm0, xmm8
-	movdqu	xmm1, XMMWORD PTR [r8+rdi]
-	pxor	xmm5, xmm1
-	movdqu	XMMWORD PTR [r8+rdi], xmm0
-	paddq	xmm1, xmm4
-	xor	r8d, 16
-	mov	eax, r8d
-	xor	rax, 32
-	movdqu	xmm0, XMMWORD PTR [r8+rdi]
-	pxor	xmm5, xmm0
-	movdqu	XMMWORD PTR [r8+rdi], xmm1
-	paddq	xmm0, xmm10
-	movdqu	XMMWORD PTR [rax+rdi], xmm0
-	movdqa	xmm0, xmm5
-	pxor	xmm0, xmm8
-	movdqu	XMMWORD PTR [rcx+rdi], xmm0
-	movd	rdi, xmm5
-	movd	rcx, xmm14
-	mov	ebp, edi
-	mov	r8, QWORD PTR [rcx+rsi]
-	mov	r10, QWORD PTR [rcx+rsi+8]
-	lea	r9, QWORD PTR [rcx+rsi]
-	xor	esi, 16
-
-	movd xmm0, rsp
-	movd xmm1, rsi
-	movd xmm2, rdi
-	movd xmm11, rbp
-	movd xmm12, r15
-	movd xmm13, rdx
-	mov [rsp+104], rcx
-	mov [rsp+112], r9
-
-	mov ebx, DWORD PTR [rsp+16]
-	mov esi, DWORD PTR [rsp+20]
-	mov edi, DWORD PTR [rsp+24]
-	mov ebp, DWORD PTR [rsp+28]
-
-	lea	eax, [ebx+esi]
-	lea	edx, [edi+ebp]
-	shl rdx, 32
-	or	rax, rdx
-	xor r8, rax
-
-	movd esp, xmm3
-	pextrd r15d, xmm3, 2
-	movd eax, xmm7
-	movd edx, xmm9
-	pextrd r9d, xmm9, 2
-
-FN_PREFIX(CryptonightR_template_double_part2):
-
-	mov eax, edi
-	mov edx, ebp
-	shl rdx, 32
-	or rax, rdx
-	xor r14, rax
-
-	mov eax, ebx
-	mov edx, esi
-	shl rdx, 32
-	or rax, rdx
-	xor r12, rax
-
-	movd rsp, xmm0
-	mov DWORD PTR [rsp+16], ebx
-	mov DWORD PTR [rsp+20], esi
-	mov DWORD PTR [rsp+24], edi
-	mov DWORD PTR [rsp+28], ebp
-
-	movd rsi, xmm1
-	movd rdi, xmm2
-	movd rbp, xmm11
-	movd r15, xmm12
-	movd rdx, xmm13
-	mov rcx, [rsp+104]
-	mov r9, [rsp+112]
-
-	mov rbx, r8
-	mov	rax, r8
-	mul	rdx
-	and	ebp, 2097136
-	mov	r8, rax
-	movdqu	xmm1, XMMWORD PTR [rcx+rsi]
-	pxor	xmm6, xmm1
-	xor	esi, 48
-	paddq	xmm1, xmm7
-	movdqu	xmm2, XMMWORD PTR [rsi+rcx]
-	pxor	xmm6, xmm2
-	paddq	xmm2, xmm3
-	movdqu	XMMWORD PTR [rsi+rcx], xmm1
-	xor	esi, 16
-	mov	eax, esi
-	mov	rsi, rcx
-	movdqu	xmm0, XMMWORD PTR [rax+rcx]
-	pxor	xmm6, xmm0
-	movdqu	XMMWORD PTR [rax+rcx], xmm2
-	paddq	xmm0, xmm9
-	add	r12, r8
-	xor	rax, 32
-	add	r14, rdx
-	movdqa	xmm9, xmm7
-	movdqa	xmm7, xmm6
-	movdqu	XMMWORD PTR [rax+rcx], xmm0
-	mov	QWORD PTR [r9+8], r12
-	xor	r12, r10
-	mov	QWORD PTR [r9], r14
-	movd rcx, xmm15
-	xor	r14, rbx
-	mov	r10d, ebp
-	mov	ebx, r14d
-	xor	ebp, 16
-	and	ebx, 2097136
-	mov	r8, QWORD PTR [r10+rcx]
-	mov	r9, QWORD PTR [r10+rcx+8]
-
-	movd xmm0, rsp
-	movd xmm1, rbx
-	movd xmm2, rsi
-	movd xmm11, rdi
-	movd xmm12, rbp
-	movd xmm13, r15
-	mov [rsp+104], rcx
-	mov [rsp+112], r9
-
-	mov ebx, DWORD PTR [rsp]
-	mov esi, DWORD PTR [rsp+4]
-	mov edi, DWORD PTR [rsp+8]
-	mov ebp, DWORD PTR [rsp+12]
-
-	lea	eax, [ebx+esi]
-	lea	edx, [edi+ebp]
-	shl rdx, 32
-	or	rax, rdx
-
-	xor r8, rax
-	movd xmm3, r8
-
-	movd esp, xmm4
-	pextrd r15d, xmm4, 2
-	movd eax, xmm8
-	movd edx, xmm10
-	pextrd r9d, xmm10, 2
-
-FN_PREFIX(CryptonightR_template_double_part3):
-
-	movd r15, xmm13
-
-	mov eax, edi
-	mov edx, ebp
-	shl rdx, 32
-	or rax, rdx
-	xor r15, rax
-
-	mov eax, ebx
-	mov edx, esi
-	shl rdx, 32
-	or rax, rdx
-	xor r13, rax
-
-	movd rsp, xmm0
-	mov DWORD PTR [rsp], ebx
-	mov DWORD PTR [rsp+4], esi
-	mov DWORD PTR [rsp+8], edi
-	mov DWORD PTR [rsp+12], ebp
-
-	movd rbx, xmm1
-	movd rsi, xmm2
-	movd rdi, xmm11
-	movd rbp, xmm12
-	mov rcx, [rsp+104]
-	mov r9, [rsp+112]
-
-	mov rax, r8
-	mul	rdi
-	mov	rdi, rcx
-	mov	r8, rax
-	movdqu	xmm1, XMMWORD PTR [rbp+rcx]
-	pxor xmm5, xmm1
-	xor	ebp, 48
-	paddq	xmm1, xmm8
-	add	r13, r8
-	movdqu	xmm2, XMMWORD PTR [rbp+rcx]
-	pxor xmm5, xmm2
-	add	r15, rdx
-	movdqu	XMMWORD PTR [rbp+rcx], xmm1
-	paddq	xmm2, xmm4
-	xor	ebp, 16
-	mov	eax, ebp
-	xor	rax, 32
-	movdqu	xmm0, XMMWORD PTR [rbp+rcx]
-	pxor xmm5, xmm0
-	movdqu	XMMWORD PTR [rbp+rcx], xmm2
-	paddq	xmm0, xmm10
-	movdqu	XMMWORD PTR [rax+rcx], xmm0
-	movd rax, xmm3
-	movdqa	xmm10, xmm8
-	mov	QWORD PTR [r10+rcx], r15
-	movdqa	xmm8, xmm5
-	xor	r15, rax
-	mov	QWORD PTR [r10+rcx+8], r13
-	mov	r8d, r15d
-	xor	r13, r9
-	and	r8d, 2097136
-	dec r11d
-	jnz	FN_PREFIX(CryptonightR_template_double_mainloop)
-
-FN_PREFIX(CryptonightR_template_double_part4):
-
-	mov	rbx, QWORD PTR [rsp+400]
-	movaps	xmm6, XMMWORD PTR [rsp+160]
-	movaps	xmm7, XMMWORD PTR [rsp+176]
-	movaps	xmm8, XMMWORD PTR [rsp+192]
-	movaps	xmm9, XMMWORD PTR [rsp+208]
-	movaps	xmm10, XMMWORD PTR [rsp+224]
-	movaps	xmm11, XMMWORD PTR [rsp+240]
-	movaps	xmm12, XMMWORD PTR [rsp+256]
-	movaps	xmm13, XMMWORD PTR [rsp+272]
-	movaps	xmm14, XMMWORD PTR [rsp+288]
-	movaps	xmm15, XMMWORD PTR [rsp+304]
-	add	rsp, 320
-	pop	r15
-	pop	r14
-	pop	r13
-	pop	r12
-	pop	rdi
-	pop	rsi
-	pop	rbp
-	ret	0
-FN_PREFIX(CryptonightR_template_double_end):

src/crypto/asm/win64/CryptonightR_template_win.inc

@@ -70,29 +70,30 @@ CryptonightR_template_mainloop:
 
 	aesenc	xmm5, xmm4
 
-	mov	r12d, r9d
+	mov	r13d, r9d
 	mov	eax, r9d
 	xor	r9d, 48
-	xor	r12d, 16
+	xor	r13d, 16
 	xor	eax, 32
 	movdqu	xmm0, XMMWORD PTR [r9+r11]
 	movaps xmm3, xmm0
-	movdqu	xmm2, XMMWORD PTR [r12+r11]
+	movdqu	xmm2, XMMWORD PTR [r13+r11]
 	movdqu	xmm1, XMMWORD PTR [rax+r11]
 	pxor xmm0, xmm2
 	pxor xmm5, xmm1
 	pxor xmm5, xmm0
-	paddq	xmm3, xmm7
-	paddq	xmm2, xmm6
-	paddq	xmm1, xmm4
-	movdqu	XMMWORD PTR [r12+r11], xmm3
-	movdqu	XMMWORD PTR [rax+r11], xmm2
-	movdqu	XMMWORD PTR [r9+r11], xmm1
 
 	movd	r12, xmm5
 	movd	r10d, xmm5
 	and	r10d, 2097136
 
+	paddq	xmm3, xmm7
+	paddq	xmm2, xmm6
+	paddq	xmm1, xmm4
+	movdqu	XMMWORD PTR [r13+r11], xmm3
+	movdqu	XMMWORD PTR [rax+r11], xmm2
+	movdqu	XMMWORD PTR [r9+r11], xmm1
+
 	movdqa	xmm0, xmm5
 	pxor	xmm0, xmm6
 	movdqu	XMMWORD PTR [rdx], xmm0
@@ -102,14 +103,16 @@ CryptonightR_template_mainloop:
 	shl rdx, 32
 	or	r13, rdx
 
-	xor	r13, QWORD PTR [r10+r11]
-	mov	r14, QWORD PTR [r10+r11+8]
-
 	movd eax, xmm6
 	movd edx, xmm7
 	pextrd r9d, xmm7, 2
 
+	xor	r13, QWORD PTR [r10+r11]
+	mov	r14, QWORD PTR [r10+r11+8]
+
 CryptonightR_template_part2:
+	lea	rcx, [r10+r11]
+
 	mov eax, edi
 	mov edx, ebp
 	shl rdx, 32
@@ -124,6 +127,8 @@ CryptonightR_template_part2:
 
 	mov	rax, r13
 	mul	r12
+	add	r15, rax
+	add	rsp, rdx
 
 	mov	r9d, r10d
 	mov	r12d, r10d
@@ -145,13 +150,10 @@ CryptonightR_template_part2:
 	movdqu	XMMWORD PTR [r10+r11], xmm3
 
 	movdqa	xmm7, xmm6
-	add	r15, rax
+	mov	QWORD PTR [rcx], rsp
-	add	rsp, rdx
-	xor	r10, 48
-	mov	QWORD PTR [r10+r11], rsp
 	xor	rsp, r13
 	mov	r9d, esp
-	mov	QWORD PTR [r10+r11+8], r15
+	mov	QWORD PTR [rcx+8], r15
 	and	r9d, 2097136
 	xor	r15, r14
 	movdqa	xmm6, xmm5

src/crypto/asm/win64/CryptonightWOW_soft_aes_template.inc

@@ -1,266 +0,0 @@
-PUBLIC FN_PREFIX(CryptonightWOW_soft_aes_template_part1)
-PUBLIC FN_PREFIX(CryptonightWOW_soft_aes_template_mainloop)
-PUBLIC FN_PREFIX(CryptonightWOW_soft_aes_template_part2)
-PUBLIC FN_PREFIX(CryptonightWOW_soft_aes_template_part3)
-PUBLIC FN_PREFIX(CryptonightWOW_soft_aes_template_end)
-
-ALIGN(64)
-FN_PREFIX(CryptonightWOW_soft_aes_template_part1):
-	mov	QWORD PTR [rsp+8], rcx
-	push	rbx
-	push	rbp
-	push	rsi
-	push	rdi
-	push	r12
-	push	r13
-	push	r14
-	push	r15
-	sub	rsp, 232
-
-	mov	eax, [rcx+96]
-	mov	ebx, [rcx+100]
-	mov	esi, [rcx+104]
-	mov	edx, [rcx+108]
-	mov [rsp+144], eax
-	mov [rsp+148], ebx
-	mov [rsp+152], esi
-	mov [rsp+156], edx
-
-	mov	rax, QWORD PTR [rcx+48]
-	mov	r10, rcx
-	xor	rax, QWORD PTR [rcx+16]
-	mov	r8, QWORD PTR [rcx+32]
-	xor	r8, QWORD PTR [rcx]
-	mov	r9, QWORD PTR [rcx+40]
-	xor	r9, QWORD PTR [rcx+8]
-	movd	xmm4, rax
-	mov	rdx, QWORD PTR [rcx+56]
-	xor	rdx, QWORD PTR [rcx+24]
-	mov	r11, QWORD PTR [rcx+224]
-	mov	rcx, QWORD PTR [rcx+88]
-	xor	rcx, QWORD PTR [r10+72]
-	mov	rax, QWORD PTR [r10+80]
-	movd	xmm0, rdx
-	xor	rax, QWORD PTR [r10+64]
-
-	movaps	XMMWORD PTR [rsp+16], xmm6
-	movaps	XMMWORD PTR [rsp+32], xmm7
-	movaps	XMMWORD PTR [rsp+48], xmm8
-	movaps	XMMWORD PTR [rsp+64], xmm9
-	movaps	XMMWORD PTR [rsp+80], xmm10
-	movaps	XMMWORD PTR [rsp+96], xmm11
-	movaps	XMMWORD PTR [rsp+112], xmm12
-	movaps	XMMWORD PTR [rsp+128], xmm13
-
-	movd	xmm5, rax
-
-	mov	rax, r8
-	punpcklqdq xmm4, xmm0
-	and	eax, 2097136
-	movd	xmm10, QWORD PTR [r10+96]
-	movd	xmm0, rcx
-	mov	rcx, QWORD PTR [r10+104]
-	xorps	xmm9, xmm9
-	mov	QWORD PTR [rsp+328], rax
-	movd	xmm12, r11
-	mov	QWORD PTR [rsp+320], r9
-	punpcklqdq xmm5, xmm0
-	movd xmm13, rcx
-	mov r12d, 524288
-
-	ALIGN(64)
-FN_PREFIX(CryptonightWOW_soft_aes_template_mainloop):
-	movd xmm11, r12d
-	mov	r12, QWORD PTR [r10+272]
-	lea	r13, QWORD PTR [rax+r11]
-	mov	esi, DWORD PTR [r13]
-	movd	xmm0, r9
-	mov	r10d, DWORD PTR [r13+4]
-	movd	xmm7, r8
-	mov	ebp, DWORD PTR [r13+12]
-	mov	r14d, DWORD PTR [r13+8]
-	mov	rdx, QWORD PTR [rsp+328]
-	movzx	ecx, sil
-	shr	esi, 8
-	punpcklqdq xmm7, xmm0
-	mov	r15d, DWORD PTR [r12+rcx*4]
-	movzx	ecx, r10b
-	shr	r10d, 8
-	mov	edi, DWORD PTR [r12+rcx*4]
-	movzx	ecx, r14b
-	shr	r14d, 8
-	mov	ebx, DWORD PTR [r12+rcx*4]
-	movzx	ecx, bpl
-	shr	ebp, 8
-	mov	r9d, DWORD PTR [r12+rcx*4]
-	movzx	ecx, r10b
-	shr	r10d, 8
-	xor	r15d, DWORD PTR [r12+rcx*4+1024]
-	movzx	ecx, r14b
-	shr	r14d, 8
-	mov	eax, r14d
-	shr	eax, 8
-	xor	edi, DWORD PTR [r12+rcx*4+1024]
-	add	eax, 256
-	movzx	ecx, bpl
-	shr	ebp, 8
-	xor	ebx, DWORD PTR [r12+rcx*4+1024]
-	movzx	ecx, sil
-	shr	esi, 8
-	xor	r9d, DWORD PTR [r12+rcx*4+1024]
-	add	r12, 2048
-	movzx	ecx, r10b
-	shr	r10d, 8
-	add	r10d, 256
-	mov	r11d, DWORD PTR [r12+rax*4]
-	xor	r11d, DWORD PTR [r12+rcx*4]
-	xor	r11d, r9d
-	movzx	ecx, sil
-	mov	r10d, DWORD PTR [r12+r10*4]
-	shr	esi, 8
-	add	esi, 256
-	xor	r10d, DWORD PTR [r12+rcx*4]
-	movzx	ecx, bpl
-	xor	r10d, ebx
-	shr	ebp, 8
-	movd	xmm1, r11d
-	add	ebp, 256
-	movd	r11, xmm12
-	mov	r9d, DWORD PTR [r12+rcx*4]
-	xor	r9d, DWORD PTR [r12+rsi*4]
-	mov	eax, DWORD PTR [r12+rbp*4]
-	xor	r9d, edi
-	movzx	ecx, r14b
-	movd	xmm0, r10d
-	movd	xmm2, r9d
-	xor	eax, DWORD PTR [r12+rcx*4]
-	mov	rcx, rdx
-	xor	eax, r15d
-	punpckldq xmm2, xmm1
-	xor	rcx, 16
-	movd	xmm6, eax
-	mov	rax, rdx
-	punpckldq xmm6, xmm0
-	xor	rax, 32
-	punpckldq xmm6, xmm2
-	xor	rdx, 48
-	movdqu	xmm2, XMMWORD PTR [rcx+r11]
-	pxor	xmm6, xmm7
-	paddq	xmm2, xmm4
-	movdqu	xmm1, XMMWORD PTR [rax+r11]
-	movdqu	xmm0, XMMWORD PTR [rdx+r11]
-	paddq	xmm0, xmm5
-	movdqu	XMMWORD PTR [rcx+r11], xmm0
-	movdqu	XMMWORD PTR [rax+r11], xmm2
-	movd rcx, xmm13
-	paddq	xmm1, xmm7
-	movdqu	XMMWORD PTR [rdx+r11], xmm1
-	movd	rdi, xmm6
-	mov	r10, rdi
-	and	r10d, 2097136
-	movdqa	xmm0, xmm6
-	pxor	xmm0, xmm4
-	movdqu	XMMWORD PTR [r13], xmm0
-
-	mov ebx, [rsp+144]
-	mov ebp, [rsp+152]
-	add ebx, [rsp+148]
-	add ebp, [rsp+156]
-	shl rbp, 32
-	or rbx, rbp
-
-	xor rbx, QWORD PTR [r10+r11]
-	lea	r14, QWORD PTR [r10+r11]
-	mov	rbp, QWORD PTR [r14+8]
-
-	mov [rsp+160], rbx
-	mov [rsp+168], rdi
-	mov [rsp+176], rbp
-	mov [rsp+184], r10
-	mov r10, rsp
-
-	mov ebx, [rsp+144]
-	mov esi, [rsp+148]
-	mov edi, [rsp+152]
-	mov ebp, [rsp+156]
-
-	movd esp, xmm7
-	movaps xmm0, xmm7
-	psrldq xmm0, 8
-	movd r15d, xmm0
-	movd eax, xmm4
-	movd edx, xmm5
-
-FN_PREFIX(CryptonightWOW_soft_aes_template_part2):
-	mov rsp, r10
-	mov [rsp+144], ebx
-	mov [rsp+148], esi
-	mov [rsp+152], edi
-	mov [rsp+156], ebp
-
-	mov rbx, [rsp+160]
-	mov rdi, [rsp+168]
-	mov rbp, [rsp+176]
-	mov r10, [rsp+184]
-
-	mov	r9, r10
-	xor	r9, 16
-	mov	rcx, r10
-	xor	rcx, 32
-	xor	r10, 48
-	mov	rax, rbx
-	mul	rdi
-	movdqu	xmm2, XMMWORD PTR [r9+r11]
-	movdqu	xmm1, XMMWORD PTR [rcx+r11]
-	paddq	xmm1, xmm7
-	movd	xmm0, rax
-	movd	xmm3, rdx
-	xor	rax, QWORD PTR [r11+rcx+8]
-	xor	rdx, QWORD PTR [rcx+r11]
-	punpcklqdq xmm3, xmm0
-	add	r8, rdx
-	movdqu	xmm0, XMMWORD PTR [r10+r11]
-	pxor	xmm2, xmm3
-	paddq	xmm0, xmm5
-	paddq	xmm2, xmm4
-	movdqu	XMMWORD PTR [r9+r11], xmm0
-	movdqa	xmm5, xmm4
-	mov	r9, QWORD PTR [rsp+320]
-	movdqa	xmm4, xmm6
-	add	r9, rax
-	movdqu	XMMWORD PTR [rcx+r11], xmm2
-	movdqu	XMMWORD PTR [r10+r11], xmm1
-	mov	r10, QWORD PTR [rsp+304]
-	movd r12d, xmm11
-	mov	QWORD PTR [r14], r8
-	xor	r8, rbx
-	mov	rax, r8
-	mov	QWORD PTR [r14+8], r9
-	and	eax, 2097136
-	xor	r9, rbp
-	mov	QWORD PTR [rsp+320], r9
-	mov	QWORD PTR [rsp+328], rax
-	sub	r12d, 1
-	jne	FN_PREFIX(CryptonightWOW_soft_aes_template_mainloop)
-
-FN_PREFIX(CryptonightWOW_soft_aes_template_part3):
-	movaps	xmm6, XMMWORD PTR [rsp+16]
-	movaps	xmm7, XMMWORD PTR [rsp+32]
-	movaps	xmm8, XMMWORD PTR [rsp+48]
-	movaps	xmm9, XMMWORD PTR [rsp+64]
-	movaps	xmm10, XMMWORD PTR [rsp+80]
-	movaps	xmm11, XMMWORD PTR [rsp+96]
-	movaps	xmm12, XMMWORD PTR [rsp+112]
-	movaps	xmm13, XMMWORD PTR [rsp+128]
-
-	add	rsp, 232
-	pop	r15
-	pop	r14
-	pop	r13
-	pop	r12
-	pop	rdi
-	pop	rsi
-	pop	rbp
-	pop	rbx
-	ret
-FN_PREFIX(CryptonightWOW_soft_aes_template_end):

src/crypto/asm/win64/CryptonightWOW_template.inc

@@ -1,486 +0,0 @@
-PUBLIC FN_PREFIX(CryptonightWOW_template_part1)
-PUBLIC FN_PREFIX(CryptonightWOW_template_mainloop)
-PUBLIC FN_PREFIX(CryptonightWOW_template_part2)
-PUBLIC FN_PREFIX(CryptonightWOW_template_part3)
-PUBLIC FN_PREFIX(CryptonightWOW_template_end)
-PUBLIC FN_PREFIX(CryptonightWOW_template_double_part1)
-PUBLIC FN_PREFIX(CryptonightWOW_template_double_mainloop)
-PUBLIC FN_PREFIX(CryptonightWOW_template_double_part2)
-PUBLIC FN_PREFIX(CryptonightWOW_template_double_part3)
-PUBLIC FN_PREFIX(CryptonightWOW_template_double_part4)
-PUBLIC FN_PREFIX(CryptonightWOW_template_double_end)
-
-ALIGN(64)
-FN_PREFIX(CryptonightWOW_template_part1):
-	mov	QWORD PTR [rsp+16], rbx
-	mov	QWORD PTR [rsp+24], rbp
-	mov	QWORD PTR [rsp+32], rsi
-	push	r10
-	push	r11
-	push	r12
-	push	r13
-	push	r14
-	push	r15
-	push	rdi
-	sub	rsp, 64
-	mov	r12, rcx
-	mov	r8, QWORD PTR [r12+32]
-	mov	rdx, r12
-	xor	r8, QWORD PTR [r12]
-	mov	r15, QWORD PTR [r12+40]
-	mov	r9, r8
-	xor	r15, QWORD PTR [r12+8]
-	mov	r11, QWORD PTR [r12+224]
-	mov	r12, QWORD PTR [r12+56]
-	xor	r12, QWORD PTR [rdx+24]
-	mov	rax, QWORD PTR [rdx+48]
-	xor	rax, QWORD PTR [rdx+16]
-	movaps	XMMWORD PTR [rsp+48], xmm6
-	movd	xmm0, r12
-	movaps	XMMWORD PTR [rsp+32], xmm7
-	movaps	XMMWORD PTR [rsp+16], xmm8
-	movaps	XMMWORD PTR [rsp], xmm9
-	mov	r12, QWORD PTR [rdx+88]
-	xor	r12, QWORD PTR [rdx+72]
-	movd	xmm6, rax
-	mov	rax, QWORD PTR [rdx+80]
-	xor	rax, QWORD PTR [rdx+64]
-	punpcklqdq xmm6, xmm0
-	and	r9d, 2097136
-	movd	xmm0, r12
-	movd	xmm7, rax
-	punpcklqdq xmm7, xmm0
-	mov r10d, r9d
-	movd	xmm9, rsp
-	mov rsp, r8
-	mov	r8d, 524288
-
-	mov	ebx, [rdx+96]
-	mov	esi, [rdx+100]
-	mov	edi, [rdx+104]
-	mov	ebp, [rdx+108]
-
-	ALIGN(64)
-FN_PREFIX(CryptonightWOW_template_mainloop):
-	movdqa	xmm5, XMMWORD PTR [r9+r11]
-	movd	xmm0, r15
-	movd	xmm4, rsp
-	punpcklqdq xmm4, xmm0
-	lea	rdx, QWORD PTR [r9+r11]
-
-	aesenc	xmm5, xmm4
-	movd	r10d, xmm5
-	and	r10d, 2097136
-
-	mov	r12d, r9d
-	mov	eax, r9d
-	xor	r9d, 48
-	xor	r12d, 16
-	xor	eax, 32
-	movdqu	xmm0, XMMWORD PTR [r9+r11]
-	movdqu	xmm2, XMMWORD PTR [r12+r11]
-	movdqu	xmm1, XMMWORD PTR [rax+r11]
-	paddq	xmm0, xmm7
-	paddq	xmm2, xmm6
-	paddq	xmm1, xmm4
-	movdqu	XMMWORD PTR [r12+r11], xmm0
-	movd	r12, xmm5
-	movdqu	XMMWORD PTR [rax+r11], xmm2
-	movdqu	XMMWORD PTR [r9+r11], xmm1
-
-	movdqa	xmm0, xmm5
-	pxor	xmm0, xmm6
-	movdqu	XMMWORD PTR [rdx], xmm0
-
-	lea	r13d, [ebx+esi]
-	lea	edx, [edi+ebp]
-	shl rdx, 32
-	or	r13, rdx
-
-	xor	r13, QWORD PTR [r10+r11]
-	mov	r14, QWORD PTR [r10+r11+8]
-
-	movd eax, xmm6
-	movd edx, xmm7
-	pextrd r9d, xmm7, 2
-
-FN_PREFIX(CryptonightWOW_template_part2):
-	mov	rax, r13
-	mul	r12
-	movd	xmm0, rax
-	movd	xmm3, rdx
-	punpcklqdq xmm3, xmm0
-
-	mov	r9d, r10d
-	mov	r12d, r10d
-	xor	r9d, 16
-	xor	r12d, 32
-	xor	r10d, 48
-	movdqa	xmm1, XMMWORD PTR [r12+r11]
-	xor	rdx, QWORD PTR [r12+r11]
-	xor	rax, QWORD PTR [r11+r12+8]
-	movdqa	xmm2, XMMWORD PTR [r9+r11]
-	pxor	xmm3, xmm2
-	paddq	xmm7, XMMWORD PTR [r10+r11]
-	paddq	xmm1, xmm4
-	paddq	xmm3, xmm6
-	movdqu	XMMWORD PTR [r9+r11], xmm7
-	movdqu	XMMWORD PTR [r12+r11], xmm3
-	movdqu	XMMWORD PTR [r10+r11], xmm1
-
-	movdqa	xmm7, xmm6
-	add	r15, rax
-	add	rsp, rdx
-	xor	r10, 48
-	mov	QWORD PTR [r10+r11], rsp
-	xor	rsp, r13
-	mov	r9d, esp
-	mov	QWORD PTR [r10+r11+8], r15
-	and	r9d, 2097136
-	xor	r15, r14
-	movdqa	xmm6, xmm5
-	dec	r8d
-	jnz	FN_PREFIX(CryptonightWOW_template_mainloop)
-
-FN_PREFIX(CryptonightWOW_template_part3):
-	movd	rsp, xmm9
-
-	mov	rbx, QWORD PTR [rsp+136]
-	mov	rbp, QWORD PTR [rsp+144]
-	mov	rsi, QWORD PTR [rsp+152]
-	movaps	xmm6, XMMWORD PTR [rsp+48]
-	movaps	xmm7, XMMWORD PTR [rsp+32]
-	movaps	xmm8, XMMWORD PTR [rsp+16]
-	movaps	xmm9, XMMWORD PTR [rsp]
-	add	rsp, 64
-	pop	rdi
-	pop	r15
-	pop	r14
-	pop	r13
-	pop	r12
-	pop	r11
-	pop	r10
-	ret	0
-FN_PREFIX(CryptonightWOW_template_end):
-
-ALIGN(64)
-FN_PREFIX(CryptonightWOW_template_double_part1):
-	mov	QWORD PTR [rsp+24], rbx
-	push	rbp
-	push	rsi
-	push	rdi
-	push	r12
-	push	r13
-	push	r14
-	push	r15
-	sub	rsp, 320
-	mov	r14, QWORD PTR [rcx+32]
-	mov	r8, rcx
-	xor	r14, QWORD PTR [rcx]
-	mov	r12, QWORD PTR [rcx+40]
-	mov	ebx, r14d
-	mov	rsi, QWORD PTR [rcx+224]
-	and	ebx, 2097136
-	xor	r12, QWORD PTR [rcx+8]
-	mov	rcx, QWORD PTR [rcx+56]
-	xor	rcx, QWORD PTR [r8+24]
-	mov	rax, QWORD PTR [r8+48]
-	xor	rax, QWORD PTR [r8+16]
-	mov	r15, QWORD PTR [rdx+32]
-	xor	r15, QWORD PTR [rdx]
-	movd	xmm0, rcx
-	mov	rcx, QWORD PTR [r8+88]
-	xor	rcx, QWORD PTR [r8+72]
-	mov	r13, QWORD PTR [rdx+40]
-	mov	rdi, QWORD PTR [rdx+224]
-	xor	r13, QWORD PTR [rdx+8]
-	movaps	XMMWORD PTR [rsp+160], xmm6
-	movaps	XMMWORD PTR [rsp+176], xmm7
-	movaps	XMMWORD PTR [rsp+192], xmm8
-	movaps	XMMWORD PTR [rsp+208], xmm9
-	movaps	XMMWORD PTR [rsp+224], xmm10
-	movaps	XMMWORD PTR [rsp+240], xmm11
-	movaps	XMMWORD PTR [rsp+256], xmm12
-	movaps	XMMWORD PTR [rsp+272], xmm13
-	movaps	XMMWORD PTR [rsp+288], xmm14
-	movaps	XMMWORD PTR [rsp+304], xmm15
-	movd	xmm7, rax
-	mov	rax, QWORD PTR [r8+80]
-	xor	rax, QWORD PTR [r8+64]
-
-	movaps xmm1, XMMWORD PTR [rdx+96]
-	movaps xmm2, XMMWORD PTR [r8+96]
-	movaps XMMWORD PTR [rsp], xmm1
-	movaps XMMWORD PTR [rsp+16], xmm2
-
-	mov	r8d, r15d
-	punpcklqdq xmm7, xmm0
-	movd	xmm0, rcx
-	mov	rcx, QWORD PTR [rdx+56]
-	xor	rcx, QWORD PTR [rdx+24]
-	movd	xmm9, rax
-	mov	QWORD PTR [rsp+128], rsi
-	mov	rax, QWORD PTR [rdx+48]
-	xor	rax, QWORD PTR [rdx+16]
-	punpcklqdq xmm9, xmm0
-	movd	xmm0, rcx
-	mov	rcx, QWORD PTR [rdx+88]
-	xor	rcx, QWORD PTR [rdx+72]
-	movd	xmm8, rax
-	mov	QWORD PTR [rsp+136], rdi
-	mov	rax, QWORD PTR [rdx+80]
-	xor	rax, QWORD PTR [rdx+64]
-	punpcklqdq xmm8, xmm0
-	and	r8d, 2097136
-	movd	xmm0, rcx
-	mov	r11d, 524288
-	movd	xmm10, rax
-	punpcklqdq xmm10, xmm0
-	
-	movd xmm14, QWORD PTR [rsp+128]
-	movd xmm15, QWORD PTR [rsp+136]
-
-	ALIGN(64)
-FN_PREFIX(CryptonightWOW_template_double_mainloop):
-	movdqu	xmm6, XMMWORD PTR [rbx+rsi]
-	movd	xmm0, r12
-	mov	ecx, ebx
-	movd	xmm3, r14
-	punpcklqdq xmm3, xmm0
-	xor	ebx, 16
-	aesenc	xmm6, xmm3
-	movd	rdx, xmm6
-	movd	xmm4, r15
-	movdqu	xmm0, XMMWORD PTR [rbx+rsi]
-	xor	ebx, 48
-	paddq	xmm0, xmm7
-	movdqu	xmm1, XMMWORD PTR [rbx+rsi]
-	movdqu	XMMWORD PTR [rbx+rsi], xmm0
-	paddq	xmm1, xmm3
-	xor	ebx, 16
-	mov	eax, ebx
-	xor	rax, 32
-	movdqu	xmm0, XMMWORD PTR [rbx+rsi]
-	movdqu	XMMWORD PTR [rbx+rsi], xmm1
-	paddq	xmm0, xmm9
-	movdqu	XMMWORD PTR [rax+rsi], xmm0
-	movdqa	xmm0, xmm6
-	pxor	xmm0, xmm7
-	movdqu	XMMWORD PTR [rcx+rsi], xmm0
-	mov	esi, edx
-	movdqu	xmm5, XMMWORD PTR [r8+rdi]
-	and	esi, 2097136
-	mov	ecx, r8d
-	movd	xmm0, r13
-	punpcklqdq xmm4, xmm0
-	xor	r8d, 16
-	aesenc	xmm5, xmm4
-	movdqu	xmm0, XMMWORD PTR [r8+rdi]
-	xor	r8d, 48
-	paddq	xmm0, xmm8
-	movdqu	xmm1, XMMWORD PTR [r8+rdi]
-	movdqu	XMMWORD PTR [r8+rdi], xmm0
-	paddq	xmm1, xmm4
-	xor	r8d, 16
-	mov	eax, r8d
-	xor	rax, 32
-	movdqu	xmm0, XMMWORD PTR [r8+rdi]
-	movdqu	XMMWORD PTR [r8+rdi], xmm1
-	paddq	xmm0, xmm10
-	movdqu	XMMWORD PTR [rax+rdi], xmm0
-	movdqa	xmm0, xmm5
-	pxor	xmm0, xmm8
-	movdqu	XMMWORD PTR [rcx+rdi], xmm0
-	movd	rdi, xmm5
-	movd	rcx, xmm14
-	mov	ebp, edi
-	mov	r8, QWORD PTR [rcx+rsi]
-	mov	r10, QWORD PTR [rcx+rsi+8]
-	lea	r9, QWORD PTR [rcx+rsi]
-	xor	esi, 16
-
-	movd xmm0, rsp
-	movd xmm1, rsi
-	movd xmm2, rdi
-	movd xmm11, rbp
-	movd xmm12, r15
-	movd xmm13, rdx
-	mov [rsp+104], rcx
-	mov [rsp+112], r9
-
-	mov ebx, DWORD PTR [rsp+16]
-	mov esi, DWORD PTR [rsp+20]
-	mov edi, DWORD PTR [rsp+24]
-	mov ebp, DWORD PTR [rsp+28]
-
-	lea	eax, [ebx+esi]
-	lea	edx, [edi+ebp]
-	shl rdx, 32
-	or	rax, rdx
-	xor r8, rax
-
-	movd esp, xmm3
-	pextrd r15d, xmm3, 2
-	movd eax, xmm7
-	movd edx, xmm9
-	pextrd r9d, xmm9, 2
-
-FN_PREFIX(CryptonightWOW_template_double_part2):
-
-	movd rsp, xmm0
-	mov DWORD PTR [rsp+16], ebx
-	mov DWORD PTR [rsp+20], esi
-	mov DWORD PTR [rsp+24], edi
-	mov DWORD PTR [rsp+28], ebp
-
-	movd rsi, xmm1
-	movd rdi, xmm2
-	movd rbp, xmm11
-	movd r15, xmm12
-	movd rdx, xmm13
-	mov rcx, [rsp+104]
-	mov r9, [rsp+112]
-
-	mov rbx, r8
-	mov	rax, r8
-	mul	rdx
-	and	ebp, 2097136
-	mov	r8, rax
-	movd	xmm1, rdx
-	movd	xmm0, r8
-	punpcklqdq xmm1, xmm0
-	pxor	xmm1, XMMWORD PTR [rcx+rsi]
-	xor	esi, 48
-	paddq	xmm1, xmm7
-	movdqu	xmm2, XMMWORD PTR [rsi+rcx]
-	xor	rdx, QWORD PTR [rsi+rcx]
-	paddq	xmm2, xmm3
-	xor	r8, QWORD PTR [rsi+rcx+8]
-	movdqu	XMMWORD PTR [rsi+rcx], xmm1
-	xor	esi, 16
-	mov	eax, esi
-	mov	rsi, rcx
-	movdqu	xmm0, XMMWORD PTR [rax+rcx]
-	movdqu	XMMWORD PTR [rax+rcx], xmm2
-	paddq	xmm0, xmm9
-	add	r12, r8
-	xor	rax, 32
-	add	r14, rdx
-	movdqa	xmm9, xmm7
-	movdqa	xmm7, xmm6
-	movdqu	XMMWORD PTR [rax+rcx], xmm0
-	mov	QWORD PTR [r9+8], r12
-	xor	r12, r10
-	mov	QWORD PTR [r9], r14
-	movd rcx, xmm15
-	xor	r14, rbx
-	mov	r10d, ebp
-	mov	ebx, r14d
-	xor	ebp, 16
-	and	ebx, 2097136
-	mov	r8, QWORD PTR [r10+rcx]
-	mov	r9, QWORD PTR [r10+rcx+8]
-
-	movd xmm0, rsp
-	movd xmm1, rbx
-	movd xmm2, rsi
-	movd xmm11, rdi
-	movd xmm12, rbp
-	movd xmm13, r15
-	mov [rsp+104], rcx
-	mov [rsp+112], r9
-
-	mov ebx, DWORD PTR [rsp]
-	mov esi, DWORD PTR [rsp+4]
-	mov edi, DWORD PTR [rsp+8]
-	mov ebp, DWORD PTR [rsp+12]
-
-	lea	eax, [ebx+esi]
-	lea	edx, [edi+ebp]
-	shl rdx, 32
-	or	rax, rdx
-
-	xor r8, rax
-	movd xmm3, r8
-
-	movd esp, xmm4
-	pextrd r15d, xmm4, 2
-	movd eax, xmm8
-	movd edx, xmm10
-	pextrd r9d, xmm10, 2
-
-FN_PREFIX(CryptonightWOW_template_double_part3):
-
-	movd rsp, xmm0
-	mov DWORD PTR [rsp], ebx
-	mov DWORD PTR [rsp+4], esi
-	mov DWORD PTR [rsp+8], edi
-	mov DWORD PTR [rsp+12], ebp
-
-	movd rbx, xmm1
-	movd rsi, xmm2
-	movd rdi, xmm11
-	movd rbp, xmm12
-	movd r15, xmm13
-	mov rcx, [rsp+104]
-	mov r9, [rsp+112]
-
-	mov rax, r8
-	mul	rdi
-	movd	xmm1, rdx
-	movd	xmm0, rax
-	punpcklqdq xmm1, xmm0
-	mov	rdi, rcx
-	mov	r8, rax
-	pxor	xmm1, XMMWORD PTR [rbp+rcx]
-	xor	ebp, 48
-	paddq	xmm1, xmm8
-	xor	r8, QWORD PTR [rbp+rcx+8]
-	xor	rdx, QWORD PTR [rbp+rcx]
-	add	r13, r8
-	movdqu	xmm2, XMMWORD PTR [rbp+rcx]
-	add	r15, rdx
-	movdqu	XMMWORD PTR [rbp+rcx], xmm1
-	paddq	xmm2, xmm4
-	xor	ebp, 16
-	mov	eax, ebp
-	xor	rax, 32
-	movdqu	xmm0, XMMWORD PTR [rbp+rcx]
-	movdqu	XMMWORD PTR [rbp+rcx], xmm2
-	paddq	xmm0, xmm10
-	movdqu	XMMWORD PTR [rax+rcx], xmm0
-	movd rax, xmm3
-	movdqa	xmm10, xmm8
-	mov	QWORD PTR [r10+rcx], r15
-	movdqa	xmm8, xmm5
-	xor	r15, rax
-	mov	QWORD PTR [r10+rcx+8], r13
-	mov	r8d, r15d
-	xor	r13, r9
-	and	r8d, 2097136
-	dec r11d
-	jnz	FN_PREFIX(CryptonightWOW_template_double_mainloop)
-
-FN_PREFIX(CryptonightWOW_template_double_part4):
-
-	mov	rbx, QWORD PTR [rsp+400]
-	movaps	xmm6, XMMWORD PTR [rsp+160]
-	movaps	xmm7, XMMWORD PTR [rsp+176]
-	movaps	xmm8, XMMWORD PTR [rsp+192]
-	movaps	xmm9, XMMWORD PTR [rsp+208]
-	movaps	xmm10, XMMWORD PTR [rsp+224]
-	movaps	xmm11, XMMWORD PTR [rsp+240]
-	movaps	xmm12, XMMWORD PTR [rsp+256]
-	movaps	xmm13, XMMWORD PTR [rsp+272]
-	movaps	xmm14, XMMWORD PTR [rsp+288]
-	movaps	xmm15, XMMWORD PTR [rsp+304]
-	add	rsp, 320
-	pop	r15
-	pop	r14
-	pop	r13
-	pop	r12
-	pop	rdi
-	pop	rsi
-	pop	rbp
-	ret	0
-FN_PREFIX(CryptonightWOW_template_double_end):

src/crypto/asm/win64/cn2/cnv2_rwz_double_main_loop.inc

@@ -0,0 +1,410 @@
+	mov	rax, rsp
+	push	rbx
+	push	rbp
+	push	rsi
+	push	rdi
+	push	r12
+	push	r13
+	push	r14
+	push	r15
+	sub	rsp, 184
+
+	stmxcsr DWORD PTR [rsp+272]
+	mov DWORD PTR [rsp+276], 24448
+	ldmxcsr DWORD PTR [rsp+276]
+
+	mov	r13, QWORD PTR [rcx+224]
+	mov	r9, rdx
+	mov	r10, QWORD PTR [rcx+32]
+	mov	r8, rcx
+	xor	r10, QWORD PTR [rcx]
+	mov	r14d, 393216
+	mov	r11, QWORD PTR [rcx+40]
+	xor	r11, QWORD PTR [rcx+8]
+	mov	rsi, QWORD PTR [rdx+224]
+	mov	rdx, QWORD PTR [rcx+56]
+	xor	rdx, QWORD PTR [rcx+24]
+	mov	rdi, QWORD PTR [r9+32]
+	xor	rdi, QWORD PTR [r9]
+	mov	rbp, QWORD PTR [r9+40]
+	xor	rbp, QWORD PTR [r9+8]
+	movd	xmm0, rdx
+	movaps	XMMWORD PTR [rax-88], xmm6
+	movaps	XMMWORD PTR [rax-104], xmm7
+	movaps	XMMWORD PTR [rax-120], xmm8
+	movaps	XMMWORD PTR [rsp+112], xmm9
+	movaps	XMMWORD PTR [rsp+96], xmm10
+	movaps	XMMWORD PTR [rsp+80], xmm11
+	movaps	XMMWORD PTR [rsp+64], xmm12
+	movaps	XMMWORD PTR [rsp+48], xmm13
+	movaps	XMMWORD PTR [rsp+32], xmm14
+	movaps	XMMWORD PTR [rsp+16], xmm15
+	mov	rdx, r10
+	movd	xmm4, QWORD PTR [r8+96]
+	and	edx, 2097136
+	mov	rax, QWORD PTR [rcx+48]
+	xorps	xmm13, xmm13
+	xor	rax, QWORD PTR [rcx+16]
+	mov	rcx, QWORD PTR [rcx+88]
+	xor	rcx, QWORD PTR [r8+72]
+	movd	xmm5, QWORD PTR [r8+104]
+	movd	xmm7, rax
+
+	mov eax, 1
+	shl rax, 52
+	movd xmm14, rax
+	punpcklqdq xmm14, xmm14
+
+	mov eax, 1023
+	shl rax, 52
+	movd xmm12, rax
+	punpcklqdq xmm12, xmm12
+
+	mov	rax, QWORD PTR [r8+80]
+	xor	rax, QWORD PTR [r8+64]
+	punpcklqdq xmm7, xmm0
+	movd	xmm0, rcx
+	mov	rcx, QWORD PTR [r9+56]
+	xor	rcx, QWORD PTR [r9+24]
+	movd	xmm3, rax
+	mov	rax, QWORD PTR [r9+48]
+	xor	rax, QWORD PTR [r9+16]
+	punpcklqdq xmm3, xmm0
+	movd	xmm0, rcx
+	mov	QWORD PTR [rsp], r13
+	mov	rcx, QWORD PTR [r9+88]
+	xor	rcx, QWORD PTR [r9+72]
+	movd	xmm6, rax
+	mov	rax, QWORD PTR [r9+80]
+	xor	rax, QWORD PTR [r9+64]
+	punpcklqdq xmm6, xmm0
+	movd	xmm0, rcx
+	mov	QWORD PTR [rsp+256], r10
+	mov	rcx, rdi
+	mov	QWORD PTR [rsp+264], r11
+	movd	xmm8, rax
+	and	ecx, 2097136
+	punpcklqdq xmm8, xmm0
+	movd	xmm0, QWORD PTR [r9+96]
+	punpcklqdq xmm4, xmm0
+	movd	xmm0, QWORD PTR [r9+104]
+	lea	r8, QWORD PTR [rcx+rsi]
+	movdqu	xmm11, XMMWORD PTR [r8]
+	punpcklqdq xmm5, xmm0
+	lea	r9, QWORD PTR [rdx+r13]
+	movdqu	xmm15, XMMWORD PTR [r9]
+
+	ALIGN(64)
+rwz_main_loop_double:
+	movdqu	xmm9, xmm15
+	mov eax, edx
+	mov ebx, edx
+	xor eax, 16
+	xor ebx, 32
+	xor edx, 48
+
+	movd	xmm0, r11
+	movd	xmm2, r10
+	punpcklqdq xmm2, xmm0
+	aesenc	xmm9, xmm2
+
+	movdqu	xmm0, XMMWORD PTR [rdx+r13]
+	movdqu	xmm1, XMMWORD PTR [rbx+r13]
+	paddq	xmm0, xmm7
+	paddq	xmm1, xmm2
+	movdqu	XMMWORD PTR [rbx+r13], xmm0
+	movdqu	xmm0, XMMWORD PTR [rax+r13]
+	movdqu	XMMWORD PTR [rdx+r13], xmm1
+	paddq	xmm0, xmm3
+	movdqu	XMMWORD PTR [rax+r13], xmm0
+
+	movd	r11, xmm9
+	mov	edx, r11d
+	and	edx, 2097136
+	movdqa	xmm0, xmm9
+	pxor	xmm0, xmm7
+	movdqu	XMMWORD PTR [r9], xmm0
+
+	lea	rbx, QWORD PTR [rdx+r13]
+	mov	r10, QWORD PTR [rdx+r13]
+
+	movdqu	xmm10, xmm11
+	movd	xmm0, rbp
+	movd	xmm11, rdi
+	punpcklqdq xmm11, xmm0
+	aesenc	xmm10, xmm11
+
+	mov eax, ecx
+	mov r12d, ecx
+	xor eax, 16
+	xor r12d, 32
+	xor ecx, 48
+
+	movdqu	xmm0, XMMWORD PTR [rcx+rsi]
+	paddq	xmm0, xmm6
+	movdqu	xmm1, XMMWORD PTR [r12+rsi]
+	movdqu	XMMWORD PTR [r12+rsi], xmm0
+	paddq	xmm1, xmm11
+	movdqu	xmm0, XMMWORD PTR [rax+rsi]
+	movdqu	XMMWORD PTR [rcx+rsi], xmm1
+	paddq	xmm0, xmm8
+	movdqu	XMMWORD PTR [rax+rsi], xmm0
+
+	movd	rcx, xmm10
+	and	ecx, 2097136
+
+	movdqa	xmm0, xmm10
+	pxor	xmm0, xmm6
+	movdqu	XMMWORD PTR [r8], xmm0
+	mov r12, QWORD PTR [rcx+rsi]
+
+	mov	r9, QWORD PTR [rbx+8]
+
+	xor edx, 16
+	mov r8d, edx
+	mov r15d, edx
+
+	movd	rdx, xmm5
+	shl	rdx, 32
+	movd	rax, xmm4
+	xor	rdx, rax
+	xor	r10, rdx
+	mov	rax, r10
+	mul	r11
+	mov r11d, r8d
+	xor r11d, 48
+	movd xmm0, rdx
+	xor rdx, [r11+r13]
+	movd xmm1, rax
+	xor rax, [r11+r13+8]
+	punpcklqdq xmm0, xmm1
+
+	pxor xmm0, XMMWORD PTR [r8+r13]
+	movdqu	xmm1, XMMWORD PTR [r11+r13]
+	paddq	xmm0, xmm3
+	paddq	xmm1, xmm2
+	movdqu	XMMWORD PTR [r8+r13], xmm0
+	xor	r8d, 32
+	movdqu	xmm0, XMMWORD PTR [r8+r13]
+	movdqu	XMMWORD PTR [r8+r13], xmm1
+	paddq	xmm0, xmm7
+	movdqu	XMMWORD PTR [r11+r13], xmm0
+
+	mov	r11, QWORD PTR [rsp+256]
+	add	r11, rdx
+	mov	rdx, QWORD PTR [rsp+264]
+	add	rdx, rax
+	mov	QWORD PTR [rbx], r11
+	xor	r11, r10
+	mov	QWORD PTR [rbx+8], rdx
+	xor	rdx, r9
+	mov	QWORD PTR [rsp+256], r11
+	and	r11d, 2097136
+	mov	QWORD PTR [rsp+264], rdx
+	mov	QWORD PTR [rsp+8], r11
+	lea	r15, QWORD PTR [r11+r13]
+	movdqu xmm15, XMMWORD PTR [r11+r13]
+	lea	r13, QWORD PTR [rsi+rcx]
+	movdqa	xmm0, xmm5
+	psrldq	xmm0, 8
+	movaps	xmm2, xmm13
+	movd	r10, xmm0
+	psllq	xmm5, 1
+	shl	r10, 32
+	movdqa	xmm0, xmm9
+	psrldq	xmm0, 8
+	movdqa	xmm1, xmm10
+	movd	r11, xmm0
+	psrldq	xmm1, 8
+	movd	r8, xmm1
+	psrldq	xmm4, 8
+	movaps	xmm0, xmm13
+	movd	rax, xmm4
+	xor	r10, rax
+	movaps	xmm1, xmm13
+	xor	r10, r12
+	lea	rax, QWORD PTR [r11+1]
+	shr	rax, 1
+	movdqa	xmm3, xmm9
+	punpcklqdq xmm3, xmm10
+	paddq	xmm5, xmm3
+	movd	rdx, xmm5
+	psrldq	xmm5, 8
+	cvtsi2sd xmm2, rax
+	or	edx, -2147483647
+	lea	rax, QWORD PTR [r8+1]
+	shr	rax, 1
+	movd	r9, xmm5
+	cvtsi2sd xmm0, rax
+	or	r9d, -2147483647
+	cvtsi2sd xmm1, rdx
+	unpcklpd xmm2, xmm0
+	movaps	xmm0, xmm13
+	cvtsi2sd xmm0, r9
+	unpcklpd xmm1, xmm0
+	divpd	xmm2, xmm1
+	paddq	xmm2, xmm14
+	cvttsd2si rax, xmm2
+	psrldq	xmm2, 8
+	mov	rbx, rax
+	imul	rax, rdx
+	sub	r11, rax
+	js	rwz_div_fix_1
+rwz_div_fix_1_ret:
+
+	cvttsd2si rdx, xmm2
+	mov	rax, rdx
+	imul	rax, r9
+	movd	xmm2, r11d
+	movd	xmm4, ebx
+	sub	r8, rax
+	js	rwz_div_fix_2
+rwz_div_fix_2_ret:
+
+	movd	xmm1, r8d
+	movd	xmm0, edx
+	punpckldq xmm2, xmm1
+	punpckldq xmm4, xmm0
+	punpckldq xmm4, xmm2
+	paddq	xmm3, xmm4
+	movdqa	xmm0, xmm3
+	psrlq	xmm0, 12
+	paddq	xmm0, xmm12
+	sqrtpd	xmm1, xmm0
+	movd	r9, xmm1
+	movdqa xmm5, xmm1
+	psrlq xmm5, 19
+	test	r9, 524287
+	je	rwz_sqrt_fix_1
+rwz_sqrt_fix_1_ret:
+
+	movd r9, xmm10
+	psrldq	xmm1, 8
+	movd	r8, xmm1
+	test	r8, 524287
+	je	rwz_sqrt_fix_2
+rwz_sqrt_fix_2_ret:
+
+	mov r12d, ecx
+	mov r8d, ecx
+	xor r12d, 16
+	xor r8d, 32
+	xor ecx, 48
+	mov	rax, r10
+	mul	r9
+	movd xmm0, rax
+	movd xmm3, rdx
+	punpcklqdq xmm3, xmm0
+
+	movdqu	xmm0, XMMWORD PTR [r12+rsi]
+	pxor xmm0, xmm3
+	movdqu	xmm1, XMMWORD PTR [r8+rsi]
+	xor rdx, [r8+rsi]
+	xor rax, [r8+rsi+8]
+	movdqu	xmm3, XMMWORD PTR [rcx+rsi]
+	paddq	xmm3, xmm6
+	paddq	xmm1, xmm11
+	paddq	xmm0, xmm8
+	movdqu	XMMWORD PTR [r8+rsi], xmm3
+	movdqu	XMMWORD PTR [rcx+rsi], xmm1
+	movdqu	XMMWORD PTR [r12+rsi], xmm0
+
+	add	rdi, rdx
+	mov	QWORD PTR [r13], rdi
+	xor	rdi, r10
+	mov	ecx, edi
+	and	ecx, 2097136
+	lea	r8, QWORD PTR [rcx+rsi]
+
+	mov rdx, QWORD PTR [r13+8]	
+	add	rbp, rax
+	mov	QWORD PTR [r13+8], rbp
+	movdqu xmm11, XMMWORD PTR [rcx+rsi]
+	xor	rbp, rdx
+	mov	r13, QWORD PTR [rsp]
+	movdqa	xmm3, xmm7
+	mov	rdx, QWORD PTR [rsp+8]
+	movdqa	xmm8, xmm6
+	mov	r10, QWORD PTR [rsp+256]
+	movdqa	xmm7, xmm9
+	mov	r11, QWORD PTR [rsp+264]
+	movdqa	xmm6, xmm10
+	mov	r9, r15
+	dec r14d
+	jne	rwz_main_loop_double
+
+	ldmxcsr DWORD PTR [rsp+272]
+	movaps	xmm13, XMMWORD PTR [rsp+48]
+	lea	r11, QWORD PTR [rsp+184]
+	movaps	xmm6, XMMWORD PTR [r11-24]
+	movaps	xmm7, XMMWORD PTR [r11-40]
+	movaps	xmm8, XMMWORD PTR [r11-56]
+	movaps	xmm9, XMMWORD PTR [r11-72]
+	movaps	xmm10, XMMWORD PTR [r11-88]
+	movaps	xmm11, XMMWORD PTR [r11-104]
+	movaps	xmm12, XMMWORD PTR [r11-120]
+	movaps	xmm14, XMMWORD PTR [rsp+32]
+	movaps	xmm15, XMMWORD PTR [rsp+16]
+	mov	rsp, r11
+	pop	r15
+	pop	r14
+	pop	r13
+	pop	r12
+	pop	rdi
+	pop	rsi
+	pop	rbp
+	pop	rbx
+	jmp rwz_cnv2_double_mainloop_asm_endp
+
+rwz_div_fix_1:
+	dec	rbx
+	add	r11, rdx
+	jmp	rwz_div_fix_1_ret
+
+rwz_div_fix_2:
+	dec	rdx
+	add	r8, r9
+	jmp	rwz_div_fix_2_ret
+
+rwz_sqrt_fix_1:
+	movd	r8, xmm3
+	movdqa xmm0, xmm5
+	psrldq xmm0, 8
+	dec	r9
+	mov r11d, -1022
+	shl r11, 32
+	mov	rax, r9
+	shr	r9, 19
+	shr	rax, 20
+	mov	rdx, r9
+	sub	rdx, rax
+	lea	rdx, [rdx+r11+1]
+	add	rax, r11
+	imul	rdx, rax
+	sub	rdx, r8
+	adc	r9, 0
+	movd xmm5, r9
+	punpcklqdq xmm5, xmm0
+	jmp	rwz_sqrt_fix_1_ret
+
+rwz_sqrt_fix_2:
+	psrldq	xmm3, 8
+	movd	r11, xmm3
+	dec	r8
+	mov ebx, -1022
+	shl rbx, 32
+	mov	rax, r8
+	shr	r8, 19
+	shr	rax, 20
+	mov	rdx, r8
+	sub	rdx, rax
+	lea	rdx, [rdx+rbx+1]
+	add	rax, rbx
+	imul	rdx, rax
+	sub	rdx, r11
+	adc	r8, 0
+	movd xmm0, r8
+	punpcklqdq xmm5, xmm0
+	jmp	rwz_sqrt_fix_2_ret
+
+rwz_cnv2_double_mainloop_asm_endp:

src/crypto/asm/win64/cn2/cnv2_rwz_main_loop.inc

@@ -0,0 +1,186 @@
+	mov	 QWORD PTR [rsp+24], rbx
+	push	 rbp
+	push	 rsi
+	push	 rdi
+	push	 r12
+	push	 r13
+	push	 r14
+	push	 r15
+	sub	 rsp, 80
+
+	stmxcsr DWORD PTR [rsp]
+	mov DWORD PTR [rsp+4], 24448
+	ldmxcsr DWORD PTR [rsp+4]
+
+	mov	 rax, QWORD PTR [rcx+48]
+	mov	 r9, rcx
+	xor	 rax, QWORD PTR [rcx+16]
+	mov	 esi, 393216
+	mov	 r8, QWORD PTR [rcx+32]
+	mov	 r13d, -2147483647
+	xor	 r8, QWORD PTR [rcx]
+	mov	 r11, QWORD PTR [rcx+40]
+	mov	 r10, r8
+	mov	 rdx, QWORD PTR [rcx+56]
+	movd	 xmm4, rax
+	xor	 rdx, QWORD PTR [rcx+24]
+	xor	 r11, QWORD PTR [rcx+8]
+	mov	 rbx, QWORD PTR [rcx+224]
+	mov	 rax, QWORD PTR [r9+80]
+	xor	 rax, QWORD PTR [r9+64]
+	movd	 xmm0, rdx
+	mov	 rcx, QWORD PTR [rcx+88]
+	xor	 rcx, QWORD PTR [r9+72]
+	movd	 xmm3, QWORD PTR [r9+104]
+	movaps	 XMMWORD PTR [rsp+64], xmm6
+	movaps	 XMMWORD PTR [rsp+48], xmm7
+	movaps	 XMMWORD PTR [rsp+32], xmm8
+	and	 r10d, 2097136
+	movd	 xmm5, rax
+
+	xor eax, eax
+	mov QWORD PTR [rsp+16], rax
+
+	mov ax, 1023
+	shl rax, 52
+	movd xmm8, rax
+	mov r15, QWORD PTR [r9+96]
+	punpcklqdq xmm4, xmm0
+	movd	 xmm0, rcx
+	punpcklqdq xmm5, xmm0
+	movdqu	 xmm6, XMMWORD PTR [r10+rbx]
+
+	ALIGN(64)
+rwz_main_loop:
+	lea	 rdx, QWORD PTR [r10+rbx]
+	mov	 ecx, r10d
+	mov	 eax, r10d
+	mov rdi, r15
+	xor	 ecx, 16
+	xor	 eax, 32
+	xor	 r10d, 48
+	movd	 xmm0, r11
+	movd	 xmm7, r8
+	punpcklqdq xmm7, xmm0
+	aesenc	 xmm6, xmm7
+	movd	 rbp, xmm6
+	mov	 r9, rbp
+	and	 r9d, 2097136
+	movdqu	 xmm0, XMMWORD PTR [rcx+rbx]
+	movdqu	 xmm1, XMMWORD PTR [rax+rbx]
+	movdqu	 xmm2, XMMWORD PTR [r10+rbx]
+	paddq	 xmm0, xmm5
+	paddq	 xmm1, xmm7
+	paddq	 xmm2, xmm4
+	movdqu	 XMMWORD PTR [rcx+rbx], xmm0
+	movdqu	 XMMWORD PTR [rax+rbx], xmm2
+	movdqu	 XMMWORD PTR [r10+rbx], xmm1
+	mov r10, r9
+	xor r10d, 32
+	movd	 rcx, xmm3
+	mov	 rax, rcx
+	shl	 rax, 32
+	xor	 rdi, rax
+	movdqa	 xmm0, xmm6
+	pxor	 xmm0, xmm4
+	movdqu	 XMMWORD PTR [rdx], xmm0
+	xor	 rdi, QWORD PTR [r9+rbx]
+	lea	 r14, QWORD PTR [r9+rbx]
+	mov	 r12, QWORD PTR [r14+8]
+	xor	 edx, edx
+	lea	 r9d, DWORD PTR [ecx+ecx]
+	add	 r9d, ebp
+	movdqa	 xmm0, xmm6
+	psrldq	 xmm0, 8
+	or	 r9d, r13d
+	movd	 rax, xmm0
+	div	 r9
+	xorps xmm3, xmm3
+	mov	 eax, eax
+	shl	 rdx, 32
+	add	 rdx, rax
+	lea	 r9, QWORD PTR [rdx+rbp]
+	mov r15, rdx
+	mov	 rax, r9
+	shr	 rax, 12
+	movd	 xmm0, rax
+	paddq	 xmm0, xmm8
+	sqrtsd	 xmm3, xmm0
+	psubq	 xmm3, XMMWORD PTR [rsp+16]
+	movd	 rdx, xmm3
+	test	 edx, 524287
+	je	 rwz_sqrt_fixup
+	psrlq	 xmm3, 19
+rwz_sqrt_fixup_ret:
+
+	mov	 ecx, r10d
+	mov	 rax, rdi
+	mul	 rbp
+	movd xmm2, rdx
+	xor rdx, [rcx+rbx]
+	add	 r8, rdx
+	mov	 QWORD PTR [r14], r8
+	xor	 r8, rdi
+	mov edi, r8d
+	and edi, 2097136
+	movd xmm0, rax
+	xor rax, [rcx+rbx+8]
+	add	 r11, rax
+	mov	 QWORD PTR [r14+8], r11
+	punpcklqdq xmm2, xmm0
+
+	mov	 r9d, r10d
+	xor	 r9d, 48
+	xor	 r10d, 16
+	pxor	 xmm2, XMMWORD PTR [r9+rbx]
+	movdqu	 xmm0, XMMWORD PTR [r10+rbx]
+	paddq	 xmm0, xmm4
+	movdqu	 xmm1, XMMWORD PTR [rcx+rbx]
+	paddq	 xmm2, xmm5
+	paddq	 xmm1, xmm7
+	movdqa	 xmm5, xmm4
+	movdqu	 XMMWORD PTR [r9+rbx], xmm2
+	movdqa	 xmm4, xmm6
+	movdqu	 XMMWORD PTR [rcx+rbx], xmm0
+	movdqu	 XMMWORD PTR [r10+rbx], xmm1
+	movdqu xmm6, [rdi+rbx]
+	mov	 r10d, edi
+	xor	 r11, r12
+	dec rsi
+	jne	 rwz_main_loop
+
+	ldmxcsr DWORD PTR [rsp]
+	mov	 rbx, QWORD PTR [rsp+160]
+	movaps	 xmm6, XMMWORD PTR [rsp+64]
+	movaps	 xmm7, XMMWORD PTR [rsp+48]
+	movaps	 xmm8, XMMWORD PTR [rsp+32]
+	add	 rsp, 80
+	pop	 r15
+	pop	 r14
+	pop	 r13
+	pop	 r12
+	pop	 rdi
+	pop	 rsi
+	pop	 rbp
+	jmp cnv2_rwz_main_loop_endp
+
+rwz_sqrt_fixup:
+	dec	 rdx
+	mov r13d, -1022
+	shl r13, 32
+	mov	 rax, rdx
+	shr	 rdx, 19
+	shr	 rax, 20
+	mov	 rcx, rdx
+	sub	 rcx, rax
+	add	 rax, r13
+	not r13
+	sub	 rcx, r13
+	mov	 r13d, -2147483647
+	imul	 rcx, rax
+	sub	 rcx, r9
+	adc	 rdx, 0
+	movd	 xmm3, rdx
+	jmp	 rwz_sqrt_fixup_ret
+
+cnv2_rwz_main_loop_endp:

src/crypto/asm/win64/cn_main_loop.S

@@ -5,6 +5,8 @@
 .global cnv2_mainloop_ryzen_asm
 .global cnv2_mainloop_bulldozer_asm
 .global cnv2_double_mainloop_sandybridge_asm
+.global cnv2_rwz_mainloop_asm
+.global cnv2_rwz_double_mainloop_asm
 
 ALIGN(64)
 cnv2_mainloop_ivybridge_asm:
@@ -29,3 +31,15 @@ cnv2_double_mainloop_sandybridge_asm:
 	#include "../cn2/cnv2_double_main_loop_sandybridge.inc"
 	ret 0
 	mov eax, 3735929054
+
+ALIGN(64)
+cnv2_rwz_mainloop_asm:
+	#include "cn2/cnv2_rwz_main_loop.inc"
+	ret 0
+	mov eax, 3735929054
+
+ALIGN(64)
+cnv2_rwz_double_mainloop_asm:
+	#include "cn2/cnv2_rwz_double_main_loop.inc"
+	ret 0
+	mov eax, 3735929054

src/crypto/asm/win64/cn_main_loop.asm

@@ -3,6 +3,8 @@ PUBLIC cnv2_mainloop_ivybridge_asm
 PUBLIC cnv2_mainloop_ryzen_asm
 PUBLIC cnv2_mainloop_bulldozer_asm
 PUBLIC cnv2_double_mainloop_sandybridge_asm
+PUBLIC cnv2_rwz_mainloop_asm
+PUBLIC cnv2_rwz_double_mainloop_asm
 
 ALIGN 64
 cnv2_mainloop_ivybridge_asm PROC
@@ -32,5 +34,19 @@ cnv2_double_mainloop_sandybridge_asm PROC
 	mov eax, 3735929054
 cnv2_double_mainloop_sandybridge_asm ENDP
 
+ALIGN(64)
+cnv2_rwz_mainloop_asm PROC
+	INCLUDE cn2/cnv2_rwz_main_loop.inc
+	ret 0
+	mov eax, 3735929054
+cnv2_rwz_mainloop_asm ENDP
+
+ALIGN(64)
+cnv2_rwz_double_mainloop_asm PROC
+	INCLUDE cn2/cnv2_rwz_double_main_loop.inc
+	ret 0
+	mov eax, 3735929054
+cnv2_rwz_double_mainloop_asm ENDP
+
 _TEXT_CNV2_MAINLOOP ENDS
 END

src/interfaces/IWorker.h

@@ -5,7 +5,8 @@
  * Copyright 2014-2016 Wolf9466    <https://github.com/OhGodAPet>
  * Copyright 2016      Jay D Dee   <jayddee246@gmail.com>
  * Copyright 2017-2018 XMR-Stak    <https://github.com/fireice-uk>, <https://github.com/psychocrypt>
- * Copyright 2016-2018 XMRig       <https://github.com/xmrig>, <support@xmrig.com>
+ * Copyright 2018-2019 SChernykh   <https://github.com/SChernykh>
+ * Copyright 2016-2019 XMRig       <https://github.com/xmrig>, <support@xmrig.com>
  *
  *   This program is free software: you can redistribute it and/or modify
  *   it under the terms of the GNU General Public License as published by
@@ -21,8 +22,8 @@
  *   along with this program. If not, see <http://www.gnu.org/licenses/>.
  */
 
-#ifndef __IWORKER_H__
+#ifndef XMRIG_IWORKER_H
-#define __IWORKER_H__
+#define XMRIG_IWORKER_H
 
 
 #include <stdint.h>
@@ -31,7 +32,7 @@
 class IWorker
 {
 public:
-    virtual ~IWorker() {}
+    virtual ~IWorker() = default;
 
     virtual bool selfTest()            = 0;
     virtual size_t id() const          = 0;
@@ -41,4 +42,4 @@ public:
 };
 
 
-#endif // __IWORKER_H__
+#endif // XMRIG_IWORKER_H

src/version.h

@@ -28,7 +28,7 @@
 #define APP_ID        "xmrig"
 #define APP_NAME      "XMRig"
 #define APP_DESC      "XMRig CPU miner"
-#define APP_VERSION   "2.13.1"
+#define APP_VERSION   "2.13.2-dev"
 #define APP_DOMAIN    "xmrig.com"
 #define APP_SITE      "www.xmrig.com"
 #define APP_COPYRIGHT "Copyright (C) 2016-2019 xmrig.com"
@@ -36,7 +36,7 @@
 
 #define APP_VER_MAJOR  2
 #define APP_VER_MINOR  13
-#define APP_VER_PATCH  1
+#define APP_VER_PATCH  2
 
 #ifdef _MSC_VER
 #   if (_MSC_VER >= 1910)

src/workers/CpuThread.cpp

@@ -106,6 +106,16 @@ xmrig::CpuThread::cn_mainloop_fun        cn_trtl_mainloop_ryzen_asm
 xmrig::CpuThread::cn_mainloop_fun        cn_trtl_mainloop_bulldozer_asm             = nullptr;
 xmrig::CpuThread::cn_mainloop_double_fun cn_trtl_double_mainloop_sandybridge_asm    = nullptr;
 
+xmrig::CpuThread::cn_mainloop_fun        cn_zls_mainloop_ivybridge_asm              = nullptr;
+xmrig::CpuThread::cn_mainloop_fun        cn_zls_mainloop_ryzen_asm                  = nullptr;
+xmrig::CpuThread::cn_mainloop_fun        cn_zls_mainloop_bulldozer_asm              = nullptr;
+xmrig::CpuThread::cn_mainloop_double_fun cn_zls_double_mainloop_sandybridge_asm     = nullptr;
+
+xmrig::CpuThread::cn_mainloop_fun        cn_double_mainloop_ivybridge_asm           = nullptr;
+xmrig::CpuThread::cn_mainloop_fun        cn_double_mainloop_ryzen_asm               = nullptr;
+xmrig::CpuThread::cn_mainloop_fun        cn_double_mainloop_bulldozer_asm           = nullptr;
+xmrig::CpuThread::cn_mainloop_double_fun cn_double_double_mainloop_sandybridge_asm  = nullptr;
+
 
 void xmrig::CpuThread::patchAsmVariants()
 {
@@ -122,6 +132,16 @@ void xmrig::CpuThread::patchAsmVariants()
     cn_trtl_mainloop_bulldozer_asm              = reinterpret_cast<cn_mainloop_fun>         (base + 0x6000);
     cn_trtl_double_mainloop_sandybridge_asm     = reinterpret_cast<cn_mainloop_double_fun>  (base + 0x7000);
 
+    cn_zls_mainloop_ivybridge_asm               = reinterpret_cast<cn_mainloop_fun>         (base + 0x8000);
+    cn_zls_mainloop_ryzen_asm                   = reinterpret_cast<cn_mainloop_fun>         (base + 0x9000);
+    cn_zls_mainloop_bulldozer_asm               = reinterpret_cast<cn_mainloop_fun>         (base + 0xA000);
+    cn_zls_double_mainloop_sandybridge_asm      = reinterpret_cast<cn_mainloop_double_fun>  (base + 0xB000);
+
+    cn_double_mainloop_ivybridge_asm            = reinterpret_cast<cn_mainloop_fun>         (base + 0xC000);
+    cn_double_mainloop_ryzen_asm                = reinterpret_cast<cn_mainloop_fun>         (base + 0xD000);
+    cn_double_mainloop_bulldozer_asm            = reinterpret_cast<cn_mainloop_fun>         (base + 0xE000);
+    cn_double_double_mainloop_sandybridge_asm   = reinterpret_cast<cn_mainloop_double_fun>  (base + 0xF000);
+
     patchCode(cn_half_mainloop_ivybridge_asm,            cnv2_mainloop_ivybridge_asm,           xmrig::CRYPTONIGHT_HALF_ITER,   xmrig::CRYPTONIGHT_MASK);
     patchCode(cn_half_mainloop_ryzen_asm,                cnv2_mainloop_ryzen_asm,               xmrig::CRYPTONIGHT_HALF_ITER,   xmrig::CRYPTONIGHT_MASK);
     patchCode(cn_half_mainloop_bulldozer_asm,            cnv2_mainloop_bulldozer_asm,           xmrig::CRYPTONIGHT_HALF_ITER,   xmrig::CRYPTONIGHT_MASK);
@@ -132,6 +152,16 @@ void xmrig::CpuThread::patchAsmVariants()
     patchCode(cn_trtl_mainloop_bulldozer_asm,            cnv2_mainloop_bulldozer_asm,           xmrig::CRYPTONIGHT_TRTL_ITER,   xmrig::CRYPTONIGHT_PICO_MASK);
     patchCode(cn_trtl_double_mainloop_sandybridge_asm,   cnv2_double_mainloop_sandybridge_asm,  xmrig::CRYPTONIGHT_TRTL_ITER,   xmrig::CRYPTONIGHT_PICO_MASK);
 
+    patchCode(cn_zls_mainloop_ivybridge_asm,             cnv2_mainloop_ivybridge_asm,           xmrig::CRYPTONIGHT_ZLS_ITER,    xmrig::CRYPTONIGHT_MASK);
+    patchCode(cn_zls_mainloop_ryzen_asm,                 cnv2_mainloop_ryzen_asm,               xmrig::CRYPTONIGHT_ZLS_ITER,    xmrig::CRYPTONIGHT_MASK);
+    patchCode(cn_zls_mainloop_bulldozer_asm,             cnv2_mainloop_bulldozer_asm,           xmrig::CRYPTONIGHT_ZLS_ITER,    xmrig::CRYPTONIGHT_MASK);
+    patchCode(cn_zls_double_mainloop_sandybridge_asm,    cnv2_double_mainloop_sandybridge_asm,  xmrig::CRYPTONIGHT_ZLS_ITER,    xmrig::CRYPTONIGHT_MASK);
+
+    patchCode(cn_double_mainloop_ivybridge_asm,          cnv2_mainloop_ivybridge_asm,           xmrig::CRYPTONIGHT_DOUBLE_ITER, xmrig::CRYPTONIGHT_MASK);
+    patchCode(cn_double_mainloop_ryzen_asm,              cnv2_mainloop_ryzen_asm,               xmrig::CRYPTONIGHT_DOUBLE_ITER, xmrig::CRYPTONIGHT_MASK);
+    patchCode(cn_double_mainloop_bulldozer_asm,          cnv2_mainloop_bulldozer_asm,           xmrig::CRYPTONIGHT_DOUBLE_ITER, xmrig::CRYPTONIGHT_MASK);
+    patchCode(cn_double_double_mainloop_sandybridge_asm, cnv2_double_mainloop_sandybridge_asm,  xmrig::CRYPTONIGHT_DOUBLE_ITER, xmrig::CRYPTONIGHT_MASK);
+
     Mem::protectExecutableMemory(base, allocation_size);
     Mem::flushInstructionCache(base, allocation_size);
 }
@@ -180,6 +210,10 @@ xmrig::CpuThread::cn_hash_fun xmrig::CpuThread::fn(Algo algorithm, AlgoVariant a
         add_asm_func<CRYPTONIGHT_PICO, VARIANT_TRTL>(asm_func_map);
 #       endif
 
+        add_asm_func<CRYPTONIGHT, VARIANT_RWZ>(asm_func_map);
+        add_asm_func<CRYPTONIGHT, VARIANT_ZLS>(asm_func_map);
+        add_asm_func<CRYPTONIGHT, VARIANT_DOUBLE>(asm_func_map);
+
         asm_func_map_initialized = true;
     }
 
@@ -323,6 +357,39 @@ xmrig::CpuThread::cn_hash_fun xmrig::CpuThread::fn(Algo algorithm, AlgoVariant a
         cryptonight_quad_hash<CRYPTONIGHT,   true,  VARIANT_4>,
         cryptonight_penta_hash<CRYPTONIGHT,  true,  VARIANT_4>,
 
+        cryptonight_single_hash<CRYPTONIGHT, false, VARIANT_RWZ>,
+        cryptonight_double_hash<CRYPTONIGHT, false, VARIANT_RWZ>,
+        cryptonight_single_hash<CRYPTONIGHT, true,  VARIANT_RWZ>,
+        cryptonight_double_hash<CRYPTONIGHT, true,  VARIANT_RWZ>,
+        cryptonight_triple_hash<CRYPTONIGHT, false, VARIANT_RWZ>,
+        cryptonight_quad_hash<CRYPTONIGHT,   false, VARIANT_RWZ>,
+        cryptonight_penta_hash<CRYPTONIGHT,  false, VARIANT_RWZ>,
+        cryptonight_triple_hash<CRYPTONIGHT, true,  VARIANT_RWZ>,
+        cryptonight_quad_hash<CRYPTONIGHT,   true,  VARIANT_RWZ>,
+        cryptonight_penta_hash<CRYPTONIGHT,  true,  VARIANT_RWZ>,
+
+        cryptonight_single_hash<CRYPTONIGHT, false, VARIANT_ZLS>,
+        cryptonight_double_hash<CRYPTONIGHT, false, VARIANT_ZLS>,
+        cryptonight_single_hash<CRYPTONIGHT, true,  VARIANT_ZLS>,
+        cryptonight_double_hash<CRYPTONIGHT, true,  VARIANT_ZLS>,
+        cryptonight_triple_hash<CRYPTONIGHT, false, VARIANT_ZLS>,
+        cryptonight_quad_hash<CRYPTONIGHT,   false, VARIANT_ZLS>,
+        cryptonight_penta_hash<CRYPTONIGHT,  false, VARIANT_ZLS>,
+        cryptonight_triple_hash<CRYPTONIGHT, true,  VARIANT_ZLS>,
+        cryptonight_quad_hash<CRYPTONIGHT,   true,  VARIANT_ZLS>,
+        cryptonight_penta_hash<CRYPTONIGHT,  true,  VARIANT_ZLS>,
+
+        cryptonight_single_hash<CRYPTONIGHT, false, VARIANT_DOUBLE>,
+        cryptonight_double_hash<CRYPTONIGHT, false, VARIANT_DOUBLE>,
+        cryptonight_single_hash<CRYPTONIGHT, true,  VARIANT_DOUBLE>,
+        cryptonight_double_hash<CRYPTONIGHT, true,  VARIANT_DOUBLE>,
+        cryptonight_triple_hash<CRYPTONIGHT, false, VARIANT_DOUBLE>,
+        cryptonight_quad_hash<CRYPTONIGHT,   false, VARIANT_DOUBLE>,
+        cryptonight_penta_hash<CRYPTONIGHT,  false, VARIANT_DOUBLE>,
+        cryptonight_triple_hash<CRYPTONIGHT, true,  VARIANT_DOUBLE>,
+        cryptonight_quad_hash<CRYPTONIGHT,   true,  VARIANT_DOUBLE>,
+        cryptonight_penta_hash<CRYPTONIGHT,  true,  VARIANT_DOUBLE>,
+
 #       ifndef XMRIG_NO_AEON
         cryptonight_single_hash<CRYPTONIGHT_LITE, false, VARIANT_0>,
         cryptonight_double_hash<CRYPTONIGHT_LITE, false, VARIANT_0>,
@@ -358,6 +425,9 @@ xmrig::CpuThread::cn_hash_fun xmrig::CpuThread::fn(Algo algorithm, AlgoVariant a
         nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, // VARIANT_GPU
         nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, // VARIANT_WOW
         nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, // VARIANT_4
+        nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, // VARIANT_RWZ
+        nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, // VARIANT_ZLS
+        nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, // VARIANT_DOUBLE
 #       else
         nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, // VARIANT_0
         nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, // VARIANT_1
@@ -373,6 +443,9 @@ xmrig::CpuThread::cn_hash_fun xmrig::CpuThread::fn(Algo algorithm, AlgoVariant a
         nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, // VARIANT_GPU
         nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, // VARIANT_WOW
         nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, // VARIANT_4
+        nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, // VARIANT_RWZ
+        nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, // VARIANT_ZLS
+        nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, // VARIANT_DOUBLE
 #       endif
 
 #       ifndef XMRIG_NO_SUMO
@@ -422,6 +495,9 @@ xmrig::CpuThread::cn_hash_fun xmrig::CpuThread::fn(Algo algorithm, AlgoVariant a
         nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, // VARIANT_GPU
         nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, // VARIANT_WOW
         nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, // VARIANT_4
+        nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, // VARIANT_RWZ
+        nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, // VARIANT_ZLS
+        nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, // VARIANT_DOUBLE
 #       else
         nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, // VARIANT_0
         nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, // VARIANT_1
@@ -437,6 +513,9 @@ xmrig::CpuThread::cn_hash_fun xmrig::CpuThread::fn(Algo algorithm, AlgoVariant a
         nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, // VARIANT_GPU
         nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, // VARIANT_WOW
         nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, // VARIANT_4
+        nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, // VARIANT_RWZ
+        nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, // VARIANT_ZLS
+        nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, // VARIANT_DOUBLE
 #       endif
 
 #       ifndef XMRIG_NO_CN_PICO
@@ -465,6 +544,9 @@ xmrig::CpuThread::cn_hash_fun xmrig::CpuThread::fn(Algo algorithm, AlgoVariant a
         nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, // VARIANT_GPU
         nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, // VARIANT_WOW
         nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, // VARIANT_4
+        nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, // VARIANT_RWZ
+        nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, // VARIANT_ZLS
+        nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, // VARIANT_DOUBLE
 #       else
         nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, // VARIANT_0
         nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, // VARIANT_1
@@ -480,6 +562,9 @@ xmrig::CpuThread::cn_hash_fun xmrig::CpuThread::fn(Algo algorithm, AlgoVariant a
         nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, // VARIANT_GPU
         nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, // VARIANT_WOW
         nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, // VARIANT_4
+        nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, // VARIANT_RWZ
+        nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, // VARIANT_ZLS
+        nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, // VARIANT_DOUBLE
 #       endif
     };
 

src/workers/MultiWorker.cpp

@@ -25,7 +25,6 @@
 
 
 #include <thread>
-#include <sstream>
 
 
 #include "crypto/CryptoNight_test.h"
@@ -56,15 +55,6 @@ bool MultiWorker<N>::selfTest()
     using namespace xmrig;
 
     if (m_thread->algorithm() == CRYPTONIGHT) {
-        if (!verify2(VARIANT_WOW, test_input_WOW)) {
-            LOG_WARN("CryptonightR (Wownero) self-test failed");
-            return false;
-        }
-        if (!verify2(VARIANT_4, test_input_R)) {
-            LOG_WARN("CryptonightR self-test failed");
-            return false;
-        }
-
         const bool rc = verify(VARIANT_0,      test_output_v0)   &&
                         verify(VARIANT_1,      test_output_v1)   &&
                         verify(VARIANT_2,      test_output_v2)   &&
@@ -72,7 +62,12 @@ bool MultiWorker<N>::selfTest()
                         verify(VARIANT_MSR,    test_output_msr)  &&
                         verify(VARIANT_XAO,    test_output_xao)  &&
                         verify(VARIANT_RTO,    test_output_rto)  &&
-                        verify(VARIANT_HALF, test_output_half);
+                        verify(VARIANT_HALF,   test_output_half) &&
+                        verify2(VARIANT_WOW,   test_output_wow)  &&
+                        verify2(VARIANT_4,     test_output_r)    &&
+                        verify(VARIANT_RWZ,    test_output_rwz)  &&
+                        verify(VARIANT_ZLS,    test_output_zls)  &&
+                        verify(VARIANT_DOUBLE, test_output_double);
 
 #       ifndef XMRIG_NO_CN_GPU
         if (!rc || N > 1) {
@@ -179,61 +174,48 @@ bool MultiWorker<N>::verify(xmrig::Variant variant, const uint8_t *referenceValu
 
 
 template<size_t N>
-bool MultiWorker<N>::verify2(xmrig::Variant variant, const char *test_data)
+bool MultiWorker<N>::verify2(xmrig::Variant variant, const uint8_t *referenceValue)
 {
     xmrig::CpuThread::cn_hash_fun func = m_thread->fn(variant);
     if (!func) {
         return false;
     }
 
-    std::stringstream s(test_data);
+    for (size_t i = 0; i < (sizeof(cn_r_test_input) / sizeof(cn_r_test_input[0])); ++i) {
-    std::string expected_hex;
+        const size_t size = cn_r_test_input[i].size;
-    std::string input_hex;
+        for (size_t k = 0; k < N; ++k) {
-    uint64_t height;
+            memcpy(m_state.blob + (k * size), cn_r_test_input[i].data, size);
-    while (!s.eof())
-    {
-        uint8_t referenceValue[N * 32];
-        uint8_t input[N * 256];
-
-        s >> expected_hex;
-        s >> input_hex;
-        s >> height;
-
-        if ((expected_hex.length() != 64) || (input_hex.length() > 512))
-        {
-            return false;
         }
 
-        bool err = false;
+        func(m_state.blob, size, m_hash, m_ctx, cn_r_test_input[i].height);
 
-        for (int i = 0; i < 32; ++i)
+        for (size_t k = 0; k < N; ++k) {
-        {
+            if (memcmp(m_hash + k * 32, referenceValue + i * 32, sizeof m_hash / N) != 0) {
-            referenceValue[i] = (hf_hex2bin(expected_hex[i * 2], err) << 4) + hf_hex2bin(expected_hex[i * 2 + 1], err);
-        }
-
-        const size_t input_len = input_hex.length() / 2;
-        for (size_t i = 0; i < input_len; ++i)
-        {
-            input[i] = (hf_hex2bin(input_hex[i * 2], err) << 4) + hf_hex2bin(input_hex[i * 2 + 1], err);
-        }
-
-        if (err)
-        {
-            return false;
-        }
-
-        for (size_t i = 1; i < N; ++i)
-        {
-            memcpy(input + i * input_len, input, input_len);
-            memcpy(referenceValue + i * 32, referenceValue, 32);
-        }
-
-        func(input, input_len, m_hash, m_ctx, height);
-        if (memcmp(m_hash, referenceValue, sizeof m_hash) != 0)
-        {
                 return false;
             }
         }
+    }
+
+    return true;
+}
+
+
+template<>
+bool MultiWorker<1>::verify2(xmrig::Variant variant, const uint8_t *referenceValue)
+{
+    xmrig::CpuThread::cn_hash_fun func = m_thread->fn(variant);
+    if (!func) {
+        return false;
+    }
+
+    for (size_t i = 0; i < (sizeof(cn_r_test_input) / sizeof(cn_r_test_input[0])); ++i) {
+        func(cn_r_test_input[i].data, cn_r_test_input[i].size, m_hash, m_ctx, cn_r_test_input[i].height);
+
+        if (memcmp(m_hash, referenceValue + i * 32, sizeof m_hash) != 0) {
+            return false;
+        }
+    }
+
     return true;
 }
 

src/workers/MultiWorker.h

@@ -50,7 +50,7 @@ protected:
 private:
     bool resume(const xmrig::Job &job);
     bool verify(xmrig::Variant variant, const uint8_t *referenceValue);
-    bool verify2(xmrig::Variant variant, const char *test_data);
+    bool verify2(xmrig::Variant variant, const uint8_t *referenceValue);
     void consumeJob();
     void save(const xmrig::Job &job);