Ecosystem/xmrig
Ecosystem/xmrig
1
0
Fork 0
mirror of https://github.com/xmrig/xmrig.git synced 2025-01-10 21:04:37 +00:00
Code Issues Projects Releases Packages Wiki Activity Actions

RandomX ARMv8: optimized dataset read

Browse source 
Break dependency from readReg2 and readReg3. It should run faster on superscalar and out-of-order CPUs i.e. Apple M1.
This commit is contained in:
SChernykh 2021-05-20 21:24:28 +02:00
parent 3bfa5ea038
commit 94fecb5e92
1 changed files with 3 additions and 2 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

5
src/crypto/randomx/jit_compiler_a64_static.S
View file

@ -304,6 +304,9 @@ literal_v14: .fill 2,8,0
literal_v15: .fill 2,8,0 literal_v15: .fill 2,8,0
DECL(randomx_program_aarch64_vm_instructions_end): DECL(randomx_program_aarch64_vm_instructions_end):
# Calculate dataset pointer for dataset read
# Do it here to break false dependency from readReg2 and readReg3 (see next line)
lsr x10, x9, 32
# mx ^= r[readReg2] ^ r[readReg3]; # mx ^= r[readReg2] ^ r[readReg3];
eor x9, x9, x18 eor x9, x9, x18
@ -321,8 +324,6 @@ DECL(randomx_program_aarch64_cacheline_align_mask1):
# mx <-> ma # mx <-> ma
ror x9, x9, 32 ror x9, x9, 32
# Calculate dataset pointer for dataset read
mov w10, w9
DECL(randomx_program_aarch64_cacheline_align_mask2): DECL(randomx_program_aarch64_cacheline_align_mask2):
# Actual mask will be inserted by JIT compiler # Actual mask will be inserted by JIT compiler
and x10, x10, 1 and x10, x10, 1