github-wownero/RandomX
1
0
Fork 0
mirror of https://github.com/wownero/RandomX.git synced 2026-10-21 01:08:48 +02:00
Code Issues Packages Projects Releases Wiki Activity

Added explicit STORE instructions

Browse Source 
JIT compiler
This commit is contained in:
tevador
2019-01-27 10:52:30 +01:00
parent d2cb086221
commit 005c67f64c
27 changed files with 1751 additions and 1518 deletions
Download Patch File Download Diff File Expand all files Collapse all files
src/asm/program_epilogue_store.inc
+8 -12
View File
@@ -1,9 +1,5 @@
;# unroll VM stack
mov rsp, rdi
;# save VM register values
pop rcx
pop rcx
mov qword ptr [rcx+0], r8
mov qword ptr [rcx+8], r9
mov qword ptr [rcx+16], r10
@@ -12,12 +8,12 @@
mov qword ptr [rcx+40], r13
mov qword ptr [rcx+48], r14
mov qword ptr [rcx+56], r15
movapd xmmword ptr [rcx+64], xmm8
movapd xmmword ptr [rcx+80], xmm9
movapd xmmword ptr [rcx+96], xmm2
movapd xmmword ptr [rcx+112], xmm3
movdqa xmmword ptr [rcx+64], xmm0
movdqa xmmword ptr [rcx+80], xmm1
movdqa xmmword ptr [rcx+96], xmm2
movdqa xmmword ptr [rcx+112], xmm3
lea rcx, [rcx+64]
movapd xmmword ptr [rcx+64], xmm4
movapd xmmword ptr [rcx+80], xmm5
movapd xmmword ptr [rcx+96], xmm6
movapd xmmword ptr [rcx+112], xmm7
movdqa xmmword ptr [rcx+64], xmm4
movdqa xmmword ptr [rcx+80], xmm5
movdqa xmmword ptr [rcx+96], xmm6
movdqa xmmword ptr [rcx+112], xmm7
src/asm/program_epilogue_win64.inc
+7 -1
View File
@@ -1,6 +1,12 @@
include program_epilogue_store.inc
;# restore callee-saved registers - Microsoft x64 calling convention
movdqu xmm15, xmmword ptr [rsp]
movdqu xmm14, xmmword ptr [rsp+16]
movdqu xmm13, xmmword ptr [rsp+32]
movdqu xmm12, xmmword ptr [rsp+48]
movdqu xmm11, xmmword ptr [rsp+64]
add rsp, 80
movdqu xmm10, xmmword ptr [rsp]
movdqu xmm9, xmmword ptr [rsp+16]
movdqu xmm8, xmmword ptr [rsp+32]
@@ -17,4 +23,4 @@
pop rbx
;# program finished
ret 0
ret
src/asm/program_load_flt.inc
+14
View File
@@ -0,0 +1,14 @@
and eax, 262080
lea rcx, [rsi+rax]
cvtdq2pd xmm0, qword ptr [rcx+0]
cvtdq2pd xmm1, qword ptr [rcx+8]
cvtdq2pd xmm2, qword ptr [rcx+16]
cvtdq2pd xmm3, qword ptr [rcx+24]
cvtdq2pd xmm4, qword ptr [rcx+32]
cvtdq2pd xmm5, qword ptr [rcx+40]
cvtdq2pd xmm6, qword ptr [rcx+48]
cvtdq2pd xmm7, qword ptr [rcx+56]
andps xmm4, xmm14
andps xmm5, xmm14
andps xmm6, xmm14
andps xmm7, xmm14
src/asm/program_load_int.inc
+10
View File
@@ -0,0 +1,10 @@
and eax, 262080
lea rcx, [rsi+rax]
xor r8, qword ptr [rcx+0]
xor r9, qword ptr [rcx+8]
xor r10, qword ptr [rcx+16]
xor r11, qword ptr [rcx+24]
xor r12, qword ptr [rcx+32]
xor r13, qword ptr [rcx+40]
xor r14, qword ptr [rcx+48]
xor r15, qword ptr [rcx+56]
src/asm/program_prologue_linux.inc
+6 -5
View File
@@ -7,13 +7,14 @@
push r15
;# function arguments
mov rbx, rcx ;# loop counter
push rdi ;# RegisterFile& registerFile
mov rbp, qword ptr [rsi] ;# "mx", "ma"
mov rax, qword ptr [rsi+8] ;# uint8_t* dataset
push rax
mov rsi, rdx ;# convertible_t* scratchpad
mov rcx, rdi
mov rbp, qword ptr [rsi] ;# "mx", "ma"
mov eax, ebp ;# "mx"
mov rdi, qword ptr [rsi+8] ;# uint8_t* dataset
mov rsi, rdx ;# convertible_t* scratchpad
#include "program_prologue_load.inc"
jmp randomx_program_begin
jmp DECL(randomx_loop_begin)
src/asm/program_prologue_load.inc
+18 -25
View File
@@ -1,27 +1,20 @@
mov rdi, rsp ;# beginning of VM stack
mov ebx, 262145 ;# number of VM instructions to execute + 1
;# zero integer registers
xor r8, r8
xor r9, r9
xor r10, r10
xor r11, r11
xor r12, r12
xor r13, r13
xor r14, r14
xor r15, r15
xorps xmm10, xmm10
cmpeqpd xmm10, xmm10
psrlq xmm10, 1 ;# mask for absolute value = 0x7fffffffffffffff7fffffffffffffff
;# load constant registers
lea rcx, [rcx+120]
movapd xmm8, xmmword ptr [rcx+72]
movapd xmm9, xmmword ptr [rcx+88]
movapd xmm10, xmmword ptr [rcx+104]
movapd xmm11, xmmword ptr [rcx+120]
movapd xmm13, xmmword ptr [minDbl]
movapd xmm14, xmmword ptr [absMask]
movapd xmm15, xmmword ptr [signMask]
;# load integer registers
mov r8, qword ptr [rcx+0]
mov r9, qword ptr [rcx+8]
mov r10, qword ptr [rcx+16]
mov r11, qword ptr [rcx+24]
mov r12, qword ptr [rcx+32]
mov r13, qword ptr [rcx+40]
mov r14, qword ptr [rcx+48]
mov r15, qword ptr [rcx+56]
;# load floating point registers
movapd xmm8, xmmword ptr [rcx+64]
movapd xmm9, xmmword ptr [rcx+80]
movapd xmm2, xmmword ptr [rcx+96]
movapd xmm3, xmmword ptr [rcx+112]
lea rcx, [rcx+64]
movapd xmm4, xmmword ptr [rcx+64]
movapd xmm5, xmmword ptr [rcx+80]
movapd xmm6, xmmword ptr [rcx+96]
movapd xmm7, xmmword ptr [rcx+112]
src/asm/program_prologue_win64.inc
+14 -7
View File
@@ -13,14 +13,21 @@
movdqu xmmword ptr [rsp+32], xmm8
movdqu xmmword ptr [rsp+16], xmm9
movdqu xmmword ptr [rsp+0], xmm10
sub rsp, 80
movdqu xmmword ptr [rsp+64], xmm11
movdqu xmmword ptr [rsp+48], xmm12
movdqu xmmword ptr [rsp+32], xmm13
movdqu xmmword ptr [rsp+16], xmm14
movdqu xmmword ptr [rsp+0], xmm15
;# function arguments
push rcx ;# RegisterFile& registerFile
mov rbp, qword ptr [rdx] ;# "mx", "ma"
mov rax, qword ptr [rdx+8] ;# uint8_t* dataset
push rax
mov rsi, r8 ;# convertible_t* scratchpad
; function arguments
push rcx ; RegisterFile& registerFile
mov rbp, qword ptr [rdx] ; "mx", "ma"
mov eax, ebp ; "mx"
mov rdi, qword ptr [rdx+8] ; uint8_t* dataset
mov rsi, r8 ; convertible_t* scratchpad
mov rbx, r9 ; loop counter
include program_prologue_load.inc
jmp randomx_program_begin
jmp randomx_loop_begin
src/asm/program_read.inc
-20
View File
@@ -1,20 +0,0 @@
db 0, 0, 0, 0 ;# TransformAddress placeholder
mov rcx, qword ptr [rdi] ;# load the dataset address
xor rbp, rax ;# modify "mx"
;# prefetch cacheline "mx"
and rbp, -64 ;# align "mx" to the start of a cache line
mov edx, ebp ;# edx = mx
prefetchnta byte ptr [rcx+rdx]
;# read cacheline "ma"
ror rbp, 32 ;# swap "ma" and "mx"
mov edx, ebp ;# edx = ma
lea rcx, [rcx+rdx] ;# dataset cache line
xor r8, qword ptr [rcx+0]
xor r9, qword ptr [rcx+8]
xor r10, qword ptr [rcx+16]
xor r11, qword ptr [rcx+24]
xor r12, qword ptr [rcx+32]
xor r13, qword ptr [rcx+40]
xor r14, qword ptr [rcx+48]
xor r15, qword ptr [rcx+56]
ret
src/asm/program_read_dataset.inc
+16
View File
@@ -0,0 +1,16 @@
xor rbp, rax ;# modify "mx"
and rbp, -64 ;# align "mx" to the start of a cache line
mov edx, ebp ;# edx = mx
prefetchnta byte ptr [rdi+rdx]
ror rbp, 32 ;# swap "ma" and "mx"
mov edx, ebp ;# edx = ma
lea rcx, [rdi+rdx] ;# dataset cache line
xor r8, qword ptr [rcx+0]
xor r9, qword ptr [rcx+8]
xor r10, qword ptr [rcx+16]
xor r11, qword ptr [rcx+24]
xor r12, qword ptr [rcx+32]
xor r13, qword ptr [rcx+40]
xor r14, qword ptr [rcx+48]
xor r15, qword ptr [rcx+56]
src/asm/program_store_flt.inc
+11
View File
@@ -0,0 +1,11 @@
and eax, 262080
lea rcx, [rsi+rax]
mulpd xmm0, xmm4
mulpd xmm1, xmm5
mulpd xmm2, xmm6
mulpd xmm3, xmm7
movapd xmmword ptr [rcx+0], xmm0
movapd xmmword ptr [rcx+16], xmm1
movapd xmmword ptr [rcx+32], xmm2
movapd xmmword ptr [rcx+48], xmm3
src/asm/program_store_int.inc
+10
View File
@@ -0,0 +1,10 @@
and eax, 262080
lea rcx, [rsi+rax]
mov qword ptr [rcx+0], r8
mov qword ptr [rcx+8], r9
mov qword ptr [rcx+16], r10
mov qword ptr [rcx+24], r11
mov qword ptr [rcx+32], r12
mov qword ptr [rcx+40], r13
mov qword ptr [rcx+48], r14
mov qword ptr [rcx+56], r15
src/asm/program_xmm_constants.inc
+6
View File
@@ -0,0 +1,6 @@
minDbl:
db 0, 0, 0, 0, 0, 0, 16, 0, 0, 0, 0, 0, 0, 0, 16, 0
absMask:
db 255, 255, 255, 255, 255, 255, 255, 127, 255, 255, 255, 255, 255, 255, 255, 127
signMask:
db 0, 0, 0, 0, 0, 0, 0, 128, 0, 0, 0, 0, 0, 0, 0, 128