mirrors/xmrig
1
0
Fork 1
mirror of https://github.com/xmrig/xmrig.git synced 2026-04-18 13:22:43 -04:00
Code Issues Projects Releases Wiki Activity

RandomX v2 + commitments

Browse Source
This commit is contained in:
SChernykh
2026-01-30 00:34:42 +01:00
parent 5a80c65d31
commit 9d71358f46
26 changed files with 6532 additions and 6068 deletions
Download Patch File Download Diff File Expand all files Collapse all files

126
src/crypto/randomx/jit_compiler_x86.cpp
View File

@@ -41,6 +41,7 @@ OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
#include "crypto/randomx/reciprocal.h"
#include "crypto/randomx/superscalar.hpp"
#include "crypto/randomx/virtual_memory.hpp"
#include "crypto/randomx/soft_aes.h"
#include "crypto/rx/Profiler.h"
#ifdef XMRIG_FIX_RYZEN
@@ -116,6 +117,7 @@ namespace randomx {
#define codeLoopLoadXOP ADDR(randomx_program_loop_load_xop)
#define codeProgramStart ADDR(randomx_program_start)
#define codeReadDataset ADDR(randomx_program_read_dataset)
#define codeReadDatasetV2 ADDR(randomx_program_read_dataset_v2)
#define codeReadDatasetLightSshInit ADDR(randomx_program_read_dataset_sshash_init)
#define codeReadDatasetLightSshFin ADDR(randomx_program_read_dataset_sshash_fin)
#define codeDatasetInit ADDR(randomx_dataset_init)
@@ -125,6 +127,8 @@ namespace randomx {
#define codeDatasetInitAVX2SshLoad ADDR(randomx_dataset_init_avx2_ssh_load)
#define codeDatasetInitAVX2SshPrefetch ADDR(randomx_dataset_init_avx2_ssh_prefetch)
#define codeLoopStore ADDR(randomx_program_loop_store)
#define codeLoopStoreHardAES ADDR(randomx_program_loop_store_hard_aes)
#define codeLoopStoreSoftAES ADDR(randomx_program_loop_store_soft_aes)
#define codeLoopEnd ADDR(randomx_program_loop_end)
#define codeEpilogue ADDR(randomx_program_epilogue)
#define codeProgramEnd ADDR(randomx_program_end)
@@ -136,10 +140,13 @@ namespace randomx {
#define prologueSize (codeLoopBegin - codePrologue)
#define loopLoadSize (codeLoopLoadXOP - codeLoopLoad)
#define loopLoadXOPSize (codeProgramStart - codeLoopLoadXOP)
#define readDatasetSize (codeReadDatasetLightSshInit - codeReadDataset)
#define readDatasetSize (codeReadDatasetV2 - codeReadDataset)
#define readDatasetV2Size (codeReadDatasetLightSshInit - codeReadDatasetV2)
#define readDatasetLightInitSize (codeReadDatasetLightSshFin - codeReadDatasetLightSshInit)
#define readDatasetLightFinSize (codeLoopStore - codeReadDatasetLightSshFin)
#define loopStoreSize (codeLoopEnd - codeLoopStore)
#define loopStoreSize (codeLoopStoreHardAES - codeLoopStore)
#define loopStoreHardAESSize (codeLoopStoreSoftAES - codeLoopStoreHardAES)
#define loopStoreSoftAESSize (codeLoopEnd - codeLoopStoreSoftAES)
#define datasetInitSize (codeDatasetInitAVX2Prologue - codeDatasetInit)
#define datasetInitAVX2PrologueSize (codeDatasetInitAVX2LoopEnd - codeDatasetInitAVX2Prologue)
#define datasetInitAVX2LoopEndSize (codeDatasetInitAVX2Epilogue - codeDatasetInitAVX2LoopEnd)
@@ -223,6 +230,8 @@ namespace randomx {
JitCompilerX86::JitCompilerX86(bool hugePagesEnable, bool optimizedInitDatasetEnable) {
BranchesWithin32B = xmrig::Cpu::info()->jccErratum();
hasAES = xmrig::Cpu::info()->hasAES();
hasAVX = xmrig::Cpu::info()->hasAVX();
hasAVX2 = xmrig::Cpu::info()->hasAVX2();
@@ -341,7 +350,14 @@ namespace randomx {
vm_flags = flags;
generateProgramPrologue(prog, pcfg);
emit(codeReadDataset, readDatasetSize, code, codePos);
if (RandomX_CurrentConfig.Tweak_V2_PREFETCH) {
emit(codeReadDatasetV2, readDatasetV2Size, code, codePos);
}
else {
emit(codeReadDataset, readDatasetSize, code, codePos);
}
generateProgramEpilogue(prog, pcfg);
}
@@ -424,8 +440,15 @@ namespace randomx {
void JitCompilerX86::generateProgramPrologue(Program& prog, ProgramConfiguration& pcfg) {
codePos = ADDR(randomx_program_prologue_first_load) - ADDR(randomx_program_prologue);
*(uint32_t*)(code + codePos + 4) = RandomX_CurrentConfig.ScratchpadL3Mask64_Calculated;
*(uint32_t*)(code + codePos + 14) = RandomX_CurrentConfig.ScratchpadL3Mask64_Calculated;
if (RandomX_CurrentConfig.Tweak_V2_AES && !hasAES) {
*(uint64_t*)(code + codePos + 9) = reinterpret_cast<uint64_t>(lutEnc);
*(uint64_t*)(code + codePos + 27) = reinterpret_cast<uint64_t>(lutDec);
}
*(uint32_t*)(code + codePos + 47) = RandomX_CurrentConfig.ScratchpadL3Mask64_Calculated;
*(uint32_t*)(code + codePos + 57) = RandomX_CurrentConfig.ScratchpadL3Mask64_Calculated;
if (hasAVX) {
uint32_t* p = (uint32_t*)(code + codePos + 61);
*p = (*p & 0xFF000000U) | 0x0077F8C5U; // vzeroupper
@@ -476,8 +499,21 @@ namespace randomx {
*(uint64_t*)(code + codePos) = 0xc03349c08b49ull + (static_cast<uint64_t>(pcfg.readReg0) << 16) + (static_cast<uint64_t>(pcfg.readReg1) << 40);
codePos += 6;
emit(RandomX_CurrentConfig.codePrefetchScratchpadTweaked, RandomX_CurrentConfig.codePrefetchScratchpadTweakedSize, code, codePos);
memcpy(code + codePos, codeLoopStore, loopStoreSize);
codePos += loopStoreSize;
if (RandomX_CurrentConfig.Tweak_V2_AES) {
if (hasAES) {
memcpy(code + codePos, codeLoopStoreHardAES, loopStoreHardAESSize);
codePos += loopStoreHardAESSize;
}
else {
memcpy(code + codePos, codeLoopStoreSoftAES, loopStoreSoftAESSize);
codePos += loopStoreSoftAESSize;
}
}
else {
memcpy(code + codePos, codeLoopStore, loopStoreSize);
codePos += loopStoreSize;
}
if (BranchesWithin32B) {
const uint32_t branch_begin = static_cast<uint32_t>(codePos);
@@ -1307,7 +1343,7 @@ namespace randomx {
uint8_t* const p = code;
int32_t t = prevCFROUND;
if (t > prevFPOperation) {
if ((t > prevFPOperation) && !RandomX_CurrentConfig.Tweak_V2_CFROUND) {
if (vm_flags & RANDOMX_FLAG_AMD) {
memcpy(p + t, NOP26, 26);
}
@@ -1326,14 +1362,38 @@ namespace randomx {
*(uint32_t*)(p + pos + 3) = 0x00C8C148 + (rotate << 24);
if (vm_flags & RANDOMX_FLAG_AMD) {
*(uint64_t*)(p + pos + 7) = 0x742024443B0CE083ULL;
*(uint64_t*)(p + pos + 15) = 0x8900EB0414AE0F0AULL;
*(uint32_t*)(p + pos + 23) = 0x202444;
pos += 26;
if (RandomX_CurrentConfig.Tweak_V2_CFROUND) {
*(uint32_t*)(p + pos + 7) = 0x1375F0A8;
pos += 11;
}
else {
pos += 7;
}
*(uint64_t*)(p + pos) = 0x742024443B0CE083ULL;
*(uint64_t*)(p + pos + 8) = 0x8900EB0414AE0F0AULL;
*(uint32_t*)(p + pos + 16) = 0x202444;
pos += 19;
}
else {
*(uint64_t*)(p + pos + 7) = 0x0414AE0F0CE083ULL;
pos += 14;
pos += 7;
if (RandomX_CurrentConfig.Tweak_V2_CFROUND) {
if (BranchesWithin32B) {
const uint32_t branch_begin = static_cast<uint32_t>(pos + 2) & 31;
// If the jump crosses or touches 32-byte boundary, align it
if (branch_begin >= 30) {
const uint32_t alignment_size = 32 - branch_begin;
emit(NOPX[alignment_size - 1], alignment_size, code, pos);
}
}
*(uint32_t*)(p + pos) = 0x0775F0A8;
pos += 4;
}
*(uint64_t*)(p + pos) = 0x0414AE0F0CE083ULL;
pos += 7;
}
codePos = pos;
@@ -1343,7 +1403,7 @@ namespace randomx {
uint8_t* const p = code;
int32_t t = prevCFROUND;
if (t > prevFPOperation) {
if ((t > prevFPOperation) && !RandomX_CurrentConfig.Tweak_V2_CFROUND){
if (vm_flags & RANDOMX_FLAG_AMD) {
memcpy(p + t, NOP25, 25);
}
@@ -1361,14 +1421,38 @@ namespace randomx {
*(uint64_t*)(p + pos) = 0xC0F0FBC3C4ULL | (src << 32) | (rotate << 40);
if (vm_flags & RANDOMX_FLAG_AMD) {
*(uint64_t*)(p + pos + 6) = 0x742024443B0CE083ULL;
*(uint64_t*)(p + pos + 14) = 0x8900EB0414AE0F0AULL;
*(uint32_t*)(p + pos + 22) = 0x202444;
pos += 25;
if (RandomX_CurrentConfig.Tweak_V2_CFROUND) {
*(uint32_t*)(p + pos + 6) = 0x1375F0A8;
pos += 10;
}
else {
pos += 6;
}
*(uint64_t*)(p + pos) = 0x742024443B0CE083ULL;
*(uint64_t*)(p + pos + 8) = 0x8900EB0414AE0F0AULL;
*(uint32_t*)(p + pos + 16) = 0x202444;
pos += 19;
}
else {
*(uint64_t*)(p + pos + 6) = 0x0414AE0F0CE083ULL;
pos += 13;
pos += 6;
if (RandomX_CurrentConfig.Tweak_V2_CFROUND) {
if (BranchesWithin32B) {
const uint32_t branch_begin = static_cast<uint32_t>(pos + 2) & 31;
// If the jump crosses or touches 32-byte boundary, align it
if (branch_begin >= 30) {
const uint32_t alignment_size = 32 - branch_begin;
emit(NOPX[alignment_size - 1], alignment_size, code, pos);
}
}
*(uint32_t*)(p + pos) = 0x0775F0A8;
pos += 4;
}
*(uint64_t*)(p + pos) = 0x0414AE0F0CE083ULL;
pos += 7;
}
codePos = pos;