aaron/ispc
1
0
Fork 0
Code Issues Pull Requests Releases Wiki Activity

Merge pull request #1146 from Vsevolod-Livinskij/intr_pass

Browse Source 
Rearrange IntrinsicsOpt pass. This fixes bunch of fails, which show up from time to time due to memory corruption.
This commit is contained in:
Dmitry Babokin
2016-01-20 20:25:36 +03:00
parent 042c662d9d 6f886b2457
commit b4bbc7b1f3
1 changed files with 32 additions and 34 deletions
Download Patch File Download Diff File Expand all files Collapse all files

66
opt.cpp
View File

@@ -934,7 +934,7 @@ Optimize(llvm::Module *module, int optLevel) {
*/ */
class IntrinsicsOpt : public llvm::BasicBlockPass { class IntrinsicsOpt : public llvm::BasicBlockPass {
public: public:
IntrinsicsOpt(); IntrinsicsOpt() : BasicBlockPass(ID) {};
const char *getPassName() const { return "Intrinsics Cleanup Optimization"; } const char *getPassName() const { return "Intrinsics Cleanup Optimization"; }
bool runOnBasicBlock(llvm::BasicBlock &BB); bool runOnBasicBlock(llvm::BasicBlock &BB);
@@ -977,39 +977,6 @@ private:
char IntrinsicsOpt::ID = 0; char IntrinsicsOpt::ID = 0;
IntrinsicsOpt::IntrinsicsOpt()
: BasicBlockPass(ID) {
// All of the mask instructions we may encounter. Note that even if
// compiling for AVX, we may still encounter the regular 4-wide SSE
// MOVMSK instruction.
if (llvm::Function *ssei8Movmsk =
m->module->getFunction(llvm::Intrinsic::getName(llvm::Intrinsic::x86_sse2_pmovmskb_128))) {
maskInstructions.push_back(ssei8Movmsk);
}
if (llvm::Function *sseFloatMovmsk =
m->module->getFunction(llvm::Intrinsic::getName(llvm::Intrinsic::x86_sse_movmsk_ps))) {
maskInstructions.push_back(sseFloatMovmsk);
}
if (llvm::Function *__movmsk =
m->module->getFunction("__movmsk")) {
maskInstructions.push_back(__movmsk);
}
if (llvm::Function *avxFloatMovmsk =
m->module->getFunction(llvm::Intrinsic::getName(llvm::Intrinsic::x86_avx_movmsk_ps_256))) {
maskInstructions.push_back(avxFloatMovmsk);
}
// And all of the blend instructions
blendInstructions.push_back(BlendInstruction(
m->module->getFunction(llvm::Intrinsic::getName(llvm::Intrinsic::x86_sse41_blendvps)),
0xf, 0, 1, 2));
blendInstructions.push_back(BlendInstruction(
m->module->getFunction(llvm::Intrinsic::getName(llvm::Intrinsic::x86_avx_blendv_ps_256)),
0xff, 0, 1, 2));
}
/** Given an llvm::Value, return true if we can determine that it's an /** Given an llvm::Value, return true if we can determine that it's an
undefined value. This only makes a weak attempt at chasing this down, undefined value. This only makes a weak attempt at chasing this down,
only detecting flat-out undef values, and bitcasts of undef values. only detecting flat-out undef values, and bitcasts of undef values.
@@ -1036,6 +1003,37 @@ bool
IntrinsicsOpt::runOnBasicBlock(llvm::BasicBlock &bb) { IntrinsicsOpt::runOnBasicBlock(llvm::BasicBlock &bb) {
DEBUG_START_PASS("IntrinsicsOpt"); DEBUG_START_PASS("IntrinsicsOpt");
// We can’t initialize mask/blend function vector during pass initialization,
// as they may be optimized out by the time the pass is invoked.
// All of the mask instructions we may encounter. Note that even if
// compiling for AVX, we may still encounter the regular 4-wide SSE
// MOVMSK instruction.
if (llvm::Function *ssei8Movmsk =
m->module->getFunction(llvm::Intrinsic::getName(llvm::Intrinsic::x86_sse2_pmovmskb_128))) {
maskInstructions.push_back(ssei8Movmsk);
}
if (llvm::Function *sseFloatMovmsk =
m->module->getFunction(llvm::Intrinsic::getName(llvm::Intrinsic::x86_sse_movmsk_ps))) {
maskInstructions.push_back(sseFloatMovmsk);
}
if (llvm::Function *__movmsk =
m->module->getFunction("__movmsk")) {
maskInstructions.push_back(__movmsk);
}
if (llvm::Function *avxFloatMovmsk =
m->module->getFunction(llvm::Intrinsic::getName(llvm::Intrinsic::x86_avx_movmsk_ps_256))) {
maskInstructions.push_back(avxFloatMovmsk);
}
// And all of the blend instructions
blendInstructions.push_back(BlendInstruction(
m->module->getFunction(llvm::Intrinsic::getName(llvm::Intrinsic::x86_sse41_blendvps)),
0xf, 0, 1, 2));
blendInstructions.push_back(BlendInstruction(
m->module->getFunction(llvm::Intrinsic::getName(llvm::Intrinsic::x86_avx_blendv_ps_256)),
0xff, 0, 1, 2));
llvm::Function *avxMaskedLoad32 = llvm::Function *avxMaskedLoad32 =
m->module->getFunction(llvm::Intrinsic::getName(llvm::Intrinsic::x86_avx_maskload_ps_256)); m->module->getFunction(llvm::Intrinsic::getName(llvm::Intrinsic::x86_avx_maskload_ps_256));
llvm::Function *avxMaskedLoad64 = llvm::Function *avxMaskedLoad64 =