diff options
| m--------- | external/slang-binaries | 0 | ||||
| -rw-r--r-- | source/slang-glslang/slang-glslang.cpp | 393 |
2 files changed, 215 insertions, 178 deletions
diff --git a/external/slang-binaries b/external/slang-binaries -Subproject a1cfc4485fcfed4687afb99f51522fda7d3f24c +Subproject ef8bd16f0e40451d50b3b4761937d006e80ef5c diff --git a/source/slang-glslang/slang-glslang.cpp b/source/slang-glslang/slang-glslang.cpp index b82b21f20..e89581740 100644 --- a/source/slang-glslang/slang-glslang.cpp +++ b/source/slang-glslang/slang-glslang.cpp @@ -18,18 +18,8 @@ #include "spirv-tools/optimizer.hpp" #include "spirv-tools/libspirv.h" -#if 0 -#include <cstring> -#include <cstdlib> -#include <cctype> -#include <cmath> -#include <array> -#include <memory> -#include <thread> -#endif - #ifdef _WIN32 -#include <Windows.h> +# include <Windows.h> #endif #include <memory> @@ -113,42 +103,79 @@ static void dumpDiagnostics( dump(log.c_str(), log.length(), request.diagnosticFunc, request.diagnosticUserData, stderr); } -// Apply the SPIRV-Tools optimizer to generated SPIR-V based on the desired optimization level -// TODO: add flag for optimizing SPIR-V size as well -static void glslang_optimizeSPIRV(std::vector<unsigned int>& spirv, spv_target_env targetEnv, unsigned optimizationLevel, unsigned debugInfoType) +struct SPIRVOptimizationDiagnostic { - spvtools::Optimizer optimizer(targetEnv); - optimizer.SetMessageConsumer( - [](spv_message_level_t level, const char *source, const spv_position_t &position, const char *message) { - auto &out = std::cerr; + std::string toString() const + { + std::ostringstream out; + switch (level) { - case SPV_MSG_FATAL: - case SPV_MSG_INTERNAL_ERROR: - case SPV_MSG_ERROR: - out << "error: "; - break; - case SPV_MSG_WARNING: - out << "warning: "; - break; - case SPV_MSG_INFO: - case SPV_MSG_DEBUG: - out << "info: "; - break; - default: - break; + case SPV_MSG_FATAL: + case SPV_MSG_INTERNAL_ERROR: + case SPV_MSG_ERROR: + out << "error: "; + break; + case SPV_MSG_WARNING: + out << "warning: "; + break; + case SPV_MSG_INFO: + case SPV_MSG_DEBUG: + out << "info: "; + break; + default: + break; } - if (source) + if (source.length()) { out << source << ":"; } out << position.line << ":" << position.column << ":" << position.index << ":"; - if (message) + if (message.length()) { out << " " << message; } - out << std::endl; - }); + + return out.str(); + } + + spv_message_level_t level; + std::string source; + spv_position_t position; + std::string message; +}; + +// Apply the SPIRV-Tools optimizer to generated SPIR-V based on the desired optimization level +// TODO: add flag for optimizing SPIR-V size as well +static void glslang_optimizeSPIRV(spv_target_env targetEnv, const glslang_CompileRequest_1_1& request, std::vector<SPIRVOptimizationDiagnostic>& outDiags, std::vector<unsigned int>& ioSpirv) +{ + const auto optimizationLevel = request.optimizationLevel; + + // If there is no optimization then we are done + if (optimizationLevel == SLANG_OPTIMIZATION_LEVEL_NONE) + { + return; + } + + const auto debugInfoType = request.debugInfoType; + + spvtools::Optimizer optimizer(targetEnv); + + optimizer.SetMessageConsumer( + [&](spv_message_level_t level, const char* source, const spv_position_t& position, const char* message) { + SPIRVOptimizationDiagnostic diag; + diag.level = level; + if (source) + { + diag.source = source; + } + diag.position = position; + if (message) + { + diag.message = message; + } + outDiags.push_back(diag); + }); // If debug info is being generated, propagate // line information into all SPIR-V instructions. This avoids loss of @@ -175,16 +202,14 @@ static void glslang_optimizeSPIRV(std::vector<unsigned int>& spirv, spv_target_e // TODO confirm which passes we want to invoke for each level switch (optimizationLevel) { - case SLANG_OPTIMIZATION_LEVEL_NONE: - // Don't register any passes if our optimization level is none - break; - case SLANG_OPTIMIZATION_LEVEL_DEFAULT: - // Use a minimal set of performance settings - // If we run CreateInlineExhaustivePass, We need to run CreateMergeReturnPass first. + default: + case SLANG_OPTIMIZATION_LEVEL_DEFAULT: + { + // Use a minimal set of performance settings + // If we run CreateInlineExhaustivePass, We need to run CreateMergeReturnPass first. #if 0 - // This is the previous 'default optimization' passes setting for glslang - { + // This is the previous 'default optimization' passes setting for glslang optimizer.RegisterPass(spvtools::CreateMergeReturnPass()); optimizer.RegisterPass(spvtools::CreateInlineExhaustivePass()); optimizer.RegisterPass(spvtools::CreateAggressiveDCEPass()); @@ -192,21 +217,20 @@ static void glslang_optimizeSPIRV(std::vector<unsigned int>& spirv, spv_target_e optimizer.RegisterPass(spvtools::CreateScalarReplacementPass(100)); optimizer.RegisterPass(spvtools::CreateLocalAccessChainConvertPass()); optimizer.RegisterPass(spvtools::CreateAggressiveDCEPass()); - } #elif 1 - // 6Mb 27 secs (all passes up to 9) - // 9Mb 25 secs (all passes up to 7) - // 8Mb 15 secs (all passes) -(5,6,7) - // 6Mb 15 secs (all passes) -(6,7) - - // This list of passes takes the previous 'default optimization' - // passes (as listed above) and tries to combine them in order with the 'new' passes below. - // The issue with the passes below is that although it produces smaller SPIR-V fairly quickly - // it can cause serious problem on some drivers. - // - // Across a wide range of compilations this produced SPIR-V that is less than half size of the - // previous -O1 passes above. - { + // 6Mb 27 secs (all passes up to 9) + // 9Mb 25 secs (all passes up to 7) + // 8Mb 15 secs (all passes) -(5,6,7) + // 6Mb 15 secs (all passes) -(6,7) + + // This list of passes takes the previous 'default optimization' + // passes (as listed above) and tries to combine them in order with the 'new' passes below. + // The issue with the passes below is that although it produces smaller SPIR-V fairly quickly + // it can cause serious problem on some drivers. + // + // Across a wide range of compilations this produced SPIR-V that is less than half size of the + // previous -O1 passes above. + optimizer.RegisterPass(spvtools::CreateWrapOpKillPass()); // 1 optimizer.RegisterPass(spvtools::CreateDeadBranchElimPass()); // 2 @@ -232,15 +256,15 @@ static void glslang_optimizeSPIRV(std::vector<unsigned int>& spirv, spv_target_e optimizer.RegisterPass(spvtools::CreateAggressiveDCEPass()); optimizer.RegisterPass(spvtools::CreateVectorDCEPass()); // 9 - } + #else - // The following selection of passes was created by - // 1) Taking the list of passes from optimizer.RegisterSizePasses - // 2) Disable/enable passes to try to produce some reasonable combination of low SPIR-V output size and compilation speed - // - // For a particularly difficult glsl shader this produced 1/3 SPIR-V code (against previous -O1), in around 13th the time (against -O3 option) - // Over a wide range of compiles the SPIR-V is around 6% larger than -O3 - { + // The following selection of passes was created by + // 1) Taking the list of passes from optimizer.RegisterSizePasses + // 2) Disable/enable passes to try to produce some reasonable combination of low SPIR-V output size and compilation speed + // + // For a particularly difficult glsl shader this produced 1/3 SPIR-V code (against previous -O1), in around 13th the time (against -O3 option) + // Over a wide range of compiles the SPIR-V is around 6% larger than -O3 + // The following comments describe the path to finding this combination. The original compilation produces 18Mb SPIR-V binaries // in around 3 1/2 mins. The integer number increases with the ordering of the test. // @@ -292,81 +316,80 @@ static void glslang_optimizeSPIRV(std::vector<unsigned int>& spirv, spv_target_e optimizer.RegisterPass(spvtools::CreateSimplificationPass()); // 14 optimizer.RegisterPass(spvtools::CreateAggressiveDCEPass()); optimizer.RegisterPass(spvtools::CreateCFGCleanupPass()); - } #endif - break; - case SLANG_OPTIMIZATION_LEVEL_HIGH: - + break; + } // TODO(JS): It would be better if we had some distinction here where 'high' meant optimize 'in a reasonable time' for // a better optimization, and 'maximal' meant compilation might take a really long time... so only use it if it's really // needed. // - // Currently we just have high have the same meaning as 'maximal'. + // Currently we just have high have the same meaning as 'maximal'. + case SLANG_OPTIMIZATION_LEVEL_HIGH: + case SLANG_OPTIMIZATION_LEVEL_MAXIMAL: + { + // Use the same passes when specifying the "-O" flag in spirv-opt + // Roughly equivalent to `RegisterPerformancePasses` - case SLANG_OPTIMIZATION_LEVEL_MAXIMAL: - // Use the same passes when specifying the "-O" flag in spirv-opt - // Roughly equivalent to `RegisterPerformancePasses` + optimizer.RegisterPass(spvtools::CreateWrapOpKillPass()); + optimizer.RegisterPass(spvtools::CreateDeadBranchElimPass()); + optimizer.RegisterPass(spvtools::CreateMergeReturnPass()); + optimizer.RegisterPass(spvtools::CreateInlineExhaustivePass()); + optimizer.RegisterPass(spvtools::CreateEliminateDeadFunctionsPass()); + optimizer.RegisterPass(spvtools::CreateAggressiveDCEPass()); + optimizer.RegisterPass(spvtools::CreatePrivateToLocalPass()); + optimizer.RegisterPass(spvtools::CreateLocalSingleBlockLoadStoreElimPass()); + optimizer.RegisterPass(spvtools::CreateLocalSingleStoreElimPass()); + optimizer.RegisterPass(spvtools::CreateAggressiveDCEPass()); + optimizer.RegisterPass(spvtools::CreateScalarReplacementPass()); + optimizer.RegisterPass(spvtools::CreateLocalAccessChainConvertPass()); + optimizer.RegisterPass(spvtools::CreateLocalSingleBlockLoadStoreElimPass()); + optimizer.RegisterPass(spvtools::CreateLocalSingleStoreElimPass()); + optimizer.RegisterPass(spvtools::CreateAggressiveDCEPass()); - - optimizer.RegisterPass(spvtools::CreateWrapOpKillPass()); - optimizer.RegisterPass(spvtools::CreateDeadBranchElimPass()); - optimizer.RegisterPass(spvtools::CreateMergeReturnPass()); - optimizer.RegisterPass(spvtools::CreateInlineExhaustivePass()); - optimizer.RegisterPass(spvtools::CreateEliminateDeadFunctionsPass()); - optimizer.RegisterPass(spvtools::CreateAggressiveDCEPass()); - optimizer.RegisterPass(spvtools::CreatePrivateToLocalPass()); - optimizer.RegisterPass(spvtools::CreateLocalSingleBlockLoadStoreElimPass()); - optimizer.RegisterPass(spvtools::CreateLocalSingleStoreElimPass()); - optimizer.RegisterPass(spvtools::CreateAggressiveDCEPass()); - optimizer.RegisterPass(spvtools::CreateScalarReplacementPass()); - optimizer.RegisterPass(spvtools::CreateLocalAccessChainConvertPass()); - optimizer.RegisterPass(spvtools::CreateLocalSingleBlockLoadStoreElimPass()); - optimizer.RegisterPass(spvtools::CreateLocalSingleStoreElimPass()); - optimizer.RegisterPass(spvtools::CreateAggressiveDCEPass()); - - // We run CompactIdsPass here, because CreateLocalMultiStoreElimPass can explode - // id usage (by a factor of 10), and compacting ids here has been shown to half - // id usage with a complex shader. - optimizer.RegisterPass(spvtools::CreateCompactIdsPass()); - - // Note that CreateLocalMultiStoreElimPass really just does a SSARewritePass - optimizer.RegisterPass(spvtools::CreateLocalMultiStoreElimPass()); - - optimizer.RegisterPass(spvtools::CreateAggressiveDCEPass()); - optimizer.RegisterPass(spvtools::CreateCCPPass()); - optimizer.RegisterPass(spvtools::CreateAggressiveDCEPass()); - optimizer.RegisterPass(spvtools::CreateLoopUnrollPass(true)); - optimizer.RegisterPass(spvtools::CreateDeadBranchElimPass()); - optimizer.RegisterPass(spvtools::CreateRedundancyEliminationPass()); - optimizer.RegisterPass(spvtools::CreateCombineAccessChainsPass()); - optimizer.RegisterPass(spvtools::CreateSimplificationPass()); - optimizer.RegisterPass(spvtools::CreateScalarReplacementPass()); - optimizer.RegisterPass(spvtools::CreateLocalAccessChainConvertPass()); - optimizer.RegisterPass(spvtools::CreateLocalSingleBlockLoadStoreElimPass()); - optimizer.RegisterPass(spvtools::CreateLocalSingleStoreElimPass()); - optimizer.RegisterPass(spvtools::CreateAggressiveDCEPass()); - optimizer.RegisterPass(spvtools::CreateSSARewritePass()); - optimizer.RegisterPass(spvtools::CreateAggressiveDCEPass()); - optimizer.RegisterPass(spvtools::CreateVectorDCEPass()); - optimizer.RegisterPass(spvtools::CreateDeadInsertElimPass()); - optimizer.RegisterPass(spvtools::CreateDeadBranchElimPass()); - optimizer.RegisterPass(spvtools::CreateSimplificationPass()); - optimizer.RegisterPass(spvtools::CreateIfConversionPass()); - optimizer.RegisterPass(spvtools::CreateCopyPropagateArraysPass()); - optimizer.RegisterPass(spvtools::CreateReduceLoadSizePass()); - optimizer.RegisterPass(spvtools::CreateAggressiveDCEPass()); - optimizer.RegisterPass(spvtools::CreateBlockMergePass()); - optimizer.RegisterPass(spvtools::CreateRedundancyEliminationPass()); - optimizer.RegisterPass(spvtools::CreateDeadBranchElimPass()); - optimizer.RegisterPass(spvtools::CreateBlockMergePass()); - optimizer.RegisterPass(spvtools::CreateSimplificationPass()); - - // We again run compaction to try and ensure the final output uses ids that are in range. - // On a complex shader, this reduced the amount ids by 5. - optimizer.RegisterPass(spvtools::CreateCompactIdsPass()); - - break; + // We run CompactIdsPass here, because CreateLocalMultiStoreElimPass can explode + // id usage (by a factor of 10), and compacting ids here has been shown to half + // id usage with a complex shader. + optimizer.RegisterPass(spvtools::CreateCompactIdsPass()); + + // Note that CreateLocalMultiStoreElimPass really just does a SSARewritePass + optimizer.RegisterPass(spvtools::CreateLocalMultiStoreElimPass()); + + optimizer.RegisterPass(spvtools::CreateAggressiveDCEPass()); + optimizer.RegisterPass(spvtools::CreateCCPPass()); + optimizer.RegisterPass(spvtools::CreateAggressiveDCEPass()); + optimizer.RegisterPass(spvtools::CreateLoopUnrollPass(true)); + optimizer.RegisterPass(spvtools::CreateDeadBranchElimPass()); + optimizer.RegisterPass(spvtools::CreateRedundancyEliminationPass()); + optimizer.RegisterPass(spvtools::CreateCombineAccessChainsPass()); + optimizer.RegisterPass(spvtools::CreateSimplificationPass()); + optimizer.RegisterPass(spvtools::CreateScalarReplacementPass()); + optimizer.RegisterPass(spvtools::CreateLocalAccessChainConvertPass()); + optimizer.RegisterPass(spvtools::CreateLocalSingleBlockLoadStoreElimPass()); + optimizer.RegisterPass(spvtools::CreateLocalSingleStoreElimPass()); + optimizer.RegisterPass(spvtools::CreateAggressiveDCEPass()); + optimizer.RegisterPass(spvtools::CreateSSARewritePass()); + optimizer.RegisterPass(spvtools::CreateAggressiveDCEPass()); + optimizer.RegisterPass(spvtools::CreateVectorDCEPass()); + optimizer.RegisterPass(spvtools::CreateDeadInsertElimPass()); + optimizer.RegisterPass(spvtools::CreateDeadBranchElimPass()); + optimizer.RegisterPass(spvtools::CreateSimplificationPass()); + optimizer.RegisterPass(spvtools::CreateIfConversionPass()); + optimizer.RegisterPass(spvtools::CreateCopyPropagateArraysPass()); + optimizer.RegisterPass(spvtools::CreateReduceLoadSizePass()); + optimizer.RegisterPass(spvtools::CreateAggressiveDCEPass()); + optimizer.RegisterPass(spvtools::CreateBlockMergePass()); + optimizer.RegisterPass(spvtools::CreateRedundancyEliminationPass()); + optimizer.RegisterPass(spvtools::CreateDeadBranchElimPass()); + optimizer.RegisterPass(spvtools::CreateBlockMergePass()); + optimizer.RegisterPass(spvtools::CreateSimplificationPass()); + + // We again run compaction to try and ensure the final output uses ids that are in range. + // On a complex shader, this reduced the amount ids by 5. + optimizer.RegisterPass(spvtools::CreateCompactIdsPass()); + + break; + } } if (debugInfoType != SLANG_DEBUG_INFO_LEVEL_NONE) @@ -374,11 +397,17 @@ static void glslang_optimizeSPIRV(std::vector<unsigned int>& spirv, spv_target_e optimizer.RegisterPass(spvtools::CreateRedundantLineInfoElimPass()); } - spvOptOptions.set_run_validator(false); // Don't run the validator by default - optimizer.Run(spirv.data(), spirv.size(), &spirv, spvOptOptions); -} + // Put the output optimized spirv into optSpirv + std::vector<unsigned int> optSpirv; + + // Optimize + optimizer.Run(ioSpirv.data(), ioSpirv.size(), &optSpirv, spvOptOptions); + + // Make the ioSpirv the optimized spirv + ioSpirv.swap(optSpirv); +} static glslang::EShTargetLanguageVersion _makeTargetLanguageVersion(int majorVersion, int minorVersion) { @@ -505,7 +534,6 @@ static int glslang_compileGLSLToSPIRV(const glslang_CompileRequest_1_1& request) return 1; } - spv_target_env targetEnv = SPV_ENV_UNIVERSAL_1_2; glslang::EShTargetLanguageVersion targetLanguage = glslang::EShTargetLanguageVersion(0); @@ -563,26 +591,28 @@ static int glslang_compileGLSLToSPIRV(const glslang_CompileRequest_1_1& request) &request.sourcePath, 1); - EShMessages messages = EShMessages(EShMsgSpvRules | EShMsgVulkanRules); - - if( !shader->parse(&gResources, 110, false, messages) ) { - dumpDiagnostics(request, shader->getInfoLog()); - return 1; - } + const EShMessages messages = EShMessages(EShMsgSpvRules | EShMsgVulkanRules); - program->addShader(shader); + if (!shader->parse(&gResources, 110, false, messages)) + { + dumpDiagnostics(request, shader->getInfoLog()); + return 1; + } - if( !program->link(messages) ) - { - dumpDiagnostics(request, program->getInfoLog()); - return 1; - } + program->addShader(shader); - if( !program->mapIO() ) - { - dumpDiagnostics(request, program->getInfoLog()); - return 1; + if (!program->link(messages)) + { + dumpDiagnostics(request, program->getInfoLog()); + return 1; + } + + if (!program->mapIO()) + { + dumpDiagnostics(request, program->getInfoLog()); + return 1; + } } for(int stage = 0; stage < EShLangCount; ++stage) @@ -592,18 +622,48 @@ static int glslang_compileGLSLToSPIRV(const glslang_CompileRequest_1_1& request) continue; std::vector<unsigned int> spirv; - std::string warningsErrors; spv::SpvBuildLogger logger; + glslang::GlslangToSpv(*stageIntermediate, spirv, &logger); + int optErrorCount = 0; + if (request.optimizationLevel != SLANG_OPTIMIZATION_LEVEL_NONE) { - glslang_optimizeSPIRV(spirv, targetEnv, request.optimizationLevel, request.debugInfoType); + std::vector<SPIRVOptimizationDiagnostic> optDiags; + glslang_optimizeSPIRV(targetEnv, request, optDiags, spirv); + + { + for (const auto& diag : optDiags) + { + // TODO(JS): + // Hack to stop CapabilityRayTracingMotionBlurNV outputting an error + if (diag.level == SPV_MSG_ERROR && diag.message == "Invalid capability operand: 5341") + { + continue; + } + + // Count the number of errors + optErrorCount += int(diag.level <= SPV_MSG_ERROR); + + // Note this string does not have \n. + std::string diagString = diag.toString(); + + // Dump + dump(diagString.c_str(), diagString.length(), request.diagnosticFunc, request.diagnosticUserData, stderr); + } + } } dumpDiagnostics(request, logger.getAllMessages()); dump(spirv.data(), spirv.size() * sizeof(unsigned int), request.outputFunc, request.outputUserData, stdout); + + if (optErrorCount > 0) + { + // It's an error... + return 1; + } } return 0; @@ -740,26 +800,3 @@ int glslang_compile(glslang_CompileRequest_1_0* inRequest) request.set(*inRequest); return glslang_compile_1_1(&request); } - -#if 0 -static std::mutex g_globalMutex; - -namespace glslang { - -void InitGlobalLock() -{ - -} - -void GetGlobalLock() -{ - g_globalMutex.lock(); -} - -void ReleaseGlobalLock() -{ - g_globalMutex.unlock(); -} - -} // namespace glslang -#endif |