diff options
Diffstat (limited to 'src')
-rw-r--r-- | src/database/database.cpp | 6 | ||||
-rw-r--r-- | src/database/database.hpp | 4 | ||||
-rw-r--r-- | src/database/kernels/xgemm.hpp | 12 | ||||
-rw-r--r-- | src/database/kernels/xgemv.hpp | 200 | ||||
-rw-r--r-- | src/database/kernels/xgemv_fast.hpp | 247 | ||||
-rw-r--r-- | src/database/kernels/xgemv_fast_rot.hpp | 245 | ||||
-rw-r--r-- | src/routines/level2/xgemv.cpp | 2 | ||||
-rw-r--r-- | src/tuning/kernels/xgemv.cpp | 2 |
8 files changed, 602 insertions, 116 deletions
diff --git a/src/database/database.cpp b/src/database/database.cpp index 47f1da16..28124455 100644 --- a/src/database/database.cpp +++ b/src/database/database.cpp @@ -17,6 +17,8 @@ #include "database/kernels/xaxpy.hpp" #include "database/kernels/xdot.hpp" #include "database/kernels/xgemv.hpp" +#include "database/kernels/xgemv_fast.hpp" +#include "database/kernels/xgemv_fast_rot.hpp" #include "database/kernels/xger.hpp" #include "database/kernels/xgemm.hpp" #include "database/kernels/copy.hpp" @@ -32,8 +34,10 @@ const std::vector<Database::DatabaseEntry> Database::database = { XaxpyHalf, XaxpySingle, XaxpyDouble, XaxpyComplexSingle, XaxpyComplexDouble, XdotHalf, XdotSingle, XdotDouble, XdotComplexSingle, XdotComplexDouble, XgemvHalf, XgemvSingle, XgemvDouble, XgemvComplexSingle, XgemvComplexDouble, + XgemvFastHalf, XgemvFastSingle, XgemvFastDouble, XgemvFastComplexSingle, XgemvFastComplexDouble, + XgemvFastRotHalf, XgemvFastRotSingle, XgemvFastRotDouble, XgemvFastRotComplexSingle, XgemvFastRotComplexDouble, XgerHalf, XgerSingle, XgerDouble, XgerComplexSingle, XgerComplexDouble, - XgemmHalf, XgemmSingle, XgemmDouble, XgemmComplexSingle, XgemmComplexDouble, + /* XgemmHalf, */ XgemmSingle, XgemmDouble, XgemmComplexSingle, XgemmComplexDouble, CopyHalf, CopySingle, CopyDouble, CopyComplexSingle, CopyComplexDouble, PadHalf, PadSingle, PadDouble, PadComplexSingle, PadComplexDouble, TransposeHalf, TransposeSingle, TransposeDouble, TransposeComplexSingle, TransposeComplexDouble, diff --git a/src/database/database.hpp b/src/database/database.hpp index e84357dc..2fd96411 100644 --- a/src/database/database.hpp +++ b/src/database/database.hpp @@ -71,8 +71,10 @@ class Database { static const DatabaseEntry XaxpyHalf, XaxpySingle, XaxpyDouble, XaxpyComplexSingle, XaxpyComplexDouble; static const DatabaseEntry XdotHalf, XdotSingle, XdotDouble, XdotComplexSingle, XdotComplexDouble; static const DatabaseEntry XgemvHalf, XgemvSingle, XgemvDouble, XgemvComplexSingle, XgemvComplexDouble; + static const DatabaseEntry XgemvFastHalf, XgemvFastSingle, XgemvFastDouble, XgemvFastComplexSingle, XgemvFastComplexDouble; + static const DatabaseEntry XgemvFastRotHalf, XgemvFastRotSingle, XgemvFastRotDouble, XgemvFastRotComplexSingle, XgemvFastRotComplexDouble; static const DatabaseEntry XgerHalf, XgerSingle, XgerDouble, XgerComplexSingle, XgerComplexDouble; - static const DatabaseEntry XgemmHalf, XgemmSingle, XgemmDouble, XgemmComplexSingle, XgemmComplexDouble; + static const DatabaseEntry /* XgemmHalf, */ XgemmSingle, XgemmDouble, XgemmComplexSingle, XgemmComplexDouble; static const DatabaseEntry CopyHalf, CopySingle, CopyDouble, CopyComplexSingle, CopyComplexDouble; static const DatabaseEntry PadHalf, PadSingle, PadDouble, PadComplexSingle, PadComplexDouble; static const DatabaseEntry TransposeHalf, TransposeSingle, TransposeDouble, TransposeComplexSingle, TransposeComplexDouble; diff --git a/src/database/kernels/xgemm.hpp b/src/database/kernels/xgemm.hpp index 736f2695..61b7ff05 100644 --- a/src/database/kernels/xgemm.hpp +++ b/src/database/kernels/xgemm.hpp @@ -14,18 +14,6 @@ namespace clblast { // ================================================================================================= -const Database::DatabaseEntry Database::XgemmHalf = { - "Xgemm", Precision::kHalf, { - { // Default - kDeviceTypeAll, "default", { - { "default", { {"KWG",16}, {"KWI",2}, {"MDIMA",8}, {"MDIMC",8}, {"MWG",32}, {"NDIMB",8}, {"NDIMC",8}, {"NWG",64}, {"SA",0}, {"SB",0}, {"STRM",0}, {"STRN",0}, {"VWM",1}, {"VWN",1} } }, - } - }, - } -}; - -// ================================================================================================= - const Database::DatabaseEntry Database::XgemmSingle = { "Xgemm", Precision::kSingle, { { // AMD GPUs diff --git a/src/database/kernels/xgemv.hpp b/src/database/kernels/xgemv.hpp index 3aa1863f..6d680b06 100644 --- a/src/database/kernels/xgemv.hpp +++ b/src/database/kernels/xgemv.hpp @@ -18,13 +18,13 @@ const Database::DatabaseEntry Database::XgemvHalf = { "Xgemv", Precision::kHalf, { { // Intel GPUs kDeviceTypeGPU, "Intel", { - { "Intel(R) HD Graphics Skylake ULT GT2", { {"WGS1",128}, {"WPT1",1}, {"VW2",2}, {"WGS2",128}, {"WPT2",2}, {"VW3",1}, {"WGS3",64}, {"WPT3",1} } }, - { "default", { {"WGS1",128}, {"WPT1",1}, {"VW2",2}, {"WGS2",128}, {"WPT2",2}, {"VW3",1}, {"WGS3",64}, {"WPT3",1} } }, + { "Intel(R) HD Graphics Skylake ULT GT2", { {"WGS1",128}, {"WPT1",1} } }, + { "default", { {"WGS1",128}, {"WPT1",1} } }, } }, { // Default kDeviceTypeAll, "default", { - { "default", { {"WGS1",128}, {"WPT1",1}, {"VW2",2}, {"WGS2",128}, {"WPT2",2}, {"VW3",1}, {"WGS3",64}, {"WPT3",1} } }, + { "default", { {"WGS1",128}, {"WPT1",1} } }, } }, } @@ -36,57 +36,57 @@ const Database::DatabaseEntry Database::XgemvSingle = { "Xgemv", Precision::kSingle, { { // AMD GPUs kDeviceTypeGPU, "AMD", { - { "AMD Radeon R9 M370X Compute Engine", { {"WGS1",128}, {"WPT1",1}, {"VW2",1}, {"WGS2",128}, {"WPT2",1}, {"VW3",8}, {"WGS3",16}, {"WPT3",16} } }, - { "Hawaii", { {"WGS1",128}, {"WPT1",1}, {"VW2",1}, {"WGS2",64}, {"WPT2",1}, {"VW3",1}, {"WGS3",64}, {"WPT3",1} } }, - { "Oland", { {"WGS1",128}, {"WPT1",1}, {"VW2",1}, {"WGS2",64}, {"WPT2",1}, {"VW3",4}, {"WGS3",256}, {"WPT3",4} } }, - { "Pitcairn", { {"WGS1",256}, {"WPT1",1}, {"VW2",1}, {"WGS2",64}, {"WPT2",1}, {"VW3",1}, {"WGS3",64}, {"WPT3",1} } }, - { "Tahiti", { {"WGS1",256}, {"WPT1",1}, {"VW2",1}, {"WGS2",64}, {"WPT2",1}, {"VW3",1}, {"WGS3",64}, {"WPT3",1} } }, - { "default", { {"WGS1",128}, {"WPT1",1}, {"VW2",1}, {"WGS2",64}, {"WPT2",1}, {"VW3",1}, {"WGS3",64}, {"WPT3",1} } }, + { "AMD Radeon R9 M370X Compute Engine", { {"WGS1",128}, {"WPT1",1} } }, + { "Hawaii", { {"WGS1",128}, {"WPT1",1} } }, + { "Oland", { {"WGS1",128}, {"WPT1",1} } }, + { "Pitcairn", { {"WGS1",256}, {"WPT1",1} } }, + { "Tahiti", { {"WGS1",256}, {"WPT1",1} } }, + { "default", { {"WGS1",128}, {"WPT1",1} } }, } }, { // Intel CPUs kDeviceTypeCPU, "Intel", { - { "Intel(R) Core(TM) i5-6200U CPU @ 2.30GHz", { {"WGS1",64}, {"WPT1",1}, {"VW2",4}, {"WGS2",128}, {"WPT2",4}, {"VW3",1}, {"WGS3",64}, {"WPT3",1} } }, - { "Intel(R) Core(TM) i7-3770 CPU @ 3.40GHz", { {"WGS1",64}, {"WPT1",4}, {"VW2",1}, {"WGS2",64}, {"WPT2",4}, {"VW3",2}, {"WGS3",64}, {"WPT3",4} } }, - { "default", { {"WGS1",64}, {"WPT1",1}, {"VW2",1}, {"WGS2",64}, {"WPT2",4}, {"VW3",1}, {"WGS3",64}, {"WPT3",1} } }, + { "Intel(R) Core(TM) i5-6200U CPU @ 2.30GHz", { {"WGS1",64}, {"WPT1",1} } }, + { "Intel(R) Core(TM) i7-3770 CPU @ 3.40GHz", { {"WGS1",64}, {"WPT1",4} } }, + { "default", { {"WGS1",64}, {"WPT1",1} } }, } }, { // Intel GPUs kDeviceTypeGPU, "Intel", { - { "Intel(R) HD Graphics 530", { {"WGS1",256}, {"WPT1",1}, {"VW2",1}, {"WGS2",256}, {"WPT2",1}, {"VW3",1}, {"WGS3",64}, {"WPT3",1} } }, - { "Intel(R) HD Graphics Haswell Ultrabook GT2 Mobile", { {"WGS1",256}, {"WPT1",1}, {"VW2",4}, {"WGS2",128}, {"WPT2",4}, {"VW3",4}, {"WGS3",256}, {"WPT3",4} } }, - { "Intel(R) HD Graphics Skylake ULT GT2", { {"WGS1",64}, {"WPT1",1}, {"VW2",1}, {"WGS2",256}, {"WPT2",1}, {"VW3",4}, {"WGS3",64}, {"WPT3",4} } }, - { "Iris", { {"WGS1",64}, {"WPT1",2}, {"VW2",1}, {"WGS2",128}, {"WPT2",2}, {"VW3",4}, {"WGS3",64}, {"WPT3",8} } }, - { "Iris Pro", { {"WGS1",256}, {"WPT1",2}, {"VW2",1}, {"WGS2",128}, {"WPT2",2}, {"VW3",4}, {"WGS3",64}, {"WPT3",4} } }, - { "default", { {"WGS1",64}, {"WPT1",1}, {"VW2",1}, {"WGS2",128}, {"WPT2",1}, {"VW3",1}, {"WGS3",64}, {"WPT3",1} } }, + { "Intel(R) HD Graphics 530", { {"WGS1",256}, {"WPT1",1} } }, + { "Intel(R) HD Graphics Haswell Ultrabook GT2 Mobile", { {"WGS1",256}, {"WPT1",1} } }, + { "Intel(R) HD Graphics Skylake ULT GT2", { {"WGS1",64}, {"WPT1",1} } }, + { "Iris", { {"WGS1",64}, {"WPT1",2} } }, + { "Iris Pro", { {"WGS1",256}, {"WPT1",2} } }, + { "default", { {"WGS1",64}, {"WPT1",1} } }, } }, { // Intel accelerators kDeviceTypeAccelerator, "Intel", { - { "Intel(R) Many Integrated Core Acceleration Card", { {"WGS1",64}, {"WPT1",1}, {"VW2",1}, {"WGS2",64}, {"WPT2",1}, {"VW3",1}, {"WGS3",64}, {"WPT3",1} } }, - { "default", { {"WGS1",64}, {"WPT1",1}, {"VW2",1}, {"WGS2",64}, {"WPT2",1}, {"VW3",1}, {"WGS3",64}, {"WPT3",1} } }, + { "Intel(R) Many Integrated Core Acceleration Card", { {"WGS1",64}, {"WPT1",1} } }, + { "default", { {"WGS1",64}, {"WPT1",1} } }, } }, { // NVIDIA GPUs kDeviceTypeGPU, "NVIDIA", { - { "GRID K520", { {"WGS1",256}, {"WPT1",1}, {"VW2",2}, {"WGS2",256}, {"WPT2",2}, {"VW3",2}, {"WGS3",64}, {"WPT3",2} } }, - { "GeForce GTX 1070", { {"WGS1",128}, {"WPT1",1}, {"VW2",1}, {"WGS2",256}, {"WPT2",1}, {"VW3",2}, {"WGS3",64}, {"WPT3",2} } }, - { "GeForce GTX 480", { {"WGS1",64}, {"WPT1",1}, {"VW2",1}, {"WGS2",128}, {"WPT2",1}, {"VW3",2}, {"WGS3",64}, {"WPT3",2} } }, - { "GeForce GTX 670", { {"WGS1",64}, {"WPT1",1}, {"VW2",2}, {"WGS2",256}, {"WPT2",2}, {"VW3",2}, {"WGS3",64}, {"WPT3",2} } }, - { "GeForce GTX 680", { {"WGS1",256}, {"WPT1",1}, {"VW2",1}, {"WGS2",128}, {"WPT2",1}, {"VW3",2}, {"WGS3",128}, {"WPT3",2} } }, - { "GeForce GTX 750", { {"WGS1",256}, {"WPT1",1}, {"VW2",1}, {"WGS2",256}, {"WPT2",1}, {"VW3",2}, {"WGS3",128}, {"WPT3",2} } }, - { "GeForce GTX 750 Ti", { {"WGS1",256}, {"WPT1",1}, {"VW2",1}, {"WGS2",256}, {"WPT2",1}, {"VW3",4}, {"WGS3",128}, {"WPT3",4} } }, - { "GeForce GTX 980", { {"WGS1",128}, {"WPT1",1}, {"VW2",1}, {"WGS2",256}, {"WPT2",1}, {"VW3",1}, {"WGS3",128}, {"WPT3",1} } }, - { "GeForce GTX TITAN", { {"WGS1",256}, {"WPT1",1}, {"VW2",1}, {"WGS2",256}, {"WPT2",1}, {"VW3",1}, {"WGS3",256}, {"WPT3",1} } }, - { "GeForce GTX TITAN X", { {"WGS1",256}, {"WPT1",1}, {"VW2",1}, {"WGS2",64}, {"WPT2",1}, {"VW3",1}, {"WGS3",128}, {"WPT3",1} } }, - { "Tesla K20m", { {"WGS1",128}, {"WPT1",1}, {"VW2",1}, {"WGS2",256}, {"WPT2",1}, {"VW3",1}, {"WGS3",256}, {"WPT3",1} } }, - { "Tesla K40m", { {"WGS1",256}, {"WPT1",1}, {"VW2",1}, {"WGS2",256}, {"WPT2",1}, {"VW3",1}, {"WGS3",128}, {"WPT3",1} } }, - { "default", { {"WGS1",64}, {"WPT1",1}, {"VW2",1}, {"WGS2",64}, {"WPT2",1}, {"VW3",1}, {"WGS3",64}, {"WPT3",1} } }, + { "GRID K520", { {"WGS1",256}, {"WPT1",1} } }, + { "GeForce GTX 1070", { {"WGS1",128}, {"WPT1",1} } }, + { "GeForce GTX 480", { {"WGS1",64}, {"WPT1",1} } }, + { "GeForce GTX 670", { {"WGS1",64}, {"WPT1",1} } }, + { "GeForce GTX 680", { {"WGS1",256}, {"WPT1",1} } }, + { "GeForce GTX 750", { {"WGS1",256}, {"WPT1",1} } }, + { "GeForce GTX 750 Ti", { {"WGS1",256}, {"WPT1",1} } }, + { "GeForce GTX 980", { {"WGS1",128}, {"WPT1",1} } }, + { "GeForce GTX TITAN", { {"WGS1",256}, {"WPT1",1} } }, + { "GeForce GTX TITAN X", { {"WGS1",256}, {"WPT1",1} } }, + { "Tesla K20m", { {"WGS1",128}, {"WPT1",1} } }, + { "Tesla K40m", { {"WGS1",256}, {"WPT1",1} } }, + { "default", { {"WGS1",64}, {"WPT1",1} } }, } }, { // Default kDeviceTypeAll, "default", { - { "default", { {"WGS1",64}, {"WPT1",1}, {"VW2",1}, {"WGS2",64}, {"WPT2",1}, {"VW3",1}, {"WGS3",64}, {"WPT3",1} } }, + { "default", { {"WGS1",64}, {"WPT1",1} } }, } }, } @@ -98,53 +98,53 @@ const Database::DatabaseEntry Database::XgemvComplexSingle = { "Xgemv", Precision::kComplexSingle, { { // AMD GPUs kDeviceTypeGPU, "AMD", { - { "AMD Radeon R9 M370X Compute Engine", { {"WGS1",64}, {"WPT1",1}, {"VW2",2}, {"WGS2",256}, {"WPT2",2}, {"VW3",1}, {"WGS3",128}, {"WPT3",1} } }, - { "Hawaii", { {"WGS1",64}, {"WPT1",1}, {"VW2",1}, {"WGS2",64}, {"WPT2",1}, {"VW3",1}, {"WGS3",64}, {"WPT3",1} } }, - { "Oland", { {"WGS1",64}, {"WPT1",1}, {"VW2",1}, {"WGS2",64}, {"WPT2",1}, {"VW3",4}, {"WGS3",64}, {"WPT3",4} } }, - { "Pitcairn", { {"WGS1",64}, {"WPT1",1}, {"VW2",1}, {"WGS2",64}, {"WPT2",1}, {"VW3",2}, {"WGS3",64}, {"WPT3",2} } }, - { "Tahiti", { {"WGS1",64}, {"WPT1",1}, {"VW2",1}, {"WGS2",128}, {"WPT2",1}, {"VW3",1}, {"WGS3",64}, {"WPT3",1} } }, - { "default", { {"WGS1",64}, {"WPT1",1}, {"VW2",1}, {"WGS2",64}, {"WPT2",1}, {"VW3",1}, {"WGS3",64}, {"WPT3",1} } }, + { "AMD Radeon R9 M370X Compute Engine", { {"WGS1",64}, {"WPT1",1} } }, + { "Hawaii", { {"WGS1",64}, {"WPT1",1} } }, + { "Oland", { {"WGS1",64}, {"WPT1",1} } }, + { "Pitcairn", { {"WGS1",64}, {"WPT1",1} } }, + { "Tahiti", { {"WGS1",64}, {"WPT1",1} } }, + { "default", { {"WGS1",64}, {"WPT1",1} } }, } }, { // Intel CPUs kDeviceTypeCPU, "Intel", { - { "Intel(R) Core(TM) i5-6200U CPU @ 2.30GHz", { {"WGS1",128}, {"WPT1",1}, {"VW2",1}, {"WGS2",128}, {"WPT2",2}, {"VW3",1}, {"WGS3",64}, {"WPT3",1} } }, - { "Intel(R) Core(TM) i7-3770 CPU @ 3.40GHz", { {"WGS1",64}, {"WPT1",4}, {"VW2",4}, {"WGS2",64}, {"WPT2",4}, {"VW3",2}, {"WGS3",64}, {"WPT3",2} } }, - { "default", { {"WGS1",64}, {"WPT1",1}, {"VW2",1}, {"WGS2",64}, {"WPT2",2}, {"VW3",1}, {"WGS3",64}, {"WPT3",1} } }, + { "Intel(R) Core(TM) i5-6200U CPU @ 2.30GHz", { {"WGS1",128}, {"WPT1",1} } }, + { "Intel(R) Core(TM) i7-3770 CPU @ 3.40GHz", { {"WGS1",64}, {"WPT1",4} } }, + { "default", { {"WGS1",64}, {"WPT1",1} } }, } }, { // Intel GPUs kDeviceTypeGPU, "Intel", { - { "Intel(R) HD Graphics 530", { {"WGS1",64}, {"WPT1",1}, {"VW2",2}, {"WGS2",128}, {"WPT2",2}, {"VW3",4}, {"WGS3",64}, {"WPT3",4} } }, - { "Intel(R) HD Graphics Haswell Ultrabook GT2 Mobile", { {"WGS1",128}, {"WPT1",1}, {"VW2",2}, {"WGS2",128}, {"WPT2",2}, {"VW3",4}, {"WGS3",128}, {"WPT3",4} } }, - { "Intel(R) HD Graphics Skylake ULT GT2", { {"WGS1",64}, {"WPT1",1}, {"VW2",1}, {"WGS2",64}, {"WPT2",1}, {"VW3",4}, {"WGS3",64}, {"WPT3",4} } }, - { "Iris", { {"WGS1",256}, {"WPT1",1}, {"VW2",1}, {"WGS2",64}, {"WPT2",1}, {"VW3",1}, {"WGS3",64}, {"WPT3",1} } }, - { "Iris Pro", { {"WGS1",64}, {"WPT1",1}, {"VW2",4}, {"WGS2",128}, {"WPT2",4}, {"VW3",2}, {"WGS3",64}, {"WPT3",2} } }, - { "default", { {"WGS1",64}, {"WPT1",1}, {"VW2",1}, {"WGS2",64}, {"WPT2",1}, {"VW3",1}, {"WGS3",64}, {"WPT3",1} } }, + { "Intel(R) HD Graphics 530", { {"WGS1",64}, {"WPT1",1} } }, + { "Intel(R) HD Graphics Haswell Ultrabook GT2 Mobile", { {"WGS1",128}, {"WPT1",1} } }, + { "Intel(R) HD Graphics Skylake ULT GT2", { {"WGS1",64}, {"WPT1",1} } }, + { "Iris", { {"WGS1",256}, {"WPT1",1} } }, + { "Iris Pro", { {"WGS1",64}, {"WPT1",1} } }, + { "default", { {"WGS1",64}, {"WPT1",1} } }, } }, { // Intel accelerators kDeviceTypeAccelerator, "Intel", { - { "Intel(R) Many Integrated Core Acceleration Card", { {"WGS1",64}, {"WPT1",1}, {"VW2",1}, {"WGS2",64}, {"WPT2",1}, {"VW3",1}, {"WGS3",64}, {"WPT3",1} } }, - { "default", { {"WGS1",64}, {"WPT1",1}, {"VW2",1}, {"WGS2",64}, {"WPT2",1}, {"VW3",1}, {"WGS3",64}, {"WPT3",1} } }, + { "Intel(R) Many Integrated Core Acceleration Card", { {"WGS1",64}, {"WPT1",1} } }, + { "default", { {"WGS1",64}, {"WPT1",1} } }, } }, { // NVIDIA GPUs kDeviceTypeGPU, "NVIDIA", { - { "GRID K520", { {"WGS1",256}, {"WPT1",1}, {"VW2",1}, {"WGS2",256}, {"WPT2",1}, {"VW3",1}, {"WGS3",64}, {"WPT3",1} } }, - { "GeForce GTX 1070", { {"WGS1",64}, {"WPT1",1}, {"VW2",1}, {"WGS2",64}, {"WPT2",1} } }, - { "GeForce GTX 480", { {"WGS1",64}, {"WPT1",1}, {"VW2",1}, {"WGS2",64}, {"WPT2",1}, {"VW3",1}, {"WGS3",64}, {"WPT3",1} } }, - { "GeForce GTX 670", { {"WGS1",64}, {"WPT1",1}, {"VW2",1}, {"WGS2",64}, {"WPT2",1}, {"VW3",1}, {"WGS3",64}, {"WPT3",1} } }, - { "GeForce GTX 680", { {"WGS1",64}, {"WPT1",1}, {"VW2",1}, {"WGS2",64}, {"WPT2",1}, {"VW3",1}, {"WGS3",64}, {"WPT3",1} } }, + { "GRID K520", { {"WGS1",256}, {"WPT1",1} } }, + { "GeForce GTX 1070", { {"WGS1",64}, {"WPT1",1} } }, + { "GeForce GTX 480", { {"WGS1",64}, {"WPT1",1} } }, + { "GeForce GTX 670", { {"WGS1",64}, {"WPT1",1} } }, + { "GeForce GTX 680", { {"WGS1",64}, {"WPT1",1} } }, { "GeForce GTX 750", { {"WGS1",128}, {"WPT1",1} } }, - { "GeForce GTX 750 Ti", { {"WGS1",64}, {"WPT1",1}, {"VW2",1}, {"WGS2",64}, {"WPT2",1} } }, + { "GeForce GTX 750 Ti", { {"WGS1",64}, {"WPT1",1} } }, { "GeForce GTX TITAN", { {"WGS1",256}, {"WPT1",1} } }, - { "default", { {"WGS1",64}, {"WPT1",1}, {"VW2",1}, {"WGS2",64}, {"WPT2",1}, {"VW3",1}, {"WGS3",64}, {"WPT3",1} } }, + { "default", { {"WGS1",64}, {"WPT1",1} } }, } }, { // Default kDeviceTypeAll, "default", { - { "default", { {"WGS1",64}, {"WPT1",1}, {"VW2",1}, {"WGS2",64}, {"WPT2",1}, {"VW3",1}, {"WGS3",64}, {"WPT3",1} } }, + { "default", { {"WGS1",64}, {"WPT1",1} } }, } }, } @@ -156,47 +156,47 @@ const Database::DatabaseEntry Database::XgemvDouble = { "Xgemv", Precision::kDouble, { { // AMD GPUs kDeviceTypeGPU, "AMD", { - { "AMD Radeon R9 M370X Compute Engine", { {"WGS1",64}, {"WPT1",1}, {"VW2",1}, {"WGS2",256}, {"WPT2",1}, {"VW3",1}, {"WGS3",128}, {"WPT3",1} } }, - { "Hawaii", { {"WGS1",128}, {"WPT1",1}, {"VW2",1}, {"WGS2",64}, {"WPT2",1}, {"VW3",1}, {"WGS3",64}, {"WPT3",1} } }, - { "Oland", { {"WGS1",256}, {"WPT1",1}, {"VW2",1}, {"WGS2",64}, {"WPT2",1}, {"VW3",4}, {"WGS3",256}, {"WPT3",4} } }, - { "Pitcairn", { {"WGS1",256}, {"WPT1",1}, {"VW2",1}, {"WGS2",64}, {"WPT2",1}, {"VW3",2}, {"WGS3",64}, {"WPT3",2} } }, - { "Tahiti", { {"WGS1",256}, {"WPT1",1}, {"VW2",1}, {"WGS2",64}, {"WPT2",1}, {"VW3",2}, {"WGS3",64}, {"WPT3",2} } }, - { "default", { {"WGS1",64}, {"WPT1",1}, {"VW2",1}, {"WGS2",64}, {"WPT2",1}, {"VW3",1}, {"WGS3",64}, {"WPT3",1} } }, + { "AMD Radeon R9 M370X Compute Engine", { {"WGS1",64}, {"WPT1",1} } }, + { "Hawaii", { {"WGS1",128}, {"WPT1",1} } }, + { "Oland", { {"WGS1",256}, {"WPT1",1} } }, + { "Pitcairn", { {"WGS1",256}, {"WPT1",1} } }, + { "Tahiti", { {"WGS1",256}, {"WPT1",1} } }, + { "default", { {"WGS1",64}, {"WPT1",1} } }, } }, { // Intel CPUs kDeviceTypeCPU, "Intel", { - { "Intel(R) Core(TM) i5-6200U CPU @ 2.30GHz", { {"WGS1",64}, {"WPT1",2}, {"VW2",4}, {"WGS2",128}, {"WPT2",4}, {"VW3",1}, {"WGS3",64}, {"WPT3",1} } }, - { "Intel(R) Core(TM) i7-3770 CPU @ 3.40GHz", { {"WGS1",64}, {"WPT1",4}, {"VW2",1}, {"WGS2",64}, {"WPT2",4}, {"VW3",1}, {"WGS3",64}, {"WPT3",2} } }, - { "default", { {"WGS1",64}, {"WPT1",2}, {"VW2",1}, {"WGS2",64}, {"WPT2",4}, {"VW3",1}, {"WGS3",64}, {"WPT3",1} } }, + { "Intel(R) Core(TM) i5-6200U CPU @ 2.30GHz", { {"WGS1",64}, {"WPT1",2} } }, + { "Intel(R) Core(TM) i7-3770 CPU @ 3.40GHz", { {"WGS1",64}, {"WPT1",4} } }, + { "default", { {"WGS1",64}, {"WPT1",2} } }, } }, { // Intel accelerators kDeviceTypeAccelerator, "Intel", { - { "Intel(R) Many Integrated Core Acceleration Card", { {"WGS1",64}, {"WPT1",1}, {"VW2",1}, {"WGS2",64}, {"WPT2",1}, {"VW3",1}, {"WGS3",64}, {"WPT3",1} } }, - { "default", { {"WGS1",64}, {"WPT1",1}, {"VW2",1}, {"WGS2",64}, {"WPT2",1}, {"VW3",1}, {"WGS3",64}, {"WPT3",1} } }, + { "Intel(R) Many Integrated Core Acceleration Card", { {"WGS1",64}, {"WPT1",1} } }, + { "default", { {"WGS1",64}, {"WPT1",1} } }, } }, { // NVIDIA GPUs kDeviceTypeGPU, "NVIDIA", { - { "GRID K520", { {"WGS1",128}, {"WPT1",1}, {"VW2",1}, {"WGS2",256}, {"WPT2",1}, {"VW3",1}, {"WGS3",64}, {"WPT3",1} } }, - { "GeForce GTX 1070", { {"WGS1",64}, {"WPT1",1}, {"VW2",1}, {"WGS2",256}, {"WPT2",1}, {"VW3",1}, {"WGS3",128}, {"WPT3",1} } }, - { "GeForce GTX 480", { {"WGS1",256}, {"WPT1",1}, {"VW2",1}, {"WGS2",64}, {"WPT2",1}, {"VW3",1}, {"WGS3",64}, {"WPT3",1} } }, - { "GeForce GTX 670", { {"WGS1",128}, {"WPT1",1}, {"VW2",1}, {"WGS2",128}, {"WPT2",1}, {"VW3",1}, {"WGS3",64}, {"WPT3",1} } }, - { "GeForce GTX 680", { {"WGS1",128}, {"WPT1",1}, {"VW2",1}, {"WGS2",128}, {"WPT2",1}, {"VW3",2}, {"WGS3",128}, {"WPT3",2} } }, - { "GeForce GTX 750", { {"WGS1",64}, {"WPT1",1}, {"VW2",2}, {"WGS2",256}, {"WPT2",2}, {"VW3",2}, {"WGS3",64}, {"WPT3",2} } }, - { "GeForce GTX 750 Ti", { {"WGS1",64}, {"WPT1",1}, {"VW2",1}, {"WGS2",256}, {"WPT2",1}, {"VW3",2}, {"WGS3",256}, {"WPT3",2} } }, - { "GeForce GTX 980", { {"WGS1",64}, {"WPT1",1}, {"VW2",1}, {"WGS2",64}, {"WPT2",1}, {"VW3",1}, {"WGS3",64}, {"WPT3",1} } }, - { "GeForce GTX TITAN", { {"WGS1",256}, {"WPT1",1}, {"VW2",1}, {"WGS2",256}, {"WPT2",1}, {"VW3",1}, {"WGS3",256}, {"WPT3",1} } }, - { "GeForce GTX TITAN X", { {"WGS1",64}, {"WPT1",1}, {"VW2",1}, {"WGS2",128}, {"WPT2",1}, {"VW3",1}, {"WGS3",128}, {"WPT3",1} } }, - { "Tesla K20m", { {"WGS1",256}, {"WPT1",1}, {"VW2",1}, {"WGS2",128}, {"WPT2",1}, {"VW3",1}, {"WGS3",64}, {"WPT3",1} } }, - { "Tesla K40m", { {"WGS1",256}, {"WPT1",1}, {"VW2",1}, {"WGS2",256}, {"WPT2",1}, {"VW3",1}, {"WGS3",256}, {"WPT3",1} } }, - { "default", { {"WGS1",64}, {"WPT1",1}, {"VW2",1}, {"WGS2",64}, {"WPT2",1}, {"VW3",1}, {"WGS3",64}, {"WPT3",1} } }, + { "GRID K520", { {"WGS1",128}, {"WPT1",1} } }, + { "GeForce GTX 1070", { {"WGS1",64}, {"WPT1",1} } }, + { "GeForce GTX 480", { {"WGS1",256}, {"WPT1",1} } }, + { "GeForce GTX 670", { {"WGS1",128}, {"WPT1",1} } }, + { "GeForce GTX 680", { {"WGS1",128}, {"WPT1",1} } }, + { "GeForce GTX 750", { {"WGS1",64}, {"WPT1",1} } }, + { "GeForce GTX 750 Ti", { {"WGS1",64}, {"WPT1",1} } }, + { "GeForce GTX 980", { {"WGS1",64}, {"WPT1",1} } }, + { "GeForce GTX TITAN", { {"WGS1",256}, {"WPT1",1} } }, + { "GeForce GTX TITAN X", { {"WGS1",64}, {"WPT1",1} } }, + { "Tesla K20m", { {"WGS1",256}, {"WPT1",1} } }, + { "Tesla K40m", { {"WGS1",256}, {"WPT1",1} } }, + { "default", { {"WGS1",64}, {"WPT1",1} } }, } }, { // Default kDeviceTypeAll, "default", { - { "default", { {"WGS1",64}, {"WPT1",1}, {"VW2",1}, {"WGS2",64}, {"WPT2",1}, {"VW3",1}, {"WGS3",64}, {"WPT3",1} } }, + { "default", { {"WGS1",64}, {"WPT1",1} } }, } }, } @@ -208,38 +208,38 @@ const Database::DatabaseEntry Database::XgemvComplexDouble = { "Xgemv", Precision::kComplexDouble, { { // AMD GPUs kDeviceTypeGPU, "AMD", { - { "AMD Radeon R9 M370X Compute Engine", { {"WGS1",64}, {"WPT1",1}, {"VW2",1}, {"WGS2",256}, {"WPT2",1}, {"VW3",1}, {"WGS3",128}, {"WPT3",1} } }, - { "Hawaii", { {"WGS1",64}, {"WPT1",1}, {"VW2",1}, {"WGS2",64}, {"WPT2",1}, {"VW3",1}, {"WGS3",64}, {"WPT3",1} } }, - { "Oland", { {"WGS1",256}, {"WPT1",1}, {"VW2",1}, {"WGS2",256}, {"WPT2",1}, {"VW3",1}, {"WGS3",256}, {"WPT3",1} } }, - { "Pitcairn", { {"WGS1",256}, {"WPT1",1}, {"VW2",1}, {"WGS2",64}, {"WPT2",1}, {"VW3",1}, {"WGS3",64}, {"WPT3",1} } }, - { "Tahiti", { {"WGS1",256}, {"WPT1",1}, {"VW2",1}, {"WGS2",64}, {"WPT2",1}, {"VW3",1}, {"WGS3",64}, {"WPT3",1} } }, - { "default", { {"WGS1",64}, {"WPT1",1}, {"VW2",1}, {"WGS2",64}, {"WPT2",1}, {"VW3",1}, {"WGS3",64}, {"WPT3",1} } }, + { "AMD Radeon R9 M370X Compute Engine", { {"WGS1",64}, {"WPT1",1} } }, + { "Hawaii", { {"WGS1",64}, {"WPT1",1} } }, + { "Oland", { {"WGS1",256}, {"WPT1",1} } }, + { "Pitcairn", { {"WGS1",256}, {"WPT1",1} } }, + { "Tahiti", { {"WGS1",256}, {"WPT1",1} } }, + { "default", { {"WGS1",64}, {"WPT1",1} } }, } }, { // Intel CPUs kDeviceTypeCPU, "Intel", { - { "Intel(R) Core(TM) i5-6200U CPU @ 2.30GHz", { {"WGS1",64}, {"WPT1",1}, {"VW2",2}, {"WGS2",64}, {"WPT2",4}, {"VW3",1}, {"WGS3",256}, {"WPT3",1} } }, - { "Intel(R) Core(TM) i7-3770 CPU @ 3.40GHz", { {"WGS1",64}, {"WPT1",4}, {"VW2",4}, {"WGS2",64}, {"WPT2",4}, {"VW3",2}, {"WGS3",256}, {"WPT3",2} } }, - { "default", { {"WGS1",64}, {"WPT1",1}, {"VW2",2}, {"WGS2",64}, {"WPT2",4}, {"VW3",1}, {"WGS3",256}, {"WPT3",1} } }, + { "Intel(R) Core(TM) i5-6200U CPU @ 2.30GHz", { {"WGS1",64}, {"WPT1",1} } }, + { "Intel(R) Core(TM) i7-3770 CPU @ 3.40GHz", { {"WGS1",64}, {"WPT1",4} } }, + { "default", { {"WGS1",64}, {"WPT1",1} } }, } }, { // Intel accelerators kDeviceTypeAccelerator, "Intel", { - { "Intel(R) Many Integrated Core Acceleration Card", { {"WGS1",64}, {"WPT1",1}, {"VW2",1}, {"WGS2",64}, {"WPT2",1}, {"VW3",1}, {"WGS3",64}, {"WPT3",1} } }, - { "default", { {"WGS1",64}, {"WPT1",1}, {"VW2",1}, {"WGS2",64}, {"WPT2",1}, {"VW3",1}, {"WGS3",64}, {"WPT3",1} } }, + { "Intel(R) Many Integrated Core Acceleration Card", { {"WGS1",64}, {"WPT1",1} } }, + { "default", { {"WGS1",64}, {"WPT1",1} } }, } }, { // NVIDIA GPUs kDeviceTypeGPU, "NVIDIA", { - { "GRID K520", { {"WGS1",128}, {"WPT1",1}, {"VW2",1}, {"WGS2",128}, {"WPT2",1}, {"VW3",1}, {"WGS3",256}, {"WPT3",1} } }, - { "GeForce GTX 480", { {"WGS1",64}, {"WPT1",1}, {"VW2",1}, {"WGS2",64}, {"WPT2",1}, {"VW3",1}, {"WGS3",64}, {"WPT3",1} } }, - { "GeForce GTX 670", { {"WGS1",128}, {"WPT1",1}, {"VW2",1}, {"WGS2",64}, {"WPT2",1}, {"VW3",1}, {"WGS3",64}, {"WPT3",1} } }, - { "default", { {"WGS1",64}, {"WPT1",1}, {"VW2",1}, {"WGS2",64}, {"WPT2",1}, {"VW3",1}, {"WGS3",64}, {"WPT3",1} } }, + { "GRID K520", { {"WGS1",128}, {"WPT1",1} } }, + { "GeForce GTX 480", { {"WGS1",64}, {"WPT1",1} } }, + { "GeForce GTX 670", { {"WGS1",128}, {"WPT1",1} } }, + { "default", { {"WGS1",64}, {"WPT1",1} } }, } }, { // Default kDeviceTypeAll, "default", { - { "default", { {"WGS1",64}, {"WPT1",1}, {"VW2",1}, {"WGS2",64}, {"WPT2",1}, {"VW3",1}, {"WGS3",64}, {"WPT3",1} } }, + { "default", { {"WGS1",64}, {"WPT1",1} } }, } }, } diff --git a/src/database/kernels/xgemv_fast.hpp b/src/database/kernels/xgemv_fast.hpp new file mode 100644 index 00000000..65b15030 --- /dev/null +++ b/src/database/kernels/xgemv_fast.hpp @@ -0,0 +1,247 @@ + +// ================================================================================================= +// This file is part of the CLBlast project. The project is licensed under Apache Version 2.0. This +// project loosely follows the Google C++ styleguide and uses a tab-size of two spaces and a max- +// width of 100 characters per line. +// +// Author(s): +// Database generator <database.py> +// +// This file populates the database with best-found tuning parameters for the 'Xgemv_Fast' kernels. +// +// ================================================================================================= + +namespace clblast { +// ================================================================================================= + +const Database::DatabaseEntry Database::XgemvFastHalf = { + "XgemvFast", Precision::kHalf, { + { // Intel GPUs + kDeviceTypeGPU, "Intel", { + { "Intel(R) HD Graphics Skylake ULT GT2", { {"VW2",2}, {"WGS2",128}, {"WPT2",2} } }, + { "default", { {"VW2",2}, {"WGS2",128}, {"WPT2",2} } }, + } + }, + { // Default + kDeviceTypeAll, "default", { + { "default", { {"VW2",2}, {"WGS2",128}, {"WPT2",2} } }, + } + }, + } +}; + +// ================================================================================================= + +const Database::DatabaseEntry Database::XgemvFastSingle = { + "XgemvFast", Precision::kSingle, { + { // AMD GPUs + kDeviceTypeGPU, "AMD", { + { "AMD Radeon R9 M370X Compute Engine", { {"VW2",1}, {"WGS2",128}, {"WPT2",1} } }, + { "Hawaii", { {"VW2",1}, {"WGS2",64}, {"WPT2",1} } }, + { "Oland", { {"VW2",1}, {"WGS2",64}, {"WPT2",1} } }, + { "Pitcairn", { {"VW2",1}, {"WGS2",64}, {"WPT2",1} } }, + { "Tahiti", { {"VW2",1}, {"WGS2",64}, {"WPT2",1} } }, + { "default", { {"VW2",1}, {"WGS2",64}, {"WPT2",1} } }, + } + }, + { // Intel CPUs + kDeviceTypeCPU, "Intel", { + { "Intel(R) Core(TM) i5-6200U CPU @ 2.30GHz", { {"VW2",4}, {"WGS2",128}, {"WPT2",4} } }, + { "Intel(R) Core(TM) i7-3770 CPU @ 3.40GHz", { {"VW2",1}, {"WGS2",64}, {"WPT2",4} } }, + { "default", { {"VW2",1}, {"WGS2",64}, {"WPT2",4} } }, + } + }, + { // Intel GPUs + kDeviceTypeGPU, "Intel", { + { "Intel(R) HD Graphics 530", { {"VW2",1}, {"WGS2",256}, {"WPT2",1} } }, + { "Intel(R) HD Graphics Haswell Ultrabook GT2 Mobile", { {"VW2",4}, {"WGS2",128}, {"WPT2",4} } }, + { "Intel(R) HD Graphics Skylake ULT GT2", { {"VW2",1}, {"WGS2",256}, {"WPT2",1} } }, + { "Iris", { {"VW2",1}, {"WGS2",128}, {"WPT2",2} } }, + { "Iris Pro", { {"VW2",1}, {"WGS2",128}, {"WPT2",2} } }, + { "default", { {"VW2",1}, {"WGS2",128}, {"WPT2",1} } }, + } + }, + { // Intel accelerators + kDeviceTypeAccelerator, "Intel", { + { "Intel(R) Many Integrated Core Acceleration Card", { {"VW2",1}, {"WGS2",64}, {"WPT2",1} } }, + { "default", { {"VW2",1}, {"WGS2",64}, {"WPT2",1} } }, + } + }, + { // NVIDIA GPUs + kDeviceTypeGPU, "NVIDIA", { + { "GRID K520", { {"VW2",2}, {"WGS2",256}, {"WPT2",2} } }, + { "GeForce GTX 1070", { {"VW2",1}, {"WGS2",256}, {"WPT2",1} } }, + { "GeForce GTX 480", { {"VW2",1}, {"WGS2",128}, {"WPT2",1} } }, + { "GeForce GTX 670", { {"VW2",2}, {"WGS2",256}, {"WPT2",2} } }, + { "GeForce GTX 680", { {"VW2",1}, {"WGS2",128}, {"WPT2",1} } }, + { "GeForce GTX 750", { {"VW2",1}, {"WGS2",256}, {"WPT2",1} } }, + { "GeForce GTX 750 Ti", { {"VW2",1}, {"WGS2",256}, {"WPT2",1} } }, + { "GeForce GTX 980", { {"VW2",1}, {"WGS2",256}, {"WPT2",1} } }, + { "GeForce GTX TITAN", { {"VW2",1}, {"WGS2",256}, {"WPT2",1} } }, + { "GeForce GTX TITAN X", { {"VW2",1}, {"WGS2",64}, {"WPT2",1} } }, + { "Tesla K20m", { {"VW2",1}, {"WGS2",256}, {"WPT2",1} } }, + { "Tesla K40m", { {"VW2",1}, {"WGS2",256}, {"WPT2",1} } }, + { "default", { {"VW2",1}, {"WGS2",64}, {"WPT2",1} } }, + } + }, + { // Default + kDeviceTypeAll, "default", { + { "default", { {"VW2",1}, {"WGS2",64}, {"WPT2",1} } }, + } + }, + } +}; + +// ================================================================================================= + +const Database::DatabaseEntry Database::XgemvFastComplexSingle = { + "XgemvFast", Precision::kComplexSingle, { + { // AMD GPUs + kDeviceTypeGPU, "AMD", { + { "AMD Radeon R9 M370X Compute Engine", { {"VW2",2}, {"WGS2",256}, {"WPT2",2} } }, + { "Hawaii", { {"VW2",1}, {"WGS2",64}, {"WPT2",1} } }, + { "Oland", { {"VW2",1}, {"WGS2",64}, {"WPT2",1} } }, + { "Pitcairn", { {"VW2",1}, {"WGS2",64}, {"WPT2",1} } }, + { "Tahiti", { {"VW2",1}, {"WGS2",128}, {"WPT2",1} } }, + { "default", { {"VW2",1}, {"WGS2",64}, {"WPT2",1} } }, + } + }, + { // Intel CPUs + kDeviceTypeCPU, "Intel", { + { "Intel(R) Core(TM) i5-6200U CPU @ 2.30GHz", { {"VW2",1}, {"WGS2",128}, {"WPT2",2} } }, + { "Intel(R) Core(TM) i7-3770 CPU @ 3.40GHz", { {"VW2",4}, {"WGS2",64}, {"WPT2",4} } }, + { "default", { {"VW2",1}, {"WGS2",64}, {"WPT2",2} } }, + } + }, + { // Intel GPUs + kDeviceTypeGPU, "Intel", { + { "Intel(R) HD Graphics 530", { {"VW2",2}, {"WGS2",128}, {"WPT2",2} } }, + { "Intel(R) HD Graphics Haswell Ultrabook GT2 Mobile", { {"VW2",2}, {"WGS2",128}, {"WPT2",2} } }, + { "Intel(R) HD Graphics Skylake ULT GT2", { {"VW2",1}, {"WGS2",64}, {"WPT2",1} } }, + { "Iris", { {"VW2",1}, {"WGS2",64}, {"WPT2",1} } }, + { "Iris Pro", { {"VW2",4}, {"WGS2",128}, {"WPT2",4} } }, + { "default", { {"VW2",1}, {"WGS2",64}, {"WPT2",1} } }, + } + }, + { // Intel accelerators + kDeviceTypeAccelerator, "Intel", { + { "Intel(R) Many Integrated Core Acceleration Card", { {"VW2",1}, {"WGS2",64}, {"WPT2",1} } }, + { "default", { {"VW2",1}, {"WGS2",64}, {"WPT2",1} } }, + } + }, + { // NVIDIA GPUs + kDeviceTypeGPU, "NVIDIA", { + { "GRID K520", { {"VW2",1}, {"WGS2",256}, {"WPT2",1} } }, + { "GeForce GTX 1070", { {"VW2",1}, {"WGS2",64}, {"WPT2",1} } }, + { "GeForce GTX 480", { {"VW2",1}, {"WGS2",64}, {"WPT2",1} } }, + { "GeForce GTX 670", { {"VW2",1}, {"WGS2",64}, {"WPT2",1} } }, + { "GeForce GTX 680", { {"VW2",1}, {"WGS2",64}, {"WPT2",1} } }, + { "GeForce GTX 750 Ti", { {"VW2",1}, {"WGS2",64}, {"WPT2",1} } }, + { "default", { {"VW2",1}, {"WGS2",64}, {"WPT2",1} } }, + } + }, + { // Default + kDeviceTypeAll, "default", { + { "default", { {"VW2",1}, {"WGS2",64}, {"WPT2",1} } }, + } + }, + } +}; + +// ================================================================================================= + +const Database::DatabaseEntry Database::XgemvFastDouble = { + "XgemvFast", Precision::kDouble, { + { // AMD GPUs + kDeviceTypeGPU, "AMD", { + { "AMD Radeon R9 M370X Compute Engine", { {"VW2",1}, {"WGS2",256}, {"WPT2",1} } }, + { "Hawaii", { {"VW2",1}, {"WGS2",64}, {"WPT2",1} } }, + { "Oland", { {"VW2",1}, {"WGS2",64}, {"WPT2",1} } }, + { "Pitcairn", { {"VW2",1}, {"WGS2",64}, {"WPT2",1} } }, + { "Tahiti", { {"VW2",1}, {"WGS2",64}, {"WPT2",1} } }, + { "default", { {"VW2",1}, {"WGS2",64}, {"WPT2",1} } }, + } + }, + { // Intel CPUs + kDeviceTypeCPU, "Intel", { + { "Intel(R) Core(TM) i5-6200U CPU @ 2.30GHz", { {"VW2",4}, {"WGS2",128}, {"WPT2",4} } }, + { "Intel(R) Core(TM) i7-3770 CPU @ 3.40GHz", { {"VW2",1}, {"WGS2",64}, {"WPT2",4} } }, + { "default", { {"VW2",1}, {"WGS2",64}, {"WPT2",4} } }, + } + }, + { // Intel accelerators + kDeviceTypeAccelerator, "Intel", { + { "Intel(R) Many Integrated Core Acceleration Card", { {"VW2",1}, {"WGS2",64}, {"WPT2",1} } }, + { "default", { {"VW2",1}, {"WGS2",64}, {"WPT2",1} } }, + } + }, + { // NVIDIA GPUs + kDeviceTypeGPU, "NVIDIA", { + { "GRID K520", { {"VW2",1}, {"WGS2",256}, {"WPT2",1} } }, + { "GeForce GTX 1070", { {"VW2",1}, {"WGS2",256}, {"WPT2",1} } }, + { "GeForce GTX 480", { {"VW2",1}, {"WGS2",64}, {"WPT2",1} } }, + { "GeForce GTX 670", { {"VW2",1}, {"WGS2",128}, {"WPT2",1} } }, + { "GeForce GTX 680", { {"VW2",1}, {"WGS2",128}, {"WPT2",1} } }, + { "GeForce GTX 750", { {"VW2",2}, {"WGS2",256}, {"WPT2",2} } }, + { "GeForce GTX 750 Ti", { {"VW2",1}, {"WGS2",256}, {"WPT2",1} } }, + { "GeForce GTX 980", { {"VW2",1}, {"WGS2",64}, {"WPT2",1} } }, + { "GeForce GTX TITAN", { {"VW2",1}, {"WGS2",256}, {"WPT2",1} } }, + { "GeForce GTX TITAN X", { {"VW2",1}, {"WGS2",128}, {"WPT2",1} } }, + { "Tesla K20m", { {"VW2",1}, {"WGS2",128}, {"WPT2",1} } }, + { "Tesla K40m", { {"VW2",1}, {"WGS2",256}, {"WPT2",1} } }, + { "default", { {"VW2",1}, {"WGS2",64}, {"WPT2",1} } }, + } + }, + { // Default + kDeviceTypeAll, "default", { + { "default", { {"VW2",1}, {"WGS2",64}, {"WPT2",1} } }, + } + }, + } +}; + +// ================================================================================================= + +const Database::DatabaseEntry Database::XgemvFastComplexDouble = { + "XgemvFast", Precision::kComplexDouble, { + { // AMD GPUs + kDeviceTypeGPU, "AMD", { + { "AMD Radeon R9 M370X Compute Engine", { {"VW2",1}, {"WGS2",256}, {"WPT2",1} } }, + { "Hawaii", { {"VW2",1}, {"WGS2",64}, {"WPT2",1} } }, + { "Oland", { {"VW2",1}, {"WGS2",256}, {"WPT2",1} } }, + { "Pitcairn", { {"VW2",1}, {"WGS2",64}, {"WPT2",1} } }, + { "Tahiti", { {"VW2",1}, {"WGS2",64}, {"WPT2",1} } }, + { "default", { {"VW2",1}, {"WGS2",64}, {"WPT2",1} } }, + } + }, + { // Intel CPUs + kDeviceTypeCPU, "Intel", { + { "Intel(R) Core(TM) i5-6200U CPU @ 2.30GHz", { {"VW2",2}, {"WGS2",64}, {"WPT2",4} } }, + { "Intel(R) Core(TM) i7-3770 CPU @ 3.40GHz", { {"VW2",4}, {"WGS2",64}, {"WPT2",4} } }, + { "default", { {"VW2",2}, {"WGS2",64}, {"WPT2",4} } }, + } + }, + { // Intel accelerators + kDeviceTypeAccelerator, "Intel", { + { "Intel(R) Many Integrated Core Acceleration Card", { {"VW2",1}, {"WGS2",64}, {"WPT2",1} } }, + { "default", { {"VW2",1}, {"WGS2",64}, {"WPT2",1} } }, + } + }, + { // NVIDIA GPUs + kDeviceTypeGPU, "NVIDIA", { + { "GRID K520", { {"VW2",1}, {"WGS2",128}, {"WPT2",1} } }, + { "GeForce GTX 480", { {"VW2",1}, {"WGS2",64}, {"WPT2",1} } }, + { "GeForce GTX 670", { {"VW2",1}, {"WGS2",64}, {"WPT2",1} } }, + { "default", { {"VW2",1}, {"WGS2",64}, {"WPT2",1} } }, + } + }, + { // Default + kDeviceTypeAll, "default", { + { "default", { {"VW2",1}, {"WGS2",64}, {"WPT2",1} } }, + } + }, + } +}; + +// ================================================================================================= +} // namespace clblast diff --git a/src/database/kernels/xgemv_fast_rot.hpp b/src/database/kernels/xgemv_fast_rot.hpp new file mode 100644 index 00000000..ee866e26 --- /dev/null +++ b/src/database/kernels/xgemv_fast_rot.hpp @@ -0,0 +1,245 @@ + +// ================================================================================================= +// This file is part of the CLBlast project. The project is licensed under Apache Version 2.0. This +// project loosely follows the Google C++ styleguide and uses a tab-size of two spaces and a max- +// width of 100 characters per line. +// +// Author(s): +// Database generator <database.py> +// +// This file populates the database with best-found tuning parameters for the 'Xgemv_Fast_Rot' kernels. +// +// ================================================================================================= + +namespace clblast { +// ================================================================================================= + +const Database::DatabaseEntry Database::XgemvFastRotHalf = { + "XgemvFastRot", Precision::kHalf, { + { // Intel GPUs + kDeviceTypeGPU, "Intel", { + { "Intel(R) HD Graphics Skylake ULT GT2", { {"VW3",1}, {"WGS3",64}, {"WPT3",1} } }, + { "default", { {"VW3",1}, {"WGS3",64}, {"WPT3",1} } }, + } + }, + { // Default + kDeviceTypeAll, "default", { + { "default", { {"VW3",1}, {"WGS3",64}, {"WPT3",1} } }, + } + }, + } +}; + +// ================================================================================================= + +const Database::DatabaseEntry Database::XgemvFastRotSingle = { + "XgemvFastRot", Precision::kSingle, { + { // AMD GPUs + kDeviceTypeGPU, "AMD", { + { "AMD Radeon R9 M370X Compute Engine", { {"VW3",1}, {"WGS3",64}, {"WPT3",1} } }, + { "Hawaii", { {"VW3",1}, {"WGS3",64}, {"WPT3",1} } }, + { "Oland", { {"VW3",4}, {"WGS3",256}, {"WPT3",4} } }, + { "Pitcairn", { {"VW3",1}, {"WGS3",64}, {"WPT3",1} } }, + { "Tahiti", { {"VW3",1}, {"WGS3",64}, {"WPT3",1} } }, + { "default", { {"VW3",1}, {"WGS3",64}, {"WPT3",1} } }, + } + }, + { // Intel CPUs + kDeviceTypeCPU, "Intel", { + { "Intel(R) Core(TM) i5-6200U CPU @ 2.30GHz", { {"VW3",1}, {"WGS3",64}, {"WPT3",1} } }, + { "Intel(R) Core(TM) i7-3770 CPU @ 3.40GHz", { {"VW3",2}, {"WGS3",64}, {"WPT3",4} } }, + { "default", { {"VW3",1}, {"WGS3",64}, {"WPT3",1} } }, + } + }, + { // Intel GPUs + kDeviceTypeGPU, "Intel", { + { "Intel(R) HD Graphics 530", { {"VW3",1}, {"WGS3",64}, {"WPT3",1} } }, + { "Intel(R) HD Graphics Haswell Ultrabook GT2 Mobile", { {"VW3",4}, {"WGS3",256}, {"WPT3",4} } }, + { "Intel(R) HD Graphics Skylake ULT GT2", { {"VW3",4}, {"WGS3",64}, {"WPT3",4} } }, + { "Iris", { {"VW3",4}, {"WGS3",64}, {"WPT3",8} } }, + { "Iris Pro", { {"VW3",4}, {"WGS3",64}, {"WPT3",4} } }, + { "default", { {"VW3",1}, {"WGS3",64}, {"WPT3",1} } }, + } + }, + { // Intel accelerators + kDeviceTypeAccelerator, "Intel", { + { "Intel(R) Many Integrated Core Acceleration Card", { {"VW3",1}, {"WGS3",64}, {"WPT3",1} } }, + { "default", { {"VW3",1}, {"WGS3",64}, {"WPT3",1} } }, + } + }, + { // NVIDIA GPUs + kDeviceTypeGPU, "NVIDIA", { + { "GRID K520", { {"VW3",2}, {"WGS3",64}, {"WPT3",2} } }, + { "GeForce GTX 1070", { {"VW3",2}, {"WGS3",64}, {"WPT3",2} } }, + { "GeForce GTX 480", { {"VW3",2}, {"WGS3",64}, {"WPT3",2} } }, + { "GeForce GTX 670", { {"VW3",2}, {"WGS3",64}, {"WPT3",2} } }, + { "GeForce GTX 680", { {"VW3",2}, {"WGS3",128}, {"WPT3",2} } }, + { "GeForce GTX 750", { {"VW3",2}, {"WGS3",128}, {"WPT3",2} } }, + { "GeForce GTX 750 Ti", { {"VW3",4}, {"WGS3",128}, {"WPT3",4} } }, + { "GeForce GTX 980", { {"VW3",1}, {"WGS3",128}, {"WPT3",1} } }, + { "GeForce GTX TITAN", { {"VW3",1}, {"WGS3",256}, {"WPT3",1} } }, + { "GeForce GTX TITAN X", { {"VW3",1}, {"WGS3",128}, {"WPT3",1} } }, + { "Tesla K20m", { {"VW3",1}, {"WGS3",256}, {"WPT3",1} } }, + { "Tesla K40m", { {"VW3",1}, {"WGS3",128}, {"WPT3",1} } }, + { "default", { {"VW3",1}, {"WGS3",64}, {"WPT3",1} } }, + } + }, + { // Default + kDeviceTypeAll, "default", { + { "default", { {"VW3",1}, {"WGS3",64}, {"WPT3",1} } }, + } + }, + } +}; + +// ================================================================================================= + +const Database::DatabaseEntry Database::XgemvFastRotComplexSingle = { + "XgemvFastRot", Precision::kComplexSingle, { + { // AMD GPUs + kDeviceTypeGPU, "AMD", { + { "AMD Radeon R9 M370X Compute Engine", { {"VW3",1}, {"WGS3",128}, {"WPT3",1} } }, + { "Hawaii", { {"VW3",1}, {"WGS3",64}, {"WPT3",1} } }, + { "Oland", { {"VW3",4}, {"WGS3",64}, {"WPT3",4} } }, + { "Pitcairn", { {"VW3",2}, {"WGS3",64}, {"WPT3",2} } }, + { "Tahiti", { {"VW3",1}, {"WGS3",64}, {"WPT3",1} } }, + { "default", { {"VW3",1}, {"WGS3",64}, {"WPT3",1} } }, + } + }, + { // Intel CPUs + kDeviceTypeCPU, "Intel", { + { "Intel(R) Core(TM) i5-6200U CPU @ 2.30GHz", { {"VW3",1}, {"WGS3",64}, {"WPT3",1} } }, + { "Intel(R) Core(TM) i7-3770 CPU @ 3.40GHz", { {"VW3",2}, {"WGS3",64}, {"WPT3",2} } }, + { "default", { {"VW3",1}, {"WGS3",64}, {"WPT3",1} } }, + } + }, + { // Intel GPUs + kDeviceTypeGPU, "Intel", { + { "Intel(R) HD Graphics 530", { {"VW3",4}, {"WGS3",64}, {"WPT3",4} } }, + { "Intel(R) HD Graphics Haswell Ultrabook GT2 Mobile", { {"VW3",4}, {"WGS3",128}, {"WPT3",4} } }, + { "Intel(R) HD Graphics Skylake ULT GT2", { {"VW3",4}, {"WGS3",64}, {"WPT3",4} } }, + { "Iris", { {"VW3",1}, {"WGS3",64}, {"WPT3",1} } }, + { "Iris Pro", { {"VW3",2}, {"WGS3",64}, {"WPT3",2} } }, + { "default", { {"VW3",1}, {"WGS3",64}, {"WPT3",1} } }, + } + }, + { // Intel accelerators + kDeviceTypeAccelerator, "Intel", { + { "Intel(R) Many Integrated Core Acceleration Card", { {"VW3",1}, {"WGS3",64}, {"WPT3",1} } }, + { "default", { {"VW3",1}, {"WGS3",64}, {"WPT3",1} } }, + } + }, + { // NVIDIA GPUs + kDeviceTypeGPU, "NVIDIA", { + { "GRID K520", { {"VW3",1}, {"WGS3",64}, {"WPT3",1} } }, + { "GeForce GTX 480", { {"VW3",1}, {"WGS3",64}, {"WPT3",1} } }, + { "GeForce GTX 670", { {"VW3",1}, {"WGS3",64}, {"WPT3",1} } }, + { "GeForce GTX 680", { {"VW3",1}, {"WGS3",64}, {"WPT3",1} } }, + { "default", { {"VW3",1}, {"WGS3",64}, {"WPT3",1} } }, + } + }, + { // Default + kDeviceTypeAll, "default", { + { "default", { {"VW3",1}, {"WGS3",64}, {"WPT3",1} } }, + } + }, + } +}; + +// ================================================================================================= + +const Database::DatabaseEntry Database::XgemvFastRotDouble = { + "XgemvFastRot", Precision::kDouble, { + { // AMD GPUs + kDeviceTypeGPU, "AMD", { + { "AMD Radeon R9 M370X Compute Engine", { {"VW3",1}, {"WGS3",128}, {"WPT3",1} } }, + { "Hawaii", { {"VW3",1}, {"WGS3",64}, {"WPT3",1} } }, + { "Oland", { {"VW3",4}, {"WGS3",256}, {"WPT3",4} } }, + { "Pitcairn", { {"VW3",2}, {"WGS3",64}, {"WPT3",2} } }, + { "Tahiti", { {"VW3",2}, {"WGS3",64}, {"WPT3",2} } }, + { "default", { {"VW3",1}, {"WGS3",64}, {"WPT3",1} } }, + } + }, + { // Intel CPUs + kDeviceTypeCPU, "Intel", { + { "Intel(R) Core(TM) i5-6200U CPU @ 2.30GHz", { {"VW3",1}, {"WGS3",64}, {"WPT3",1} } }, + { "Intel(R) Core(TM) i7-3770 CPU @ 3.40GHz", { {"VW3",1}, {"WGS3",64}, {"WPT3",2} } }, + { "default", { {"VW3",1}, {"WGS3",64}, {"WPT3",1} } }, + } + }, + { // Intel accelerators + kDeviceTypeAccelerator, "Intel", { + { "Intel(R) Many Integrated Core Acceleration Card", { {"VW3",1}, {"WGS3",64}, {"WPT3",1} } }, + { "default", { {"VW3",1}, {"WGS3",64}, {"WPT3",1} } }, + } + }, + { // NVIDIA GPUs + kDeviceTypeGPU, "NVIDIA", { + { "GRID K520", { {"VW3",1}, {"WGS3",64}, {"WPT3",1} } }, + { "GeForce GTX 1070", { {"VW3",1}, {"WGS3",128}, {"WPT3",1} } }, + { "GeForce GTX 480", { {"VW3",1}, {"WGS3",64}, {"WPT3",1} } }, + { "GeForce GTX 670", { {"VW3",1}, {"WGS3",64}, {"WPT3",1} } }, + { "GeForce GTX 680", { {"VW3",2}, {"WGS3",128}, {"WPT3",2} } }, + { "GeForce GTX 750", { {"VW3",2}, {"WGS3",64}, {"WPT3",2} } }, + { "GeForce GTX 750 Ti", { {"VW3",2}, {"WGS3",256}, {"WPT3",2} } }, + { "GeForce GTX 980", { {"VW3",1}, {"WGS3",64}, {"WPT3",1} } }, + { "GeForce GTX TITAN", { {"VW3",1}, {"WGS3",256}, {"WPT3",1} } }, + { "GeForce GTX TITAN X", { {"VW3",1}, {"WGS3",128}, {"WPT3",1} } }, + { "Tesla K20m", { {"VW3",1}, {"WGS3",64}, {"WPT3",1} } }, + { "Tesla K40m", { {"VW3",1}, {"WGS3",256}, {"WPT3",1} } }, + { "default", { {"VW3",1}, {"WGS3",64}, {"WPT3",1} } }, + } + }, + { // Default + kDeviceTypeAll, "default", { + { "default", { {"VW3",1}, {"WGS3",64}, {"WPT3",1} } }, + } + }, + } +}; + +// ================================================================================================= + +const Database::DatabaseEntry Database::XgemvFastRotComplexDouble = { + "XgemvFastRot", Precision::kComplexDouble, { + { // AMD GPUs + kDeviceTypeGPU, "AMD", { + { "AMD Radeon R9 M370X Compute Engine", { {"VW3",1}, {"WGS3",128}, {"WPT3",1} } }, + { "Hawaii", { {"VW3",1}, {"WGS3",64}, {"WPT3",1} } }, + { "Oland", { {"VW3",1}, {"WGS3",256}, {"WPT3",1} } }, + { "Pitcairn", { {"VW3",1}, {"WGS3",64}, {"WPT3",1} } }, + { "Tahiti", { {"VW3",1}, {"WGS3",64}, {"WPT3",1} } }, + { "default", { {"VW3",1}, {"WGS3",64}, {"WPT3",1} } }, + } + }, + { // Intel CPUs + kDeviceTypeCPU, "Intel", { + { "Intel(R) Core(TM) i5-6200U CPU @ 2.30GHz", { {"VW3",1}, {"WGS3",256}, {"WPT3",1} } }, + { "Intel(R) Core(TM) i7-3770 CPU @ 3.40GHz", { {"VW3",2}, {"WGS3",256}, {"WPT3",2} } }, + { "default", { {"VW3",1}, {"WGS3",256}, {"WPT3",1} } }, + } + }, + { // Intel accelerators + kDeviceTypeAccelerator, "Intel", { + { "Intel(R) Many Integrated Core Acceleration Card", { {"VW3",1}, {"WGS3",64}, {"WPT3",1} } }, + { "default", { {"VW3",1}, {"WGS3",64}, {"WPT3",1} } }, + } + }, + { // NVIDIA GPUs + kDeviceTypeGPU, "NVIDIA", { + { "GRID K520", { {"VW3",1}, {"WGS3",256}, {"WPT3",1} } }, + { "GeForce GTX 480", { {"VW3",1}, {"WGS3",64}, {"WPT3",1} } }, + { "GeForce GTX 670", { {"VW3",1}, {"WGS3",64}, {"WPT3",1} } }, + { "default", { {"VW3",1}, {"WGS3",64}, {"WPT3",1} } }, + } + }, + { // Default + kDeviceTypeAll, "default", { + { "default", { {"VW3",1}, {"WGS3",64}, {"WPT3",1} } }, + } + }, + } +}; + +// ================================================================================================= +} // namespace clblast diff --git a/src/routines/level2/xgemv.cpp b/src/routines/level2/xgemv.cpp index e4d407c8..4e32ba41 100644 --- a/src/routines/level2/xgemv.cpp +++ b/src/routines/level2/xgemv.cpp @@ -22,7 +22,7 @@ namespace clblast { // Constructor: forwards to base class constructor template <typename T> Xgemv<T>::Xgemv(Queue &queue, EventPointer event, const std::string &name): - Routine(queue, event, name, {"Pad", "Xgemv"}, PrecisionValue<T>()) { + Routine(queue, event, name, {"Pad", "Xgemv", "XgemvFast", "XgemvFastRot"}, PrecisionValue<T>()) { source_string_ = #include "../../kernels/level2/xgemv.opencl" #include "../../kernels/level2/xgemv_fast.opencl" diff --git a/src/tuning/kernels/xgemv.cpp b/src/tuning/kernels/xgemv.cpp index 96d4a5f2..7229602d 100644 --- a/src/tuning/kernels/xgemv.cpp +++ b/src/tuning/kernels/xgemv.cpp @@ -29,7 +29,7 @@ class TuneXgemv { public: // The representative kernel and the source code - static std::string KernelFamily() { return "xgemv_"+std::to_string(V); } + static std::string KernelFamily() { return (V==1) ? "xgemv" : ((V==2) ? "xgemv_fast" : "xgemv_fast_rot"); } static std::string KernelName() { return (V==1) ? "Xgemv" : ((V==2) ? "XgemvFast" : "XgemvFastRot"); } static std::string GetSources() { return |