summaryrefslogtreecommitdiff
path: root/src/database/kernels
diff options
context:
space:
mode:
authorCedric Nugteren <web@cedricnugteren.nl>2016-10-16 11:43:03 +0200
committerCedric Nugteren <web@cedricnugteren.nl>2016-10-16 11:43:03 +0200
commit9331442a56f2969e5d4650bf0c31dd84473a4085 (patch)
tree997aca9619e211fab040bde658e2f0f1519d2171 /src/database/kernels
parent8d5747aa54b88812ef4060328e3befdb13f3f45a (diff)
parent53deed298facc005f71e9ec2a96d5c4e90d826af (diff)
Merge branch 'development' into netlib_blas_api
Diffstat (limited to 'src/database/kernels')
-rw-r--r--src/database/kernels/copy.hpp24
-rw-r--r--src/database/kernels/pad.hpp24
-rw-r--r--src/database/kernels/padtranspose.hpp16
-rw-r--r--src/database/kernels/transpose.hpp18
-rw-r--r--src/database/kernels/xaxpy.hpp20
-rw-r--r--src/database/kernels/xdot.hpp18
-rw-r--r--src/database/kernels/xgemm.hpp24
-rw-r--r--src/database/kernels/xgemm_direct.hpp138
-rw-r--r--src/database/kernels/xgemv.hpp20
-rw-r--r--src/database/kernels/xgemv_fast.hpp21
-rw-r--r--src/database/kernels/xgemv_fast_rot.hpp20
-rw-r--r--src/database/kernels/xger.hpp22
12 files changed, 278 insertions, 87 deletions
diff --git a/src/database/kernels/copy.hpp b/src/database/kernels/copy.hpp
index a6b7dfe8..16aa6b3f 100644
--- a/src/database/kernels/copy.hpp
+++ b/src/database/kernels/copy.hpp
@@ -12,9 +12,10 @@
// =================================================================================================
namespace clblast {
+namespace database {
// =================================================================================================
-const Database::DatabaseEntry Database::CopyHalf = {
+const Database::DatabaseEntry CopyHalf = {
"Copy", Precision::kHalf, {
{ // Intel GPUs
kDeviceTypeGPU, "Intel", {
@@ -33,7 +34,7 @@ const Database::DatabaseEntry Database::CopyHalf = {
// =================================================================================================
-const Database::DatabaseEntry Database::CopySingle = {
+const Database::DatabaseEntry CopySingle = {
"Copy", Precision::kSingle, {
{ // AMD GPUs
kDeviceTypeGPU, "AMD", {
@@ -64,6 +65,7 @@ const Database::DatabaseEntry Database::CopySingle = {
{ "Intel(R) HD Graphics 530", { {"COPY_DIMX",8}, {"COPY_DIMY",8}, {"COPY_VW",2}, {"COPY_WPT",1} } },
{ "Intel(R) HD Graphics 5500 BroadWell U-Processor GT2", { {"COPY_DIMX",32}, {"COPY_DIMY",16}, {"COPY_VW",4}, {"COPY_WPT",1} } },
{ "Intel(R) HD Graphics Haswell Ultrabook GT2 Mobile", { {"COPY_DIMX",32}, {"COPY_DIMY",16}, {"COPY_VW",4}, {"COPY_WPT",1} } },
+ { "Intel(R) HD Graphics IvyBridge M GT2", { {"COPY_DIMX",16}, {"COPY_DIMY",8}, {"COPY_VW",2}, {"COPY_WPT",1} } },
{ "Intel(R) HD Graphics Skylake ULT GT2", { {"COPY_DIMX",32}, {"COPY_DIMY",8}, {"COPY_VW",2}, {"COPY_WPT",1} } },
{ "Iris", { {"COPY_DIMX",16}, {"COPY_DIMY",8}, {"COPY_VW",1}, {"COPY_WPT",2} } },
{ "Iris Pro", { {"COPY_DIMX",32}, {"COPY_DIMY",8}, {"COPY_VW",4}, {"COPY_WPT",4} } },
@@ -84,7 +86,7 @@ const Database::DatabaseEntry Database::CopySingle = {
{ "GeForce GTX 670", { {"COPY_DIMX",16}, {"COPY_DIMY",32}, {"COPY_VW",4}, {"COPY_WPT",1} } },
{ "GeForce GTX 680", { {"COPY_DIMX",32}, {"COPY_DIMY",16}, {"COPY_VW",4}, {"COPY_WPT",1} } },
{ "GeForce GTX 750", { {"COPY_DIMX",32}, {"COPY_DIMY",8}, {"COPY_VW",2}, {"COPY_WPT",2} } },
- { "GeForce GTX 750 Ti", { {"COPY_DIMX",32}, {"COPY_DIMY",8}, {"COPY_VW",2}, {"COPY_WPT",1} } },
+ { "GeForce GTX 750 Ti", { {"COPY_DIMX",16}, {"COPY_DIMY",32}, {"COPY_VW",2}, {"COPY_WPT",2} } },
{ "GeForce GTX 980", { {"COPY_DIMX",32}, {"COPY_DIMY",16}, {"COPY_VW",1}, {"COPY_WPT",1} } },
{ "GeForce GTX TITAN", { {"COPY_DIMX",32}, {"COPY_DIMY",8}, {"COPY_VW",2}, {"COPY_WPT",4} } },
{ "GeForce GTX TITAN X", { {"COPY_DIMX",32}, {"COPY_DIMY",8}, {"COPY_VW",1}, {"COPY_WPT",2} } },
@@ -103,7 +105,7 @@ const Database::DatabaseEntry Database::CopySingle = {
// =================================================================================================
-const Database::DatabaseEntry Database::CopyComplexSingle = {
+const Database::DatabaseEntry CopyComplexSingle = {
"Copy", Precision::kComplexSingle, {
{ // AMD GPUs
kDeviceTypeGPU, "AMD", {
@@ -128,6 +130,7 @@ const Database::DatabaseEntry Database::CopyComplexSingle = {
{ "Intel(R) HD Graphics 530", { {"COPY_DIMX",16}, {"COPY_DIMY",8}, {"COPY_VW",2}, {"COPY_WPT",1} } },
{ "Intel(R) HD Graphics 5500 BroadWell U-Processor GT2", { {"COPY_DIMX",16}, {"COPY_DIMY",16}, {"COPY_VW",2}, {"COPY_WPT",2} } },
{ "Intel(R) HD Graphics Haswell Ultrabook GT2 Mobile", { {"COPY_DIMX",8}, {"COPY_DIMY",8}, {"COPY_VW",1}, {"COPY_WPT",1} } },
+ { "Intel(R) HD Graphics IvyBridge M GT2", { {"COPY_DIMX",8}, {"COPY_DIMY",32}, {"COPY_VW",2}, {"COPY_WPT",4} } },
{ "Intel(R) HD Graphics Skylake ULT GT2", { {"COPY_DIMX",8}, {"COPY_DIMY",8}, {"COPY_VW",4}, {"COPY_WPT",4} } },
{ "Iris", { {"COPY_DIMX",16}, {"COPY_DIMY",8}, {"COPY_VW",1}, {"COPY_WPT",2} } },
{ "Iris Pro", { {"COPY_DIMX",32}, {"COPY_DIMY",16}, {"COPY_VW",1}, {"COPY_WPT",4} } },
@@ -147,7 +150,7 @@ const Database::DatabaseEntry Database::CopyComplexSingle = {
{ "GeForce GTX 480", { {"COPY_DIMX",16}, {"COPY_DIMY",16}, {"COPY_VW",1}, {"COPY_WPT",1} } },
{ "GeForce GTX 670", { {"COPY_DIMX",16}, {"COPY_DIMY",8}, {"COPY_VW",1}, {"COPY_WPT",1} } },
{ "GeForce GTX 750", { {"COPY_DIMX",16}, {"COPY_DIMY",8}, {"COPY_VW",1}, {"COPY_WPT",2} } },
- { "GeForce GTX 750 Ti", { {"COPY_DIMX",32}, {"COPY_DIMY",16}, {"COPY_VW",1}, {"COPY_WPT",1} } },
+ { "GeForce GTX 750 Ti", { {"COPY_DIMX",16}, {"COPY_DIMY",32}, {"COPY_VW",1}, {"COPY_WPT",1} } },
{ "GeForce GTX 980", { {"COPY_DIMX",8}, {"COPY_DIMY",8}, {"COPY_VW",1}, {"COPY_WPT",1} } },
{ "GeForce GTX TITAN X", { {"COPY_DIMX",16}, {"COPY_DIMY",8}, {"COPY_VW",1}, {"COPY_WPT",1} } },
{ "Tesla K20m", { {"COPY_DIMX",8}, {"COPY_DIMY",8}, {"COPY_VW",1}, {"COPY_WPT",4} } },
@@ -165,7 +168,7 @@ const Database::DatabaseEntry Database::CopyComplexSingle = {
// =================================================================================================
-const Database::DatabaseEntry Database::CopyDouble = {
+const Database::DatabaseEntry CopyDouble = {
"Copy", Precision::kDouble, {
{ // AMD GPUs
kDeviceTypeGPU, "AMD", {
@@ -205,13 +208,13 @@ const Database::DatabaseEntry Database::CopyDouble = {
{ "GeForce GTX 670", { {"COPY_DIMX",8}, {"COPY_DIMY",8}, {"COPY_VW",2}, {"COPY_WPT",1} } },
{ "GeForce GTX 680", { {"COPY_DIMX",16}, {"COPY_DIMY",32}, {"COPY_VW",2}, {"COPY_WPT",1} } },
{ "GeForce GTX 750", { {"COPY_DIMX",8}, {"COPY_DIMY",16}, {"COPY_VW",2}, {"COPY_WPT",1} } },
- { "GeForce GTX 750 Ti", { {"COPY_DIMX",8}, {"COPY_DIMY",8}, {"COPY_VW",2}, {"COPY_WPT",1} } },
+ { "GeForce GTX 750 Ti", { {"COPY_DIMX",16}, {"COPY_DIMY",8}, {"COPY_VW",2}, {"COPY_WPT",1} } },
{ "GeForce GTX 980", { {"COPY_DIMX",32}, {"COPY_DIMY",8}, {"COPY_VW",2}, {"COPY_WPT",1} } },
{ "GeForce GTX TITAN", { {"COPY_DIMX",16}, {"COPY_DIMY",32}, {"COPY_VW",2}, {"COPY_WPT",2} } },
{ "GeForce GTX TITAN X", { {"COPY_DIMX",32}, {"COPY_DIMY",16}, {"COPY_VW",1}, {"COPY_WPT",1} } },
{ "Tesla K20m", { {"COPY_DIMX",8}, {"COPY_DIMY",8}, {"COPY_VW",2}, {"COPY_WPT",1} } },
{ "Tesla K40m", { {"COPY_DIMX",8}, {"COPY_DIMY",8}, {"COPY_VW",2}, {"COPY_WPT",2} } },
- { "default", { {"COPY_DIMX",32}, {"COPY_DIMY",32}, {"COPY_VW",2}, {"COPY_WPT",1} } },
+ { "default", { {"COPY_DIMX",32}, {"COPY_DIMY",16}, {"COPY_VW",2}, {"COPY_WPT",1} } },
}
},
{ // Default
@@ -224,7 +227,7 @@ const Database::DatabaseEntry Database::CopyDouble = {
// =================================================================================================
-const Database::DatabaseEntry Database::CopyComplexDouble = {
+const Database::DatabaseEntry CopyComplexDouble = {
"Copy", Precision::kComplexDouble, {
{ // AMD GPUs
kDeviceTypeGPU, "AMD", {
@@ -264,7 +267,7 @@ const Database::DatabaseEntry Database::CopyComplexDouble = {
{ "GeForce GTX 670", { {"COPY_DIMX",16}, {"COPY_DIMY",8}, {"COPY_VW",1}, {"COPY_WPT",1} } },
{ "GeForce GTX 680", { {"COPY_DIMX",8}, {"COPY_DIMY",8}, {"COPY_VW",1}, {"COPY_WPT",1} } },
{ "GeForce GTX 750", { {"COPY_DIMX",32}, {"COPY_DIMY",8}, {"COPY_VW",1}, {"COPY_WPT",1} } },
- { "GeForce GTX 750 Ti", { {"COPY_DIMX",32}, {"COPY_DIMY",16}, {"COPY_VW",1}, {"COPY_WPT",1} } },
+ { "GeForce GTX 750 Ti", { {"COPY_DIMX",16}, {"COPY_DIMY",16}, {"COPY_VW",1}, {"COPY_WPT",1} } },
{ "GeForce GTX 980", { {"COPY_DIMX",8}, {"COPY_DIMY",8}, {"COPY_VW",1}, {"COPY_WPT",1} } },
{ "GeForce GTX TITAN", { {"COPY_DIMX",16}, {"COPY_DIMY",16}, {"COPY_VW",1}, {"COPY_WPT",1} } },
{ "GeForce GTX TITAN X", { {"COPY_DIMX",16}, {"COPY_DIMY",8}, {"COPY_VW",1}, {"COPY_WPT",1} } },
@@ -282,4 +285,5 @@ const Database::DatabaseEntry Database::CopyComplexDouble = {
};
// =================================================================================================
+} // namespace database
} // namespace clblast
diff --git a/src/database/kernels/pad.hpp b/src/database/kernels/pad.hpp
index 3cfabaf4..6c5e0c2f 100644
--- a/src/database/kernels/pad.hpp
+++ b/src/database/kernels/pad.hpp
@@ -12,9 +12,10 @@
// =================================================================================================
namespace clblast {
+namespace database {
// =================================================================================================
-const Database::DatabaseEntry Database::PadHalf = {
+const Database::DatabaseEntry PadHalf = {
"Pad", Precision::kHalf, {
{ // Intel GPUs
kDeviceTypeGPU, "Intel", {
@@ -33,7 +34,7 @@ const Database::DatabaseEntry Database::PadHalf = {
// =================================================================================================
-const Database::DatabaseEntry Database::PadSingle = {
+const Database::DatabaseEntry PadSingle = {
"Pad", Precision::kSingle, {
{ // AMD GPUs
kDeviceTypeGPU, "AMD", {
@@ -64,6 +65,7 @@ const Database::DatabaseEntry Database::PadSingle = {
{ "Intel(R) HD Graphics 530", { {"PAD_DIMX",32}, {"PAD_DIMY",8}, {"PAD_WPTX",2}, {"PAD_WPTY",4} } },
{ "Intel(R) HD Graphics 5500 BroadWell U-Processor GT2", { {"PAD_DIMX",32}, {"PAD_DIMY",8}, {"PAD_WPTX",2}, {"PAD_WPTY",4} } },
{ "Intel(R) HD Graphics Haswell Ultrabook GT2 Mobile", { {"PAD_DIMX",16}, {"PAD_DIMY",8}, {"PAD_WPTX",1}, {"PAD_WPTY",2} } },
+ { "Intel(R) HD Graphics IvyBridge M GT2", { {"PAD_DIMX",16}, {"PAD_DIMY",8}, {"PAD_WPTX",4}, {"PAD_WPTY",1} } },
{ "Intel(R) HD Graphics Skylake ULT GT2", { {"PAD_DIMX",16}, {"PAD_DIMY",8}, {"PAD_WPTX",2}, {"PAD_WPTY",1} } },
{ "Iris", { {"PAD_DIMX",32}, {"PAD_DIMY",16}, {"PAD_WPTX",2}, {"PAD_WPTY",1} } },
{ "Iris Pro", { {"PAD_DIMX",16}, {"PAD_DIMY",8}, {"PAD_WPTX",2}, {"PAD_WPTY",1} } },
@@ -84,7 +86,7 @@ const Database::DatabaseEntry Database::PadSingle = {
{ "GeForce GTX 670", { {"PAD_DIMX",32}, {"PAD_DIMY",8}, {"PAD_WPTX",2}, {"PAD_WPTY",2} } },
{ "GeForce GTX 680", { {"PAD_DIMX",16}, {"PAD_DIMY",8}, {"PAD_WPTX",4}, {"PAD_WPTY",1} } },
{ "GeForce GTX 750", { {"PAD_DIMX",32}, {"PAD_DIMY",16}, {"PAD_WPTX",4}, {"PAD_WPTY",2} } },
- { "GeForce GTX 750 Ti", { {"PAD_DIMX",32}, {"PAD_DIMY",8}, {"PAD_WPTX",4}, {"PAD_WPTY",1} } },
+ { "GeForce GTX 750 Ti", { {"PAD_DIMX",16}, {"PAD_DIMY",8}, {"PAD_WPTX",4}, {"PAD_WPTY",1} } },
{ "GeForce GTX 980", { {"PAD_DIMX",16}, {"PAD_DIMY",8}, {"PAD_WPTX",1}, {"PAD_WPTY",1} } },
{ "GeForce GTX TITAN", { {"PAD_DIMX",32}, {"PAD_DIMY",8}, {"PAD_WPTX",2}, {"PAD_WPTY",1} } },
{ "GeForce GTX TITAN X", { {"PAD_DIMX",16}, {"PAD_DIMY",16}, {"PAD_WPTX",1}, {"PAD_WPTY",1} } },
@@ -103,7 +105,7 @@ const Database::DatabaseEntry Database::PadSingle = {
// =================================================================================================
-const Database::DatabaseEntry Database::PadComplexSingle = {
+const Database::DatabaseEntry PadComplexSingle = {
"Pad", Precision::kComplexSingle, {
{ // AMD GPUs
kDeviceTypeGPU, "AMD", {
@@ -134,10 +136,11 @@ const Database::DatabaseEntry Database::PadComplexSingle = {
{ "Intel(R) HD Graphics 530", { {"PAD_DIMX",8}, {"PAD_DIMY",8}, {"PAD_WPTX",1}, {"PAD_WPTY",2} } },
{ "Intel(R) HD Graphics 5500 BroadWell U-Processor GT2", { {"PAD_DIMX",8}, {"PAD_DIMY",8}, {"PAD_WPTX",1}, {"PAD_WPTY",1} } },
{ "Intel(R) HD Graphics Haswell Ultrabook GT2 Mobile", { {"PAD_DIMX",8}, {"PAD_DIMY",8}, {"PAD_WPTX",1}, {"PAD_WPTY",1} } },
+ { "Intel(R) HD Graphics IvyBridge M GT2", { {"PAD_DIMX",32}, {"PAD_DIMY",8}, {"PAD_WPTX",1}, {"PAD_WPTY",1} } },
{ "Intel(R) HD Graphics Skylake ULT GT2", { {"PAD_DIMX",32}, {"PAD_DIMY",8}, {"PAD_WPTX",2}, {"PAD_WPTY",4} } },
{ "Iris", { {"PAD_DIMX",32}, {"PAD_DIMY",16}, {"PAD_WPTX",2}, {"PAD_WPTY",4} } },
{ "Iris Pro", { {"PAD_DIMX",32}, {"PAD_DIMY",8}, {"PAD_WPTX",2}, {"PAD_WPTY",1} } },
- { "default", { {"PAD_DIMX",32}, {"PAD_DIMY",8}, {"PAD_WPTX",1}, {"PAD_WPTY",4} } },
+ { "default", { {"PAD_DIMX",32}, {"PAD_DIMY",8}, {"PAD_WPTX",1}, {"PAD_WPTY",2} } },
}
},
{ // Intel accelerators
@@ -154,13 +157,13 @@ const Database::DatabaseEntry Database::PadComplexSingle = {
{ "GeForce GTX 670", { {"PAD_DIMX",16}, {"PAD_DIMY",8}, {"PAD_WPTX",1}, {"PAD_WPTY",2} } },
{ "GeForce GTX 680", { {"PAD_DIMX",16}, {"PAD_DIMY",32}, {"PAD_WPTX",1}, {"PAD_WPTY",2} } },
{ "GeForce GTX 750", { {"PAD_DIMX",32}, {"PAD_DIMY",8}, {"PAD_WPTX",2}, {"PAD_WPTY",1} } },
- { "GeForce GTX 750 Ti", { {"PAD_DIMX",16}, {"PAD_DIMY",16}, {"PAD_WPTX",1}, {"PAD_WPTY",1} } },
+ { "GeForce GTX 750 Ti", { {"PAD_DIMX",16}, {"PAD_DIMY",8}, {"PAD_WPTX",1}, {"PAD_WPTY",1} } },
{ "GeForce GTX 980", { {"PAD_DIMX",16}, {"PAD_DIMY",16}, {"PAD_WPTX",1}, {"PAD_WPTY",1} } },
{ "GeForce GTX TITAN", { {"PAD_DIMX",16}, {"PAD_DIMY",8}, {"PAD_WPTX",2}, {"PAD_WPTY",1} } },
{ "GeForce GTX TITAN X", { {"PAD_DIMX",16}, {"PAD_DIMY",8}, {"PAD_WPTX",1}, {"PAD_WPTY",1} } },
{ "Tesla K20m", { {"PAD_DIMX",32}, {"PAD_DIMY",8}, {"PAD_WPTX",1}, {"PAD_WPTY",2} } },
{ "Tesla K40m", { {"PAD_DIMX",16}, {"PAD_DIMY",8}, {"PAD_WPTX",1}, {"PAD_WPTY",1} } },
- { "default", { {"PAD_DIMX",16}, {"PAD_DIMY",8}, {"PAD_WPTX",2}, {"PAD_WPTY",1} } },
+ { "default", { {"PAD_DIMX",16}, {"PAD_DIMY",8}, {"PAD_WPTX",1}, {"PAD_WPTY",1} } },
}
},
{ // Default
@@ -173,7 +176,7 @@ const Database::DatabaseEntry Database::PadComplexSingle = {
// =================================================================================================
-const Database::DatabaseEntry Database::PadDouble = {
+const Database::DatabaseEntry PadDouble = {
"Pad", Precision::kDouble, {
{ // AMD GPUs
kDeviceTypeGPU, "AMD", {
@@ -232,7 +235,7 @@ const Database::DatabaseEntry Database::PadDouble = {
// =================================================================================================
-const Database::DatabaseEntry Database::PadComplexDouble = {
+const Database::DatabaseEntry PadComplexDouble = {
"Pad", Precision::kComplexDouble, {
{ // AMD GPUs
kDeviceTypeGPU, "AMD", {
@@ -272,7 +275,7 @@ const Database::DatabaseEntry Database::PadComplexDouble = {
{ "GeForce GTX 670", { {"PAD_DIMX",32}, {"PAD_DIMY",8}, {"PAD_WPTX",1}, {"PAD_WPTY",1} } },
{ "GeForce GTX 680", { {"PAD_DIMX",8}, {"PAD_DIMY",8}, {"PAD_WPTX",1}, {"PAD_WPTY",1} } },
{ "GeForce GTX 750", { {"PAD_DIMX",8}, {"PAD_DIMY",8}, {"PAD_WPTX",1}, {"PAD_WPTY",1} } },
- { "GeForce GTX 750 Ti", { {"PAD_DIMX",32}, {"PAD_DIMY",32}, {"PAD_WPTX",1}, {"PAD_WPTY",1} } },
+ { "GeForce GTX 750 Ti", { {"PAD_DIMX",16}, {"PAD_DIMY",32}, {"PAD_WPTX",1}, {"PAD_WPTY",1} } },
{ "GeForce GTX 980", { {"PAD_DIMX",16}, {"PAD_DIMY",16}, {"PAD_WPTX",1}, {"PAD_WPTY",1} } },
{ "GeForce GTX TITAN", { {"PAD_DIMX",8}, {"PAD_DIMY",32}, {"PAD_WPTX",1}, {"PAD_WPTY",2} } },
{ "GeForce GTX TITAN X", { {"PAD_DIMX",16}, {"PAD_DIMY",8}, {"PAD_WPTX",1}, {"PAD_WPTY",1} } },
@@ -290,4 +293,5 @@ const Database::DatabaseEntry Database::PadComplexDouble = {
};
// =================================================================================================
+} // namespace database
} // namespace clblast
diff --git a/src/database/kernels/padtranspose.hpp b/src/database/kernels/padtranspose.hpp
index 88bd4ea7..4003ec6d 100644
--- a/src/database/kernels/padtranspose.hpp
+++ b/src/database/kernels/padtranspose.hpp
@@ -12,9 +12,10 @@
// =================================================================================================
namespace clblast {
+namespace database {
// =================================================================================================
-const Database::DatabaseEntry Database::PadtransposeHalf = {
+const Database::DatabaseEntry PadtransposeHalf = {
"Padtranspose", Precision::kHalf, {
{ // Intel GPUs
kDeviceTypeGPU, "Intel", {
@@ -33,7 +34,7 @@ const Database::DatabaseEntry Database::PadtransposeHalf = {
// =================================================================================================
-const Database::DatabaseEntry Database::PadtransposeSingle = {
+const Database::DatabaseEntry PadtransposeSingle = {
"Padtranspose", Precision::kSingle, {
{ // AMD GPUs
kDeviceTypeGPU, "AMD", {
@@ -64,6 +65,7 @@ const Database::DatabaseEntry Database::PadtransposeSingle = {
{ "Intel(R) HD Graphics 530", { {"PADTRA_PAD",1}, {"PADTRA_TILE",16}, {"PADTRA_WPT",2} } },
{ "Intel(R) HD Graphics 5500 BroadWell U-Processor GT2", { {"PADTRA_PAD",0}, {"PADTRA_TILE",16}, {"PADTRA_WPT",4} } },
{ "Intel(R) HD Graphics Haswell Ultrabook GT2 Mobile", { {"PADTRA_PAD",1}, {"PADTRA_TILE",16}, {"PADTRA_WPT",2} } },
+ { "Intel(R) HD Graphics IvyBridge M GT2", { {"PADTRA_PAD",0}, {"PADTRA_TILE",16}, {"PADTRA_WPT",4} } },
{ "Intel(R) HD Graphics Skylake ULT GT2", { {"PADTRA_PAD",1}, {"PADTRA_TILE",16}, {"PADTRA_WPT",2} } },
{ "Iris", { {"PADTRA_PAD",1}, {"PADTRA_TILE",16}, {"PADTRA_WPT",2} } },
{ "Iris Pro", { {"PADTRA_PAD",1}, {"PADTRA_TILE",16}, {"PADTRA_WPT",2} } },
@@ -103,7 +105,7 @@ const Database::DatabaseEntry Database::PadtransposeSingle = {
// =================================================================================================
-const Database::DatabaseEntry Database::PadtransposeComplexSingle = {
+const Database::DatabaseEntry PadtransposeComplexSingle = {
"Padtranspose", Precision::kComplexSingle, {
{ // AMD GPUs
kDeviceTypeGPU, "AMD", {
@@ -134,6 +136,7 @@ const Database::DatabaseEntry Database::PadtransposeComplexSingle = {
{ "Intel(R) HD Graphics 530", { {"PADTRA_PAD",1}, {"PADTRA_TILE",16}, {"PADTRA_WPT",2} } },
{ "Intel(R) HD Graphics 5500 BroadWell U-Processor GT2", { {"PADTRA_PAD",0}, {"PADTRA_TILE",16}, {"PADTRA_WPT",2} } },
{ "Intel(R) HD Graphics Haswell Ultrabook GT2 Mobile", { {"PADTRA_PAD",1}, {"PADTRA_TILE",16}, {"PADTRA_WPT",2} } },
+ { "Intel(R) HD Graphics IvyBridge M GT2", { {"PADTRA_PAD",0}, {"PADTRA_TILE",16}, {"PADTRA_WPT",2} } },
{ "Intel(R) HD Graphics Skylake ULT GT2", { {"PADTRA_PAD",0}, {"PADTRA_TILE",16}, {"PADTRA_WPT",4} } },
{ "Iris", { {"PADTRA_PAD",0}, {"PADTRA_TILE",16}, {"PADTRA_WPT",2} } },
{ "Iris Pro", { {"PADTRA_PAD",1}, {"PADTRA_TILE",16}, {"PADTRA_WPT",2} } },
@@ -173,7 +176,7 @@ const Database::DatabaseEntry Database::PadtransposeComplexSingle = {
// =================================================================================================
-const Database::DatabaseEntry Database::PadtransposeDouble = {
+const Database::DatabaseEntry PadtransposeDouble = {
"Padtranspose", Precision::kDouble, {
{ // AMD GPUs
kDeviceTypeGPU, "AMD", {
@@ -232,7 +235,7 @@ const Database::DatabaseEntry Database::PadtransposeDouble = {
// =================================================================================================
-const Database::DatabaseEntry Database::PadtransposeComplexDouble = {
+const Database::DatabaseEntry PadtransposeComplexDouble = {
"Padtranspose", Precision::kComplexDouble, {
{ // AMD GPUs
kDeviceTypeGPU, "AMD", {
@@ -272,7 +275,7 @@ const Database::DatabaseEntry Database::PadtransposeComplexDouble = {
{ "GeForce GTX 670", { {"PADTRA_PAD",1}, {"PADTRA_TILE",16}, {"PADTRA_WPT",1} } },
{ "GeForce GTX 680", { {"PADTRA_PAD",1}, {"PADTRA_TILE",32}, {"PADTRA_WPT",1} } },
{ "GeForce GTX 750", { {"PADTRA_PAD",1}, {"PADTRA_TILE",16}, {"PADTRA_WPT",1} } },
- { "GeForce GTX 750 Ti", { {"PADTRA_PAD",1}, {"PADTRA_TILE",16}, {"PADTRA_WPT",2} } },
+ { "GeForce GTX 750 Ti", { {"PADTRA_PAD",1}, {"PADTRA_TILE",8}, {"PADTRA_WPT",2} } },
{ "GeForce GTX 980", { {"PADTRA_PAD",0}, {"PADTRA_TILE",16}, {"PADTRA_WPT",1} } },
{ "GeForce GTX TITAN", { {"PADTRA_PAD",1}, {"PADTRA_TILE",16}, {"PADTRA_WPT",1} } },
{ "GeForce GTX TITAN X", { {"PADTRA_PAD",1}, {"PADTRA_TILE",32}, {"PADTRA_WPT",1} } },
@@ -290,4 +293,5 @@ const Database::DatabaseEntry Database::PadtransposeComplexDouble = {
};
// =================================================================================================
+} // namespace database
} // namespace clblast
diff --git a/src/database/kernels/transpose.hpp b/src/database/kernels/transpose.hpp
index 0e1b608e..c5ea50c2 100644
--- a/src/database/kernels/transpose.hpp
+++ b/src/database/kernels/transpose.hpp
@@ -12,9 +12,10 @@
// =================================================================================================
namespace clblast {
+namespace database {
// =================================================================================================
-const Database::DatabaseEntry Database::TransposeHalf = {
+const Database::DatabaseEntry TransposeHalf = {
"Transpose", Precision::kHalf, {
{ // Intel GPUs
kDeviceTypeGPU, "Intel", {
@@ -33,7 +34,7 @@ const Database::DatabaseEntry Database::TransposeHalf = {
// =================================================================================================
-const Database::DatabaseEntry Database::TransposeSingle = {
+const Database::DatabaseEntry TransposeSingle = {
"Transpose", Precision::kSingle, {
{ // AMD GPUs
kDeviceTypeGPU, "AMD", {
@@ -64,6 +65,7 @@ const Database::DatabaseEntry Database::TransposeSingle = {
{ "Intel(R) HD Graphics 530", { {"TRA_DIM",16}, {"TRA_PAD",0}, {"TRA_SHUFFLE",1}, {"TRA_WPT",4} } },
{ "Intel(R) HD Graphics 5500 BroadWell U-Processor GT2", { {"TRA_DIM",16}, {"TRA_PAD",0}, {"TRA_SHUFFLE",0}, {"TRA_WPT",4} } },
{ "Intel(R) HD Graphics Haswell Ultrabook GT2 Mobile", { {"TRA_DIM",16}, {"TRA_PAD",0}, {"TRA_SHUFFLE",0}, {"TRA_WPT",4} } },
+ { "Intel(R) HD Graphics IvyBridge M GT2", { {"TRA_DIM",8}, {"TRA_PAD",0}, {"TRA_SHUFFLE",1}, {"TRA_WPT",4} } },
{ "Intel(R) HD Graphics Skylake ULT GT2", { {"TRA_DIM",8}, {"TRA_PAD",0}, {"TRA_SHUFFLE",1}, {"TRA_WPT",4} } },
{ "Iris", { {"TRA_DIM",8}, {"TRA_PAD",1}, {"TRA_SHUFFLE",0}, {"TRA_WPT",4} } },
{ "Iris Pro", { {"TRA_DIM",16}, {"TRA_PAD",1}, {"TRA_SHUFFLE",0}, {"TRA_WPT",4} } },
@@ -103,7 +105,7 @@ const Database::DatabaseEntry Database::TransposeSingle = {
// =================================================================================================
-const Database::DatabaseEntry Database::TransposeComplexSingle = {
+const Database::DatabaseEntry TransposeComplexSingle = {
"Transpose", Precision::kComplexSingle, {
{ // AMD GPUs
kDeviceTypeGPU, "AMD", {
@@ -134,6 +136,7 @@ const Database::DatabaseEntry Database::TransposeComplexSingle = {
{ "Intel(R) HD Graphics 530", { {"TRA_DIM",16}, {"TRA_PAD",1}, {"TRA_SHUFFLE",1}, {"TRA_WPT",2} } },
{ "Intel(R) HD Graphics 5500 BroadWell U-Processor GT2", { {"TRA_DIM",8}, {"TRA_PAD",0}, {"TRA_SHUFFLE",0}, {"TRA_WPT",2} } },
{ "Intel(R) HD Graphics Haswell Ultrabook GT2 Mobile", { {"TRA_DIM",8}, {"TRA_PAD",0}, {"TRA_SHUFFLE",0}, {"TRA_WPT",2} } },
+ { "Intel(R) HD Graphics IvyBridge M GT2", { {"TRA_DIM",16}, {"TRA_PAD",1}, {"TRA_SHUFFLE",1}, {"TRA_WPT",2} } },
{ "Intel(R) HD Graphics Skylake ULT GT2", { {"TRA_DIM",16}, {"TRA_PAD",1}, {"TRA_SHUFFLE",1}, {"TRA_WPT",2} } },
{ "Iris", { {"TRA_DIM",8}, {"TRA_PAD",0}, {"TRA_SHUFFLE",0}, {"TRA_WPT",2} } },
{ "Iris Pro", { {"TRA_DIM",16}, {"TRA_PAD",1}, {"TRA_SHUFFLE",0}, {"TRA_WPT",2} } },
@@ -159,7 +162,7 @@ const Database::DatabaseEntry Database::TransposeComplexSingle = {
},
{ // Default
kDeviceTypeAll, "default", {
- { "default", { {"TRA_DIM",16}, {"TRA_PAD",1}, {"TRA_SHUFFLE",0}, {"TRA_WPT",1} } },
+ { "default", { {"TRA_DIM",8}, {"TRA_PAD",1}, {"TRA_SHUFFLE",1}, {"TRA_WPT",2} } },
}
},
}
@@ -167,7 +170,7 @@ const Database::DatabaseEntry Database::TransposeComplexSingle = {
// =================================================================================================
-const Database::DatabaseEntry Database::TransposeDouble = {
+const Database::DatabaseEntry TransposeDouble = {
"Transpose", Precision::kDouble, {
{ // AMD GPUs
kDeviceTypeGPU, "AMD", {
@@ -207,7 +210,7 @@ const Database::DatabaseEntry Database::TransposeDouble = {
{ "GeForce GTX 670", { {"TRA_DIM",16}, {"TRA_PAD",1}, {"TRA_SHUFFLE",1}, {"TRA_WPT",2} } },
{ "GeForce GTX 680", { {"TRA_DIM",16}, {"TRA_PAD",1}, {"TRA_SHUFFLE",1}, {"TRA_WPT",2} } },
{ "GeForce GTX 750", { {"TRA_DIM",16}, {"TRA_PAD",1}, {"TRA_SHUFFLE",0}, {"TRA_WPT",1} } },
- { "GeForce GTX 750 Ti", { {"TRA_DIM",32}, {"TRA_PAD",1}, {"TRA_SHUFFLE",0}, {"TRA_WPT",1} } },
+ { "GeForce GTX 750 Ti", { {"TRA_DIM",32}, {"TRA_PAD",1}, {"TRA_SHUFFLE",0}, {"TRA_WPT",2} } },
{ "GeForce GTX 980", { {"TRA_DIM",16}, {"TRA_PAD",0}, {"TRA_SHUFFLE",0}, {"TRA_WPT",2} } },
{ "GeForce GTX TITAN", { {"TRA_DIM",8}, {"TRA_PAD",0}, {"TRA_SHUFFLE",0}, {"TRA_WPT",2} } },
{ "GeForce GTX TITAN X", { {"TRA_DIM",32}, {"TRA_PAD",1}, {"TRA_SHUFFLE",0}, {"TRA_WPT",1} } },
@@ -226,7 +229,7 @@ const Database::DatabaseEntry Database::TransposeDouble = {
// =================================================================================================
-const Database::DatabaseEntry Database::TransposeComplexDouble = {
+const Database::DatabaseEntry TransposeComplexDouble = {
"Transpose", Precision::kComplexDouble, {
{ // AMD GPUs
kDeviceTypeGPU, "AMD", {
@@ -278,4 +281,5 @@ const Database::DatabaseEntry Database::TransposeComplexDouble = {
};
// =================================================================================================
+} // namespace database
} // namespace clblast
diff --git a/src/database/kernels/xaxpy.hpp b/src/database/kernels/xaxpy.hpp
index 9c1bcd99..60471bef 100644
--- a/src/database/kernels/xaxpy.hpp
+++ b/src/database/kernels/xaxpy.hpp
@@ -12,9 +12,10 @@
// =================================================================================================
namespace clblast {
+namespace database {
// =================================================================================================
-const Database::DatabaseEntry Database::XaxpyHalf = {
+const Database::DatabaseEntry XaxpyHalf = {
"Xaxpy", Precision::kHalf, {
{ // Intel GPUs
kDeviceTypeGPU, "Intel", {
@@ -33,7 +34,7 @@ const Database::DatabaseEntry Database::XaxpyHalf = {
// =================================================================================================
-const Database::DatabaseEntry Database::XaxpySingle = {
+const Database::DatabaseEntry XaxpySingle = {
"Xaxpy", Precision::kSingle, {
{ // AMD GPUs
kDeviceTypeGPU, "AMD", {
@@ -64,6 +65,7 @@ const Database::DatabaseEntry Database::XaxpySingle = {
{ "Intel(R) HD Graphics 530", { {"VW",1}, {"WGS",128}, {"WPT",1} } },
{ "Intel(R) HD Graphics 5500 BroadWell U-Processor GT2", { {"VW",1}, {"WGS",256}, {"WPT",1} } },
{ "Intel(R) HD Graphics Haswell Ultrabook GT2 Mobile", { {"VW",1}, {"WGS",64}, {"WPT",1} } },
+ { "Intel(R) HD Graphics IvyBridge M GT2", { {"VW",1}, {"WGS",64}, {"WPT",1} } },
{ "Intel(R) HD Graphics Skylake ULT GT2", { {"VW",1}, {"WGS",512}, {"WPT",2} } },
{ "Iris", { {"VW",1}, {"WGS",64}, {"WPT",1} } },
{ "Iris Pro", { {"VW",1}, {"WGS",128}, {"WPT",2} } },
@@ -84,7 +86,7 @@ const Database::DatabaseEntry Database::XaxpySingle = {
{ "GeForce GTX 670", { {"VW",2}, {"WGS",64}, {"WPT",1} } },
{ "GeForce GTX 680", { {"VW",1}, {"WGS",128}, {"WPT",1} } },
{ "GeForce GTX 750", { {"VW",1}, {"WGS",64}, {"WPT",1} } },
- { "GeForce GTX 750 Ti", { {"VW",1}, {"WGS",1024}, {"WPT",1} } },
+ { "GeForce GTX 750 Ti", { {"VW",2}, {"WGS",64}, {"WPT",1} } },
{ "GeForce GTX 980", { {"VW",1}, {"WGS",1024}, {"WPT",1} } },
{ "GeForce GTX TITAN", { {"VW",4}, {"WGS",256}, {"WPT",1} } },
{ "GeForce GTX TITAN X", { {"VW",1}, {"WGS",64}, {"WPT",1} } },
@@ -103,7 +105,7 @@ const Database::DatabaseEntry Database::XaxpySingle = {
// =================================================================================================
-const Database::DatabaseEntry Database::XaxpyComplexSingle = {
+const Database::DatabaseEntry XaxpyComplexSingle = {
"Xaxpy", Precision::kComplexSingle, {
{ // AMD GPUs
kDeviceTypeGPU, "AMD", {
@@ -134,6 +136,7 @@ const Database::DatabaseEntry Database::XaxpyComplexSingle = {
{ "Intel(R) HD Graphics 530", { {"VW",4}, {"WGS",64}, {"WPT",2} } },
{ "Intel(R) HD Graphics 5500 BroadWell U-Processor GT2", { {"VW",1}, {"WGS",64}, {"WPT",1} } },
{ "Intel(R) HD Graphics Haswell Ultrabook GT2 Mobile", { {"VW",1}, {"WGS",64}, {"WPT",1} } },
+ { "Intel(R) HD Graphics IvyBridge M GT2", { {"VW",1}, {"WGS",64}, {"WPT",1} } },
{ "Intel(R) HD Graphics Skylake ULT GT2", { {"VW",2}, {"WGS",512}, {"WPT",1} } },
{ "Iris", { {"VW",2}, {"WGS",128}, {"WPT",1} } },
{ "Iris Pro", { {"VW",1}, {"WGS",256}, {"WPT",8} } },
@@ -173,7 +176,7 @@ const Database::DatabaseEntry Database::XaxpyComplexSingle = {
// =================================================================================================
-const Database::DatabaseEntry Database::XaxpyDouble = {
+const Database::DatabaseEntry XaxpyDouble = {
"Xaxpy", Precision::kDouble, {
{ // AMD GPUs
kDeviceTypeGPU, "AMD", {
@@ -213,7 +216,7 @@ const Database::DatabaseEntry Database::XaxpyDouble = {
{ "GeForce GTX 670", { {"VW",1}, {"WGS",64}, {"WPT",1} } },
{ "GeForce GTX 680", { {"VW",1}, {"WGS",64}, {"WPT",1} } },
{ "GeForce GTX 750", { {"VW",1}, {"WGS",128}, {"WPT",1} } },
- { "GeForce GTX 750 Ti", { {"VW",1}, {"WGS",64}, {"WPT",1} } },
+ { "GeForce GTX 750 Ti", { {"VW",1}, {"WGS",256}, {"WPT",2} } },
{ "GeForce GTX 980", { {"VW",1}, {"WGS",256}, {"WPT",1} } },
{ "GeForce GTX TITAN", { {"VW",2}, {"WGS",1024}, {"WPT",1} } },
{ "GeForce GTX TITAN X", { {"VW",1}, {"WGS",512}, {"WPT",1} } },
@@ -232,7 +235,7 @@ const Database::DatabaseEntry Database::XaxpyDouble = {
// =================================================================================================
-const Database::DatabaseEntry Database::XaxpyComplexDouble = {
+const Database::DatabaseEntry XaxpyComplexDouble = {
"Xaxpy", Precision::kComplexDouble, {
{ // AMD GPUs
kDeviceTypeGPU, "AMD", {
@@ -272,7 +275,7 @@ const Database::DatabaseEntry Database::XaxpyComplexDouble = {
{ "GeForce GTX 670", { {"VW",1}, {"WGS",256}, {"WPT",1} } },
{ "GeForce GTX 680", { {"VW",1}, {"WGS",64}, {"WPT",1} } },
{ "GeForce GTX 750", { {"VW",1}, {"WGS",1024}, {"WPT",1} } },
- { "GeForce GTX 750 Ti", { {"VW",1}, {"WGS",256}, {"WPT",2} } },
+ { "GeForce GTX 750 Ti", { {"VW",1}, {"WGS",64}, {"WPT",2} } },
{ "GeForce GTX 980", { {"VW",1}, {"WGS",1024}, {"WPT",1} } },
{ "GeForce GTX TITAN", { {"VW",1}, {"WGS",64}, {"WPT",4} } },
{ "GeForce GTX TITAN X", { {"VW",1}, {"WGS",1024}, {"WPT",1} } },
@@ -290,4 +293,5 @@ const Database::DatabaseEntry Database::XaxpyComplexDouble = {
};
// =================================================================================================
+} // namespace database
} // namespace clblast
diff --git a/src/database/kernels/xdot.hpp b/src/database/kernels/xdot.hpp
index 987a990d..686b2839 100644
--- a/src/database/kernels/xdot.hpp
+++ b/src/database/kernels/xdot.hpp
@@ -12,9 +12,10 @@
// =================================================================================================
namespace clblast {
+namespace database {
// =================================================================================================
-const Database::DatabaseEntry Database::XdotHalf = {
+const Database::DatabaseEntry XdotHalf = {
"Xdot", Precision::kHalf, {
{ // Intel GPUs
kDeviceTypeGPU, "Intel", {
@@ -33,7 +34,7 @@ const Database::DatabaseEntry Database::XdotHalf = {
// =================================================================================================
-const Database::DatabaseEntry Database::XdotSingle = {
+const Database::DatabaseEntry XdotSingle = {
"Xdot", Precision::kSingle, {
{ // AMD GPUs
kDeviceTypeGPU, "AMD", {
@@ -55,6 +56,7 @@ const Database::DatabaseEntry Database::XdotSingle = {
{ "Intel(R) HD Graphics 530", { {"WGS1",64}, {"WGS2",32} } },
{ "Intel(R) HD Graphics 5500 BroadWell U-Processor GT2", { {"WGS1",256}, {"WGS2",32} } },
{ "Intel(R) HD Graphics Haswell Ultrabook GT2 Mobile", { {"WGS1",64}, {"WGS2",32} } },
+ { "Intel(R) HD Graphics IvyBridge M GT2", { {"WGS1",512}, {"WGS2",128} } },
{ "Intel(R) HD Graphics Skylake ULT GT2", { {"WGS1",64}, {"WGS2",32} } },
{ "Iris Pro", { {"WGS1",512}, {"WGS2",64} } },
{ "default", { {"WGS1",64}, {"WGS2",32} } },
@@ -68,6 +70,7 @@ const Database::DatabaseEntry Database::XdotSingle = {
{ "GeForce GTX 670", { {"WGS1",512}, {"WGS2",1024} } },
{ "GeForce GTX 680", { {"WGS1",128}, {"WGS2",128} } },
{ "GeForce GTX 750", { {"WGS1",128}, {"WGS2",32} } },
+ { "GeForce GTX 750 Ti", { {"WGS1",64}, {"WGS2",32} } },
{ "GeForce GTX 980", { {"WGS1",256}, {"WGS2",32} } },
{ "GeForce GTX TITAN X", { {"WGS1",256}, {"WGS2",32} } },
{ "Tesla K20m", { {"WGS1",1024}, {"WGS2",32} } },
@@ -84,7 +87,7 @@ const Database::DatabaseEntry Database::XdotSingle = {
// =================================================================================================
-const Database::DatabaseEntry Database::XdotComplexSingle = {
+const Database::DatabaseEntry XdotComplexSingle = {
"Xdot", Precision::kComplexSingle, {
{ // AMD GPUs
kDeviceTypeGPU, "AMD", {
@@ -106,6 +109,7 @@ const Database::DatabaseEntry Database::XdotComplexSingle = {
{ "Intel(R) HD Graphics 530", { {"WGS1",256}, {"WGS2",32} } },
{ "Intel(R) HD Graphics 5500 BroadWell U-Processor GT2", { {"WGS1",256}, {"WGS2",32} } },
{ "Intel(R) HD Graphics Haswell Ultrabook GT2 Mobile", { {"WGS1",32}, {"WGS2",32} } },
+ { "Intel(R) HD Graphics IvyBridge M GT2", { {"WGS1",512}, {"WGS2",32} } },
{ "Intel(R) HD Graphics Skylake ULT GT2", { {"WGS1",32}, {"WGS2",32} } },
{ "Iris Pro", { {"WGS1",32}, {"WGS2",32} } },
{ "default", { {"WGS1",32}, {"WGS2",32} } },
@@ -119,6 +123,7 @@ const Database::DatabaseEntry Database::XdotComplexSingle = {
{ "GeForce GTX 670", { {"WGS1",256}, {"WGS2",32} } },
{ "GeForce GTX 680", { {"WGS1",128}, {"WGS2",64} } },
{ "GeForce GTX 750", { {"WGS1",64}, {"WGS2",32} } },
+ { "GeForce GTX 750 Ti", { {"WGS1",64}, {"WGS2",32} } },
{ "GeForce GTX 980", { {"WGS1",256}, {"WGS2",64} } },
{ "GeForce GTX TITAN X", { {"WGS1",256}, {"WGS2",32} } },
{ "Tesla K20m", { {"WGS1",512}, {"WGS2",32} } },
@@ -135,7 +140,7 @@ const Database::DatabaseEntry Database::XdotComplexSingle = {
// =================================================================================================
-const Database::DatabaseEntry Database::XdotDouble = {
+const Database::DatabaseEntry XdotDouble = {
"Xdot", Precision::kDouble, {
{ // AMD GPUs
kDeviceTypeGPU, "AMD", {
@@ -160,6 +165,7 @@ const Database::DatabaseEntry Database::XdotDouble = {
{ "GeForce GTX 670", { {"WGS1",256}, {"WGS2",32} } },
{ "GeForce GTX 680", { {"WGS1",128}, {"WGS2",64} } },
{ "GeForce GTX 750", { {"WGS1",64}, {"WGS2",256} } },
+ { "GeForce GTX 750 Ti", { {"WGS1",128}, {"WGS2",64} } },
{ "GeForce GTX 980", { {"WGS1",128}, {"WGS2",32} } },
{ "GeForce GTX TITAN X", { {"WGS1",256}, {"WGS2",32} } },
{ "Tesla K20m", { {"WGS1",512}, {"WGS2",32} } },
@@ -176,7 +182,7 @@ const Database::DatabaseEntry Database::XdotDouble = {
// =================================================================================================
-const Database::DatabaseEntry Database::XdotComplexDouble = {
+const Database::DatabaseEntry XdotComplexDouble = {
"Xdot", Precision::kComplexDouble, {
{ // AMD GPUs
kDeviceTypeGPU, "AMD", {
@@ -201,6 +207,7 @@ const Database::DatabaseEntry Database::XdotComplexDouble = {
{ "GeForce GTX 670", { {"WGS1",512}, {"WGS2",128} } },
{ "GeForce GTX 680", { {"WGS1",256}, {"WGS2",64} } },
{ "GeForce GTX 750", { {"WGS1",256}, {"WGS2",32} } },
+ { "GeForce GTX 750 Ti", { {"WGS1",64}, {"WGS2",32} } },
{ "GeForce GTX 980", { {"WGS1",64}, {"WGS2",32} } },
{ "GeForce GTX TITAN X", { {"WGS1",128}, {"WGS2",32} } },
{ "Tesla K20m", { {"WGS1",128}, {"WGS2",32} } },
@@ -216,4 +223,5 @@ const Database::DatabaseEntry Database::XdotComplexDouble = {
};
// =================================================================================================
+} // namespace database
} // namespace clblast
diff --git a/src/database/kernels/xgemm.hpp b/src/database/kernels/xgemm.hpp
index d19c55b5..8303fa83 100644
--- a/src/database/kernels/xgemm.hpp
+++ b/src/database/kernels/xgemm.hpp
@@ -12,9 +12,10 @@
// =================================================================================================
namespace clblast {
+namespace database {
// =================================================================================================
-const Database::DatabaseEntry Database::XgemmHalf = {
+const Database::DatabaseEntry XgemmHalf = {
"Xgemm", Precision::kHalf, {
{ // Default
kDeviceTypeAll, "default", {
@@ -26,7 +27,7 @@ const Database::DatabaseEntry Database::XgemmHalf = {
// =================================================================================================
-const Database::DatabaseEntry Database::XgemmSingle = {
+const Database::DatabaseEntry XgemmSingle = {
"Xgemm", Precision::kSingle, {
{ // AMD GPUs
kDeviceTypeGPU, "AMD", {
@@ -57,9 +58,10 @@ const Database::DatabaseEntry Database::XgemmSingle = {
{ "Intel(R) HD Graphics 530", { {"KWG",32}, {"KWI",2}, {"MDIMA",8}, {"MDIMC",8}, {"MWG",128}, {"NDIMB",32}, {"NDIMC",16}, {"NWG",64}, {"SA",0}, {"SB",0}, {"STRM",1}, {"STRN",0}, {"VWM",4}, {"VWN",2} } },
{ "Intel(R) HD Graphics 5500 BroadWell U-Processor GT2", { {"KWG",32}, {"KWI",8}, {"MDIMA",8}, {"MDIMC",8}, {"MWG",64}, {"NDIMB",32}, {"NDIMC",16}, {"NWG",64}, {"SA",1}, {"SB",1}, {"STRM",1}, {"STRN",1}, {"VWM",4}, {"VWN",2} } },
{ "Intel(R) HD Graphics Haswell Ultrabook GT2 Mobile", { {"KWG",16}, {"KWI",2}, {"MDIMA",16}, {"MDIMC",8}, {"MWG",32}, {"NDIMB",8}, {"NDIMC",16}, {"NWG",128}, {"SA",1}, {"SB",1}, {"STRM",1}, {"STRN",1}, {"VWM",2}, {"VWN",4} } },
+ { "Intel(R) HD Graphics IvyBridge M GT2", { {"KWG",32}, {"KWI",2}, {"MDIMA",16}, {"MDIMC",16}, {"MWG",64}, {"NDIMB",16}, {"NDIMC",8}, {"NWG",64}, {"SA",1}, {"SB",1}, {"STRM",1}, {"STRN",0}, {"VWM",2}, {"VWN",4} } },
{ "Intel(R) HD Graphics Skylake ULT GT2", { {"KWG",32}, {"KWI",8}, {"MDIMA",16}, {"MDIMC",16}, {"MWG",64}, {"NDIMB",16}, {"NDIMC",16}, {"NWG",128}, {"SA",0}, {"SB",0}, {"STRM",0}, {"STRN",1}, {"VWM",1}, {"VWN",8} } },
{ "Iris", { {"KWG",16}, {"KWI",8}, {"MDIMA",16}, {"MDIMC",8}, {"MWG",128}, {"NDIMB",32}, {"NDIMC",16}, {"NWG",64}, {"SA",1}, {"SB",1}, {"STRM",1}, {"STRN",1}, {"VWM",4}, {"VWN",1} } },
- { "Iris Pro", { {"KWG",32}, {"KWI",8}, {"MDIMA",16}, {"MDIMC",8}, {"MWG",64}, {"NDIMB",8}, {"NDIMC",8}, {"NWG",64}, {"SA",1}, {"SB",0}, {"STRM",1}, {"STRN",0}, {"VWM",4}, {"VWN",4} } },
+ { "Iris Pro", { {"KWG",16}, {"KWI",2}, {"MDIMA",16}, {"MDIMC",8}, {"MWG",64}, {"NDIMB",32}, {"NDIMC",32}, {"NWG",128}, {"SA",1}, {"SB",1}, {"STRM",1}, {"STRN",0}, {"VWM",4}, {"VWN",4} } },
{ "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} } },
}
},
@@ -77,7 +79,7 @@ const Database::DatabaseEntry Database::XgemmSingle = {
{ "GeForce GTX 670", { {"KWG",16}, {"KWI",2}, {"MDIMA",8}, {"MDIMC",8}, {"MWG",64}, {"NDIMB",16}, {"NDIMC",16}, {"NWG",64}, {"SA",1}, {"SB",1}, {"STRM",1}, {"STRN",0}, {"VWM",2}, {"VWN",4} } },
{ "GeForce GTX 680", { {"KWG",32}, {"KWI",8}, {"MDIMA",8}, {"MDIMC",16}, {"MWG",64}, {"NDIMB",32}, {"NDIMC",16}, {"NWG",128}, {"SA",1}, {"SB",1}, {"STRM",0}, {"STRN",0}, {"VWM",4}, {"VWN",2} } },
{ "GeForce GTX 750", { {"KWG",16}, {"KWI",2}, {"MDIMA",16}, {"MDIMC",16}, {"MWG",64}, {"NDIMB",32}, {"NDIMC",8}, {"NWG",128}, {"SA",1}, {"SB",1}, {"STRM",1}, {"STRN",1}, {"VWM",1}, {"VWN",2} } },
- { "GeForce GTX 750 Ti", { {"KWG",16}, {"KWI",2}, {"MDIMA",32}, {"MDIMC",16}, {"MWG",128}, {"NDIMB",16}, {"NDIMC",8}, {"NWG",128}, {"SA",1}, {"SB",1}, {"STRM",1}, {"STRN",1}, {"VWM",4}, {"VWN",4} } },
+ { "GeForce GTX 750 Ti", { {"KWG",16}, {"KWI",2}, {"MDIMA",16}, {"MDIMC",16}, {"MWG",128}, {"NDIMB",32}, {"NDIMC",8}, {"NWG",64}, {"SA",1}, {"SB",1}, {"STRM",0}, {"STRN",1}, {"VWM",8}, {"VWN",2} } },
{ "GeForce GTX 980", { {"KWG",16}, {"KWI",2}, {"MDIMA",16}, {"MDIMC",16}, {"MWG",64}, {"NDIMB",16}, {"NDIMC",8}, {"NWG",128}, {"SA",1}, {"SB",1}, {"STRM",1}, {"STRN",0}, {"VWM",4}, {"VWN",8} } },
{ "GeForce GTX TITAN", { {"KWG",16}, {"KWI",8}, {"MDIMA",32}, {"MDIMC",16}, {"MWG",64}, {"NDIMB",8}, {"NDIMC",8}, {"NWG",64}, {"SA",1}, {"SB",1}, {"STRM",1}, {"STRN",0}, {"VWM",2}, {"VWN",2} } },
{ "GeForce GTX TITAN X", { {"KWG",16}, {"KWI",2}, {"MDIMA",8}, {"MDIMC",16}, {"MWG",128}, {"NDIMB",8}, {"NDIMC",8}, {"NWG",128}, {"SA",1}, {"SB",1}, {"STRM",1}, {"STRN",1}, {"VWM",4}, {"VWN",8} } },
@@ -96,7 +98,7 @@ const Database::DatabaseEntry Database::XgemmSingle = {
// =================================================================================================
-const Database::DatabaseEntry Database::XgemmComplexSingle = {
+const Database::DatabaseEntry XgemmComplexSingle = {
"Xgemm", Precision::kComplexSingle, {
{ // AMD GPUs
kDeviceTypeGPU, "AMD", {
@@ -127,6 +129,7 @@ const Database::DatabaseEntry Database::XgemmComplexSingle = {
{ "Intel(R) HD Graphics 530", { {"KWG",16}, {"KWI",8}, {"MDIMA",8}, {"MDIMC",16}, {"MWG",64}, {"NDIMB",32}, {"NDIMC",8}, {"NWG",32}, {"SA",0}, {"SB",0}, {"STRM",0}, {"STRN",0}, {"VWM",2}, {"VWN",1} } },
{ "Intel(R) HD Graphics 5500 BroadWell U-Processor GT2", { {"KWG",16}, {"KWI",8}, {"MDIMA",8}, {"MDIMC",8}, {"MWG",32}, {"NDIMB",16}, {"NDIMC",16}, {"NWG",64}, {"SA",1}, {"SB",0}, {"STRM",0}, {"STRN",0}, {"VWM",4}, {"VWN",4} } },
{ "Intel(R) HD Graphics Haswell Ultrabook GT2 Mobile", { {"KWG",32}, {"KWI",8}, {"MDIMA",16}, {"MDIMC",16}, {"MWG",64}, {"NDIMB",16}, {"NDIMC",16}, {"NWG",64}, {"SA",1}, {"SB",1}, {"STRM",1}, {"STRN",1}, {"VWM",2}, {"VWN",1} } },
+ { "Intel(R) HD Graphics IvyBridge M GT2", { {"KWG",32}, {"KWI",2}, {"MDIMA",16}, {"MDIMC",16}, {"MWG",64}, {"NDIMB",16}, {"NDIMC",16}, {"NWG",64}, {"SA",1}, {"SB",1}, {"STRM",0}, {"STRN",0}, {"VWM",4}, {"VWN",4} } },
{ "Intel(R) HD Graphics Skylake ULT GT2", { {"KWG",32}, {"KWI",8}, {"MDIMA",8}, {"MDIMC",8}, {"MWG",32}, {"NDIMB",32}, {"NDIMC",16}, {"NWG",32}, {"SA",1}, {"SB",0}, {"STRM",0}, {"STRN",1}, {"VWM",4}, {"VWN",1} } },
{ "Iris", { {"KWG",32}, {"KWI",8}, {"MDIMA",32}, {"MDIMC",16}, {"MWG",64}, {"NDIMB",8}, {"NDIMC",16}, {"NWG",64}, {"SA",1}, {"SB",0}, {"STRM",1}, {"STRN",0}, {"VWM",1}, {"VWN",1} } },
{ "Iris Pro", { {"KWG",16}, {"KWI",2}, {"MDIMA",8}, {"MDIMC",8}, {"MWG",32}, {"NDIMB",32}, {"NDIMC",8}, {"NWG",32}, {"SA",1}, {"SB",1}, {"STRM",1}, {"STRN",1}, {"VWM",1}, {"VWN",1} } },
@@ -147,7 +150,7 @@ const Database::DatabaseEntry Database::XgemmComplexSingle = {
{ "GeForce GTX 670", { {"KWG",16}, {"KWI",2}, {"MDIMA",32}, {"MDIMC",32}, {"MWG",64}, {"NDIMB",32}, {"NDIMC",8}, {"NWG",32}, {"SA",1}, {"SB",1}, {"STRM",1}, {"STRN",1}, {"VWM",1}, {"VWN",1} } },
{ "GeForce GTX 680", { {"KWG",16}, {"KWI",2}, {"MDIMA",32}, {"MDIMC",16}, {"MWG",64}, {"NDIMB",32}, {"NDIMC",32}, {"NWG",128}, {"SA",1}, {"SB",0}, {"STRM",0}, {"STRN",0}, {"VWM",2}, {"VWN",2} } },
{ "GeForce GTX 750", { {"KWG",16}, {"KWI",8}, {"MDIMA",16}, {"MDIMC",16}, {"MWG",64}, {"NDIMB",16}, {"NDIMC",16}, {"NWG",64}, {"SA",1}, {"SB",1}, {"STRM",1}, {"STRN",0}, {"VWM",2}, {"VWN",2} } },
- { "GeForce GTX 750 Ti", { {"KWG",16}, {"KWI",2}, {"MDIMA",32}, {"MDIMC",8}, {"MWG",128}, {"NDIMB",16}, {"NDIMC",32}, {"NWG",128}, {"SA",1}, {"SB",1}, {"STRM",1}, {"STRN",0}, {"VWM",1}, {"VWN",4} } },
+ { "GeForce GTX 750 Ti", { {"KWG",16}, {"KWI",2}, {"MDIMA",16}, {"MDIMC",8}, {"MWG",32}, {"NDIMB",32}, {"NDIMC",16}, {"NWG",64}, {"SA",1}, {"SB",1}, {"STRM",1}, {"STRN",0}, {"VWM",1}, {"VWN",2} } },
{ "GeForce GTX 980", { {"KWG",32}, {"KWI",8}, {"MDIMA",32}, {"MDIMC",32}, {"MWG",64}, {"NDIMB",16}, {"NDIMC",16}, {"NWG",64}, {"SA",1}, {"SB",1}, {"STRM",1}, {"STRN",0}, {"VWM",2}, {"VWN",1} } },
{ "GeForce GTX TITAN", { {"KWG",16}, {"KWI",8}, {"MDIMA",16}, {"MDIMC",16}, {"MWG",64}, {"NDIMB",32}, {"NDIMC",16}, {"NWG",64}, {"SA",1}, {"SB",1}, {"STRM",1}, {"STRN",0}, {"VWM",1}, {"VWN",1} } },
{ "GeForce GTX TITAN X", { {"KWG",16}, {"KWI",2}, {"MDIMA",8}, {"MDIMC",8}, {"MWG",64}, {"NDIMB",8}, {"NDIMC",8}, {"NWG",32}, {"SA",1}, {"SB",0}, {"STRM",1}, {"STRN",1}, {"VWM",1}, {"VWN",4} } },
@@ -166,7 +169,7 @@ const Database::DatabaseEntry Database::XgemmComplexSingle = {
// =================================================================================================
-const Database::DatabaseEntry Database::XgemmDouble = {
+const Database::DatabaseEntry XgemmDouble = {
"Xgemm", Precision::kDouble, {
{ // AMD GPUs
kDeviceTypeGPU, "AMD", {
@@ -206,7 +209,7 @@ const Database::DatabaseEntry Database::XgemmDouble = {
{ "GeForce GTX 670", { {"KWG",32}, {"KWI",8}, {"MDIMA",16}, {"MDIMC",32}, {"MWG",128}, {"NDIMB",16}, {"NDIMC",8}, {"NWG",32}, {"SA",0}, {"SB",1}, {"STRM",1}, {"STRN",0}, {"VWM",1}, {"VWN",1} } },
{ "GeForce GTX 680", { {"KWG",32}, {"KWI",8}, {"MDIMA",8}, {"MDIMC",8}, {"MWG",32}, {"NDIMB",16}, {"NDIMC",32}, {"NWG",128}, {"SA",1}, {"SB",0}, {"STRM",0}, {"STRN",1}, {"VWM",2}, {"VWN",4} } },
{ "GeForce GTX 750", { {"KWG",32}, {"KWI",8}, {"MDIMA",16}, {"MDIMC",32}, {"MWG",64}, {"NDIMB",16}, {"NDIMC",8}, {"NWG",128}, {"SA",0}, {"SB",0}, {"STRM",0}, {"STRN",1}, {"VWM",2}, {"VWN",1} } },
- { "GeForce GTX 750 Ti", { {"KWG",16}, {"KWI",2}, {"MDIMA",16}, {"MDIMC",8}, {"MWG",32}, {"NDIMB",16}, {"NDIMC",32}, {"NWG",32}, {"SA",0}, {"SB",0}, {"STRM",1}, {"STRN",1}, {"VWM",2}, {"VWN",1} } },
+ { "GeForce GTX 750 Ti", { {"KWG",32}, {"KWI",2}, {"MDIMA",8}, {"MDIMC",8}, {"MWG",32}, {"NDIMB",16}, {"NDIMC",16}, {"NWG",32}, {"SA",0}, {"SB",0}, {"STRM",0}, {"STRN",0}, {"VWM",4}, {"VWN",2} } },
{ "GeForce GTX 980", { {"KWG",32}, {"KWI",8}, {"MDIMA",16}, {"MDIMC",8}, {"MWG",64}, {"NDIMB",32}, {"NDIMC",32}, {"NWG",128}, {"SA",0}, {"SB",0}, {"STRM",1}, {"STRN",0}, {"VWM",2}, {"VWN",4} } },
{ "GeForce GTX TITAN", { {"KWG",16}, {"KWI",8}, {"MDIMA",16}, {"MDIMC",8}, {"MWG",32}, {"NDIMB",16}, {"NDIMC",32}, {"NWG",128}, {"SA",1}, {"SB",1}, {"STRM",1}, {"STRN",1}, {"VWM",2}, {"VWN",2} } },
{ "GeForce GTX TITAN X", { {"KWG",16}, {"KWI",8}, {"MDIMA",16}, {"MDIMC",16}, {"MWG",16}, {"NDIMB",16}, {"NDIMC",16}, {"NWG",16}, {"SA",0}, {"SB",0}, {"STRM",0}, {"STRN",0}, {"VWM",1}, {"VWN",1} } },
@@ -225,7 +228,7 @@ const Database::DatabaseEntry Database::XgemmDouble = {
// =================================================================================================
-const Database::DatabaseEntry Database::XgemmComplexDouble = {
+const Database::DatabaseEntry XgemmComplexDouble = {
"Xgemm", Precision::kComplexDouble, {
{ // AMD GPUs
kDeviceTypeGPU, "AMD", {
@@ -265,7 +268,7 @@ const Database::DatabaseEntry Database::XgemmComplexDouble = {
{ "GeForce GTX 670", { {"KWG",32}, {"KWI",8}, {"MDIMA",16}, {"MDIMC",8}, {"MWG",16}, {"NDIMB",16}, {"NDIMC",32}, {"NWG",64}, {"SA",1}, {"SB",0}, {"STRM",0}, {"STRN",1}, {"VWM",1}, {"VWN",2} } },
{ "GeForce GTX 680", { {"KWG",16}, {"KWI",8}, {"MDIMA",16}, {"MDIMC",8}, {"MWG",64}, {"NDIMB",16}, {"NDIMC",32}, {"NWG",32}, {"SA",0}, {"SB",1}, {"STRM",1}, {"STRN",0}, {"VWM",1}, {"VWN",1} } },
{ "GeForce GTX 750", { {"KWG",32}, {"KWI",2}, {"MDIMA",8}, {"MDIMC",32}, {"MWG",32}, {"NDIMB",8}, {"NDIMC",8}, {"NWG",64}, {"SA",0}, {"SB",0}, {"STRM",1}, {"STRN",0}, {"VWM",1}, {"VWN",4} } },
- { "GeForce GTX 750 Ti", { {"KWG",16}, {"KWI",2}, {"MDIMA",16}, {"MDIMC",16}, {"MWG",16}, {"NDIMB",8}, {"NDIMC",8}, {"NWG",32}, {"SA",0}, {"SB",0}, {"STRM",0}, {"STRN",0}, {"VWM",1}, {"VWN",4} } },
+ { "GeForce GTX 750 Ti", { {"KWG",32}, {"KWI",2}, {"MDIMA",8}, {"MDIMC",8}, {"MWG",16}, {"NDIMB",8}, {"NDIMC",8}, {"NWG",32}, {"SA",0}, {"SB",0}, {"STRM",0}, {"STRN",0}, {"VWM",1}, {"VWN",1} } },
{ "GeForce GTX 980", { {"KWG",16}, {"KWI",2}, {"MDIMA",16}, {"MDIMC",8}, {"MWG",32}, {"NDIMB",8}, {"NDIMC",16}, {"NWG",128}, {"SA",0}, {"SB",0}, {"STRM",1}, {"STRN",1}, {"VWM",2}, {"VWN",2} } },
{ "GeForce GTX TITAN X", { {"KWG",32}, {"KWI",8}, {"MDIMA",16}, {"MDIMC",16}, {"MWG",128}, {"NDIMB",16}, {"NDIMC",16}, {"NWG",32}, {"SA",0}, {"SB",0}, {"STRM",1}, {"STRN",0}, {"VWM",1}, {"VWN",1} } },
{ "Tesla K20m", { {"KWG",32}, {"KWI",2}, {"MDIMA",32}, {"MDIMC",8}, {"MWG",32}, {"NDIMB",16}, {"NDIMC",16}, {"NWG",64}, {"SA",0}, {"SB",0}, {"STRM",1}, {"STRN",0}, {"VWM",1}, {"VWN",1} } },
@@ -282,4 +285,5 @@ const Database::DatabaseEntry Database::XgemmComplexDouble = {
};
// =================================================================================================
+} // namespace database
} // namespace clblast
diff --git a/src/database/kernels/xgemm_direct.hpp b/src/database/kernels/xgemm_direct.hpp
new file mode 100644
index 00000000..89499cc6
--- /dev/null
+++ b/src/database/kernels/xgemm_direct.hpp
@@ -0,0 +1,138 @@
+
+// =================================================================================================
+// 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 'Xgemm_Direct' kernels.
+//
+// =================================================================================================
+
+namespace clblast {
+namespace database {
+// =================================================================================================
+
+const Database::DatabaseEntry XgemmDirectHalf = {
+ "XgemmDirect", Precision::kHalf, {
+ { // Default
+ kDeviceTypeAll, "default", {
+ { "default", { {"KWID",2}, {"MDIMAD",8}, {"MDIMCD",8}, {"NDIMBD",8}, {"NDIMCD",8}, {"PADA",1}, {"PADB",1}, {"VWMD",4}, {"VWND",4}, {"WGD",32} } },
+ }
+ },
+ }
+};
+
+// =================================================================================================
+
+const Database::DatabaseEntry XgemmDirectSingle = {
+ "XgemmDirect", Precision::kSingle, {
+ { // AMD GPUs
+ kDeviceTypeGPU, "AMD", {
+ { "AMD Radeon R9 M370X Compute Engine", { {"KWID",2}, {"MDIMAD",8}, {"MDIMCD",8}, {"NDIMBD",8}, {"NDIMCD",8}, {"PADA",1}, {"PADB",1}, {"VWMD",2}, {"VWND",2}, {"WGD",32} } },
+ { "default", { {"KWID",2}, {"MDIMAD",8}, {"MDIMCD",8}, {"NDIMBD",8}, {"NDIMCD",8}, {"PADA",1}, {"PADB",1}, {"VWMD",2}, {"VWND",2}, {"WGD",32} } },
+ }
+ },
+ { // Intel GPUs
+ kDeviceTypeGPU, "Intel", {
+ { "Iris Pro", { {"KWID",2}, {"MDIMAD",16}, {"MDIMCD",16}, {"NDIMBD",8}, {"NDIMCD",8}, {"PADA",1}, {"PADB",1}, {"VWMD",2}, {"VWND",4}, {"WGD",32} } },
+ { "default", { {"KWID",2}, {"MDIMAD",16}, {"MDIMCD",16}, {"NDIMBD",8}, {"NDIMCD",8}, {"PADA",1}, {"PADB",1}, {"VWMD",2}, {"VWND",4}, {"WGD",32} } },
+ }
+ },
+ { // NVIDIA GPUs
+ kDeviceTypeGPU, "NVIDIA", {
+ { "GeForce GTX 750 Ti", { {"KWID",2}, {"MDIMAD",8}, {"MDIMCD",8}, {"NDIMBD",8}, {"NDIMCD",8}, {"PADA",1}, {"PADB",1}, {"VWMD",4}, {"VWND",2}, {"WGD",32} } },
+ { "default", { {"KWID",2}, {"MDIMAD",8}, {"MDIMCD",8}, {"NDIMBD",8}, {"NDIMCD",8}, {"PADA",1}, {"PADB",1}, {"VWMD",4}, {"VWND",2}, {"WGD",32} } },
+ }
+ },
+ { // Default
+ kDeviceTypeAll, "default", {
+ { "default", { {"KWID",2}, {"MDIMAD",8}, {"MDIMCD",8}, {"NDIMBD",8}, {"NDIMCD",8}, {"PADA",1}, {"PADB",1}, {"VWMD",4}, {"VWND",4}, {"WGD",32} } },
+ }
+ },
+ }
+};
+
+// =================================================================================================
+
+const Database::DatabaseEntry XgemmDirectComplexSingle = {
+ "XgemmDirect", Precision::kComplexSingle, {
+ { // AMD GPUs
+ kDeviceTypeGPU, "AMD", {
+ { "AMD Radeon R9 M370X Compute Engine", { {"KWID",2}, {"MDIMAD",16}, {"MDIMCD",16}, {"NDIMBD",16}, {"NDIMCD",16}, {"PADA",1}, {"PADB",1}, {"VWMD",1}, {"VWND",1}, {"WGD",16} } },
+ { "default", { {"KWID",2}, {"MDIMAD",16}, {"MDIMCD",16}, {"NDIMBD",16}, {"NDIMCD",16}, {"PADA",1}, {"PADB",1}, {"VWMD",1}, {"VWND",1}, {"WGD",16} } },
+ }
+ },
+ { // Intel GPUs
+ kDeviceTypeGPU, "Intel", {
+ { "Iris Pro", { {"KWID",2}, {"MDIMAD",16}, {"MDIMCD",16}, {"NDIMBD",8}, {"NDIMCD",8}, {"PADA",1}, {"PADB",1}, {"VWMD",2}, {"VWND",2}, {"WGD",32} } },
+ { "default", { {"KWID",2}, {"MDIMAD",16}, {"MDIMCD",16}, {"NDIMBD",8}, {"NDIMCD",8}, {"PADA",1}, {"PADB",1}, {"VWMD",2}, {"VWND",2}, {"WGD",32} } },
+ }
+ },
+ { // NVIDIA GPUs
+ kDeviceTypeGPU, "NVIDIA", {
+ { "GeForce GTX 750 Ti", { {"KWID",16}, {"MDIMAD",8}, {"MDIMCD",8}, {"NDIMBD",16}, {"NDIMCD",8}, {"PADA",1}, {"PADB",1}, {"VWMD",2}, {"VWND",1}, {"WGD",16} } },
+ { "default", { {"KWID",16}, {"MDIMAD",8}, {"MDIMCD",8}, {"NDIMBD",16}, {"NDIMCD",8}, {"PADA",1}, {"PADB",1}, {"VWMD",2}, {"VWND",1}, {"WGD",16} } },
+ }
+ },
+ { // Default
+ kDeviceTypeAll, "default", {
+ { "default", { {"KWID",2}, {"MDIMAD",16}, {"MDIMCD",16}, {"NDIMBD",8}, {"NDIMCD",8}, {"PADA",1}, {"PADB",1}, {"VWMD",2}, {"VWND",2}, {"WGD",32} } },
+ }
+ },
+ }
+};
+
+// =================================================================================================
+
+const Database::DatabaseEntry XgemmDirectDouble = {
+ "XgemmDirect", Precision::kDouble, {
+ { // AMD GPUs
+ kDeviceTypeGPU, "AMD", {
+ { "AMD Radeon R9 M370X Compute Engine", { {"KWID",2}, {"MDIMAD",16}, {"MDIMCD",16}, {"NDIMBD",8}, {"NDIMCD",8}, {"PADA",1}, {"PADB",1}, {"VWMD",1}, {"VWND",1}, {"WGD",16} } },
+ { "default", { {"KWID",2}, {"MDIMAD",16}, {"MDIMCD",16}, {"NDIMBD",8}, {"NDIMCD",8}, {"PADA",1}, {"PADB",1}, {"VWMD",1}, {"VWND",1}, {"WGD",16} } },
+ }
+ },
+ { // NVIDIA GPUs
+ kDeviceTypeGPU, "NVIDIA", {
+ { "GeForce GTX 750 Ti", { {"KWID",2}, {"MDIMAD",8}, {"MDIMCD",8}, {"NDIMBD",8}, {"NDIMCD",8}, {"PADA",1}, {"PADB",1}, {"VWMD",2}, {"VWND",4}, {"WGD",32} } },
+ { "default", { {"KWID",2}, {"MDIMAD",8}, {"MDIMCD",8}, {"NDIMBD",8}, {"NDIMCD",8}, {"PADA",1}, {"PADB",1}, {"VWMD",2}, {"VWND",4}, {"WGD",32} } },
+ }
+ },
+ { // Default
+ kDeviceTypeAll, "default", {
+ { "default", { {"KWID",2}, {"MDIMAD",16}, {"MDIMCD",16}, {"NDIMBD",8}, {"NDIMCD",8}, {"PADA",1}, {"PADB",1}, {"VWMD",1}, {"VWND",1}, {"WGD",16} } },
+ }
+ },
+ }
+};
+
+// =================================================================================================
+
+const Database::DatabaseEntry XgemmDirectComplexDouble = {
+ "XgemmDirect", Precision::kComplexDouble, {
+ { // AMD GPUs
+ kDeviceTypeGPU, "AMD", {
+ { "AMD Radeon R9 M370X Compute Engine", { {"KWID",2}, {"MDIMAD",16}, {"MDIMCD",16}, {"NDIMBD",16}, {"NDIMCD",16}, {"PADA",1}, {"PADB",1}, {"VWMD",1}, {"VWND",1}, {"WGD",16} } },
+ { "default", { {"KWID",2}, {"MDIMAD",16}, {"MDIMCD",16}, {"NDIMBD",16}, {"NDIMCD",16}, {"PADA",1}, {"PADB",1}, {"VWMD",1}, {"VWND",1}, {"WGD",16} } },
+ }
+ },
+ { // NVIDIA GPUs
+ kDeviceTypeGPU, "NVIDIA", {
+ { "GeForce GTX 750 Ti", { {"KWID",2}, {"MDIMAD",32}, {"MDIMCD",32}, {"NDIMBD",8}, {"NDIMCD",8}, {"PADA",1}, {"PADB",1}, {"VWMD",1}, {"VWND",1}, {"WGD",32} } },
+ { "default", { {"KWID",2}, {"MDIMAD",32}, {"MDIMCD",32}, {"NDIMBD",8}, {"NDIMCD",8}, {"PADA",1}, {"PADB",1}, {"VWMD",1}, {"VWND",1}, {"WGD",32} } },
+ }
+ },
+ { // Default
+ kDeviceTypeAll, "default", {
+ { "default", { {"KWID",2}, {"MDIMAD",8}, {"MDIMCD",8}, {"NDIMBD",8}, {"NDIMCD",8}, {"PADA",1}, {"PADB",1}, {"VWMD",1}, {"VWND",2}, {"WGD",16} } },
+ }
+ },
+ }
+};
+
+// =================================================================================================
+} // namespace database
+} // namespace clblast
diff --git a/src/database/kernels/xgemv.hpp b/src/database/kernels/xgemv.hpp
index e5e8845e..90355b96 100644
--- a/src/database/kernels/xgemv.hpp
+++ b/src/database/kernels/xgemv.hpp
@@ -12,9 +12,10 @@
// =================================================================================================
namespace clblast {
+namespace database {
// =================================================================================================
-const Database::DatabaseEntry Database::XgemvHalf = {
+const Database::DatabaseEntry XgemvHalf = {
"Xgemv", Precision::kHalf, {
{ // Intel GPUs
kDeviceTypeGPU, "Intel", {
@@ -33,7 +34,7 @@ const Database::DatabaseEntry Database::XgemvHalf = {
// =================================================================================================
-const Database::DatabaseEntry Database::XgemvSingle = {
+const Database::DatabaseEntry XgemvSingle = {
"Xgemv", Precision::kSingle, {
{ // AMD GPUs
kDeviceTypeGPU, "AMD", {
@@ -57,6 +58,7 @@ const Database::DatabaseEntry Database::XgemvSingle = {
{ "Intel(R) HD Graphics 530", { {"WGS1",256}, {"WPT1",1} } },
{ "Intel(R) HD Graphics 5500 BroadWell U-Processor GT2", { {"WGS1",64}, {"WPT1",1} } },
{ "Intel(R) HD Graphics Haswell Ultrabook GT2 Mobile", { {"WGS1",64}, {"WPT1",1} } },
+ { "Intel(R) HD Graphics IvyBridge M GT2", { {"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} } },
@@ -77,7 +79,7 @@ const Database::DatabaseEntry Database::XgemvSingle = {
{ "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 750 Ti", { {"WGS1",32}, {"WPT1",1} } },
{ "GeForce GTX 980", { {"WGS1",128}, {"WPT1",1} } },
{ "GeForce GTX TITAN", { {"WGS1",256}, {"WPT1",1} } },
{ "GeForce GTX TITAN X", { {"WGS1",256}, {"WPT1",1} } },
@@ -96,7 +98,7 @@ const Database::DatabaseEntry Database::XgemvSingle = {
// =================================================================================================
-const Database::DatabaseEntry Database::XgemvComplexSingle = {
+const Database::DatabaseEntry XgemvComplexSingle = {
"Xgemv", Precision::kComplexSingle, {
{ // AMD GPUs
kDeviceTypeGPU, "AMD", {
@@ -120,6 +122,7 @@ const Database::DatabaseEntry Database::XgemvComplexSingle = {
{ "Intel(R) HD Graphics 530", { {"WGS1",64}, {"WPT1",1} } },
{ "Intel(R) HD Graphics 5500 BroadWell U-Processor GT2", { {"WGS1",64}, {"WPT1",1} } },
{ "Intel(R) HD Graphics Haswell Ultrabook GT2 Mobile", { {"WGS1",128}, {"WPT1",1} } },
+ { "Intel(R) HD Graphics IvyBridge M GT2", { {"WGS1",256}, {"WPT1",1} } },
{ "Intel(R) HD Graphics Skylake ULT GT2", { {"WGS1",64}, {"WPT1",1} } },
{ "Iris", { {"WGS1",256}, {"WPT1",1} } },
{ "Iris Pro", { {"WGS1",64}, {"WPT1",1} } },
@@ -140,7 +143,7 @@ const Database::DatabaseEntry Database::XgemvComplexSingle = {
{ "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} } },
+ { "GeForce GTX 750 Ti", { {"WGS1",32}, {"WPT1",1} } },
{ "GeForce GTX TITAN", { {"WGS1",256}, {"WPT1",1} } },
{ "default", { {"WGS1",64}, {"WPT1",1} } },
}
@@ -155,7 +158,7 @@ const Database::DatabaseEntry Database::XgemvComplexSingle = {
// =================================================================================================
-const Database::DatabaseEntry Database::XgemvDouble = {
+const Database::DatabaseEntry XgemvDouble = {
"Xgemv", Precision::kDouble, {
{ // AMD GPUs
kDeviceTypeGPU, "AMD", {
@@ -188,7 +191,7 @@ const Database::DatabaseEntry Database::XgemvDouble = {
{ "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 750 Ti", { {"WGS1",32}, {"WPT1",1} } },
{ "GeForce GTX 980", { {"WGS1",64}, {"WPT1",1} } },
{ "GeForce GTX TITAN", { {"WGS1",256}, {"WPT1",1} } },
{ "GeForce GTX TITAN X", { {"WGS1",64}, {"WPT1",1} } },
@@ -207,7 +210,7 @@ const Database::DatabaseEntry Database::XgemvDouble = {
// =================================================================================================
-const Database::DatabaseEntry Database::XgemvComplexDouble = {
+const Database::DatabaseEntry XgemvComplexDouble = {
"Xgemv", Precision::kComplexDouble, {
{ // AMD GPUs
kDeviceTypeGPU, "AMD", {
@@ -249,4 +252,5 @@ const Database::DatabaseEntry Database::XgemvComplexDouble = {
};
// =================================================================================================
+} // namespace database
} // namespace clblast
diff --git a/src/database/kernels/xgemv_fast.hpp b/src/database/kernels/xgemv_fast.hpp
index 52af628c..8e6254ac 100644
--- a/src/database/kernels/xgemv_fast.hpp
+++ b/src/database/kernels/xgemv_fast.hpp
@@ -12,9 +12,10 @@
// =================================================================================================
namespace clblast {
+namespace database {
// =================================================================================================
-const Database::DatabaseEntry Database::XgemvFastHalf = {
+const Database::DatabaseEntry XgemvFastHalf = {
"XgemvFast", Precision::kHalf, {
{ // Intel GPUs
kDeviceTypeGPU, "Intel", {
@@ -33,7 +34,7 @@ const Database::DatabaseEntry Database::XgemvFastHalf = {
// =================================================================================================
-const Database::DatabaseEntry Database::XgemvFastSingle = {
+const Database::DatabaseEntry XgemvFastSingle = {
"XgemvFast", Precision::kSingle, {
{ // AMD GPUs
kDeviceTypeGPU, "AMD", {
@@ -57,10 +58,11 @@ const Database::DatabaseEntry Database::XgemvFastSingle = {
{ "Intel(R) HD Graphics 530", { {"VW2",1}, {"WGS2",256}, {"WPT2",1} } },
{ "Intel(R) HD Graphics 5500 BroadWell U-Processor GT2", { {"VW2",2}, {"WGS2",32}, {"WPT2",2} } },
{ "Intel(R) HD Graphics Haswell Ultrabook GT2 Mobile", { {"VW2",4}, {"WGS2",128}, {"WPT2",4} } },
+ { "Intel(R) HD Graphics IvyBridge M GT2", { {"VW2",1}, {"WGS2",64}, {"WPT2",2} } },
{ "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",2}, {"WGS2",64}, {"WPT2",2} } },
+ { "default", { {"VW2",1}, {"WGS2",64}, {"WPT2",2} } },
}
},
{ // Intel accelerators
@@ -77,7 +79,7 @@ const Database::DatabaseEntry Database::XgemvFastSingle = {
{ "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 750 Ti", { {"VW2",2}, {"WGS2",32}, {"WPT2",2} } },
{ "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} } },
@@ -96,7 +98,7 @@ const Database::DatabaseEntry Database::XgemvFastSingle = {
// =================================================================================================
-const Database::DatabaseEntry Database::XgemvFastComplexSingle = {
+const Database::DatabaseEntry XgemvFastComplexSingle = {
"XgemvFast", Precision::kComplexSingle, {
{ // AMD GPUs
kDeviceTypeGPU, "AMD", {
@@ -120,6 +122,7 @@ const Database::DatabaseEntry Database::XgemvFastComplexSingle = {
{ "Intel(R) HD Graphics 530", { {"VW2",2}, {"WGS2",128}, {"WPT2",2} } },
{ "Intel(R) HD Graphics 5500 BroadWell U-Processor GT2", { {"VW2",1}, {"WGS2",32}, {"WPT2",2} } },
{ "Intel(R) HD Graphics Haswell Ultrabook GT2 Mobile", { {"VW2",2}, {"WGS2",128}, {"WPT2",2} } },
+ { "Intel(R) HD Graphics IvyBridge M GT2", { {"VW2",1}, {"WGS2",32}, {"WPT2",4} } },
{ "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} } },
@@ -139,7 +142,6 @@ const Database::DatabaseEntry Database::XgemvFastComplexSingle = {
{ "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} } },
}
},
@@ -153,7 +155,7 @@ const Database::DatabaseEntry Database::XgemvFastComplexSingle = {
// =================================================================================================
-const Database::DatabaseEntry Database::XgemvFastDouble = {
+const Database::DatabaseEntry XgemvFastDouble = {
"XgemvFast", Precision::kDouble, {
{ // AMD GPUs
kDeviceTypeGPU, "AMD", {
@@ -186,7 +188,7 @@ const Database::DatabaseEntry Database::XgemvFastDouble = {
{ "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 750 Ti", { {"VW2",1}, {"WGS2",32}, {"WPT2",2} } },
{ "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} } },
@@ -205,7 +207,7 @@ const Database::DatabaseEntry Database::XgemvFastDouble = {
// =================================================================================================
-const Database::DatabaseEntry Database::XgemvFastComplexDouble = {
+const Database::DatabaseEntry XgemvFastComplexDouble = {
"XgemvFast", Precision::kComplexDouble, {
{ // AMD GPUs
kDeviceTypeGPU, "AMD", {
@@ -247,4 +249,5 @@ const Database::DatabaseEntry Database::XgemvFastComplexDouble = {
};
// =================================================================================================
+} // namespace database
} // namespace clblast
diff --git a/src/database/kernels/xgemv_fast_rot.hpp b/src/database/kernels/xgemv_fast_rot.hpp
index 328094e1..8fe45e01 100644
--- a/src/database/kernels/xgemv_fast_rot.hpp
+++ b/src/database/kernels/xgemv_fast_rot.hpp
@@ -12,9 +12,10 @@
// =================================================================================================
namespace clblast {
+namespace database {
// =================================================================================================
-const Database::DatabaseEntry Database::XgemvFastRotHalf = {
+const Database::DatabaseEntry XgemvFastRotHalf = {
"XgemvFastRot", Precision::kHalf, {
{ // Default
kDeviceTypeAll, "default", {
@@ -26,7 +27,7 @@ const Database::DatabaseEntry Database::XgemvFastRotHalf = {
// =================================================================================================
-const Database::DatabaseEntry Database::XgemvFastRotSingle = {
+const Database::DatabaseEntry XgemvFastRotSingle = {
"XgemvFastRot", Precision::kSingle, {
{ // AMD GPUs
kDeviceTypeGPU, "AMD", {
@@ -44,6 +45,7 @@ const Database::DatabaseEntry Database::XgemvFastRotSingle = {
kDeviceTypeGPU, "Intel", {
{ "Intel(R) HD Graphics 5500 BroadWell U-Processor GT2", { {"VW3",8}, {"WGS3",64}, {"WPT3",32} } },
{ "Intel(R) HD Graphics Haswell Ultrabook GT2 Mobile", { {"VW3",4}, {"WGS3",64}, {"WPT3",16} } },
+ { "Intel(R) HD Graphics IvyBridge M GT2", { {"VW3",2}, {"WGS3",32}, {"WPT3",16} } },
{ "Intel(R) HD Graphics Skylake ULT GT2", { {"VW3",4}, {"WGS3",128}, {"WPT3",16} } },
{ "Iris Pro", { {"VW3",4}, {"WGS3",32}, {"WPT3",16} } },
{ "default", { {"VW3",8}, {"WGS3",32}, {"WPT3",32} } },
@@ -51,8 +53,9 @@ const Database::DatabaseEntry Database::XgemvFastRotSingle = {
},
{ // NVIDIA GPUs
kDeviceTypeGPU, "NVIDIA", {
+ { "GeForce GTX 750 Ti", { {"VW3",8}, {"WGS3",32}, {"WPT3",32} } },
{ "GeForce GTX TITAN", { {"VW3",1}, {"WGS3",16}, {"WPT3",16} } },
- { "default", { {"VW3",1}, {"WGS3",16}, {"WPT3",16} } },
+ { "default", { {"VW3",8}, {"WGS3",32}, {"WPT3",32} } },
}
},
{ // Default
@@ -65,7 +68,7 @@ const Database::DatabaseEntry Database::XgemvFastRotSingle = {
// =================================================================================================
-const Database::DatabaseEntry Database::XgemvFastRotComplexSingle = {
+const Database::DatabaseEntry XgemvFastRotComplexSingle = {
"XgemvFastRot", Precision::kComplexSingle, {
{ // AMD GPUs
kDeviceTypeGPU, "AMD", {
@@ -83,6 +86,7 @@ const Database::DatabaseEntry Database::XgemvFastRotComplexSingle = {
kDeviceTypeGPU, "Intel", {
{ "Intel(R) HD Graphics 5500 BroadWell U-Processor GT2", { {"VW3",2}, {"WGS3",16}, {"WPT3",16} } },
{ "Intel(R) HD Graphics Haswell Ultrabook GT2 Mobile", { {"VW3",4}, {"WGS3",128}, {"WPT3",8} } },
+ { "Intel(R) HD Graphics IvyBridge M GT2", { {"VW3",4}, {"WGS3",32}, {"WPT3",8} } },
{ "Intel(R) HD Graphics Skylake ULT GT2", { {"VW3",2}, {"WGS3",32}, {"WPT3",16} } },
{ "Iris Pro", { {"VW3",4}, {"WGS3",16}, {"WPT3",16} } },
{ "default", { {"VW3",2}, {"WGS3",32}, {"WPT3",8} } },
@@ -98,7 +102,7 @@ const Database::DatabaseEntry Database::XgemvFastRotComplexSingle = {
// =================================================================================================
-const Database::DatabaseEntry Database::XgemvFastRotDouble = {
+const Database::DatabaseEntry XgemvFastRotDouble = {
"XgemvFastRot", Precision::kDouble, {
{ // AMD GPUs
kDeviceTypeGPU, "AMD", {
@@ -114,8 +118,9 @@ const Database::DatabaseEntry Database::XgemvFastRotDouble = {
},
{ // NVIDIA GPUs
kDeviceTypeGPU, "NVIDIA", {
+ { "GeForce GTX 750 Ti", { {"VW3",4}, {"WGS3",32}, {"WPT3",16} } },
{ "GeForce GTX TITAN", { {"VW3",1}, {"WGS3",16}, {"WPT3",16} } },
- { "default", { {"VW3",1}, {"WGS3",16}, {"WPT3",16} } },
+ { "default", { {"VW3",4}, {"WGS3",32}, {"WPT3",16} } },
}
},
{ // Default
@@ -128,7 +133,7 @@ const Database::DatabaseEntry Database::XgemvFastRotDouble = {
// =================================================================================================
-const Database::DatabaseEntry Database::XgemvFastRotComplexDouble = {
+const Database::DatabaseEntry XgemvFastRotComplexDouble = {
"XgemvFastRot", Precision::kComplexDouble, {
{ // AMD GPUs
kDeviceTypeGPU, "AMD", {
@@ -151,4 +156,5 @@ const Database::DatabaseEntry Database::XgemvFastRotComplexDouble = {
};
// =================================================================================================
+} // namespace database
} // namespace clblast
diff --git a/src/database/kernels/xger.hpp b/src/database/kernels/xger.hpp
index 3e9c25c1..f2fc2a9a 100644
--- a/src/database/kernels/xger.hpp
+++ b/src/database/kernels/xger.hpp
@@ -12,9 +12,10 @@
// =================================================================================================
namespace clblast {
+namespace database {
// =================================================================================================
-const Database::DatabaseEntry Database::XgerHalf = {
+const Database::DatabaseEntry XgerHalf = {
"Xger", Precision::kHalf, {
{ // Intel GPUs
kDeviceTypeGPU, "Intel", {
@@ -33,7 +34,7 @@ const Database::DatabaseEntry Database::XgerHalf = {
// =================================================================================================
-const Database::DatabaseEntry Database::XgerSingle = {
+const Database::DatabaseEntry XgerSingle = {
"Xger", Precision::kSingle, {
{ // AMD GPUs
kDeviceTypeGPU, "AMD", {
@@ -63,6 +64,7 @@ const Database::DatabaseEntry Database::XgerSingle = {
{ "Intel(R) HD Graphics 530", { {"WGS1",32}, {"WGS2",1}, {"WPT",2} } },
{ "Intel(R) HD Graphics 5500 BroadWell U-Processor GT2", { {"WGS1",256}, {"WGS2",2}, {"WPT",2} } },
{ "Intel(R) HD Graphics Haswell Ultrabook GT2 Mobile", { {"WGS1",128}, {"WGS2",1}, {"WPT",2} } },
+ { "Intel(R) HD Graphics IvyBridge M GT2", { {"WGS1",64}, {"WGS2",1}, {"WPT",4} } },
{ "Intel(R) HD Graphics Skylake ULT GT2", { {"WGS1",8}, {"WGS2",8}, {"WPT",4} } },
{ "Iris Pro", { {"WGS1",64}, {"WGS2",1}, {"WPT",4} } },
{ "default", { {"WGS1",32}, {"WGS2",4}, {"WPT",2} } },
@@ -76,8 +78,9 @@ const Database::DatabaseEntry Database::XgerSingle = {
{ "GeForce GTX 670", { {"WGS1",32}, {"WGS2",8}, {"WPT",2} } },
{ "GeForce GTX 680", { {"WGS1",128}, {"WGS2",1}, {"WPT",4} } },
{ "GeForce GTX 750", { {"WGS1",64}, {"WGS2",16}, {"WPT",4} } },
+ { "GeForce GTX 750 Ti", { {"WGS1",64}, {"WGS2",1}, {"WPT",2} } },
{ "GeForce GTX TITAN", { {"WGS1",32}, {"WGS2",4}, {"WPT",2} } },
- { "default", { {"WGS1",256}, {"WGS2",1}, {"WPT",4} } },
+ { "default", { {"WGS1",128}, {"WGS2",1}, {"WPT",2} } },
}
},
{ // Default
@@ -90,7 +93,7 @@ const Database::DatabaseEntry Database::XgerSingle = {
// =================================================================================================
-const Database::DatabaseEntry Database::XgerComplexSingle = {
+const Database::DatabaseEntry XgerComplexSingle = {
"Xger", Precision::kComplexSingle, {
{ // AMD GPUs
kDeviceTypeGPU, "AMD", {
@@ -120,6 +123,7 @@ const Database::DatabaseEntry Database::XgerComplexSingle = {
{ "Intel(R) HD Graphics 530", { {"WGS1",32}, {"WGS2",1}, {"WPT",2} } },
{ "Intel(R) HD Graphics 5500 BroadWell U-Processor GT2", { {"WGS1",128}, {"WGS2",2}, {"WPT",1} } },
{ "Intel(R) HD Graphics Haswell Ultrabook GT2 Mobile", { {"WGS1",512}, {"WGS2",1}, {"WPT",1} } },
+ { "Intel(R) HD Graphics IvyBridge M GT2", { {"WGS1",256}, {"WGS2",1}, {"WPT",2} } },
{ "Intel(R) HD Graphics Skylake ULT GT2", { {"WGS1",128}, {"WGS2",4}, {"WPT",2} } },
{ "Iris Pro", { {"WGS1",16}, {"WGS2",2}, {"WPT",4} } },
{ "default", { {"WGS1",64}, {"WGS2",1}, {"WPT",2} } },
@@ -133,6 +137,7 @@ const Database::DatabaseEntry Database::XgerComplexSingle = {
{ "GeForce GTX 670", { {"WGS1",16}, {"WGS2",32}, {"WPT",2} } },
{ "GeForce GTX 680", { {"WGS1",32}, {"WGS2",4}, {"WPT",2} } },
{ "GeForce GTX 750", { {"WGS1",32}, {"WGS2",16}, {"WPT",4} } },
+ { "GeForce GTX 750 Ti", { {"WGS1",32}, {"WGS2",8}, {"WPT",2} } },
{ "GeForce GTX TITAN", { {"WGS1",16}, {"WGS2",16}, {"WPT",2} } },
{ "default", { {"WGS1",64}, {"WGS2",2}, {"WPT",2} } },
}
@@ -147,7 +152,7 @@ const Database::DatabaseEntry Database::XgerComplexSingle = {
// =================================================================================================
-const Database::DatabaseEntry Database::XgerDouble = {
+const Database::DatabaseEntry XgerDouble = {
"Xger", Precision::kDouble, {
{ // AMD GPUs
kDeviceTypeGPU, "AMD", {
@@ -180,8 +185,9 @@ const Database::DatabaseEntry Database::XgerDouble = {
{ "GeForce GTX 670", { {"WGS1",32}, {"WGS2",32}, {"WPT",2} } },
{ "GeForce GTX 680", { {"WGS1",128}, {"WGS2",4}, {"WPT",2} } },
{ "GeForce GTX 750", { {"WGS1",256}, {"WGS2",2}, {"WPT",2} } },
+ { "GeForce GTX 750 Ti", { {"WGS1",32}, {"WGS2",16}, {"WPT",1} } },
{ "GeForce GTX TITAN", { {"WGS1",16}, {"WGS2",8}, {"WPT",2} } },
- { "default", { {"WGS1",256}, {"WGS2",2}, {"WPT",2} } },
+ { "default", { {"WGS1",64}, {"WGS2",2}, {"WPT",2} } },
}
},
{ // Default
@@ -194,7 +200,7 @@ const Database::DatabaseEntry Database::XgerDouble = {
// =================================================================================================
-const Database::DatabaseEntry Database::XgerComplexDouble = {
+const Database::DatabaseEntry XgerComplexDouble = {
"Xger", Precision::kComplexDouble, {
{ // AMD GPUs
kDeviceTypeGPU, "AMD", {
@@ -227,6 +233,7 @@ const Database::DatabaseEntry Database::XgerComplexDouble = {
{ "GeForce GTX 670", { {"WGS1",8}, {"WGS2",16}, {"WPT",2} } },
{ "GeForce GTX 680", { {"WGS1",8}, {"WGS2",16}, {"WPT",1} } },
{ "GeForce GTX 750", { {"WGS1",8}, {"WGS2",32}, {"WPT",4} } },
+ { "GeForce GTX 750 Ti", { {"WGS1",32}, {"WGS2",8}, {"WPT",2} } },
{ "GeForce GTX TITAN", { {"WGS1",32}, {"WGS2",4}, {"WPT",2} } },
{ "default", { {"WGS1",16}, {"WGS2",8}, {"WPT",2} } },
}
@@ -240,4 +247,5 @@ const Database::DatabaseEntry Database::XgerComplexDouble = {
};
// =================================================================================================
+} // namespace database
} // namespace clblast