summaryrefslogtreecommitdiff
path: root/src/database/kernels/padtranspose
diff options
context:
space:
mode:
authorCedric Nugteren <web@cedricnugteren.nl>2017-09-06 21:50:42 +0200
committerCedric Nugteren <web@cedricnugteren.nl>2017-09-06 21:50:42 +0200
commit20da5e33a86eda746c17cbdb7bfd295d9f92f074 (patch)
treed35e7091ddc8bbd81d581c4bd49468c6329111fd /src/database/kernels/padtranspose
parentbb947890dec90712c92028c20234eafd48e6fa3e (diff)
Split the database files over multiple directories and files; first step towards separate compilation
Diffstat (limited to 'src/database/kernels/padtranspose')
-rw-r--r--src/database/kernels/padtranspose/padtranspose.hpp14
-rw-r--r--src/database/kernels/padtranspose/padtranspose_16.hpp43
-rw-r--r--src/database/kernels/padtranspose/padtranspose_32.hpp100
-rw-r--r--src/database/kernels/padtranspose/padtranspose_3232.hpp100
-rw-r--r--src/database/kernels/padtranspose/padtranspose_64.hpp80
-rw-r--r--src/database/kernels/padtranspose/padtranspose_6464.hpp80
6 files changed, 417 insertions, 0 deletions
diff --git a/src/database/kernels/padtranspose/padtranspose.hpp b/src/database/kernels/padtranspose/padtranspose.hpp
new file mode 100644
index 00000000..c395653a
--- /dev/null
+++ b/src/database/kernels/padtranspose/padtranspose.hpp
@@ -0,0 +1,14 @@
+
+// =================================================================================================
+// This file is part of the CLBlast project. The project is licensed under Apache Version 2.0. It
+// is auto-generated by the 'scripts/database/database.py' Python script.
+//
+// This file populates the database with best-found tuning parameters for the 'Padtranspose' kernels.
+//
+// =================================================================================================
+
+#include "database/kernels/padtranspose/padtranspose_16.hpp"
+#include "database/kernels/padtranspose/padtranspose_32.hpp"
+#include "database/kernels/padtranspose/padtranspose_3232.hpp"
+#include "database/kernels/padtranspose/padtranspose_64.hpp"
+#include "database/kernels/padtranspose/padtranspose_6464.hpp"
diff --git a/src/database/kernels/padtranspose/padtranspose_16.hpp b/src/database/kernels/padtranspose/padtranspose_16.hpp
new file mode 100644
index 00000000..e815ced0
--- /dev/null
+++ b/src/database/kernels/padtranspose/padtranspose_16.hpp
@@ -0,0 +1,43 @@
+
+// =================================================================================================
+// This file is part of the CLBlast project. The project is licensed under Apache Version 2.0. It
+// is auto-generated by the 'scripts/database/database.py' Python script.
+//
+// This file populates the database with best-found tuning parameters for the 'Padtranspose16' kernels.
+//
+// =================================================================================================
+
+namespace clblast {
+namespace database {
+
+const DatabaseEntry PadtransposeHalf = {
+ "Padtranspose", Precision::kHalf, {"PADTRA_PAD", "PADTRA_TILE", "PADTRA_WPT"}, {
+ { // AMD GPUs
+ kDeviceTypeGPU, "AMD", {
+ { "Ellesmere", { 0, 16, 4 } },
+ { "default", { 0, 16, 4 } },
+ }
+ },
+ { // Intel GPUs
+ kDeviceTypeGPU, "Intel", {
+ { "Intel(R) HD Graphics 5500 BroadWell U-Processor GT2", { 0, 8, 1 } },
+ { "Intel(R) HD Graphics Skylake ULT GT2", { 1, 8, 2 } },
+ { "default", { 0, 8, 1 } },
+ }
+ },
+ { // QUALCOMM GPUs
+ kDeviceTypeGPU, "QUALCOMM", {
+ { "QUALCOMM Adreno(TM)", { 0, 8, 8 } },
+ { "default", { 0, 8, 8 } },
+ }
+ },
+ { // Default
+ kDeviceTypeAll, "default", {
+ { "default", { 0, 8, 1 } },
+ }
+ },
+ }
+};
+
+} // namespace database
+} // namespace clblast
diff --git a/src/database/kernels/padtranspose/padtranspose_32.hpp b/src/database/kernels/padtranspose/padtranspose_32.hpp
new file mode 100644
index 00000000..ca04b01e
--- /dev/null
+++ b/src/database/kernels/padtranspose/padtranspose_32.hpp
@@ -0,0 +1,100 @@
+
+// =================================================================================================
+// This file is part of the CLBlast project. The project is licensed under Apache Version 2.0. It
+// is auto-generated by the 'scripts/database/database.py' Python script.
+//
+// This file populates the database with best-found tuning parameters for the 'Padtranspose32' kernels.
+//
+// =================================================================================================
+
+namespace clblast {
+namespace database {
+
+const DatabaseEntry PadtransposeSingle = {
+ "Padtranspose", Precision::kSingle, {"PADTRA_PAD", "PADTRA_TILE", "PADTRA_WPT"}, {
+ { // AMD GPUs
+ kDeviceTypeGPU, "AMD", {
+ { "AMD Radeon R9 M370X Compute Engine", { 0, 16, 4 } },
+ { "ATI Radeon HD 6750M", { 1, 16, 1 } },
+ { "Ellesmere", { 1, 8, 4 } },
+ { "Fiji", { 0, 16, 2 } },
+ { "Hawaii", { 1, 16, 4 } },
+ { "Oland", { 0, 16, 4 } },
+ { "Pitcairn", { 0, 16, 4 } },
+ { "Tahiti", { 0, 16, 4 } },
+ { "Tonga", { 0, 16, 2 } },
+ { "Turks", { 1, 16, 1 } },
+ { "default", { 0, 16, 4 } },
+ }
+ },
+ { // ARM GPUs
+ kDeviceTypeGPU, "ARM", {
+ { "Mali-T628", { 0, 8, 2 } },
+ { "default", { 0, 8, 2 } },
+ }
+ },
+ { // Intel CPUs
+ kDeviceTypeCPU, "Intel", {
+ { "Intel(R) Core(TM) i7-2670QM CPU @ 2.20GHz", { 0, 8, 8 } },
+ { "Intel(R) Core(TM) i5-6200U CPU @ 2.30GHz", { 0, 16, 1 } },
+ { "Intel(R) Core(TM) i7 CPU 920 @ 2.67GHz", { 0, 32, 1 } },
+ { "Intel(R) Core(TM) i7-3770 CPU @ 3.40GHz", { 0, 8, 8 } },
+ { "Intel(R) Core(TM) i7-4790K CPU @ 4.00GHz", { 0, 8, 8 } },
+ { "Intel(R) Core(TM) i7-5930K CPU @ 3.50GHz", { 0, 32, 1 } },
+ { "default", { 0, 8, 8 } },
+ }
+ },
+ { // Intel GPUs
+ kDeviceTypeGPU, "Intel", {
+ { "Intel(R) HD Graphics 530", { 1, 16, 2 } },
+ { "Intel(R) HD Graphics 5500 BroadWell U-Processor GT2", { 0, 16, 4 } },
+ { "Intel(R) HD Graphics Haswell Ultrabook GT2 Mobile", { 1, 16, 2 } },
+ { "Intel(R) HD Graphics IvyBridge M GT2", { 0, 16, 4 } },
+ { "Intel(R) HD Graphics Skylake ULT GT2", { 1, 16, 2 } },
+ { "Iris", { 1, 16, 2 } },
+ { "Iris Pro", { 1, 16, 2 } },
+ { "default", { 1, 16, 2 } },
+ }
+ },
+ { // Intel accelerators
+ kDeviceTypeAccelerator, "Intel", {
+ { "Intel(R) Many Integrated Core Acceleration Card", { 0, 16, 2 } },
+ { "default", { 0, 16, 2 } },
+ }
+ },
+ { // NVIDIA GPUs
+ kDeviceTypeGPU, "NVIDIA", {
+ { "GRID K520", { 1, 32, 2 } },
+ { "GeForce GTX 1070", { 0, 16, 1 } },
+ { "GeForce GTX 1080", { 1, 16, 2 } },
+ { "GeForce GTX 480", { 1, 16, 2 } },
+ { "GeForce GTX 670", { 1, 32, 2 } },
+ { "GeForce GTX 680", { 1, 16, 2 } },
+ { "GeForce GTX 750", { 1, 32, 2 } },
+ { "GeForce GTX 750 Ti", { 1, 32, 2 } },
+ { "GeForce GTX 980", { 0, 16, 1 } },
+ { "GeForce GTX TITAN", { 1, 16, 2 } },
+ { "GeForce GTX TITAN Black", { 1, 32, 2 } },
+ { "GeForce GTX TITAN X", { 1, 32, 1 } },
+ { "TITAN X (Pascal)", { 1, 16, 2 } },
+ { "Tesla K20m", { 1, 16, 2 } },
+ { "Tesla K40m", { 1, 32, 2 } },
+ { "default", { 1, 32, 2 } },
+ }
+ },
+ { // QUALCOMM GPUs
+ kDeviceTypeGPU, "QUALCOMM", {
+ { "QUALCOMM Adreno(TM)", { 0, 8, 2 } },
+ { "default", { 0, 8, 2 } },
+ }
+ },
+ { // Default
+ kDeviceTypeAll, "default", {
+ { "default", { 1, 16, 2 } },
+ }
+ },
+ }
+};
+
+} // namespace database
+} // namespace clblast
diff --git a/src/database/kernels/padtranspose/padtranspose_3232.hpp b/src/database/kernels/padtranspose/padtranspose_3232.hpp
new file mode 100644
index 00000000..bc9425da
--- /dev/null
+++ b/src/database/kernels/padtranspose/padtranspose_3232.hpp
@@ -0,0 +1,100 @@
+
+// =================================================================================================
+// This file is part of the CLBlast project. The project is licensed under Apache Version 2.0. It
+// is auto-generated by the 'scripts/database/database.py' Python script.
+//
+// This file populates the database with best-found tuning parameters for the 'Padtranspose3232' kernels.
+//
+// =================================================================================================
+
+namespace clblast {
+namespace database {
+
+const DatabaseEntry PadtransposeComplexSingle = {
+ "Padtranspose", Precision::kComplexSingle, {"PADTRA_PAD", "PADTRA_TILE", "PADTRA_WPT"}, {
+ { // AMD GPUs
+ kDeviceTypeGPU, "AMD", {
+ { "AMD Radeon R9 M370X Compute Engine", { 0, 16, 4 } },
+ { "ATI Radeon HD 6750M", { 1, 16, 1 } },
+ { "Ellesmere", { 0, 8, 4 } },
+ { "Fiji", { 1, 16, 2 } },
+ { "Hawaii", { 0, 16, 2 } },
+ { "Oland", { 0, 8, 4 } },
+ { "Pitcairn", { 0, 8, 4 } },
+ { "Tahiti", { 0, 16, 2 } },
+ { "Tonga", { 0, 16, 2 } },
+ { "Turks", { 0, 16, 4 } },
+ { "default", { 0, 8, 4 } },
+ }
+ },
+ { // ARM GPUs
+ kDeviceTypeGPU, "ARM", {
+ { "Mali-T628", { 1, 16, 2 } },
+ { "default", { 1, 16, 2 } },
+ }
+ },
+ { // Intel CPUs
+ kDeviceTypeCPU, "Intel", {
+ { "Intel(R) Core(TM) i7-2670QM CPU @ 2.20GHz", { 0, 8, 8 } },
+ { "Intel(R) Core(TM) i5-6200U CPU @ 2.30GHz", { 1, 8, 4 } },
+ { "Intel(R) Core(TM) i7 CPU 920 @ 2.67GHz", { 0, 8, 4 } },
+ { "Intel(R) Core(TM) i7-3770 CPU @ 3.40GHz", { 0, 8, 8 } },
+ { "Intel(R) Core(TM) i7-4790K CPU @ 4.00GHz", { 0, 8, 8 } },
+ { "Intel(R) Core(TM) i7-5930K CPU @ 3.50GHz", { 0, 8, 4 } },
+ { "default", { 0, 8, 8 } },
+ }
+ },
+ { // Intel GPUs
+ kDeviceTypeGPU, "Intel", {
+ { "Intel(R) HD Graphics 530", { 1, 16, 2 } },
+ { "Intel(R) HD Graphics 5500 BroadWell U-Processor GT2", { 0, 16, 2 } },
+ { "Intel(R) HD Graphics Haswell Ultrabook GT2 Mobile", { 1, 16, 2 } },
+ { "Intel(R) HD Graphics IvyBridge M GT2", { 0, 16, 2 } },
+ { "Intel(R) HD Graphics Skylake ULT GT2", { 0, 16, 4 } },
+ { "Iris", { 0, 16, 2 } },
+ { "Iris Pro", { 1, 16, 2 } },
+ { "default", { 1, 16, 2 } },
+ }
+ },
+ { // Intel accelerators
+ kDeviceTypeAccelerator, "Intel", {
+ { "Intel(R) Many Integrated Core Acceleration Card", { 1, 16, 1 } },
+ { "default", { 1, 16, 1 } },
+ }
+ },
+ { // NVIDIA GPUs
+ kDeviceTypeGPU, "NVIDIA", {
+ { "GRID K520", { 1, 16, 1 } },
+ { "GeForce GTX 1070", { 1, 16, 1 } },
+ { "GeForce GTX 1080", { 0, 8, 1 } },
+ { "GeForce GTX 480", { 1, 16, 1 } },
+ { "GeForce GTX 670", { 1, 16, 1 } },
+ { "GeForce GTX 680", { 1, 16, 1 } },
+ { "GeForce GTX 750", { 1, 16, 2 } },
+ { "GeForce GTX 750 Ti", { 1, 16, 1 } },
+ { "GeForce GTX 980", { 0, 16, 1 } },
+ { "GeForce GTX TITAN", { 1, 16, 1 } },
+ { "GeForce GTX TITAN Black", { 0, 16, 1 } },
+ { "GeForce GTX TITAN X", { 1, 32, 1 } },
+ { "TITAN X (Pascal)", { 1, 8, 1 } },
+ { "Tesla K20m", { 0, 16, 1 } },
+ { "Tesla K40m", { 1, 16, 1 } },
+ { "default", { 1, 16, 1 } },
+ }
+ },
+ { // QUALCOMM GPUs
+ kDeviceTypeGPU, "QUALCOMM", {
+ { "QUALCOMM Adreno(TM)", { 0, 8, 4 } },
+ { "default", { 0, 8, 4 } },
+ }
+ },
+ { // Default
+ kDeviceTypeAll, "default", {
+ { "default", { 1, 8, 2 } },
+ }
+ },
+ }
+};
+
+} // namespace database
+} // namespace clblast
diff --git a/src/database/kernels/padtranspose/padtranspose_64.hpp b/src/database/kernels/padtranspose/padtranspose_64.hpp
new file mode 100644
index 00000000..bdfe9788
--- /dev/null
+++ b/src/database/kernels/padtranspose/padtranspose_64.hpp
@@ -0,0 +1,80 @@
+
+// =================================================================================================
+// This file is part of the CLBlast project. The project is licensed under Apache Version 2.0. It
+// is auto-generated by the 'scripts/database/database.py' Python script.
+//
+// This file populates the database with best-found tuning parameters for the 'Padtranspose64' kernels.
+//
+// =================================================================================================
+
+namespace clblast {
+namespace database {
+
+const DatabaseEntry PadtransposeDouble = {
+ "Padtranspose", Precision::kDouble, {"PADTRA_PAD", "PADTRA_TILE", "PADTRA_WPT"}, {
+ { // AMD GPUs
+ kDeviceTypeGPU, "AMD", {
+ { "AMD Radeon R9 M370X Compute Engine", { 0, 16, 4 } },
+ { "Ellesmere", { 0, 16, 4 } },
+ { "Fiji", { 0, 16, 2 } },
+ { "Hawaii", { 0, 16, 2 } },
+ { "Oland", { 0, 16, 4 } },
+ { "Pitcairn", { 0, 8, 4 } },
+ { "Tahiti", { 1, 16, 2 } },
+ { "Tonga", { 0, 8, 2 } },
+ { "default", { 0, 16, 4 } },
+ }
+ },
+ { // ARM GPUs
+ kDeviceTypeGPU, "ARM", {
+ { "Mali-T628", { 0, 16, 2 } },
+ { "default", { 0, 16, 2 } },
+ }
+ },
+ { // Intel CPUs
+ kDeviceTypeCPU, "Intel", {
+ { "Intel(R) Core(TM) i7-2670QM CPU @ 2.20GHz", { 0, 8, 8 } },
+ { "Intel(R) Core(TM) i5-6200U CPU @ 2.30GHz", { 1, 8, 4 } },
+ { "Intel(R) Core(TM) i7 CPU 920 @ 2.67GHz", { 0, 64, 1 } },
+ { "Intel(R) Core(TM) i7-3770 CPU @ 3.40GHz", { 0, 8, 8 } },
+ { "Intel(R) Core(TM) i7-4790K CPU @ 4.00GHz", { 0, 8, 8 } },
+ { "Intel(R) Core(TM) i7-5930K CPU @ 3.50GHz", { 1, 32, 1 } },
+ { "default", { 1, 8, 4 } },
+ }
+ },
+ { // Intel accelerators
+ kDeviceTypeAccelerator, "Intel", {
+ { "Intel(R) Many Integrated Core Acceleration Card", { 0, 16, 1 } },
+ { "default", { 0, 16, 1 } },
+ }
+ },
+ { // NVIDIA GPUs
+ kDeviceTypeGPU, "NVIDIA", {
+ { "GRID K520", { 1, 16, 1 } },
+ { "GeForce GTX 1070", { 1, 16, 1 } },
+ { "GeForce GTX 1080", { 0, 8, 1 } },
+ { "GeForce GTX 480", { 1, 16, 1 } },
+ { "GeForce GTX 670", { 1, 16, 1 } },
+ { "GeForce GTX 680", { 1, 16, 1 } },
+ { "GeForce GTX 750", { 1, 16, 2 } },
+ { "GeForce GTX 750 Ti", { 1, 32, 2 } },
+ { "GeForce GTX 980", { 1, 32, 1 } },
+ { "GeForce GTX TITAN", { 0, 16, 1 } },
+ { "GeForce GTX TITAN Black", { 0, 16, 1 } },
+ { "GeForce GTX TITAN X", { 1, 32, 1 } },
+ { "TITAN X (Pascal)", { 0, 8, 1 } },
+ { "Tesla K20m", { 0, 16, 1 } },
+ { "Tesla K40m", { 1, 16, 1 } },
+ { "default", { 1, 16, 1 } },
+ }
+ },
+ { // Default
+ kDeviceTypeAll, "default", {
+ { "default", { 1, 16, 2 } },
+ }
+ },
+ }
+};
+
+} // namespace database
+} // namespace clblast
diff --git a/src/database/kernels/padtranspose/padtranspose_6464.hpp b/src/database/kernels/padtranspose/padtranspose_6464.hpp
new file mode 100644
index 00000000..c839ab2c
--- /dev/null
+++ b/src/database/kernels/padtranspose/padtranspose_6464.hpp
@@ -0,0 +1,80 @@
+
+// =================================================================================================
+// This file is part of the CLBlast project. The project is licensed under Apache Version 2.0. It
+// is auto-generated by the 'scripts/database/database.py' Python script.
+//
+// This file populates the database with best-found tuning parameters for the 'Padtranspose6464' kernels.
+//
+// =================================================================================================
+
+namespace clblast {
+namespace database {
+
+const DatabaseEntry PadtransposeComplexDouble = {
+ "Padtranspose", Precision::kComplexDouble, {"PADTRA_PAD", "PADTRA_TILE", "PADTRA_WPT"}, {
+ { // AMD GPUs
+ kDeviceTypeGPU, "AMD", {
+ { "AMD Radeon R9 M370X Compute Engine", { 0, 8, 4 } },
+ { "Ellesmere", { 0, 8, 4 } },
+ { "Fiji", { 0, 8, 2 } },
+ { "Hawaii", { 0, 8, 4 } },
+ { "Oland", { 0, 8, 4 } },
+ { "Pitcairn", { 0, 8, 4 } },
+ { "Tahiti", { 0, 8, 2 } },
+ { "Tonga", { 0, 8, 2 } },
+ { "default", { 0, 8, 4 } },
+ }
+ },
+ { // ARM GPUs
+ kDeviceTypeGPU, "ARM", {
+ { "Mali-T628", { 0, 8, 1 } },
+ { "default", { 0, 8, 1 } },
+ }
+ },
+ { // Intel CPUs
+ kDeviceTypeCPU, "Intel", {
+ { "Intel(R) Core(TM) i7-2670QM CPU @ 2.20GHz", { 0, 8, 4 } },
+ { "Intel(R) Core(TM) i5-6200U CPU @ 2.30GHz", { 1, 8, 2 } },
+ { "Intel(R) Core(TM) i7 CPU 920 @ 2.67GHz", { 1, 16, 2 } },
+ { "Intel(R) Core(TM) i7-3770 CPU @ 3.40GHz", { 1, 8, 4 } },
+ { "Intel(R) Core(TM) i7-4790K CPU @ 4.00GHz", { 0, 8, 4 } },
+ { "Intel(R) Core(TM) i7-5930K CPU @ 3.50GHz", { 1, 8, 4 } },
+ { "default", { 0, 8, 4 } },
+ }
+ },
+ { // Intel accelerators
+ kDeviceTypeAccelerator, "Intel", {
+ { "Intel(R) Many Integrated Core Acceleration Card", { 0, 16, 1 } },
+ { "default", { 0, 16, 1 } },
+ }
+ },
+ { // NVIDIA GPUs
+ kDeviceTypeGPU, "NVIDIA", {
+ { "GRID K520", { 1, 16, 1 } },
+ { "GeForce GTX 1070", { 1, 16, 1 } },
+ { "GeForce GTX 1080", { 1, 8, 1 } },
+ { "GeForce GTX 480", { 1, 16, 1 } },
+ { "GeForce GTX 670", { 1, 16, 1 } },
+ { "GeForce GTX 680", { 1, 32, 1 } },
+ { "GeForce GTX 750", { 1, 16, 1 } },
+ { "GeForce GTX 750 Ti", { 1, 8, 2 } },
+ { "GeForce GTX 980", { 0, 16, 1 } },
+ { "GeForce GTX TITAN", { 1, 16, 1 } },
+ { "GeForce GTX TITAN Black", { 0, 16, 1 } },
+ { "GeForce GTX TITAN X", { 1, 32, 1 } },
+ { "TITAN X (Pascal)", { 1, 8, 1 } },
+ { "Tesla K20m", { 1, 16, 1 } },
+ { "Tesla K40m", { 1, 16, 1 } },
+ { "default", { 1, 16, 1 } },
+ }
+ },
+ { // Default
+ kDeviceTypeAll, "default", {
+ { "default", { 0, 8, 2 } },
+ }
+ },
+ }
+};
+
+} // namespace database
+} // namespace clblast