summaryrefslogtreecommitdiff
path: root/src/database/kernels
diff options
context:
space:
mode:
authorCedric Nugteren <web@cedricnugteren.nl>2017-05-11 22:53:52 -0700
committerCedric Nugteren <web@cedricnugteren.nl>2017-05-11 22:53:52 -0700
commit71933c341190fc17c390344b31614f0f46a67ad4 (patch)
tree788e040690474f0482712ff381f3029b457eb6b6 /src/database/kernels
parentd67455fdb861319da9d78772b6b65d10a66fe4b4 (diff)
Added tuning results for the AMD Radeon Fiji GPU
Diffstat (limited to 'src/database/kernels')
-rw-r--r--src/database/kernels/copy.hpp10
-rw-r--r--src/database/kernels/pad.hpp6
-rw-r--r--src/database/kernels/padtranspose.hpp6
-rw-r--r--src/database/kernels/transpose.hpp6
-rw-r--r--src/database/kernels/xaxpy.hpp6
-rw-r--r--src/database/kernels/xdot.hpp10
-rw-r--r--src/database/kernels/xgemm.hpp8
-rw-r--r--src/database/kernels/xgemm_direct.hpp6
-rw-r--r--src/database/kernels/xgemv.hpp4
-rw-r--r--src/database/kernels/xgemv_fast.hpp4
-rw-r--r--src/database/kernels/xgemv_fast_rot.hpp8
-rw-r--r--src/database/kernels/xger.hpp6
12 files changed, 64 insertions, 16 deletions
diff --git a/src/database/kernels/copy.hpp b/src/database/kernels/copy.hpp
index 89d5a3a6..fd8be5e6 100644
--- a/src/database/kernels/copy.hpp
+++ b/src/database/kernels/copy.hpp
@@ -47,13 +47,14 @@ const Database::DatabaseEntry CopySingle = {
{ "AMD Radeon R9 M370X Compute Engine", { {"COPY_DIMX",32}, {"COPY_DIMY",8}, {"COPY_VW",4}, {"COPY_WPT",1} } },
{ "ATI Radeon HD 6750M", { {"COPY_DIMX",16}, {"COPY_DIMY",8}, {"COPY_VW",2}, {"COPY_WPT",1} } },
{ "Ellesmere", { {"COPY_DIMX",8}, {"COPY_DIMY",8}, {"COPY_VW",4}, {"COPY_WPT",8} } },
+ { "Fiji", { {"COPY_DIMX",16}, {"COPY_DIMY",16}, {"COPY_VW",1}, {"COPY_WPT",2} } },
{ "Hawaii", { {"COPY_DIMX",32}, {"COPY_DIMY",8}, {"COPY_VW",2}, {"COPY_WPT",2} } },
{ "Oland", { {"COPY_DIMX",32}, {"COPY_DIMY",8}, {"COPY_VW",4}, {"COPY_WPT",2} } },
{ "Pitcairn", { {"COPY_DIMX",8}, {"COPY_DIMY",16}, {"COPY_VW",4}, {"COPY_WPT",1} } },
{ "Tahiti", { {"COPY_DIMX",32}, {"COPY_DIMY",8}, {"COPY_VW",2}, {"COPY_WPT",2} } },
{ "Tonga", { {"COPY_DIMX",32}, {"COPY_DIMY",8}, {"COPY_VW",4}, {"COPY_WPT",4} } },
{ "Turks", { {"COPY_DIMX",8}, {"COPY_DIMY",8}, {"COPY_VW",4}, {"COPY_WPT",2} } },
- { "default", { {"COPY_DIMX",8}, {"COPY_DIMY",8}, {"COPY_VW",4}, {"COPY_WPT",1} } },
+ { "default", { {"COPY_DIMX",8}, {"COPY_DIMY",16}, {"COPY_VW",4}, {"COPY_WPT",1} } },
}
},
{ // ARM GPUs
@@ -112,7 +113,7 @@ const Database::DatabaseEntry CopySingle = {
},
{ // Default
kDeviceTypeAll, "default", {
- { "default", { {"COPY_DIMX",32}, {"COPY_DIMY",8}, {"COPY_VW",2}, {"COPY_WPT",2} } },
+ { "default", { {"COPY_DIMX",32}, {"COPY_DIMY",8}, {"COPY_VW",4}, {"COPY_WPT",4} } },
}
},
}
@@ -127,6 +128,7 @@ const Database::DatabaseEntry CopyComplexSingle = {
{ "AMD Radeon R9 M370X Compute Engine", { {"COPY_DIMX",32}, {"COPY_DIMY",8}, {"COPY_VW",1}, {"COPY_WPT",1} } },
{ "ATI Radeon HD 6750M", { {"COPY_DIMX",8}, {"COPY_DIMY",8}, {"COPY_VW",1}, {"COPY_WPT",1} } },
{ "Ellesmere", { {"COPY_DIMX",16}, {"COPY_DIMY",16}, {"COPY_VW",1}, {"COPY_WPT",4} } },
+ { "Fiji", { {"COPY_DIMX",16}, {"COPY_DIMY",8}, {"COPY_VW",1}, {"COPY_WPT",2} } },
{ "Hawaii", { {"COPY_DIMX",32}, {"COPY_DIMY",8}, {"COPY_VW",1}, {"COPY_WPT",2} } },
{ "Oland", { {"COPY_DIMX",8}, {"COPY_DIMY",16}, {"COPY_VW",1}, {"COPY_WPT",1} } },
{ "Pitcairn", { {"COPY_DIMX",8}, {"COPY_DIMY",8}, {"COPY_VW",1}, {"COPY_WPT",2} } },
@@ -198,12 +200,13 @@ const Database::DatabaseEntry CopyDouble = {
kDeviceTypeGPU, "AMD", {
{ "AMD Radeon R9 M370X Compute Engine", { {"COPY_DIMX",32}, {"COPY_DIMY",8}, {"COPY_VW",1}, {"COPY_WPT",1} } },
{ "Ellesmere", { {"COPY_DIMX",32}, {"COPY_DIMY",8}, {"COPY_VW",1}, {"COPY_WPT",4} } },
+ { "Fiji", { {"COPY_DIMX",16}, {"COPY_DIMY",8}, {"COPY_VW",1}, {"COPY_WPT",2} } },
{ "Hawaii", { {"COPY_DIMX",32}, {"COPY_DIMY",8}, {"COPY_VW",1}, {"COPY_WPT",2} } },
{ "Oland", { {"COPY_DIMX",32}, {"COPY_DIMY",8}, {"COPY_VW",2}, {"COPY_WPT",8} } },
{ "Pitcairn", { {"COPY_DIMX",32}, {"COPY_DIMY",8}, {"COPY_VW",1}, {"COPY_WPT",1} } },
{ "Tahiti", { {"COPY_DIMX",8}, {"COPY_DIMY",32}, {"COPY_VW",2}, {"COPY_WPT",1} } },
{ "Tonga", { {"COPY_DIMX",8}, {"COPY_DIMY",32}, {"COPY_VW",2}, {"COPY_WPT",4} } },
- { "default", { {"COPY_DIMX",32}, {"COPY_DIMY",8}, {"COPY_VW",1}, {"COPY_WPT",4} } },
+ { "default", { {"COPY_DIMX",16}, {"COPY_DIMY",8}, {"COPY_VW",2}, {"COPY_WPT",1} } },
}
},
{ // ARM GPUs
@@ -264,6 +267,7 @@ const Database::DatabaseEntry CopyComplexDouble = {
kDeviceTypeGPU, "AMD", {
{ "AMD Radeon R9 M370X Compute Engine", { {"COPY_DIMX",8}, {"COPY_DIMY",16}, {"COPY_VW",1}, {"COPY_WPT",1} } },
{ "Ellesmere", { {"COPY_DIMX",8}, {"COPY_DIMY",32}, {"COPY_VW",1}, {"COPY_WPT",2} } },
+ { "Fiji", { {"COPY_DIMX",8}, {"COPY_DIMY",16}, {"COPY_VW",1}, {"COPY_WPT",1} } },
{ "Hawaii", { {"COPY_DIMX",32}, {"COPY_DIMY",8}, {"COPY_VW",2}, {"COPY_WPT",8} } },
{ "Oland", { {"COPY_DIMX",8}, {"COPY_DIMY",16}, {"COPY_VW",1}, {"COPY_WPT",1} } },
{ "Pitcairn", { {"COPY_DIMX",16}, {"COPY_DIMY",8}, {"COPY_VW",1}, {"COPY_WPT",1} } },
diff --git a/src/database/kernels/pad.hpp b/src/database/kernels/pad.hpp
index c3bce0e3..4336c687 100644
--- a/src/database/kernels/pad.hpp
+++ b/src/database/kernels/pad.hpp
@@ -47,6 +47,7 @@ const Database::DatabaseEntry PadSingle = {
{ "AMD Radeon R9 M370X Compute Engine", { {"PAD_DIMX",32}, {"PAD_DIMY",8}, {"PAD_WPTX",1}, {"PAD_WPTY",1} } },
{ "ATI Radeon HD 6750M", { {"PAD_DIMX",8}, {"PAD_DIMY",16}, {"PAD_WPTX",2}, {"PAD_WPTY",1} } },
{ "Ellesmere", { {"PAD_DIMX",32}, {"PAD_DIMY",8}, {"PAD_WPTX",2}, {"PAD_WPTY",2} } },
+ { "Fiji", { {"PAD_DIMX",16}, {"PAD_DIMY",16}, {"PAD_WPTX",1}, {"PAD_WPTY",2} } },
{ "Hawaii", { {"PAD_DIMX",32}, {"PAD_DIMY",8}, {"PAD_WPTX",1}, {"PAD_WPTY",4} } },
{ "Oland", { {"PAD_DIMX",8}, {"PAD_DIMY",8}, {"PAD_WPTX",1}, {"PAD_WPTY",2} } },
{ "Pitcairn", { {"PAD_DIMX",32}, {"PAD_DIMY",8}, {"PAD_WPTX",1}, {"PAD_WPTY",2} } },
@@ -127,13 +128,14 @@ const Database::DatabaseEntry PadComplexSingle = {
{ "AMD Radeon R9 M370X Compute Engine", { {"PAD_DIMX",32}, {"PAD_DIMY",8}, {"PAD_WPTX",1}, {"PAD_WPTY",1} } },
{ "ATI Radeon HD 6750M", { {"PAD_DIMX",16}, {"PAD_DIMY",8}, {"PAD_WPTX",2}, {"PAD_WPTY",1} } },
{ "Ellesmere", { {"PAD_DIMX",16}, {"PAD_DIMY",16}, {"PAD_WPTX",2}, {"PAD_WPTY",4} } },
+ { "Fiji", { {"PAD_DIMX",16}, {"PAD_DIMY",8}, {"PAD_WPTX",1}, {"PAD_WPTY",2} } },
{ "Hawaii", { {"PAD_DIMX",32}, {"PAD_DIMY",8}, {"PAD_WPTX",1}, {"PAD_WPTY",2} } },
{ "Oland", { {"PAD_DIMX",8}, {"PAD_DIMY",32}, {"PAD_WPTX",1}, {"PAD_WPTY",1} } },
{ "Pitcairn", { {"PAD_DIMX",8}, {"PAD_DIMY",8}, {"PAD_WPTX",1}, {"PAD_WPTY",2} } },
{ "Tahiti", { {"PAD_DIMX",16}, {"PAD_DIMY",16}, {"PAD_WPTX",1}, {"PAD_WPTY",1} } },
{ "Tonga", { {"PAD_DIMX",16}, {"PAD_DIMY",8}, {"PAD_WPTX",1}, {"PAD_WPTY",2} } },
{ "Turks", { {"PAD_DIMX",16}, {"PAD_DIMY",8}, {"PAD_WPTX",4}, {"PAD_WPTY",4} } },
- { "default", { {"PAD_DIMX",32}, {"PAD_DIMY",8}, {"PAD_WPTX",1}, {"PAD_WPTY",2} } },
+ { "default", { {"PAD_DIMX",16}, {"PAD_DIMY",8}, {"PAD_WPTX",1}, {"PAD_WPTY",2} } },
}
},
{ // ARM GPUs
@@ -206,6 +208,7 @@ const Database::DatabaseEntry PadDouble = {
kDeviceTypeGPU, "AMD", {
{ "AMD Radeon R9 M370X Compute Engine", { {"PAD_DIMX",32}, {"PAD_DIMY",8}, {"PAD_WPTX",1}, {"PAD_WPTY",1} } },
{ "Ellesmere", { {"PAD_DIMX",8}, {"PAD_DIMY",32}, {"PAD_WPTX",2}, {"PAD_WPTY",1} } },
+ { "Fiji", { {"PAD_DIMX",8}, {"PAD_DIMY",16}, {"PAD_WPTX",1}, {"PAD_WPTY",2} } },
{ "Hawaii", { {"PAD_DIMX",32}, {"PAD_DIMY",8}, {"PAD_WPTX",1}, {"PAD_WPTY",2} } },
{ "Oland", { {"PAD_DIMX",8}, {"PAD_DIMY",32}, {"PAD_WPTX",1}, {"PAD_WPTY",1} } },
{ "Pitcairn", { {"PAD_DIMX",8}, {"PAD_DIMY",8}, {"PAD_WPTX",1}, {"PAD_WPTY",2} } },
@@ -272,6 +275,7 @@ const Database::DatabaseEntry PadComplexDouble = {
kDeviceTypeGPU, "AMD", {
{ "AMD Radeon R9 M370X Compute Engine", { {"PAD_DIMX",16}, {"PAD_DIMY",8}, {"PAD_WPTX",1}, {"PAD_WPTY",1} } },
{ "Ellesmere", { {"PAD_DIMX",8}, {"PAD_DIMY",16}, {"PAD_WPTX",1}, {"PAD_WPTY",2} } },
+ { "Fiji", { {"PAD_DIMX",32}, {"PAD_DIMY",8}, {"PAD_WPTX",2}, {"PAD_WPTY",1} } },
{ "Hawaii", { {"PAD_DIMX",32}, {"PAD_DIMY",8}, {"PAD_WPTX",1}, {"PAD_WPTY",1} } },
{ "Oland", { {"PAD_DIMX",8}, {"PAD_DIMY",16}, {"PAD_WPTX",2}, {"PAD_WPTY",1} } },
{ "Pitcairn", { {"PAD_DIMX",16}, {"PAD_DIMY",8}, {"PAD_WPTX",1}, {"PAD_WPTY",1} } },
diff --git a/src/database/kernels/padtranspose.hpp b/src/database/kernels/padtranspose.hpp
index 94bac195..0ca62de1 100644
--- a/src/database/kernels/padtranspose.hpp
+++ b/src/database/kernels/padtranspose.hpp
@@ -47,6 +47,7 @@ const Database::DatabaseEntry PadtransposeSingle = {
{ "AMD Radeon R9 M370X Compute Engine", { {"PADTRA_PAD",0}, {"PADTRA_TILE",16}, {"PADTRA_WPT",4} } },
{ "ATI Radeon HD 6750M", { {"PADTRA_PAD",1}, {"PADTRA_TILE",16}, {"PADTRA_WPT",1} } },
{ "Ellesmere", { {"PADTRA_PAD",1}, {"PADTRA_TILE",8}, {"PADTRA_WPT",4} } },
+ { "Fiji", { {"PADTRA_PAD",0}, {"PADTRA_TILE",16}, {"PADTRA_WPT",2} } },
{ "Hawaii", { {"PADTRA_PAD",1}, {"PADTRA_TILE",16}, {"PADTRA_WPT",4} } },
{ "Oland", { {"PADTRA_PAD",0}, {"PADTRA_TILE",16}, {"PADTRA_WPT",4} } },
{ "Pitcairn", { {"PADTRA_PAD",0}, {"PADTRA_TILE",16}, {"PADTRA_WPT",4} } },
@@ -127,6 +128,7 @@ const Database::DatabaseEntry PadtransposeComplexSingle = {
{ "AMD Radeon R9 M370X Compute Engine", { {"PADTRA_PAD",0}, {"PADTRA_TILE",16}, {"PADTRA_WPT",4} } },
{ "ATI Radeon HD 6750M", { {"PADTRA_PAD",1}, {"PADTRA_TILE",16}, {"PADTRA_WPT",1} } },
{ "Ellesmere", { {"PADTRA_PAD",0}, {"PADTRA_TILE",8}, {"PADTRA_WPT",4} } },
+ { "Fiji", { {"PADTRA_PAD",1}, {"PADTRA_TILE",16}, {"PADTRA_WPT",2} } },
{ "Hawaii", { {"PADTRA_PAD",0}, {"PADTRA_TILE",16}, {"PADTRA_WPT",2} } },
{ "Oland", { {"PADTRA_PAD",0}, {"PADTRA_TILE",8}, {"PADTRA_WPT",4} } },
{ "Pitcairn", { {"PADTRA_PAD",0}, {"PADTRA_TILE",8}, {"PADTRA_WPT",4} } },
@@ -206,6 +208,7 @@ const Database::DatabaseEntry PadtransposeDouble = {
kDeviceTypeGPU, "AMD", {
{ "AMD Radeon R9 M370X Compute Engine", { {"PADTRA_PAD",0}, {"PADTRA_TILE",16}, {"PADTRA_WPT",4} } },
{ "Ellesmere", { {"PADTRA_PAD",0}, {"PADTRA_TILE",16}, {"PADTRA_WPT",4} } },
+ { "Fiji", { {"PADTRA_PAD",0}, {"PADTRA_TILE",16}, {"PADTRA_WPT",2} } },
{ "Hawaii", { {"PADTRA_PAD",0}, {"PADTRA_TILE",16}, {"PADTRA_WPT",2} } },
{ "Oland", { {"PADTRA_PAD",0}, {"PADTRA_TILE",16}, {"PADTRA_WPT",4} } },
{ "Pitcairn", { {"PADTRA_PAD",0}, {"PADTRA_TILE",8}, {"PADTRA_WPT",4} } },
@@ -258,7 +261,7 @@ const Database::DatabaseEntry PadtransposeDouble = {
},
{ // Default
kDeviceTypeAll, "default", {
- { "default", { {"PADTRA_PAD",0}, {"PADTRA_TILE",8}, {"PADTRA_WPT",4} } },
+ { "default", { {"PADTRA_PAD",1}, {"PADTRA_TILE",16}, {"PADTRA_WPT",2} } },
}
},
}
@@ -272,6 +275,7 @@ const Database::DatabaseEntry PadtransposeComplexDouble = {
kDeviceTypeGPU, "AMD", {
{ "AMD Radeon R9 M370X Compute Engine", { {"PADTRA_PAD",0}, {"PADTRA_TILE",8}, {"PADTRA_WPT",4} } },
{ "Ellesmere", { {"PADTRA_PAD",0}, {"PADTRA_TILE",8}, {"PADTRA_WPT",4} } },
+ { "Fiji", { {"PADTRA_PAD",0}, {"PADTRA_TILE",8}, {"PADTRA_WPT",2} } },
{ "Hawaii", { {"PADTRA_PAD",0}, {"PADTRA_TILE",8}, {"PADTRA_WPT",4} } },
{ "Oland", { {"PADTRA_PAD",0}, {"PADTRA_TILE",8}, {"PADTRA_WPT",4} } },
{ "Pitcairn", { {"PADTRA_PAD",0}, {"PADTRA_TILE",8}, {"PADTRA_WPT",4} } },
diff --git a/src/database/kernels/transpose.hpp b/src/database/kernels/transpose.hpp
index 8567c725..4020e903 100644
--- a/src/database/kernels/transpose.hpp
+++ b/src/database/kernels/transpose.hpp
@@ -47,6 +47,7 @@ const Database::DatabaseEntry TransposeSingle = {
{ "AMD Radeon R9 M370X Compute Engine", { {"TRA_DIM",4}, {"TRA_PAD",0}, {"TRA_SHUFFLE",1}, {"TRA_WPT",8} } },
{ "ATI Radeon HD 6750M", { {"TRA_DIM",8}, {"TRA_PAD",0}, {"TRA_SHUFFLE",1}, {"TRA_WPT",2} } },
{ "Ellesmere", { {"TRA_DIM",16}, {"TRA_PAD",0}, {"TRA_SHUFFLE",1}, {"TRA_WPT",4} } },
+ { "Fiji", { {"TRA_DIM",16}, {"TRA_PAD",0}, {"TRA_SHUFFLE",1}, {"TRA_WPT",2} } },
{ "Hawaii", { {"TRA_DIM",4}, {"TRA_PAD",0}, {"TRA_SHUFFLE",1}, {"TRA_WPT",8} } },
{ "Oland", { {"TRA_DIM",8}, {"TRA_PAD",0}, {"TRA_SHUFFLE",1}, {"TRA_WPT",4} } },
{ "Pitcairn", { {"TRA_DIM",16}, {"TRA_PAD",0}, {"TRA_SHUFFLE",1}, {"TRA_WPT",1} } },
@@ -127,6 +128,7 @@ const Database::DatabaseEntry TransposeComplexSingle = {
{ "AMD Radeon R9 M370X Compute Engine", { {"TRA_DIM",4}, {"TRA_PAD",1}, {"TRA_SHUFFLE",1}, {"TRA_WPT",4} } },
{ "ATI Radeon HD 6750M", { {"TRA_DIM",16}, {"TRA_PAD",1}, {"TRA_SHUFFLE",1}, {"TRA_WPT",1} } },
{ "Ellesmere", { {"TRA_DIM",4}, {"TRA_PAD",0}, {"TRA_SHUFFLE",1}, {"TRA_WPT",4} } },
+ { "Fiji", { {"TRA_DIM",8}, {"TRA_PAD",1}, {"TRA_SHUFFLE",1}, {"TRA_WPT",2} } },
{ "Hawaii", { {"TRA_DIM",16}, {"TRA_PAD",0}, {"TRA_SHUFFLE",1}, {"TRA_WPT",1} } },
{ "Oland", { {"TRA_DIM",4}, {"TRA_PAD",0}, {"TRA_SHUFFLE",1}, {"TRA_WPT",2} } },
{ "Pitcairn", { {"TRA_DIM",8}, {"TRA_PAD",0}, {"TRA_SHUFFLE",1}, {"TRA_WPT",1} } },
@@ -200,6 +202,7 @@ const Database::DatabaseEntry TransposeDouble = {
kDeviceTypeGPU, "AMD", {
{ "AMD Radeon R9 M370X Compute Engine", { {"TRA_DIM",4}, {"TRA_PAD",0}, {"TRA_SHUFFLE",1}, {"TRA_WPT",4} } },
{ "Ellesmere", { {"TRA_DIM",4}, {"TRA_PAD",0}, {"TRA_SHUFFLE",1}, {"TRA_WPT",4} } },
+ { "Fiji", { {"TRA_DIM",8}, {"TRA_PAD",1}, {"TRA_SHUFFLE",1}, {"TRA_WPT",2} } },
{ "Hawaii", { {"TRA_DIM",16}, {"TRA_PAD",0}, {"TRA_SHUFFLE",1}, {"TRA_WPT",1} } },
{ "Oland", { {"TRA_DIM",8}, {"TRA_PAD",1}, {"TRA_SHUFFLE",1}, {"TRA_WPT",2} } },
{ "Pitcairn", { {"TRA_DIM",4}, {"TRA_PAD",0}, {"TRA_SHUFFLE",1}, {"TRA_WPT",2} } },
@@ -266,12 +269,13 @@ const Database::DatabaseEntry TransposeComplexDouble = {
kDeviceTypeGPU, "AMD", {
{ "AMD Radeon R9 M370X Compute Engine", { {"TRA_DIM",4}, {"TRA_PAD",1}, {"TRA_SHUFFLE",1}, {"TRA_WPT",2} } },
{ "Ellesmere", { {"TRA_DIM",16}, {"TRA_PAD",0}, {"TRA_SHUFFLE",1}, {"TRA_WPT",1} } },
+ { "Fiji", { {"TRA_DIM",16}, {"TRA_PAD",0}, {"TRA_SHUFFLE",1}, {"TRA_WPT",1} } },
{ "Hawaii", { {"TRA_DIM",4}, {"TRA_PAD",0}, {"TRA_SHUFFLE",1}, {"TRA_WPT",2} } },
{ "Oland", { {"TRA_DIM",16}, {"TRA_PAD",0}, {"TRA_SHUFFLE",1}, {"TRA_WPT",1} } },
{ "Pitcairn", { {"TRA_DIM",4}, {"TRA_PAD",0}, {"TRA_SHUFFLE",1}, {"TRA_WPT",1} } },
{ "Tahiti", { {"TRA_DIM",16}, {"TRA_PAD",0}, {"TRA_SHUFFLE",1}, {"TRA_WPT",1} } },
{ "Tonga", { {"TRA_DIM",8}, {"TRA_PAD",1}, {"TRA_SHUFFLE",1}, {"TRA_WPT",2} } },
- { "default", { {"TRA_DIM",8}, {"TRA_PAD",0}, {"TRA_SHUFFLE",1}, {"TRA_WPT",1} } },
+ { "default", { {"TRA_DIM",16}, {"TRA_PAD",0}, {"TRA_SHUFFLE",1}, {"TRA_WPT",1} } },
}
},
{ // ARM GPUs
diff --git a/src/database/kernels/xaxpy.hpp b/src/database/kernels/xaxpy.hpp
index f31e2f6e..bb384abf 100644
--- a/src/database/kernels/xaxpy.hpp
+++ b/src/database/kernels/xaxpy.hpp
@@ -47,6 +47,7 @@ const Database::DatabaseEntry XaxpySingle = {
{ "AMD Radeon R9 M370X Compute Engine", { {"VW",1}, {"WGS",128}, {"WPT",1} } },
{ "ATI Radeon HD 6750M", { {"VW",1}, {"WGS",256}, {"WPT",2} } },
{ "Ellesmere", { {"VW",1}, {"WGS",64}, {"WPT",4} } },
+ { "Fiji", { {"VW",4}, {"WGS",64}, {"WPT",1} } },
{ "Hawaii", { {"VW",2}, {"WGS",64}, {"WPT",2} } },
{ "Oland", { {"VW",1}, {"WGS",128}, {"WPT",1} } },
{ "Pitcairn", { {"VW",2}, {"WGS",128}, {"WPT",1} } },
@@ -127,6 +128,7 @@ const Database::DatabaseEntry XaxpyComplexSingle = {
{ "AMD Radeon R9 M370X Compute Engine", { {"VW",2}, {"WGS",64}, {"WPT",8} } },
{ "ATI Radeon HD 6750M", { {"VW",1}, {"WGS",64}, {"WPT",1} } },
{ "Ellesmere", { {"VW",2}, {"WGS",256}, {"WPT",1} } },
+ { "Fiji", { {"VW",1}, {"WGS",128}, {"WPT",2} } },
{ "Hawaii", { {"VW",1}, {"WGS",128}, {"WPT",2} } },
{ "Oland", { {"VW",1}, {"WGS",128}, {"WPT",1} } },
{ "Pitcairn", { {"VW",1}, {"WGS",64}, {"WPT",1} } },
@@ -206,12 +208,13 @@ const Database::DatabaseEntry XaxpyDouble = {
kDeviceTypeGPU, "AMD", {
{ "AMD Radeon R9 M370X Compute Engine", { {"VW",1}, {"WGS",256}, {"WPT",1} } },
{ "Ellesmere", { {"VW",2}, {"WGS",64}, {"WPT",4} } },
+ { "Fiji", { {"VW",2}, {"WGS",64}, {"WPT",4} } },
{ "Hawaii", { {"VW",1}, {"WGS",64}, {"WPT",2} } },
{ "Oland", { {"VW",1}, {"WGS",64}, {"WPT",1} } },
{ "Pitcairn", { {"VW",1}, {"WGS",128}, {"WPT",1} } },
{ "Tahiti", { {"VW",1}, {"WGS",64}, {"WPT",1} } },
{ "Tonga", { {"VW",1}, {"WGS",128}, {"WPT",4} } },
- { "default", { {"VW",1}, {"WGS",64}, {"WPT",1} } },
+ { "default", { {"VW",2}, {"WGS",64}, {"WPT",1} } },
}
},
{ // ARM GPUs
@@ -272,6 +275,7 @@ const Database::DatabaseEntry XaxpyComplexDouble = {
kDeviceTypeGPU, "AMD", {
{ "AMD Radeon R9 M370X Compute Engine", { {"VW",1}, {"WGS",128}, {"WPT",1} } },
{ "Ellesmere", { {"VW",1}, {"WGS",128}, {"WPT",1} } },
+ { "Fiji", { {"VW",1}, {"WGS",64}, {"WPT",1} } },
{ "Hawaii", { {"VW",2}, {"WGS",64}, {"WPT",1} } },
{ "Oland", { {"VW",1}, {"WGS",256}, {"WPT",1} } },
{ "Pitcairn", { {"VW",1}, {"WGS",128}, {"WPT",1} } },
diff --git a/src/database/kernels/xdot.hpp b/src/database/kernels/xdot.hpp
index 2bc7eeba..2555a74b 100644
--- a/src/database/kernels/xdot.hpp
+++ b/src/database/kernels/xdot.hpp
@@ -47,12 +47,13 @@ const Database::DatabaseEntry XdotSingle = {
{ "AMD Radeon R9 M370X Compute Engine", { {"WGS1",128}, {"WGS2",32} } },
{ "ATI Radeon HD 6750M", { {"WGS1",256}, {"WGS2",32} } },
{ "Ellesmere", { {"WGS1",128}, {"WGS2",32} } },
+ { "Fiji", { {"WGS1",256}, {"WGS2",32} } },
{ "Oland", { {"WGS1",256}, {"WGS2",32} } },
{ "Pitcairn", { {"WGS1",128}, {"WGS2",32} } },
{ "Tahiti", { {"WGS1",128}, {"WGS2",32} } },
{ "Tonga", { {"WGS1",64}, {"WGS2",32} } },
{ "Turks", { {"WGS1",128}, {"WGS2",64} } },
- { "default", { {"WGS1",128}, {"WGS2",32} } },
+ { "default", { {"WGS1",256}, {"WGS2",32} } },
}
},
{ // Intel CPUs
@@ -109,6 +110,7 @@ const Database::DatabaseEntry XdotComplexSingle = {
{ "AMD Radeon R9 M370X Compute Engine", { {"WGS1",64}, {"WGS2",32} } },
{ "ATI Radeon HD 6750M", { {"WGS1",256}, {"WGS2",256} } },
{ "Ellesmere", { {"WGS1",256}, {"WGS2",32} } },
+ { "Fiji", { {"WGS1",256}, {"WGS2",64} } },
{ "Oland", { {"WGS1",128}, {"WGS2",32} } },
{ "Pitcairn", { {"WGS1",256}, {"WGS2",32} } },
{ "Tahiti", { {"WGS1",64}, {"WGS2",32} } },
@@ -170,6 +172,7 @@ const Database::DatabaseEntry XdotDouble = {
kDeviceTypeGPU, "AMD", {
{ "AMD Radeon R9 M370X Compute Engine", { {"WGS1",64}, {"WGS2",128} } },
{ "Ellesmere", { {"WGS1",128}, {"WGS2",64} } },
+ { "Fiji", { {"WGS1",256}, {"WGS2",32} } },
{ "Oland", { {"WGS1",256}, {"WGS2",32} } },
{ "Pitcairn", { {"WGS1",128}, {"WGS2",32} } },
{ "Tahiti", { {"WGS1",256}, {"WGS2",32} } },
@@ -205,7 +208,7 @@ const Database::DatabaseEntry XdotDouble = {
},
{ // Default
kDeviceTypeAll, "default", {
- { "default", { {"WGS1",128}, {"WGS2",64} } },
+ { "default", { {"WGS1",256}, {"WGS2",64} } },
}
},
}
@@ -219,6 +222,7 @@ const Database::DatabaseEntry XdotComplexDouble = {
kDeviceTypeGPU, "AMD", {
{ "AMD Radeon R9 M370X Compute Engine", { {"WGS1",64}, {"WGS2",32} } },
{ "Ellesmere", { {"WGS1",256}, {"WGS2",32} } },
+ { "Fiji", { {"WGS1",256}, {"WGS2",32} } },
{ "Oland", { {"WGS1",256}, {"WGS2",32} } },
{ "Pitcairn", { {"WGS1",256}, {"WGS2",32} } },
{ "Tahiti", { {"WGS1",256}, {"WGS2",32} } },
@@ -254,7 +258,7 @@ const Database::DatabaseEntry XdotComplexDouble = {
},
{ // Default
kDeviceTypeAll, "default", {
- { "default", { {"WGS1",128}, {"WGS2",32} } },
+ { "default", { {"WGS1",256}, {"WGS2",32} } },
}
},
}
diff --git a/src/database/kernels/xgemm.hpp b/src/database/kernels/xgemm.hpp
index 49e86edc..bf296497 100644
--- a/src/database/kernels/xgemm.hpp
+++ b/src/database/kernels/xgemm.hpp
@@ -46,6 +46,7 @@ const Database::DatabaseEntry XgemmSingle = {
{ "AMD Radeon R9 M370X Compute Engine", { {"KWG",32}, {"KWI",2}, {"MDIMA",16}, {"MDIMC",16}, {"MWG",64}, {"NDIMB",8}, {"NDIMC",16}, {"NWG",128}, {"SA",0}, {"SB",0}, {"STRM",0}, {"STRN",0}, {"VWM",2}, {"VWN",8} } },
{ "ATI Radeon HD 6750M", { {"KWG",32}, {"KWI",2}, {"MDIMA",8}, {"MDIMC",16}, {"MWG",128}, {"NDIMB",8}, {"NDIMC",8}, {"NWG",128}, {"SA",0}, {"SB",0}, {"STRM",1}, {"STRN",1}, {"VWM",8}, {"VWN",8} } },
{ "Ellesmere", { {"KWG",32}, {"KWI",2}, {"MDIMA",8}, {"MDIMC",8}, {"MWG",16}, {"NDIMB",16}, {"NDIMC",16}, {"NWG",64}, {"SA",1}, {"SB",1}, {"STRM",0}, {"STRN",0}, {"VWM",1}, {"VWN",2} } },
+ { "Fiji", { {"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} } },
{ "Hawaii", { {"KWG",16}, {"KWI",2}, {"MDIMA",16}, {"MDIMC",32}, {"MWG",128}, {"NDIMB",32}, {"NDIMC",8}, {"NWG",64}, {"SA",1}, {"SB",1}, {"STRM",1}, {"STRN",1}, {"VWM",4}, {"VWN",2} } },
{ "Oland", { {"KWG",16}, {"KWI",2}, {"MDIMA",32}, {"MDIMC",16}, {"MWG",64}, {"NDIMB",32}, {"NDIMC",16}, {"NWG",128}, {"SA",1}, {"SB",1}, {"STRM",1}, {"STRN",0}, {"VWM",2}, {"VWN",4} } },
{ "Pitcairn", { {"KWG",16}, {"KWI",2}, {"MDIMA",16}, {"MDIMC",8}, {"MWG",32}, {"NDIMB",16}, {"NDIMC",16}, {"NWG",128}, {"SA",0}, {"SB",0}, {"STRM",1}, {"STRN",0}, {"VWM",1}, {"VWN",1} } },
@@ -126,13 +127,14 @@ const Database::DatabaseEntry XgemmComplexSingle = {
{ "AMD Radeon R9 M370X Compute Engine", { {"KWG",32}, {"KWI",2}, {"MDIMA",32}, {"MDIMC",32}, {"MWG",64}, {"NDIMB",8}, {"NDIMC",8}, {"NWG",64}, {"SA",0}, {"SB",0}, {"STRM",1}, {"STRN",1}, {"VWM",2}, {"VWN",8} } },
{ "ATI Radeon HD 6750M", { {"KWG",32}, {"KWI",2}, {"MDIMA",8}, {"MDIMC",8}, {"MWG",32}, {"NDIMB",16}, {"NDIMC",16}, {"NWG",64}, {"SA",1}, {"SB",1}, {"STRM",0}, {"STRN",0}, {"VWM",4}, {"VWN",4} } },
{ "Ellesmere", { {"KWG",32}, {"KWI",2}, {"MDIMA",16}, {"MDIMC",16}, {"MWG",32}, {"NDIMB",8}, {"NDIMC",8}, {"NWG",32}, {"SA",1}, {"SB",1}, {"STRM",0}, {"STRN",0}, {"VWM",1}, {"VWN",4} } },
+ { "Fiji", { {"KWG",32}, {"KWI",2}, {"MDIMA",16}, {"MDIMC",16}, {"MWG",32}, {"NDIMB",16}, {"NDIMC",16}, {"NWG",32}, {"SA",1}, {"SB",1}, {"STRM",0}, {"STRN",0}, {"VWM",1}, {"VWN",2} } },
{ "Hawaii", { {"KWG",32}, {"KWI",2}, {"MDIMA",32}, {"MDIMC",8}, {"MWG",32}, {"NDIMB",8}, {"NDIMC",16}, {"NWG",32}, {"SA",1}, {"SB",0}, {"STRM",1}, {"STRN",0}, {"VWM",1}, {"VWN",1} } },
{ "Oland", { {"KWG",32}, {"KWI",2}, {"MDIMA",16}, {"MDIMC",8}, {"MWG",32}, {"NDIMB",32}, {"NDIMC",32}, {"NWG",128}, {"SA",1}, {"SB",0}, {"STRM",0}, {"STRN",1}, {"VWM",2}, {"VWN",4} } },
{ "Pitcairn", { {"KWG",16}, {"KWI",2}, {"MDIMA",8}, {"MDIMC",8}, {"MWG",32}, {"NDIMB",8}, {"NDIMC",8}, {"NWG",32}, {"SA",0}, {"SB",1}, {"STRM",1}, {"STRN",1}, {"VWM",4}, {"VWN",2} } },
{ "Tahiti", { {"KWG",16}, {"KWI",2}, {"MDIMA",8}, {"MDIMC",8}, {"MWG",32}, {"NDIMB",8}, {"NDIMC",16}, {"NWG",32}, {"SA",1}, {"SB",0}, {"STRM",0}, {"STRN",1}, {"VWM",2}, {"VWN",1} } },
{ "Tonga", { {"KWG",16}, {"KWI",2}, {"MDIMA",32}, {"MDIMC",8}, {"MWG",64}, {"NDIMB",16}, {"NDIMC",32}, {"NWG",64}, {"SA",1}, {"SB",1}, {"STRM",1}, {"STRN",0}, {"VWM",2}, {"VWN",1} } },
{ "Turks", { {"KWG",16}, {"KWI",2}, {"MDIMA",8}, {"MDIMC",8}, {"MWG",32}, {"NDIMB",32}, {"NDIMC",8}, {"NWG",32}, {"SA",0}, {"SB",1}, {"STRM",0}, {"STRN",0}, {"VWM",2}, {"VWN",1} } },
- { "default", { {"KWG",32}, {"KWI",2}, {"MDIMA",8}, {"MDIMC",8}, {"MWG",32}, {"NDIMB",8}, {"NDIMC",8}, {"NWG",32}, {"SA",0}, {"SB",0}, {"STRM",0}, {"STRN",0}, {"VWM",4}, {"VWN",4} } },
+ { "default", { {"KWG",32}, {"KWI",2}, {"MDIMA",16}, {"MDIMC",16}, {"MWG",32}, {"NDIMB",16}, {"NDIMC",16}, {"NWG",32}, {"SA",1}, {"SB",1}, {"STRM",0}, {"STRN",0}, {"VWM",1}, {"VWN",2} } },
}
},
{ // ARM GPUs
@@ -205,6 +207,7 @@ const Database::DatabaseEntry XgemmDouble = {
kDeviceTypeGPU, "AMD", {
{ "AMD Radeon R9 M370X Compute Engine", { {"KWG",32}, {"KWI",2}, {"MDIMA",16}, {"MDIMC",16}, {"MWG",64}, {"NDIMB",8}, {"NDIMC",8}, {"NWG",32}, {"SA",0}, {"SB",0}, {"STRM",0}, {"STRN",0}, {"VWM",4}, {"VWN",4} } },
{ "Ellesmere", { {"KWG",32}, {"KWI",2}, {"MDIMA",16}, {"MDIMC",16}, {"MWG",32}, {"NDIMB",16}, {"NDIMC",16}, {"NWG",64}, {"SA",1}, {"SB",1}, {"STRM",0}, {"STRN",0}, {"VWM",2}, {"VWN",2} } },
+ { "Fiji", { {"KWG",32}, {"KWI",2}, {"MDIMA",16}, {"MDIMC",16}, {"MWG",32}, {"NDIMB",16}, {"NDIMC",16}, {"NWG",32}, {"SA",1}, {"SB",1}, {"STRM",0}, {"STRN",0}, {"VWM",2}, {"VWN",2} } },
{ "Hawaii", { {"KWG",16}, {"KWI",8}, {"MDIMA",32}, {"MDIMC",8}, {"MWG",128}, {"NDIMB",8}, {"NDIMC",8}, {"NWG",32}, {"SA",0}, {"SB",1}, {"STRM",0}, {"STRN",0}, {"VWM",1}, {"VWN",4} } },
{ "Oland", { {"KWG",16}, {"KWI",2}, {"MDIMA",8}, {"MDIMC",16}, {"MWG",64}, {"NDIMB",16}, {"NDIMC",8}, {"NWG",16}, {"SA",0}, {"SB",0}, {"STRM",1}, {"STRN",1}, {"VWM",1}, {"VWN",1} } },
{ "Pitcairn", { {"KWG",32}, {"KWI",2}, {"MDIMA",32}, {"MDIMC",16}, {"MWG",64}, {"NDIMB",8}, {"NDIMC",16}, {"NWG",32}, {"SA",0}, {"SB",0}, {"STRM",0}, {"STRN",0}, {"VWM",1}, {"VWN",2} } },
@@ -257,7 +260,7 @@ const Database::DatabaseEntry XgemmDouble = {
},
{ // Default
kDeviceTypeAll, "default", {
- { "default", { {"KWG",32}, {"KWI",2}, {"MDIMA",8}, {"MDIMC",8}, {"MWG",32}, {"NDIMB",8}, {"NDIMC",8}, {"NWG",64}, {"SA",0}, {"SB",0}, {"STRM",0}, {"STRN",0}, {"VWM",4}, {"VWN",4} } },
+ { "default", { {"KWG",32}, {"KWI",2}, {"MDIMA",8}, {"MDIMC",8}, {"MWG",64}, {"NDIMB",8}, {"NDIMC",8}, {"NWG",32}, {"SA",0}, {"SB",0}, {"STRM",0}, {"STRN",0}, {"VWM",4}, {"VWN",4} } },
}
},
}
@@ -271,6 +274,7 @@ const Database::DatabaseEntry XgemmComplexDouble = {
kDeviceTypeGPU, "AMD", {
{ "AMD Radeon R9 M370X Compute Engine", { {"KWG",32}, {"KWI",8}, {"MDIMA",8}, {"MDIMC",16}, {"MWG",32}, {"NDIMB",16}, {"NDIMC",16}, {"NWG",32}, {"SA",0}, {"SB",0}, {"STRM",1}, {"STRN",1}, {"VWM",2}, {"VWN",2} } },
{ "Ellesmere", { {"KWG",32}, {"KWI",2}, {"MDIMA",16}, {"MDIMC",16}, {"MWG",16}, {"NDIMB",16}, {"NDIMC",16}, {"NWG",16}, {"SA",1}, {"SB",1}, {"STRM",0}, {"STRN",0}, {"VWM",1}, {"VWN",1} } },
+ { "Fiji", { {"KWG",32}, {"KWI",2}, {"MDIMA",16}, {"MDIMC",16}, {"MWG",16}, {"NDIMB",16}, {"NDIMC",16}, {"NWG",16}, {"SA",1}, {"SB",1}, {"STRM",0}, {"STRN",0}, {"VWM",1}, {"VWN",1} } },
{ "Hawaii", { {"KWG",16}, {"KWI",2}, {"MDIMA",16}, {"MDIMC",16}, {"MWG",16}, {"NDIMB",16}, {"NDIMC",16}, {"NWG",32}, {"SA",1}, {"SB",0}, {"STRM",0}, {"STRN",0}, {"VWM",1}, {"VWN",2} } },
{ "Oland", { {"KWG",16}, {"KWI",2}, {"MDIMA",16}, {"MDIMC",8}, {"MWG",16}, {"NDIMB",16}, {"NDIMC",32}, {"NWG",128}, {"SA",0}, {"SB",0}, {"STRM",0}, {"STRN",0}, {"VWM",1}, {"VWN",4} } },
{ "Pitcairn", { {"KWG",32}, {"KWI",2}, {"MDIMA",16}, {"MDIMC",8}, {"MWG",32}, {"NDIMB",8}, {"NDIMC",32}, {"NWG",32}, {"SA",0}, {"SB",1}, {"STRM",1}, {"STRN",0}, {"VWM",1}, {"VWN",1} } },
diff --git a/src/database/kernels/xgemm_direct.hpp b/src/database/kernels/xgemm_direct.hpp
index 29a3fb19..b2f67b26 100644
--- a/src/database/kernels/xgemm_direct.hpp
+++ b/src/database/kernels/xgemm_direct.hpp
@@ -46,9 +46,10 @@ const Database::DatabaseEntry XgemmDirectSingle = {
{ "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} } },
{ "ATI Radeon HD 6750M", { {"KWID",8}, {"MDIMAD",8}, {"MDIMCD",16}, {"NDIMBD",8}, {"NDIMCD",8}, {"PADA",1}, {"PADB",0}, {"VWMD",2}, {"VWND",2}, {"WGD",32} } },
{ "Ellesmere", { {"KWID",2}, {"MDIMAD",8}, {"MDIMCD",8}, {"NDIMBD",32}, {"NDIMCD",32}, {"PADA",1}, {"PADB",1}, {"VWMD",2}, {"VWND",1}, {"WGD",32} } },
+ { "Fiji", { {"KWID",2}, {"MDIMAD",16}, {"MDIMCD",16}, {"NDIMBD",8}, {"NDIMCD",8}, {"PADA",1}, {"PADB",1}, {"VWMD",1}, {"VWND",1}, {"WGD",16} } },
{ "Tonga", { {"KWID",16}, {"MDIMAD",16}, {"MDIMCD",16}, {"NDIMBD",32}, {"NDIMCD",8}, {"PADA",0}, {"PADB",1}, {"VWMD",1}, {"VWND",1}, {"WGD",32} } },
{ "Turks", { {"KWID",2}, {"MDIMAD",8}, {"MDIMCD",8}, {"NDIMBD",8}, {"NDIMCD",8}, {"PADA",1}, {"PADB",1}, {"VWMD",1}, {"VWND",1}, {"WGD",16} } },
- { "default", { {"KWID",2}, {"MDIMAD",8}, {"MDIMCD",8}, {"NDIMBD",8}, {"NDIMCD",8}, {"PADA",1}, {"PADB",1}, {"VWMD",1}, {"VWND",2}, {"WGD",16} } },
+ { "default", { {"KWID",2}, {"MDIMAD",16}, {"MDIMCD",16}, {"NDIMBD",8}, {"NDIMCD",8}, {"PADA",1}, {"PADB",1}, {"VWMD",1}, {"VWND",1}, {"WGD",16} } },
}
},
{ // Intel CPUs
@@ -90,6 +91,7 @@ const Database::DatabaseEntry XgemmDirectComplexSingle = {
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} } },
{ "ATI Radeon HD 6750M", { {"KWID",2}, {"MDIMAD",8}, {"MDIMCD",8}, {"NDIMBD",8}, {"NDIMCD",8}, {"PADA",1}, {"PADB",1}, {"VWMD",1}, {"VWND",1}, {"WGD",8} } },
+ { "Fiji", { {"KWID",2}, {"MDIMAD",16}, {"MDIMCD",16}, {"NDIMBD",16}, {"NDIMCD",16}, {"PADA",1}, {"PADB",1}, {"VWMD",1}, {"VWND",1}, {"WGD",16} } },
{ "Tonga", { {"KWID",2}, {"MDIMAD",16}, {"MDIMCD",16}, {"NDIMBD",16}, {"NDIMCD",16}, {"PADA",1}, {"PADB",1}, {"VWMD",2}, {"VWND",2}, {"WGD",32} } },
{ "Turks", { {"KWID",2}, {"MDIMAD",8}, {"MDIMCD",8}, {"NDIMBD",8}, {"NDIMCD",8}, {"PADA",1}, {"PADB",1}, {"VWMD",2}, {"VWND",2}, {"WGD",16} } },
{ "default", { {"KWID",2}, {"MDIMAD",16}, {"MDIMCD",16}, {"NDIMBD",16}, {"NDIMCD",16}, {"PADA",1}, {"PADB",1}, {"VWMD",2}, {"VWND",2}, {"WGD",32} } },
@@ -134,6 +136,7 @@ const Database::DatabaseEntry XgemmDirectDouble = {
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} } },
{ "Ellesmere", { {"KWID",8}, {"MDIMAD",16}, {"MDIMCD",16}, {"NDIMBD",8}, {"NDIMCD",16}, {"PADA",1}, {"PADB",1}, {"VWMD",2}, {"VWND",1}, {"WGD",32} } },
+ { "Fiji", { {"KWID",16}, {"MDIMAD",8}, {"MDIMCD",8}, {"NDIMBD",8}, {"NDIMCD",16}, {"PADA",1}, {"PADB",1}, {"VWMD",1}, {"VWND",1}, {"WGD",16} } },
{ "Tonga", { {"KWID",2}, {"MDIMAD",16}, {"MDIMCD",16}, {"NDIMBD",16}, {"NDIMCD",16}, {"PADA",1}, {"PADB",1}, {"VWMD",1}, {"VWND",1}, {"WGD",32} } },
{ "default", { {"KWID",2}, {"MDIMAD",16}, {"MDIMCD",16}, {"NDIMBD",16}, {"NDIMCD",16}, {"PADA",1}, {"PADB",1}, {"VWMD",1}, {"VWND",1}, {"WGD",16} } },
}
@@ -170,6 +173,7 @@ const Database::DatabaseEntry XgemmDirectComplexDouble = {
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} } },
{ "Ellesmere", { {"KWID",16}, {"MDIMAD",32}, {"MDIMCD",32}, {"NDIMBD",16}, {"NDIMCD",8}, {"PADA",0}, {"PADB",0}, {"VWMD",1}, {"VWND",1}, {"WGD",32} } },
+ { "Fiji", { {"KWID",2}, {"MDIMAD",16}, {"MDIMCD",16}, {"NDIMBD",16}, {"NDIMCD",16}, {"PADA",1}, {"PADB",1}, {"VWMD",1}, {"VWND",1}, {"WGD",16} } },
{ "Tonga", { {"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} } },
}
diff --git a/src/database/kernels/xgemv.hpp b/src/database/kernels/xgemv.hpp
index d56f9acc..cb83d6f7 100644
--- a/src/database/kernels/xgemv.hpp
+++ b/src/database/kernels/xgemv.hpp
@@ -47,6 +47,7 @@ const Database::DatabaseEntry XgemvSingle = {
{ "AMD Radeon R9 M370X Compute Engine", { {"WGS1",128}, {"WPT1",1} } },
{ "ATI Radeon HD 6750M", { {"WGS1",32}, {"WPT1",1} } },
{ "Ellesmere", { {"WGS1",256}, {"WPT1",1} } },
+ { "Fiji", { {"WGS1",128}, {"WPT1",1} } },
{ "Hawaii", { {"WGS1",128}, {"WPT1",1} } },
{ "Oland", { {"WGS1",128}, {"WPT1",1} } },
{ "Pitcairn", { {"WGS1",256}, {"WPT1",1} } },
@@ -120,6 +121,7 @@ const Database::DatabaseEntry XgemvComplexSingle = {
{ "AMD Radeon R9 M370X Compute Engine", { {"WGS1",64}, {"WPT1",1} } },
{ "ATI Radeon HD 6750M", { {"WGS1",64}, {"WPT1",1} } },
{ "Ellesmere", { {"WGS1",32}, {"WPT1",1} } },
+ { "Fiji", { {"WGS1",32}, {"WPT1",1} } },
{ "Hawaii", { {"WGS1",64}, {"WPT1",1} } },
{ "Oland", { {"WGS1",64}, {"WPT1",1} } },
{ "Pitcairn", { {"WGS1",64}, {"WPT1",1} } },
@@ -188,6 +190,7 @@ const Database::DatabaseEntry XgemvDouble = {
kDeviceTypeGPU, "AMD", {
{ "AMD Radeon R9 M370X Compute Engine", { {"WGS1",64}, {"WPT1",1} } },
{ "Ellesmere", { {"WGS1",32}, {"WPT1",1} } },
+ { "Fiji", { {"WGS1",32}, {"WPT1",1} } },
{ "Hawaii", { {"WGS1",128}, {"WPT1",1} } },
{ "Oland", { {"WGS1",256}, {"WPT1",1} } },
{ "Pitcairn", { {"WGS1",256}, {"WPT1",1} } },
@@ -247,6 +250,7 @@ const Database::DatabaseEntry XgemvComplexDouble = {
kDeviceTypeGPU, "AMD", {
{ "AMD Radeon R9 M370X Compute Engine", { {"WGS1",64}, {"WPT1",1} } },
{ "Ellesmere", { {"WGS1",32}, {"WPT1",1} } },
+ { "Fiji", { {"WGS1",64}, {"WPT1",1} } },
{ "Hawaii", { {"WGS1",64}, {"WPT1",1} } },
{ "Oland", { {"WGS1",256}, {"WPT1",1} } },
{ "Pitcairn", { {"WGS1",256}, {"WPT1",1} } },
diff --git a/src/database/kernels/xgemv_fast.hpp b/src/database/kernels/xgemv_fast.hpp
index 2ab6265e..4b4acf8e 100644
--- a/src/database/kernels/xgemv_fast.hpp
+++ b/src/database/kernels/xgemv_fast.hpp
@@ -47,6 +47,7 @@ const Database::DatabaseEntry XgemvFastSingle = {
{ "AMD Radeon R9 M370X Compute Engine", { {"VW2",1}, {"WGS2",128}, {"WPT2",1} } },
{ "ATI Radeon HD 6750M", { {"VW2",2}, {"WGS2",64}, {"WPT2",2} } },
{ "Ellesmere", { {"VW2",1}, {"WGS2",64}, {"WPT2",1} } },
+ { "Fiji", { {"VW2",1}, {"WGS2",64}, {"WPT2",2} } },
{ "Hawaii", { {"VW2",1}, {"WGS2",64}, {"WPT2",1} } },
{ "Oland", { {"VW2",1}, {"WGS2",64}, {"WPT2",1} } },
{ "Pitcairn", { {"VW2",1}, {"WGS2",64}, {"WPT2",1} } },
@@ -120,6 +121,7 @@ const Database::DatabaseEntry XgemvFastComplexSingle = {
{ "AMD Radeon R9 M370X Compute Engine", { {"VW2",2}, {"WGS2",256}, {"WPT2",2} } },
{ "ATI Radeon HD 6750M", { {"VW2",1}, {"WGS2",128}, {"WPT2",1} } },
{ "Ellesmere", { {"VW2",1}, {"WGS2",64}, {"WPT2",1} } },
+ { "Fiji", { {"VW2",1}, {"WGS2",16}, {"WPT2",1} } },
{ "Hawaii", { {"VW2",1}, {"WGS2",64}, {"WPT2",1} } },
{ "Oland", { {"VW2",1}, {"WGS2",64}, {"WPT2",1} } },
{ "Pitcairn", { {"VW2",1}, {"WGS2",64}, {"WPT2",1} } },
@@ -182,6 +184,7 @@ const Database::DatabaseEntry XgemvFastDouble = {
kDeviceTypeGPU, "AMD", {
{ "AMD Radeon R9 M370X Compute Engine", { {"VW2",1}, {"WGS2",256}, {"WPT2",1} } },
{ "Ellesmere", { {"VW2",1}, {"WGS2",128}, {"WPT2",1} } },
+ { "Fiji", { {"VW2",1}, {"WGS2",32}, {"WPT2",1} } },
{ "Hawaii", { {"VW2",1}, {"WGS2",64}, {"WPT2",1} } },
{ "Oland", { {"VW2",1}, {"WGS2",64}, {"WPT2",1} } },
{ "Pitcairn", { {"VW2",1}, {"WGS2",64}, {"WPT2",1} } },
@@ -241,6 +244,7 @@ const Database::DatabaseEntry XgemvFastComplexDouble = {
kDeviceTypeGPU, "AMD", {
{ "AMD Radeon R9 M370X Compute Engine", { {"VW2",1}, {"WGS2",256}, {"WPT2",1} } },
{ "Ellesmere", { {"VW2",1}, {"WGS2",16}, {"WPT2",1} } },
+ { "Fiji", { {"VW2",1}, {"WGS2",16}, {"WPT2",1} } },
{ "Hawaii", { {"VW2",1}, {"WGS2",64}, {"WPT2",1} } },
{ "Oland", { {"VW2",1}, {"WGS2",256}, {"WPT2",1} } },
{ "Pitcairn", { {"VW2",1}, {"WGS2",64}, {"WPT2",1} } },
diff --git a/src/database/kernels/xgemv_fast_rot.hpp b/src/database/kernels/xgemv_fast_rot.hpp
index b234e27f..8088fd65 100644
--- a/src/database/kernels/xgemv_fast_rot.hpp
+++ b/src/database/kernels/xgemv_fast_rot.hpp
@@ -46,9 +46,10 @@ const Database::DatabaseEntry XgemvFastRotSingle = {
{ "AMD Radeon R9 M370X Compute Engine", { {"VW3",8}, {"WGS3",64}, {"WPT3",32} } },
{ "ATI Radeon HD 6750M", { {"VW3",8}, {"WGS3",128}, {"WPT3",16} } },
{ "Ellesmere", { {"VW3",8}, {"WGS3",32}, {"WPT3",32} } },
+ { "Fiji", { {"VW3",4}, {"WGS3",32}, {"WPT3",16} } },
{ "Tonga", { {"VW3",8}, {"WGS3",128}, {"WPT3",32} } },
{ "Turks", { {"VW3",8}, {"WGS3",128}, {"WPT3",16} } },
- { "default", { {"VW3",8}, {"WGS3",128}, {"WPT3",32} } },
+ { "default", { {"VW3",8}, {"WGS3",32}, {"WPT3",32} } },
}
},
{ // Intel CPUs
@@ -96,9 +97,10 @@ const Database::DatabaseEntry XgemvFastRotComplexSingle = {
{ "AMD Radeon R9 M370X Compute Engine", { {"VW3",8}, {"WGS3",16}, {"WPT3",16} } },
{ "ATI Radeon HD 6750M", { {"VW3",8}, {"WGS3",32}, {"WPT3",8} } },
{ "Ellesmere", { {"VW3",2}, {"WGS3",32}, {"WPT3",16} } },
+ { "Fiji", { {"VW3",4}, {"WGS3",32}, {"WPT3",32} } },
{ "Tonga", { {"VW3",4}, {"WGS3",32}, {"WPT3",32} } },
{ "Turks", { {"VW3",4}, {"WGS3",32}, {"WPT3",8} } },
- { "default", { {"VW3",8}, {"WGS3",32}, {"WPT3",32} } },
+ { "default", { {"VW3",8}, {"WGS3",16}, {"WPT3",16} } },
}
},
{ // Intel CPUs
@@ -135,6 +137,7 @@ const Database::DatabaseEntry XgemvFastRotDouble = {
kDeviceTypeGPU, "AMD", {
{ "AMD Radeon R9 M370X Compute Engine", { {"VW3",4}, {"WGS3",16}, {"WPT3",16} } },
{ "Ellesmere", { {"VW3",4}, {"WGS3",16}, {"WPT3",16} } },
+ { "Fiji", { {"VW3",4}, {"WGS3",32}, {"WPT3",32} } },
{ "Tonga", { {"VW3",4}, {"WGS3",16}, {"WPT3",16} } },
{ "default", { {"VW3",4}, {"WGS3",16}, {"WPT3",16} } },
}
@@ -173,6 +176,7 @@ const Database::DatabaseEntry XgemvFastRotComplexDouble = {
kDeviceTypeGPU, "AMD", {
{ "AMD Radeon R9 M370X Compute Engine", { {"VW3",4}, {"WGS3",32}, {"WPT3",16} } },
{ "Ellesmere", { {"VW3",4}, {"WGS3",16}, {"WPT3",16} } },
+ { "Fiji", { {"VW3",4}, {"WGS3",32}, {"WPT3",8} } },
{ "Tonga", { {"VW3",4}, {"WGS3",16}, {"WPT3",8} } },
{ "default", { {"VW3",8}, {"WGS3",32}, {"WPT3",16} } },
}
diff --git a/src/database/kernels/xger.hpp b/src/database/kernels/xger.hpp
index 5a61f0df..95dac88b 100644
--- a/src/database/kernels/xger.hpp
+++ b/src/database/kernels/xger.hpp
@@ -47,6 +47,7 @@ const Database::DatabaseEntry XgerSingle = {
{ "AMD Radeon R9 M370X Compute Engine", { {"WGS1",256}, {"WGS2",1}, {"WPT",1} } },
{ "ATI Radeon HD 6750M", { {"WGS1",16}, {"WGS2",16}, {"WPT",4} } },
{ "Ellesmere", { {"WGS1",64}, {"WGS2",4}, {"WPT",2} } },
+ { "Fiji", { {"WGS1",256}, {"WGS2",1}, {"WPT",1} } },
{ "Hawaii", { {"WGS1",64}, {"WGS2",2}, {"WPT",1} } },
{ "Oland", { {"WGS1",32}, {"WGS2",4}, {"WPT",2} } },
{ "Pitcairn", { {"WGS1",64}, {"WGS2",1}, {"WPT",1} } },
@@ -115,13 +116,14 @@ const Database::DatabaseEntry XgerComplexSingle = {
{ "AMD Radeon R9 M370X Compute Engine", { {"WGS1",64}, {"WGS2",4}, {"WPT",1} } },
{ "ATI Radeon HD 6750M", { {"WGS1",16}, {"WGS2",16}, {"WPT",1} } },
{ "Ellesmere", { {"WGS1",16}, {"WGS2",8}, {"WPT",2} } },
+ { "Fiji", { {"WGS1",128}, {"WGS2",2}, {"WPT",1} } },
{ "Hawaii", { {"WGS1",64}, {"WGS2",1}, {"WPT",2} } },
{ "Oland", { {"WGS1",4}, {"WGS2",8}, {"WPT",1} } },
{ "Pitcairn", { {"WGS1",128}, {"WGS2",2}, {"WPT",1} } },
{ "Tahiti", { {"WGS1",64}, {"WGS2",2}, {"WPT",1} } },
{ "Tonga", { {"WGS1",64}, {"WGS2",1}, {"WPT",1} } },
{ "Turks", { {"WGS1",128}, {"WGS2",2}, {"WPT",1} } },
- { "default", { {"WGS1",64}, {"WGS2",4}, {"WPT",1} } },
+ { "default", { {"WGS1",128}, {"WGS2",2}, {"WPT",1} } },
}
},
{ // ARM GPUs
@@ -182,6 +184,7 @@ const Database::DatabaseEntry XgerDouble = {
kDeviceTypeGPU, "AMD", {
{ "AMD Radeon R9 M370X Compute Engine", { {"WGS1",32}, {"WGS2",4}, {"WPT",1} } },
{ "Ellesmere", { {"WGS1",64}, {"WGS2",1}, {"WPT",4} } },
+ { "Fiji", { {"WGS1",256}, {"WGS2",1}, {"WPT",2} } },
{ "Hawaii", { {"WGS1",32}, {"WGS2",4}, {"WPT",2} } },
{ "Oland", { {"WGS1",128}, {"WGS2",1}, {"WPT",2} } },
{ "Pitcairn", { {"WGS1",64}, {"WGS2",1}, {"WPT",1} } },
@@ -237,6 +240,7 @@ const Database::DatabaseEntry XgerComplexDouble = {
kDeviceTypeGPU, "AMD", {
{ "AMD Radeon R9 M370X Compute Engine", { {"WGS1",64}, {"WGS2",1}, {"WPT",1} } },
{ "Ellesmere", { {"WGS1",8}, {"WGS2",16}, {"WPT",1} } },
+ { "Fiji", { {"WGS1",64}, {"WGS2",4}, {"WPT",2} } },
{ "Hawaii", { {"WGS1",128}, {"WGS2",1}, {"WPT",1} } },
{ "Oland", { {"WGS1",16}, {"WGS2",16}, {"WPT",2} } },
{ "Pitcairn", { {"WGS1",64}, {"WGS2",4}, {"WPT",1} } },