Added XgemvFastRot and Xgemm 16-bit tuning results: just defaults which are now automatically taken from 32-bit if there are no entries at all

2 files changed, 24 insertions, 0 deletions

diff --git a/src/database/kernels/xgemm.hpp b/src/database/kernels/xgemm.hpp
index 7e793076..d19c55b5 100644
--- a/src/database/kernels/xgemm.hpp
+++ b/src/database/kernels/xgemm.hpp
@@ -14,6 +14,18 @@
 namespace clblast {
 // =================================================================================================
 
+const Database::DatabaseEntry Database::XgemmHalf = {
+  "Xgemm", Precision::kHalf, {
+    { // Default
+      kDeviceTypeAll, "default", {
+        { "default",                                         { {"KWG",16}, {"KWI",2}, {"MDIMA",8}, {"MDIMC",8}, {"MWG",32}, {"NDIMB",8}, {"NDIMC",8}, {"NWG",16}, {"SA",0}, {"SB",0}, {"STRM",0}, {"STRN",0}, {"VWM",1}, {"VWN",1} } },
+      }
+    },
+  }
+};
+
+// =================================================================================================
+
 const Database::DatabaseEntry Database::XgemmSingle = {
   "Xgemm", Precision::kSingle, {
     { // AMD GPUs
diff --git a/src/database/kernels/xgemv_fast_rot.hpp b/src/database/kernels/xgemv_fast_rot.hpp
index 3d2e0d3a..328094e1 100644
--- a/src/database/kernels/xgemv_fast_rot.hpp
+++ b/src/database/kernels/xgemv_fast_rot.hpp
@@ -14,6 +14,18 @@
 namespace clblast {
 // =================================================================================================
 
+const Database::DatabaseEntry Database::XgemvFastRotHalf = {
+  "XgemvFastRot", Precision::kHalf, {
+    { // Default
+      kDeviceTypeAll, "default", {
+        { "default",                                         { {"VW3",8}, {"WGS3",32}, {"WPT3",32} } },
+      }
+    },
+  }
+};
+
+// =================================================================================================
+
 const Database::DatabaseEntry Database::XgemvFastRotSingle = {
   "XgemvFastRot", Precision::kSingle, {
     { // AMD GPUs