summaryrefslogtreecommitdiff
path: root/src/tuning
diff options
context:
space:
mode:
authorCNugteren <web@cedricnugteren.nl>2015-07-22 07:31:16 +0200
committerCNugteren <web@cedricnugteren.nl>2015-07-22 07:31:16 +0200
commit4dcecfe93476c2cbd1148bdbb3d327d9ad444af2 (patch)
tree2b0467ba54bcd00fcf8a3dd1ce0bd80e2d2d70c7 /src/tuning
parentd93efa31691e1a34e16865b7b96624fca670bab2 (diff)
Added workgroup shuffle option to transpose kernel for AMD GPUs
Diffstat (limited to 'src/tuning')
-rw-r--r--src/tuning/transpose.cc1
1 files changed, 1 insertions, 0 deletions
diff --git a/src/tuning/transpose.cc b/src/tuning/transpose.cc
index 46756599..8963a688 100644
--- a/src/tuning/transpose.cc
+++ b/src/tuning/transpose.cc
@@ -41,6 +41,7 @@ void TransposeTune(const Arguments<T> &args,
tuner.AddParameter(id, "TRA_DIM", {4, 8, 16, 32, 64});
tuner.AddParameter(id, "TRA_WPT", {1, 2, 4, 8, 16});
tuner.AddParameter(id, "TRA_PAD", {0, 1});
+ tuner.AddParameter(id, "TRA_SHUFFLE", {0, 1});
// Tests for a specific precision
tuner.AddParameter(id, "PRECISION", {static_cast<size_t>(args.precision)});