From 375193fe4e72b320eb63fbc6f98c24714f6970c2 Mon Sep 17 00:00:00 2001 From: Cedric Nugteren Date: Tue, 3 Oct 2017 21:55:21 +0200 Subject: Gemm in-direct implementation now uses only 1 larger instead of max 3 optional temporary buffers --- src/tuning/kernels/xgemm.cpp | 2 ++ 1 file changed, 2 insertions(+) (limited to 'src/tuning/kernels') diff --git a/src/tuning/kernels/xgemm.cpp b/src/tuning/kernels/xgemm.cpp index 7d0f3ed4..6dcdf68b 100644 --- a/src/tuning/kernels/xgemm.cpp +++ b/src/tuning/kernels/xgemm.cpp @@ -180,6 +180,8 @@ class TuneXgemm { tuner.AddArgumentInput(a_mat); tuner.AddArgumentInput(b_mat); tuner.AddArgumentOutput(c_mat); + tuner.AddArgumentScalar(0); + tuner.AddArgumentScalar(0); } }; -- cgit v1.2.3