From eb9c88ef734d693e65ec35036811363a35e6b5d3 Mon Sep 17 00:00:00 2001
From: AlexeyAB <alexeyab84@gmail.com>
Date: Sat, 14 Apr 2018 22:51:21 +0000
Subject: [PATCH] Fixed bug in Tensor Cores V100 (1. Desc in Batch norm, 2. Manually selected algo). Also fixed time measure on Linux for multi-threading.
---
src/activation_kernels.cu | 2 +-
1 files changed, 1 insertions(+), 1 deletions(-)
diff --git a/src/activation_kernels.cu b/src/activation_kernels.cu
index 994e206..d5f25a0 100644
--- a/src/activation_kernels.cu
+++ b/src/activation_kernels.cu
@@ -154,7 +154,7 @@
extern "C" void activate_array_ongpu(float *x, int n, ACTIVATION a)
{
- activate_array_kernel<<<cuda_gridsize(n), BLOCK>>>(x, n, a);
+ activate_array_kernel<<<cuda_gridsize(n), BLOCK, 0, get_cuda_stream()>>>(x, n, a);
check_error(cudaPeekAtLastError());
}
--
Gitblit v1.10.0