From 51477ab274bbb40be259844bdaab9685f693b028 Mon Sep 17 00:00:00 2001
From: AlexeyAB <alexeyab84@gmail.com>
Date: Mon, 24 Jul 2017 23:40:23 +0000
Subject: [PATCH] Increased 2.3 times performance of Yolo on CPU by using OpenMP in both darknet_no_gpu and yolo_cpp_dll_no_gpu
---
src/gemm.c | 21 +++++++++++++--------
1 files changed, 13 insertions(+), 8 deletions(-)
diff --git a/src/gemm.c b/src/gemm.c
index 3003be0..a4db8a4 100644
--- a/src/gemm.c
+++ b/src/gemm.c
@@ -151,14 +151,19 @@
C[i*ldc + j] *= BETA;
}
}
- if(!TA && !TB)
- gemm_nn(M, N, K, ALPHA,A,lda, B, ldb,C,ldc);
- else if(TA && !TB)
- gemm_tn(M, N, K, ALPHA,A,lda, B, ldb,C,ldc);
- else if(!TA && TB)
- gemm_nt(M, N, K, ALPHA,A,lda, B, ldb,C,ldc);
- else
- gemm_tt(M, N, K, ALPHA,A,lda, B, ldb,C,ldc);
+
+ int t;
+ #pragma omp parallel for
+ for (t = 0; t < M; ++t) {
+ if (!TA && !TB)
+ gemm_nn(1, N, K, ALPHA, A + t*lda, lda, B, ldb, C + t*ldc, ldc);
+ else if (TA && !TB)
+ gemm_tn(1, N, K, ALPHA, A + t, lda, B, ldb, C + t*ldc, ldc);
+ else if (!TA && TB)
+ gemm_nt(1, N, K, ALPHA, A + t*lda, lda, B, ldb, C + t*ldc, ldc);
+ else
+ gemm_tt(1, N, K, ALPHA, A + t, lda, B, ldb, C + t*ldc, ldc);
+ }
}
#ifdef GPU
--
Gitblit v1.10.0