From cb1f33c6ae840e8dc0f43518daf76e6ed01034f0 Mon Sep 17 00:00:00 2001
From: Joseph Redmon <pjreddie@gmail.com>
Date: Mon, 08 Dec 2014 19:48:57 +0000
Subject: [PATCH] Fixed race condition in server

---
 src/mini_blas.h |   80 +++++++++++++++++++++++++++++++++++-----
 1 files changed, 70 insertions(+), 10 deletions(-)

diff --git a/src/mini_blas.h b/src/mini_blas.h
index cdf3a86..ea73ed6 100644
--- a/src/mini_blas.h
+++ b/src/mini_blas.h
@@ -1,10 +1,70 @@
-void gemm(int TA, int TB, int M, int N, int K, double ALPHA, 
-                    double *A, int lda, 
-                    double *B, int ldb,
-                    double BETA,
-                    double *C, int ldc);
-void im2row(double *image, int h, int w, int c, int size, int stride, double *matrix);
-void im2col(double *image, int h, int w, int c, int size, int stride, double *matrix);
-void im2col_cpu(double* data_im, const int channels,
-        const int height, const int width, const int ksize, const int stride,
-        double* data_col);
+#include "opencl.h"
+
+void pm(int M, int N, float *A);
+void gemm(int TA, int TB, int M, int N, int K, float ALPHA, 
+                    float *A, int lda, 
+                    float *B, int ldb,
+                    float BETA,
+                    float *C, int ldc);
+float *random_matrix(int rows, int cols);
+void time_random_matrix(int TA, int TB, int m, int k, int n);
+
+#ifdef GPU
+void axpy_ongpu(int N, float ALPHA, cl_mem X, int INCX, cl_mem Y, int INCY);
+void axpy_ongpu_offset(int N, float ALPHA, cl_mem X, int OFFX, int INCX, cl_mem Y, int OFFY, int INCY);
+void copy_ongpu(int N, cl_mem X, int INCX, cl_mem Y, int INCY);
+void copy_ongpu_offset(int N, cl_mem X, int OFFX, int INCX, cl_mem Y, int OFFY, int INCY);
+void scal_ongpu(int N, float ALPHA, cl_mem X, int INCX);
+void im2col_ongpu(cl_mem data_im, int offset,
+         int channels, int height, int width,
+         int ksize, int stride, int pad, cl_mem data_col);
+
+void col2im_gpu(float *data_col,  int offset,
+         int channels,  int height,  int width,
+         int ksize,  int stride,  int pad, float *data_im);
+void col2im_ongpu(cl_mem data_col, int batch,
+        int channels, int height, int width,
+        int ksize, int stride, int pad, cl_mem data_im);
+
+void im2col_gpu(float *data_im,
+         int channels, int height, int width,
+         int ksize, int stride, int pad, float *data_col);
+
+void gemm_ongpu_offset(int TA, int TB, int M, int N, int K, float ALPHA, 
+        cl_mem A_gpu, int a_off, int lda, 
+        cl_mem B_gpu, int b_off, int ldb,
+        float BETA,
+        cl_mem C_gpu, int c_off, int ldc);
+
+void gemm_ongpu(int TA, int TB, int M, int N, int K, float ALPHA, 
+        cl_mem A_gpu, int lda, 
+        cl_mem B_gpu, int ldb,
+        float BETA,
+        cl_mem C_gpu, int ldc);
+#endif
+
+void im2col_cpu(float* data_im,
+        int channels, int height, int width,
+        int ksize, int stride, int pad, float* data_col);
+
+void col2im_cpu(float* data_col,
+        int channels, int height, int width,
+        int ksize, int stride, int pad, float* data_im);
+
+void test_blas();
+
+void gemm_gpu(int TA, int TB, int M, int N, int K, float ALPHA, 
+        float *A, int lda, 
+        float *B, int ldb,
+        float BETA,
+        float *C, int ldc);
+void gemm_cpu(int TA, int TB, int M, int N, int K, float ALPHA, 
+        float *A, int lda, 
+        float *B, int ldb,
+        float BETA,
+        float *C, int ldc);
+void axpy_cpu(int N, float ALPHA, float *X, int INCX, float *Y, int INCY);
+void copy_cpu(int N, float *X, int INCX, float *Y, int INCY);
+void scal_cpu(int N, float ALPHA, float *X, int INCX);
+float dot_cpu(int N, float *X, int INCX, float *Y, int INCY);
+void test_gpu_blas();

--
Gitblit v1.10.0