From af4e4f92dc9e5da160eb6c6870a7b38b863f1c6c Mon Sep 17 00:00:00 2001
From: Joseph Redmon <pjreddie@gmail.com>
Date: Tue, 28 Oct 2014 02:45:06 +0000
Subject: [PATCH] getting rid of sub_arrays, nvidia driver memory leak
---
src/im2col.cl | 20 ++++++++++----------
1 files changed, 10 insertions(+), 10 deletions(-)
diff --git a/src/im2col.cl b/src/im2col.cl
index 6ed5d89..8169e1a 100644
--- a/src/im2col.cl
+++ b/src/im2col.cl
@@ -1,4 +1,3 @@
-
float im2col_get_pixel(__global float *im, int height, int width, int channels,
int batch, int row, int col, int channel, int pad)
{
@@ -17,21 +16,22 @@
int c,h,w,b;
int height_col = (height - ksize) / stride + 1;
int width_col = (width - ksize) / stride + 1;
+ int channels_col = channels * ksize * ksize;
if (pad){
height_col = 1 + (height-1) / stride;
width_col = 1 + (width-1) / stride;
pad = ksize/2;
}
- int gid1 = get_global_id(0);
- b = gid1%batch;
- c = gid1/batch;
+ int id = get_global_id(0);
+ w = id % width_col;
+ id /= width_col;
+ h = id % height_col;
+ id /= height_col;
+ c = id % channels_col;
+ id /= channels_col;
+ b = id % batch;
+ id /= batch;
- int gid2 = get_global_id(1);
- h = gid2%height_col;
- w = gid2/height_col;
-
-
- int channels_col = channels * ksize * ksize;
int col_size = height_col*width_col*channels_col;
int w_offset = c % ksize;
int h_offset = (c / ksize) % ksize;
--
Gitblit v1.10.0