Commit 17df65e6 authored by Tetragramm's avatar Tetragramm

Fix the OpenCL portion to match the c++ code.

Fix an undiscovered bug in the c++ code.
parent c6772a8f
...@@ -215,7 +215,7 @@ namespace ...@@ -215,7 +215,7 @@ namespace
if (residual != 0) if (residual != 0)
{ {
int residualStep = MAX(histSize / residual, 1); int residualStep = MAX(histSize / residual, 1);
for (int i = 0; i < histSize; i += residualStep) for (int i = 0; i < histSize && residual > 0; i += residualStep, residual--)
tileHist[i]++; tileHist[i]++;
} }
} }
......
...@@ -201,7 +201,10 @@ __kernel void calcLut(__global __const uchar * src, const int srcStep, ...@@ -201,7 +201,10 @@ __kernel void calcLut(__global __const uchar * src, const int srcStep,
tHistVal += redistBatch; tHistVal += redistBatch;
int residual = totalClipped - redistBatch * 256; int residual = totalClipped - redistBatch * 256;
if (tid < residual) int rStep = 256 / residual;
if (rStep < 1)
rStep = 1;
if (tid%rStep == 0 && (tid/rStep)<residual)
++tHistVal; ++tHistVal;
} }
......
Markdown is supported
0% or
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment