COMPMID-3887: NEGEMMConvolutionLayer hangs up on num_threads>18
When store the Tensor shape information, the size was clamped to 32bit, which cause the memory allocated is not big enough.
Signed-off-by: Sheri Zhang <sheri.zhang@arm.com>
Change-Id: I9f7dfcd5595a143b0ed4f6973e20bcd9d776b673
Reviewed-on: https://review.mlplatform.org/c/ml/ComputeLibrary/+/4331
Comments-Addressed: Arm Jenkins <bsgcomp@arm.com>
Tested-by: Arm Jenkins <bsgcomp@arm.com>
Reviewed-by: Georgios Pinitas <georgios.pinitas@arm.com>
diff --git a/arm_compute/core/Window.inl b/arm_compute/core/Window.inl
index 14a432a..6100d09 100644
--- a/arm_compute/core/Window.inl
+++ b/arm_compute/core/Window.inl
@@ -197,15 +197,15 @@
{
if(d == dimension)
{
- int start = _dims[d].start();
- int end = _dims[d].end();
- const int step = _dims[d].step();
+ int start = _dims[d].start();
+ int end = _dims[d].end();
+ const int step = _dims[d].step();
const int num_it = num_iterations(d);
const int rem = num_it % total;
- int work = num_it / total;
+ int work = num_it / total;
- int it_start = work * id;
+ int it_start = work * id;
if(int(id) < rem)
{
@@ -277,7 +277,7 @@
{
for(unsigned int n = first_dimension; n < shape.num_dimensions(); ++n)
{
- set(n, Window::Dimension(0, std::max(shape[n], static_cast<uint32_t>(1))));
+ set(n, Window::Dimension(0, std::max(shape[n], static_cast<size_t>(1))));
}
}