nudles commented on a change in pull request #492: Make singa use multiple
memory pools
URL: https://github.com/apache/incubator-singa/pull/492#discussion_r309954230
##########
File path: src/core/device/platform.cc
##########
@@ -120,6 +121,26 @@ Platform::CreateCudaGPUs(const size_t num_devices, size_t
init_size) {
const vector<shared_ptr<Device>>
Platform::CreateCudaGPUsOn(const vector<int> &devices, size_t init_size) {
+
Review comment:
it would be better to add a lock here.
if two threads call this function at the same time, this function's output
is then not deterministic.
----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
For queries about this service, please contact Infrastructure at:
[email protected]
With regards,
Apache Git Services
