AI/rocm-systems
2
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

rpl_run w/o input file; queue create callback; test for n gpus and n threads

Browse Source 
Change-Id: I37157c49cf6454de591cae97b5cc43287ea95956
This commit is contained in:
Evgeny
2018-10-30 14:19:45 -05:00
parent c05bded17c
commit f977ac2fbf
12 changed files with 137 additions and 73 deletions
Download Patch File Download Diff File Expand all files Collapse all files
test/app/standalone_test.cpp
+4 -5
View File
@@ -145,8 +145,7 @@ int main() {
TEST_STATUS(status == HSA_STATUS_SUCCESS);
// Test initialization
TestHsa::SetQueue(prof_queue);
TestHsa::HsaInstantiate(0);
TestHsa::HsaInstantiate();
// Dispatching profiled kernel n-times to collect all counter groups data
const unsigned group_n = 0;
@@ -157,9 +156,9 @@ int main() {
for (unsigned ind = 0; ind < 3; ++ind) {
#if 1
const unsigned queue_ind = ind % queue_count;
TestHsa::SetQueue(queue[queue_ind]);
// ret_val = RunKernel<DummyKernel, TestAql>();
ret_val = RunKernel<SimpleConvolution, TestAql>();
hsa_queue_t* prof_queue = queue[queue_ind];
//ret_val = RunKernel<DummyKernel, TestAql>(0, NULL, NULL, prof_queue);
ret_val = RunKernel<SimpleConvolution, TestAql>(0, NULL, NULL, prof_queue);
std::cout << "run kernel, queue " << queue_ind << std::endl;
#else
sleep(3);