ROCR-Runtime

ROCR-Runtime copied to clipboard

os_linux: Support non-glibc pthread affinity

1

musl doesn't have `pthread_attr_setaffinity_np`, so we need to use `pthread_setaffinity_np` on `thread` after `pthread_create`. Note: This is not yet tested on master; it was cherry-picked from my 4.5.2 branch, hence...

jpsamaroo

Only the first GPU agent exposes device-local memory regions on a multi-GPU node (ROCr 4.3.x)

5

We try to use ROCm in a multi-GPU setup with several discrete GPUs per node. We want to allocate device-local memory on several GPUs at the same time. However, we...

utkinis

Implement timeout in case of unresponsive GPU

5

Some faulty kernels can cause the GPU to act not as expected by the runtime (such as answering in a reasonable amount of time). Currently, the runtime will loop until...

v01dXYZ

``hsa_executable_load_program_code_object`` could not succeed

1

By looking at the code, ``hsa_executable_load_program_code_object`` is equivalent to ``hsa_executable_load_agent_code_object`` with `agent := {0}`. This could not succeed as it seems the code calls `hsa_agent_get_info` with the handle. It fails...

v01dXYZ

Unitialised Segment objects: defect or feature

1

This code appends new segments to the elf image. But those segments need to be initialised using `pull`. Does the intent of the code is to do that ? If...

v01dXYZ

Please enable two factor authentication in your github account

1

@nitishjohn;@sklam;@guansong;@TermoSINteZ;@NEELMCW;@nhaustov;@kasaurov;@Kirpich30000;@arodrigx7; @amdgerrit;@skeelyamd We are going to enforce two factor authentication in (https://github.com/RadeonOpenCompute/) organization on 7th April , 2022 . Since we identified you as outside collaborator for this organization, you...

DeeptimayeeSethi

Hello, It looks like your use of CPU_ALLOC is incorrect and causing memory leaks. The use case at https://linux.die.net/man/3/cpu_alloc demonstrates the correct way to implement what you are attempting but...

thenumbernine

ROCm 4.5.x error in DiscoverGpu function?

1

Hello, We have been seeing since 4.5.0 some errors on our platform using MI100. All traces point to DiscoverGpu being the root of the problem. There indeed appears to be...

tcojean

hsa_amd_pointer_info doesn't work when calling from a function?

7

On my machine, I got `hsa_amd_pointer_info failed`. If I change `if(0)` to ` if(1)`, the call succeeds. reproducer. ``` #include #include #include #define N 100293 int checkLocked(void *ptr) { hsa_amd_pointer_info_t...

ye-luo

the 1DB Buffer shoud use Structured Buffer Used when: AddTID==0 && Stride!=0 && IdxEn==1 Out of Range if: Index(vgpr) >= NumRecords

qishilu