bybrooklyn/linux
1
0
Fork 0
mirror of https://github.com/torvalds/linux.git synced 2026-04-19 23:34:00 -04:00
Code Issues Packages Projects Releases Wiki Activity

drm/amdkfd: update SIMD distribution algo for GFXIP 9.4.2 onwards

Browse Source 
In certain cooperative group dispatch scenarios the default SPI resource
allocation may cause reduced per-CU workgroup occupancy. Set
COMPUTE_RESOURCE_LIMITS.FORCE_SIMD_DIST=1 to mitigate soft hang
scenarions.

Reviewed-by: Felix Kuehling <felix.kuehling@amd.com>
Suggested-by: Joseph Greathouse <Joseph.Greathouse@amd.com>
Signed-off-by: Rajneesh Bhardwaj <rajneesh.bhardwaj@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
This commit is contained in:
Rajneesh Bhardwaj
2024-01-31 19:33:49 -05:00
committed by Alex Deucher
parent 749f1ad0c3
commit 5869b32bbe
3 changed files with 13 additions and 1 deletions
Download Patch File Download Diff File Expand all files Collapse all files

4
drivers/gpu/drm/amd/amdkfd/kfd_process_queue_manager.c
View File

@@ -95,6 +95,7 @@ void kfd_process_dequeue_from_device(struct kfd_process_device *pdd)
int pqm_set_gws(struct process_queue_manager *pqm, unsigned int qid,
void *gws)
{
struct mqd_update_info minfo = {0};
struct kfd_node *dev = NULL;
struct process_queue_node *pqn;
struct kfd_process_device *pdd;
@@ -146,9 +147,10 @@ int pqm_set_gws(struct process_queue_manager *pqm, unsigned int qid,
}
pdd->qpd.num_gws = gws ? dev->adev->gds.gws_size : 0;
minfo.update_flag = gws ? UPDATE_FLAG_IS_GWS : 0;
return pqn->q->device->dqm->ops.update_queue(pqn->q->device->dqm,
pqn->q, NULL);
pqn->q, &minfo);
}
void kfd_process_dequeue_from_all_devices(struct kfd_process *p)