diff options
| author | Joseph Greathouse <Joseph.Greathouse@amd.com> | 2019-09-18 14:49:57 -0500 |
|---|---|---|
| committer | Alex Deucher <alexander.deucher@amd.com> | 2020-04-28 16:20:30 -0400 |
| commit | b8020b0304c8f44e5e29f0b1a04d31e0bf68d26a (patch) | |
| tree | 0fa3c05128379f6c334fbfb646d8f45efc95c1c5 /drivers/gpu/drm/amd/amdkfd/kfd_packet_manager.c | |
| parent | 29633d0e204df1e051d9036e4f493f228ac19fb4 (diff) | |
| download | linux-b8020b0304c8f44e5e29f0b1a04d31e0bf68d26a.tar.gz linux-b8020b0304c8f44e5e29f0b1a04d31e0bf68d26a.tar.bz2 linux-b8020b0304c8f44e5e29f0b1a04d31e0bf68d26a.zip | |
drm/amdkfd: Enable over-subscription with >1 GWS queue
The current GWS usage model will only allows a single GWS-enabled
process to be active on the GPU at once. This ensures that a
barrier-using kernel gets a known amount of GPU hardware, to
prevent deadlock due to inability to go beyond the GWS barrier.
The HWS watches how many GWS entries are assigned to each process,
and goes into over-subscription mode when two processes need more
than the 64 that are available. The current KFD method for working
with this is to allocate all 64 GWS entries to each GWS-capable
process.
When more than one GWS-enabled process is in the runlist, we must
make sure the runlist is in over-subscription mode, so that the
HWS gets a chained RUN_LIST packet and continues scheduling
kernels.
Signed-off-by: Joseph Greathouse <Joseph.Greathouse@amd.com>
Reviewed-by: Felix Kuehling <Felix.Kuehling@amd.com>
Signed-off-by: Felix Kuehling <Felix.Kuehling@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Diffstat (limited to 'drivers/gpu/drm/amd/amdkfd/kfd_packet_manager.c')
| -rw-r--r-- | drivers/gpu/drm/amd/amdkfd/kfd_packet_manager.c | 6 |
1 files changed, 4 insertions, 2 deletions
diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_packet_manager.c b/drivers/gpu/drm/amd/amdkfd/kfd_packet_manager.c index efdb75e7677b..685ca82d42fe 100644 --- a/drivers/gpu/drm/amd/amdkfd/kfd_packet_manager.c +++ b/drivers/gpu/drm/amd/amdkfd/kfd_packet_manager.c @@ -41,7 +41,7 @@ static void pm_calc_rlib_size(struct packet_manager *pm, unsigned int *rlib_size, bool *over_subscription) { - unsigned int process_count, queue_count, compute_queue_count; + unsigned int process_count, queue_count, compute_queue_count, gws_queue_count; unsigned int map_queue_size; unsigned int max_proc_per_quantum = 1; struct kfd_dev *dev = pm->dqm->dev; @@ -49,6 +49,7 @@ static void pm_calc_rlib_size(struct packet_manager *pm, process_count = pm->dqm->processes_count; queue_count = pm->dqm->active_queue_count; compute_queue_count = pm->dqm->active_cp_queue_count; + gws_queue_count = pm->dqm->gws_queue_count; /* check if there is over subscription * Note: the arbitration between the number of VMIDs and @@ -61,7 +62,8 @@ static void pm_calc_rlib_size(struct packet_manager *pm, max_proc_per_quantum = dev->max_proc_per_quantum; if ((process_count > max_proc_per_quantum) || - compute_queue_count > get_cp_queues_num(pm->dqm)) { + compute_queue_count > get_cp_queues_num(pm->dqm) || + gws_queue_count > 1) { *over_subscription = true; pr_debug("Over subscribed runlist\n"); } |