git/linux-net
1
0
Fork 0
mirror of https://git.kernel.org/pub/scm/linux/kernel/git/netdev/net.git/ synced 2026-04-23 17:02:06 -04:00
Code Issues Packages Projects Releases Wiki Activity

drm/amdgpu: implement lru amdgpu_queue_mgr policy for compute v4

Browse Source 
Use an LRU policy to map usermode rings to HW compute queues.

Most compute clients use one queue, and usually the first queue
available. This results in poor pipe/queue work distribution when
multiple compute apps are running. In most cases pipe 0 queue 0 is
the only queue that gets used.

In order to better distribute work across multiple HW queues, we adopt
a policy to map the usermode ring ids to the LRU HW queue.

This fixes a large majority of multi-app compute workloads sharing the
same HW queue, even though 7 other queues are available.

v2: use ring->funcs->type instead of ring->hw_ip
v3: remove amdgpu_queue_mapper_funcs
v4: change ring_lru_list_lock to spinlock, grab only once in lru_get()

Signed-off-by: Andres Rodriguez <andresx7@gmail.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
This commit is contained in:
Andres Rodriguez
2017-03-06 16:27:55 -05:00
committed by Alex Deucher
parent effd924d2f
commit 795f2813e6
5 changed files with 110 additions and 1 deletions
Download Patch File Download Diff File Expand all files Collapse all files

3
drivers/gpu/drm/amd/amdgpu/amdgpu_device.c
View File

@@ -2073,6 +2073,9 @@ int amdgpu_device_init(struct amdgpu_device *adev,
INIT_LIST_HEAD(&adev->gtt_list);
spin_lock_init(&adev->gtt_list_lock);
INIT_LIST_HEAD(&adev->ring_lru_list);
spin_lock_init(&adev->ring_lru_list_lock);
INIT_DELAYED_WORK(&adev->late_init_work, amdgpu_late_init_func_handler);
if (adev->asic_type >= CHIP_BONAIRE) {