[PATCH] drm/amdgpu: Enable tunneling on high-priority compute queues

[Michel Dänzer]

On 2023-12-08 18:59, Alex Deucher wrote:
> On Fri, Dec 8, 2023 at 12:27 PM Joshua Ashton <joshua at froggi.es> wrote:
>>
>> I heard some musings about dmabuf deadline kernel work recently, but not
>> sure if any of that is applicable to AMD.
> 
> I think something like a workload hint would be more useful.  We did a
> few patch sets to allow userspace to provide a hint to the kernel
> about the workload type so the kernel could adjust the power
> management heuristics accordingly, but there were concerns that the
> UMDs would have to maintain application lists to select which
> heuristic worked best for each application.  Maybe it would be better
> to provide a general classification?  E.g., if the GL or vulkan app
> uses these extensions, it's probably a compute type application vs
> something more graphics-y.  The usual trade-off between power and
> performance.  In general, just letting the firmware pick the clock
> based on perf counters generally seems to work the best.  Maybe a
> general workload hint set by the compositor based on the content type
> it's displaying would be a better option (video vs gaming vs desktop)?

Low clocks can be an issue even for normal desktop workloads, so doubtful that this would be a complete solution.


> The deadline stuff doesn't really align well with what we can do with
> our firmware and seems ripe for abuse.  Apps can just ask for high
> clocks all the time which is great for performance, but not great for
> power.

Maybe the firmware power management heuristics could be derived from the system power profile and fence deadline? E.g. the power profile defines the upper and lower boundaries, the upper boundary is used while there's a pending fence deadline, otherwise the lower boundary.


-- 
Earthling Michel Dänzer            |                  https://redhat.com
Libre software enthusiast          |         Mesa and Xwayland developer