From: Andres Rodriguez <andresx7-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org>
To: Oded Gabbay <oded.gabbay-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org>
Cc: Alex Deucher <Alexander.Deucher-5C7GfCeVMHo@public.gmane.org>,
jay-gJmSnxjMpeIFV7jr3Ov9Ew@public.gmane.org,
Felix Kuehling <felix.kuehling-5C7GfCeVMHo@public.gmane.org>,
amd-gfx-PD4FTy7X32lNgt0PjOBp9y5qC8QIuHrW@public.gmane.org,
John Bridgman <John.Bridgman-5C7GfCeVMHo@public.gmane.org>
Subject: Re: Change queue/pipe split between amdkfd and amdgpu
Date: Wed, 8 Feb 2017 14:47:35 -0500 [thread overview]
Message-ID: <50aecc56-f080-d343-9e49-e3955ba1212e@gmail.com> (raw)
In-Reply-To: <CAFCwf10BboWSwU9HAMjryuLw2K2ANpjC8hGgGwkJh6z8K3pR4Q-JsoAwUIsXosN+BqQ9rBEUg@public.gmane.org>
Thank you Oded.
- Andres
On 2017-02-08 02:32 PM, Oded Gabbay wrote:
> On Wed, Feb 8, 2017 at 6:23 PM, Andres Rodriguez <andresx7@gmail.com> wrote:
>> Hey Felix,
>>
>> Thanks for the pointer to the ROCm mqd commit. I like that the workarounds
>> are easy to spot. I'll add that to a new patch series I'm working on for
>> some bug-fixes for perf being lower on pipes other than pipe 0.
>>
>> I haven't tested this yet on kaveri/carrizo. I'm hoping someone with the HW
>> will be able to give it a go. I put in a few small hacks to get KFD to boot
>> but do nothing on polaris10.
>>
>> Regards,
>> Andres
>>
>>
>> On 2017-02-06 03:20 PM, Felix Kuehling wrote:
>>> Hi Andres,
>>>
>>> Thank you for tackling this task. It's more involved than I expected,
>>> mostly because I didn't have much awareness of the MQD management in
>>> amdgpu.
>>>
>>> I made one comment in a separate message about the unified MQD commit
>>> function, if you want to bring that more in line with our latest ROCm
>>> release on github.
>>>
>>> Also, were you able to test the upstream KFD with your changes on a
>>> Kaveri or Carrizo?
>>>
>>> Regards,
>>> Felix
>>>
>>>
>>> On 17-02-03 11:51 PM, Andres Rodriguez wrote:
>>>> The current queue/pipe split policy is for amdgpu to take the first pipe
>>>> of
>>>> MEC0 and leave the rest for amdkfd to use. This policy is taken as an
>>>> assumption in a few areas of the implementation.
>>>>
>>>> This patch series aims to allow for flexible/tunable queue/pipe split
>>>> policies
>>>> between kgd and kfd. It also updates the queue/pipe split policy to one
>>>> that
>>>> allows better compute app concurrency for both drivers.
>>>>
>>>> In the process some duplicate code and hardcoded constants were removed.
>>>>
>>>> Any suggestions or feedback on improvements welcome.
>>>>
>> _______________________________________________
>> amd-gfx mailing list
>> amd-gfx@lists.freedesktop.org
>> https://lists.freedesktop.org/mailman/listinfo/amd-gfx
> Hi Andres,
> I will try to find sometime to test it on my Kaveri machine.
>
> Oded
_______________________________________________
amd-gfx mailing list
amd-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/amd-gfx
next prev parent reply other threads:[~2017-02-08 19:47 UTC|newest]
Thread overview: 31+ messages / expand[flat|nested] mbox.gz Atom feed top
2017-02-04 4:51 Change queue/pipe split between amdkfd and amdgpu Andres Rodriguez
[not found] ` <20170204045142.5596-1-andresx7-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org>
2017-02-04 4:51 ` [PATCH 01/13] drm/amdgpu: refactor MQD/HQD initialization Andres Rodriguez
2017-02-04 4:51 ` [PATCH 02/13] drm/amdgpu: doorbell registers need only be set once Andres Rodriguez
[not found] ` <20170204045142.5596-3-andresx7-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org>
2017-02-06 8:31 ` Christian König
[not found] ` <7d85c562-5227-1fef-7b99-1f7543e6e69b-ANTagKRnAhcb1SvskN2V4Q@public.gmane.org>
2017-02-08 16:25 ` Andres Rodriguez
2017-02-04 4:51 ` [PATCH 03/13] drm/amdgpu: detect timeout error when deactivating hqd Andres Rodriguez
2017-02-04 4:51 ` [PATCH 04/13] drm/amdgpu: remove duplicate definition of cik_mqd Andres Rodriguez
2017-02-04 4:51 ` [PATCH 05/13] drm/amdgpu: unify MQD programming sequence for kfd and amdgpu Andres Rodriguez
[not found] ` <20170204045142.5596-6-andresx7-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org>
2017-02-06 20:16 ` Felix Kuehling
2017-02-04 4:51 ` [PATCH 06/13] drm/amdgpu: rename rdev to adev Andres Rodriguez
2017-02-04 4:51 ` [PATCH 07/13] drm/amdgpu: take ownership of per-pipe configuration Andres Rodriguez
2017-02-04 4:51 ` [PATCH 08/13] drm/radeon: take ownership of pipe initialization Andres Rodriguez
2017-02-04 4:51 ` [PATCH 09/13] drm/amdgpu: allow split of queues with kfd at queue granularity Andres Rodriguez
2017-02-04 4:51 ` [PATCH 10/13] drm/amdkfd: allow split HQD split on per-queue granularity Andres Rodriguez
2017-02-04 4:51 ` [PATCH 11/13] drm/amdgpu: remove duplicate magic constants from amdgpu_amdkfd_gfx*.c Andres Rodriguez
2017-02-04 4:51 ` [PATCH 12/13] drm/amdgpu: allocate queues horizontally across pipes Andres Rodriguez
2017-02-04 4:51 ` [PATCH 13/13] drm/amdgpu: new queue policy, take first 2 queues of each pipe Andres Rodriguez
[not found] ` <20170204045142.5596-14-andresx7-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org>
2017-02-04 12:08 ` Edward O'Callaghan
[not found] ` <86138a88-e90f-3234-7109-67ca0c427071-dczkZgxz+BNUPWh3PAxdjQ@public.gmane.org>
2017-02-06 8:35 ` Christian König
2017-02-06 20:20 ` Change queue/pipe split between amdkfd and amdgpu Felix Kuehling
[not found] ` <206b40fe-b958-8a78-623b-011712dc5ecc-5C7GfCeVMHo@public.gmane.org>
2017-02-08 16:23 ` Andres Rodriguez
[not found] ` <852b8cfe-d886-e78e-de6c-1641b107ed8f-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org>
2017-02-08 19:32 ` Oded Gabbay
[not found] ` <CAFCwf10BboWSwU9HAMjryuLw2K2ANpjC8hGgGwkJh6z8K3pR4Q-JsoAwUIsXosN+BqQ9rBEUg@public.gmane.org>
2017-02-08 19:47 ` Andres Rodriguez [this message]
[not found] ` <50aecc56-f080-d343-9e49-e3955ba1212e-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org>
2017-02-09 20:11 ` Oded Gabbay
[not found] ` <CAFCwf10L8=mwxCup6-S5Yirxit8MJEZR=rhLPF3NVnotGCSYiQ-JsoAwUIsXosN+BqQ9rBEUg@public.gmane.org>
2017-02-09 20:19 ` Andres Rodriguez
2017-02-09 20:38 ` Andres Rodriguez
[not found] ` <8e51b688-d978-d40f-8aa3-ae1090ab6a03-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org>
2017-02-10 10:35 ` Oded Gabbay
[not found] ` <CAFCwf10agJ+C0X-jvPa5jSJcR--+u3-TC6LuVQwW1+o94uGonA-JsoAwUIsXosN+BqQ9rBEUg@public.gmane.org>
2017-02-10 16:33 ` Andres Rodriguez
[not found] ` <37bd1bf7-4db6-6004-f61d-5084efd1ec83-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org>
2017-02-10 17:56 ` Oded Gabbay
[not found] ` <CAFCwf125SHM52z2UAp_Y0rRQe9UHpeWMDDpBkf1csPJyAgXTeA-JsoAwUIsXosN+BqQ9rBEUg@public.gmane.org>
2017-02-16 4:00 ` Bridgman, John
[not found] ` <BN6PR12MB1348C4F31B81516D5EAAFCA3E85A0-/b2+HYfkarQX0pEhCR5T8QdYzm3356FpvxpqHgZTriW3zl9H0oFU5g@public.gmane.org>
2017-02-16 7:14 ` Edward O'Callaghan
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=50aecc56-f080-d343-9e49-e3955ba1212e@gmail.com \
--to=andresx7-re5jqeeqqe8avxtiumwx3w@public.gmane.org \
--cc=Alexander.Deucher-5C7GfCeVMHo@public.gmane.org \
--cc=John.Bridgman-5C7GfCeVMHo@public.gmane.org \
--cc=amd-gfx-PD4FTy7X32lNgt0PjOBp9y5qC8QIuHrW@public.gmane.org \
--cc=felix.kuehling-5C7GfCeVMHo@public.gmane.org \
--cc=jay-gJmSnxjMpeIFV7jr3Ov9Ew@public.gmane.org \
--cc=oded.gabbay-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox