Re: About new backend for GPU compute ROCm in qemu

["Huang, Honglei" <honghuan@amd.com>]

On 9/7/2026 9:12 PM, Dmitry Osipenko wrote:
> On 8/20/26 13:49, Huang, Honglei wrote:
...
>>>
>>> These are great news. I suggest updating the MR description then so
>>> that potential reviewers can see the progress.
>>
>> Will update the virtio userptr and virglrenderer PR in future. But
>> unclear how compute will eventually evolve on virtio. The virgl version
>> may only support old ROCm version.
>>
>> And a single control queue is the design of the virtio GPU, and it is
>> indeed a bottleneck, modifying it is difficult. However, some other
>> virtio devices have multi queues for different usages.
>>
>> And actually such as NPUs, TPUs, or custom ASICs, no display compute
>> card all go to the virglrenderer seem like not a natural long term
>> abstraction.
> 
> It's not clear to me if you're still wanting to have current virtio-gpu
> ROCm context version landed upstream, or you now looking for a better
> alternative?
> 
> Virtio-gpu multi-queue support was in discussions over past years,
> primary driven by ChromeOS+crosvm needs. Yes, it's not a trivial task to
> support multi-queue, no progress was made on it AFAIK.
> 

Yes, we are still aiming to land the upstream ROCm native context on 
virtio-gpu.

And please take the large buffer mapping into design consider for multi 
queue support, we have a performance bottleneck here.


Regards,
Honglei