[Intel-gfx] [igt-dev] [RFC i-g-t 1/1] intel-gpu-top: Support for client stats

Tvrtko Ursulin tvrtko.ursulin at linux.intel.com
Fri Oct 25 15:38:55 UTC 2019


On 25/10/2019 16:13, Chris Wilson wrote:
> Quoting Tvrtko Ursulin (2019-10-25 15:24:10)
>> From: Tvrtko Ursulin <tvrtko.ursulin at intel.com>
>>
>> Adds support for per-client engine busyness stats i915 exports in sysfs
>> and produces output like the below:
>>
>> ==========================================================================
>> intel-gpu-top -  935/ 935 MHz;    0% RC6; 14.73 Watts;     1097 irqs/s
> 
> Could we get "gpu / pkg Watts" pretty please?

Sure, next week or so.

> Are irq/s interesting with execlists? Originally the idea was to say how
> many times clients were sleeping and being woken up. Now we interrupt
> to wipe the gpu's nose when it sneezes.
> 
>>
>>        IMC reads:     1401 MiB/s
>>       IMC writes:        4 MiB/s
>>
>>            ENGINE      BUSY                                 MI_SEMA MI_WAIT
>>       Render/3D/0   63.73% |███████████████████           |      3%      0%
>>         Blitter/0    9.53% |██▊                           |      6%      0%
>>           Video/0   39.32% |███████████▊                  |     16%      0%
>>           Video/1   15.62% |████▋                         |      0%      0%
>>    VideoEnhance/0    0.00% |                              |      0%      0%
>>
>>    PID            NAME     RCS          BCS          VCS         VECS
>>   4084        gem_wsim |█████▌     ||█          ||           ||           |
>>   4086        gem_wsim |█▌         ||           ||███        ||           |
>> ==========================================================================
>>
>> Apart from the existing physical engine utilization it now also shows
>> utilization per client and per engine class.
>>
>> Signed-off-by: Tvrtko Ursulin <tvrtko.ursulin at intel.com>
>> ---
> 
>> +#define SYSFS_CLIENTS "/sys/class/drm/card0/clients"
> 
> We need to somehow pull the right card.

Yeah, as I said RFC and reference only. :)

> Nothing shocking here. Where's the intel-gpu-overlay integration? ;)

Maybe intel-gpu-overlay should become an output plugin for intel_gpu_top. :)

Regards,

Tvrtko


More information about the Intel-gfx mailing list