Jump to content
AIDA64 Discussion Forum
12749124

Fluctuating GPU clock and power draw

Recommended Posts

Hi,

 

1) I noticed that the graphs of the iGPU's power draw (GT cores) is fluctuating.

Is this supposed to be like this when stressing it, without any other graphic cards in the system?

It looks like this: http://i.imgur.com/Ji72zII.png

 

2) Also, if I enable energysaving (standby/deepstandby) options for the iGPU (RC6 or so), the clock graph looks like this: http://i.imgur.com/AAUKz2I.png

As you can see, the clock is changing between its minimum and maximum value.

If I disable said options it looks like this: http://i.imgur.com/wtcFhFo.png

 

Im kind of confused since even when I enable energysaving options for the CPU, the clock is always a straight line under full load.

 

Could you maybe assure me that all these graphs shown in the pictures are how they actually are supposed to be?

 

Thanks in advance!

 

Regards

Share this post


Link to post
Share on other sites

The graphs look fine and normal. AIDA64 uses OpenCL kernels to stress the GPU (or GPUs). Currently OpenCL is tied to the video driver, and queuing and job prioritization is not really sophisticated in current OpenCL implementations by AMD, Intel and nVIDIA. It will take a few more years for those companies to finally decouple OpenCL from the video driver and make OpenCL a lot more useful and a lot more dependable than now. Because right now, as you can see, passing a new job to the OpenCL implementation takes too much time, and during that time the iGPU may switch power states. This whole issue is a limitation of OpenCL that cannot be fixed from those software that utilizes OpenCL for either computational or stress testing purposes.

Regards,

Fiery

Share this post


Link to post
Share on other sites

The graphs look fine and normal. AIDA64 uses OpenCL kernels to stress the GPU (or GPUs). Currently OpenCL is tied to the video driver, and queuing and job prioritization is not really sophisticated in current OpenCL implementations by AMD, Intel and nVIDIA. It will take a few more years for those companies to finally decouple OpenCL from the video driver and make OpenCL a lot more useful and a lot more dependable than now. Because right now, as you can see, passing a new job to the OpenCL implementation takes too much time, and during that time the iGPU may switch power states. This whole issue is a limitation of OpenCL that cannot be fixed from those software that utilizes OpenCL for either computational or stress testing purposes.

Regards,

Fiery

Hi, thanks for your response!
 
I played around in the BIOS a bit, en- and disabling energysaving options like those in the picture I attached.
However, no matter whether I leave it on or off, the only thing that seems to change while stressing the Intel HD4000 in AIDA is the clockrate being constant when disabling those options,
the power consumptions always stays "shivering".
I already looked in Intel's datasheet but couldnt find anything on how it is supposed to look under full load.
The reason I'm asking is because when stressing CPU and/or FPU only, the CPU IA Cores power consumption is a pretty much straight line, that's why I was assuming it should be the same for the iGPU.
 
Can I still consider my system stable after 24h of FPU + GPU (iGPU in my case) stress, even though I am kind of irritated by this whole thing?
 
Thanks!

post-29959-0-86319500-1424789228_thumb.jpg

post-29959-0-67521700-1424789235_thumb.jpg

post-29959-0-98032600-1424789241_thumb.jpg

Share this post


Link to post
Share on other sites

Can I still consider my system stable after 24h of FPU + GPU (iGPU only in my case) stress, even though I am kind of irritated by this whole thing?

Yes, you can.

Share this post


Link to post
Share on other sites

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.

Guest
Reply to this topic...

×   Pasted as rich text.   Paste as plain text instead

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.

Loading...

  • Similar Content

    • By Tamizharasu
      Aida64 crash on when start benchmark gpu its crashing when it comes to memory copy 
    • By DmitryKo
      Display - GPU tab shows incorrect shader model version 6.3 with AMD WDDM 2.7 drivers, whereas dxcapsviewer reports shader model 6.5.
      Please note you need  to pre-initialize the feature query with the highest SM version to report - so if you ask for SM 6.3, you will never get higher than that. The correct detection algorythm has to cycle through all known shader models, starting from the highest known version and going down until success (or start with 5.1 and loop up to the highest until failure). The latest Windows SDK Preview defines SM 6.6.
       
    • By xixou
      rtx 2080 super, no usage probe? (was present with the gtx 1070)
      TDP probe is ok.
       
    • By DonSalsero
      Hello,I m trying to record my gpu power in Watt (into a csv). In my pc (GTX 970) in File - Preferences - Logging , down to the "Power Values" section there is a choice "GPU1". I tried in an other pc (with Titax X) but in "Power Values" section there is only the "GPU1 TDP%" option. Do you know how can I activate the option GPU1 at "Power Values in my second pc? 
      Thank you for your time  
    • By DonSalsero
      Hello guys,I'm trying to measure the utilization of my GPU.I have 2, an Intel hd graphics (from my cpu) and a GTX 970. From File - Preferences - Logging I can only find "GPU1 Utilization" . I cant understand if GPU1 refers to GTX 970 or Intel hd graphics. (I have enabled "Wake GPUs up at AIDA64 start up")
      Thank you for your time
×
×
  • Create New...