Print 

Author Topic: GPU  (Read 6204 times)

Offline BFM_Octane

  • BFM Admin
  • *
  • Posts: 3813
GPU
« on: February 03, 2021, 12:57:11 PM »
Got desperate for a quick upgrade... (and at the right time, parts went up a few lately, that GPU seems to have gone up £100 everywhere since i bought mine)

Unknown 600w PSU
Asus A320MK motherboard
Ryzen3 3100
16Gig Corsair Vengeance
RX570 (8gb) ROG STRIX

Are these utilization spikes on idle GPU a normal thing?


Have also been getting random restarts and one or two total shutdowns lately while playing the new Resident Evil 2 and 3, during bigger battles. I have thrown another fan in at the bottom facing up to the GPU fans, just shoved in at a slight angle, no space to mount  :LOL: See what happens...

Another thing possibly to point out that maybe a cooling issue? The case is one of those fancy 3 intake fans at the front jobs, with a tempered glass sidecase. So, maybe im worrying too much about ventilation on that side... Seems like everyones got one these days.

But yeah, main thing is just wondering what might the cause be for it spiking so much on idle? Or is that normal?



Offline Taipan™

  • Posts Too Much
  • *****
  • Posts: 1297
Re: GPU
« Reply #1 on: February 03, 2021, 07:59:47 PM »
A lot of programs use GPU acceleration because of the added benefit of parallel processing so I wouldn't be too concerned about the spikes in utilisation.

Have you tried running a benchmark like 3Dmark and seeing what temperature your CPU and GPU get too? That would eliminate if it is a cooling issue or not. For a RX 570 and Ryzen 3100 they should sit ideally under 85C (185F). Though if your GPU is currently sitting at idle at 28C I doubt you have a cooling issue.

Offline BFM_Octane

  • BFM Admin
  • *
  • Posts: 3813
Re: GPU
« Reply #2 on: February 04, 2021, 03:19:41 AM »
Thanks Taipan

Started following JayzTwoCents on YT and learnt that one since posting ; Chrome and such using hardware acceleration etc. Just wanted to make sure that was what was happening i guess.

Im downloading 3dmark now.  Hoping it maybe logs stuff when the computer does end up restarting/shutting down.



Offline Taipan™

  • Posts Too Much
  • *****
  • Posts: 1297
Re: GPU
« Reply #3 on: February 04, 2021, 03:44:36 AM »
Yeah he is a good channel to follow. He is right into LN2 cooling world records at the moment haha.

Not even just chrome but even windows processes these days will use them. That is because a lot of CPU's have integrated graphics so it would be poor programming not to utilise those resources.

Gone are the days that motherboards used to have their own integrated graphics.

Offline BFM_Octane

  • BFM Admin
  • *
  • Posts: 3813
Re: GPU
« Reply #4 on: February 13, 2021, 07:18:29 AM »
Well im not really sure i got down to the exact reason.

But I got myself MSI afterburner and the manual control for the GPU fans on that is obviously better at keeping the tempurature down, well below 70 if i just leave it on full whack while i play. But not sure if thats detrimental tbh. No more restarts either way.

But before MSI Afterburner i recall tempuratures of 85 (possibly 95 even) before these restarts would occur. So, me being unsure of what temps a GPU will fail/restart/throttle etc, im left now wondering if either the automatic system wasnt kicking the fans in adequately enough to begin with, or it was just a pot luck case of the house heating was up too much and it wasnt making a difference because of that :LOL:

I guess time will tell at this point. Atleast ive got the restarts prevented.

Ive got a good few 12mm fans knocking about aswell so might do some experimenting with stacking them.
« Last Edit: February 13, 2021, 07:32:57 AM by BFM_Octane »



Offline Taipan™

  • Posts Too Much
  • *****
  • Posts: 1297
Re: GPU
« Reply #5 on: February 18, 2021, 07:48:47 PM »
Yeah anything over 85C and you will probably start getting throttling (particularly at 95C)

The fan at full speed is a potential, though annoying solution if it the fans are really loud. I guess if you are getting 95C then the cause is one of 3 things.

1. Your case fans are inadequate and heat is building in the case. Make sure you have enough fans sucking fresh air in and exhausting. To test if this is the reason for the overheat you can put a thermometer in the case and see if it gets hot in there under normal fan speed, then ramp up fans to see if it makes much difference.

2. The GPU cooling block has unseated or the thermal paste gone bad. Could remove cooling block, reapply thermal paste and pads, then reseat to see if it helps

3. The current GPU cooling was inadequate to begin with. You can purchase an after market cooler for your GPU, some thermal paste and redo the cooling on the card. My old HD5970 the factory cooling was awful so I bought a 3 fan cooling block. Halved the temps and noise. My current 1080Ti I made sure to get a 3 fan version.

Print