Having 3 issues with my new 7900 GRE

[Teddysurf]

 

So I’m having 3 issues with my new 7900 GRE so if anybody has any insight, I would appreciate it.

Issue 1. If I plug my second monitor into the integrated graphics, port of the motherboard I cannot overclock my GPU? For the past week I’ve been tearing my hair out, trying to figure out why I can’t overclock my GPU at all, as when I do, I get freezers and gray screens. But here’s the catch the freezers and grey screens would only happen when the computer was idling, and not during benchmarks or in games… In addition, when I turn the computer off, the screen will flash white very briefly. I finally plugged both monitors into the GPU and its now over clocks fine.


Issue 2. When I plugged both monitors into the new GPU, the idle wattage is between 70 and 80 Watts…. I did some research and apparently it’s because of the different refresh rates of each of the monitors as one monitor is 240 Hz and the other is 155 Hz. I made a custom profile and went through a bunch of different hz settings and the only one that worked was turning the second monitor to 60 Hz… which is not ideal. It now idles at like 30 W.

Issue 3. The memory temperature and memory Junction temperature idles at 70 to 75C. When in games, the memory temperature and memory Junction temperature never get above 93C and the GPU core itself never gets above 73C. It just seems that the memory and memory junction idle temperatures are really high for some reason, but maybe this is normal? The hotspot temperature idle is about 50C. 

So if anyone has any advice, I’m all ears. Thanks.


MBD: Asus X670E-A
CPU: Ryzen 7900X cpu
GPU Sapphire 7900 GRE Pure
SSD: 2 x WD_SN850X
PSU: Corsair RM850x Shift 850w 
RAM: 32GB Gskill

 

 

 

[Tetras]

6 minutes ago, Teddysurf said:

Issue 3. The memory temperature and memory Junction temperature idles at 70 to 75C. When in games, the memory temperature and memory Junction temperature never get above 93C and the GPU core itself never gets above 73C. It just seems that the memory and memory junction idle temperatures are really high for some reason, but maybe this is normal? The hotspot temperature idle is about 50C. 

I think you'll find that the multimonitor idle is partly due to the memory clocks being so high, so that's likely why their idle temperatures are like that. You might be able to address it by increasing the fan curve, since I assume zero-rpm is not operating if you're using 70-80 watts.

[Teddysurf]

18 minutes ago, Tetras said:

I think you'll find that the multimonitor idle is partly due to the memory clocks being so high, so that's likely why their idle temperatures are like that. You might be able to address it by increasing the fan curve, since I assume zero-rpm is not operating if you're using 70-80 watts.

First, thanks for the reply. I use a software called fancontrol (you might’ve seen it Jay’s two centz did a video on the software) to control all of the fans on my system, including the GPU. The software let you set all the system fans to ramp up to whatever the hottest component is, on my GPU I have the fans to ramp up to the gpus memory temperature.  I actually just turned the computer on while I’m writing this and the temperature is now at 64C with the second monitor set to 60 Hz. Yesterday it was idling at 70c with the second monitor said at 155 hz. I don’t know, it just seems like 70 W for idle when both monitors are set to spec is really high.

[podkall]

11 minutes ago, Tetras said:

I think you'll find that the multimonitor idle is partly due to the memory clocks being so high, so that's likely why their idle temperatures are like that. You might be able to address it by increasing the fan curve, since I assume zero-rpm is not operating if you're using 70-80 watts.

oh my card would gladly idle at 65-70W and have 0RPM as long as the temps allow it,

 

18 minutes ago, Teddysurf said:

Issue 1. If I plug my second monitor into the integrated graphics, port of the motherboard I cannot overclock my GPU? For the past week I’ve been tearing my hair out, trying to figure out why I can’t overclock my GPU at all, as when I do, I get freezers and gray screens. But here’s the catch the freezers and grey screens would only happen when the computer was idling, and not during benchmarks or in games… In addition, when I turn the computer off, the screen will flash white very briefly. I finally plugged both monitors into the GPU and its now over clocks fine.

don't OC the card,

 

modern cards "OC" themself, they will run as fast as the constraint voltage and temperatures allow them to,

 

20 minutes ago, Teddysurf said:

Issue 2. When I plugged both monitors into the new GPU, the idle wattage is between 70 and 80 Watts…. I did some research and apparently it’s because of the different refresh rates of each of the monitors as one monitor is 240 Hz and the other is 155 Hz. I made a custom profile and went through a bunch of different hz settings and the only one that worked was turning the second monitor to 60 Hz… which is not ideal. It now idles at like 30 W.

7900 GRE is high end gaming card, or near high end,

 

it's power is 260W TDP, and it's essentially 7900 XTX's chip that has certain amount of things disabled to downgrade to GRE (no you can't enable them)

 

the GPU is simply ready for task, also desktop runs at same framerate as your monitor's Hz, that's why higher Hz will make the card use more power, it's more fps on the desktop on higher Hz,

 

22 minutes ago, Teddysurf said:

Issue 3. The memory temperature and memory Junction temperature idles at 70 to 75C. When in games, the memory temperature and memory Junction temperature never get above 93C and the GPU core itself never gets above 73C. It just seems that the memory and memory junction idle temperatures are really high for some reason, but maybe this is normal? The hotspot temperature idle is about 50C. 

RDNA chips are pretty warm, not sure what their supposed idles should be,

 

but AMD cards and even some Nvidia were specifically designed to run near the TJunction of the temperature (max),

 

also cooler design matters,

 

and weather, if summer means your ambient in your room rise, that also means few temps rise in the case too.

 

5 minutes ago, Teddysurf said:

First, thanks for the reply. I use a software called fancontrol (you might’ve seen it Jay’s two centz did a video on the software) to control all of the fans on my system, including the GPU. The software let you set all the system fans to ramp up to whatever the hottest component is, on my GPU I have the fans to ramp up to the gpus memory temperature.  I actually just turned the computer on while I’m writing this and the temperature is now at 56C with the second monitor set to 60 Hz. Yesterday it was idling at 70c with the second monitor said at 155 hz. I don’t know, it just seems like 70 W for idle when both monitors are set to spec is really high.

those temps are pretty high, is that GPU new or like second-hand or?

 

these temps are bit high, but at the same time you're using a high end card, so unless the card is reaching critical temps in games or throttles so much that it's performance isn't what it should be it should be okay.

[podkall]

27 minutes ago, Teddysurf said:

MBD: Asus X670E-A
CPU: Ryzen 7900X cpu
GPU Sapphire 7900 GRE Pure
SSD: 2 x WD_SN850X
PSU: Corsair RM850x Shift 850w 
RAM: 32GB Gskill

what case?

[Teddysurf]

4 minutes ago, podkall said:

what case?

It’s the NZXT H510 flow.  I have 180 mm AIO in front of the GPU and 4 140mm fans in a push pull configuration to get more air in the case.

 

 

[Teddysurf]

18 minutes ago, podkall said:

oh my card would gladly idle at 65-70W and have 0RPM as long as the temps allow it,

 

don't OC the card,

 

modern cards "OC" themself, they will run as fast as the constraint voltage and temperatures allow them to,

 

7900 GRE is high end gaming card, or near high end,

 

it's power is 260W TDP, and it's essentially 7900 XTX's chip that has certain amount of things disabled to downgrade to GRE (no you can't enable them)

 

the GPU is simply ready for task, also desktop runs at same framerate as your monitor's Hz, that's why higher Hz will make the card use more power, it's more fps on the desktop on higher Hz,

 

RDNA chips are pretty warm, not sure what their supposed idles should be,

 

but AMD cards and even some Nvidia were specifically designed to run near the TJunction of the temperature (max),

 

also cooler design matters,

 

and weather, if summer means your ambient in your room rise, that also means few temps rise in the case too.

 

those temps are pretty high, is that GPU new or like second-hand or?

 

these temps are bit high, but at the same time you're using a high end card, so unless the card is reaching critical temps in games or throttles so much that it's performance isn't what it should be it should be okay.

I purchased the card from newegg new about a week and a half ago.

 

with all due respect if the GPU board is idling at 60 to 70 W with two monitors that’s just kind of ridiculous. I did some research and it’s a well known bug with AMD GPU’s. The problem is none of the fixes that I found on YouTube helped with the problem other than running the second second monitor at 60 Hz.

 

anyway, thank you for the advice.

[ki8aras]

38 minutes ago, Tetras said:

I think you'll find that the multimonitor idle is partly due to the memory clocks being so high, so that's likely why their idle temperatures are like that. You might be able to address it by increasing the fan curve, since I assume zero-rpm is not operating if you're using 70-80 watts.

i had a similar issue with my 7900xtx, zero rpm works but every like 20 mins or sth the fans audibly spool up

[Teddysurf]

3 minutes ago, ki8aras said:

i had a similar issue with my 7900xtx, zero rpm works but every like 20 mins or sth the fans audibly spool up

Thanks for the reply, but that’s not the problem because I have a software called fancontrol that’s running the GPU fans and every other fan in my system and they all ramp up to whatever the hottest component is. Right now I have it sent 20% at idle and the memory temperature is about 64° C.  When both monitors are at full refresh rate it goes up to about 70 C and the fans ramp up… it’s just irritating lol.  I mean nothing is thermal throttling or getting dangerous. It just seems high for idle. 

[podkall]

13 minutes ago, ki8aras said:

i had a similar issue with my 7900xtx, zero rpm works but every like 20 mins or sth the fans audibly spool up

fans will always spool up on 0rpm, because the card that runs in the PC can't stay on the same temp forever, eventually the temp will rise and the fans will spin according to their fan profile

[ki8aras]

2 minutes ago, podkall said:

fans will always spool up on 0rpm, because the card that runs in the PC can't stay on the same temp forever, eventually the temp will rise and the fans will spin according to their fan profile

idk, im pretty sure since i fixed the 100w idle i had i never heard them spool up, the pulse cooler can probably dissipate sth like 30w passively

[podkall]

8 minutes ago, ki8aras said:

idk, im pretty sure since i fixed the 100w idle i had i never heard them spool up, the pulse cooler can probably dissipate sth like 30w passively

my GPU idles on 65W, but it's an older 10 series card on a 2 fan cooler

[191x7]

42 minutes ago, Teddysurf said:

I purchased the card from newegg new about a week and a half ago.

 

with all due respect if the GPU board is idling at 60 to 70 W with two monitors that’s just kind of ridiculous. I did some research and it’s a well known bug with AMD GPU’s. The problem is none of the fixes that I found on YouTube helped with the problem other than running the second second monitor at 60 Hz.

 

anyway, thank you for the advice.

In AMD Adrenalin go to settings Displays and create a custom resolution with CVT Reduced Blanking on the second monitor. You might need to use 120 or 144 Hz.

[Teddysurf]

1 minute ago, 191x7 said:

In AMD Adrenalin go to settings Displays and create a custom resolution with CVT Reduced Blanking on the second monitor. You might need to use 120 or 144 Hz.

In my original post, I said that I already did that. I tried so many different refresh rates, including 100 Hz, 120 Hz, 142 Hz, and a few others and the only one that worked to bring down the idle wattage is 60 Hz.  

[191x7]

5 hours ago, Teddysurf said:

In my original post, I said that I already did that. I tried so many different refresh rates, including 100 Hz, 120 Hz, 142 Hz, and a few others and the only one that worked to bring down the idle wattage is 60 Hz.  

Have you tried not just adjusting the secondary but also the main?

[Teddysurf]

4 hours ago, 191x7 said:

Have you tried not just adjusting the secondary but also the main?

I have not tried it, but I did think about it and my reasoning was how many frames am I going to take off my 240 Hz monitor when I literally purchased a graphics card to get as many frames as I could… I don’t know, if I need to resort to cutting down frames my expensive main monitor I’m just gonna return the card.  Thanks for the suggestion though.

[191x7]

1 hour ago, Teddysurf said:

I have not tried it, but I did think about it and my reasoning was how many frames am I going to take off my 240 Hz monitor when I literally purchased a graphics card to get as many frames as I could… I don’t know, if I need to resort to cutting down frames my expensive main monitor I’m just gonna return the card.  Thanks for the suggestion though.

Not talking about reducing the refresh, talking about the blanking. 

[Teddysurf]

3 hours ago, 191x7 said:

Not talking about reducing the refresh, talking about the blanking. 

My apologies I went back and reread your first comment and missed the word “blanking” in the initial response. I’ve actually never heard of blanking. I’ll have to look it up, but thank you for the suggestion I will report back once I take your advice.

[191x7]

11 minutes ago, Teddysurf said:

My apologies I went back and reread your first comment and missed the word “blanking” in the initial response. I’ve actually never heard of blanking. I’ll have to look it up, but thank you for the suggestion I will report back once I take your advice.

Yeah, when you create a custom resolution in AMD Adrenalin you can change multiple options, not just the refresh.

The issue isn't on AMD's side, it's the monitor manufacturer using improper timings for a blanking type or vice versa.

AMD can't test every possible monitor on the market in all possible multi-monitor situations.

[Teddysurf]

14 hours ago, 191x7 said:

Yeah, when you create a custom resolution in AMD Adrenalin you can change multiple options, not just the refresh.

The issue isn't on AMD's side, it's the monitor manufacturer using improper timings for a blanking type or vice versa.

AMD can't test every possible monitor on the market in all possible multi-monitor situations.

I realize AMD can’t test every monitor combination, but on this particular card, I can’t use the integrated GPU and overclock the card for whatever reason?  Unfortunately the custom resolutions and reducing the blanking didn’t fix the issue. And yes, I tried a bunch of different refresh rates. In troubleshooting, I even realized I didn’t have the second monitor set to AMDs variable refresh rate, I was hoping that might be the issue, but turned out It made no difference. The problem seems to be the Vram doesn’t downclock when two monitors are hooked up to the GPU.  Anyway, thanks for trying to help. 

[Teddysurf]

OK, so I found a solution if anybody else is experiencing issues to at least one of the problems. I found a solution to the problem of I can’t use the integrated graphics and overclocked my GPU.  I found a video on YouTube that I actually paid attention to all the way through this time which said if you clock the VRAM above 2300-2350MHz the card is not stable at low power states like idle.  When I drop the voltage down to around 2300 MHz the GPU frequency is now able to be clocked all the way up to 2803 MHz stable.  In my mind, this has to be a bug as when the card originally came out, you couldn’t overclocked the GRE past 2300 MHz and AMD called it a bug. Now you can overclock it to over 2300 MHz but it’s not stable at idle clocks.
 

I have not found a solution to the memory clocking to full speed when multiple monitors are plugged into the GPU as none of the solutions I found online worked for me.

 

I think I’m going to buy better thermal pads and repad the GPU for the memory temperature issues.