• Competitor rules

    Please remember that any mention of competitors, hinting at competitors or offering to provide details of competitors will result in an account suspension. The full rules can be found under the 'Terms and Rules' link in the bottom right corner of your screen. Just don't mention competitors in any way, shape or form and you'll be OK.

The RX Vega 56 Owners Thread

Soldato
Joined
13 Jun 2009
Posts
6,847
P6: 1667MHz - 1050mV
P7: 1667Mhz - 1050mV
HBM: 950MHz
Power Limit: +50%

3DMark - https://www.3dmark.com/spy/5086826

GPU Score: 7471
That's a great result. Does it actually stay around 1667 MHz or does it drop below this?

I might try pushing mine just to see how much extra performance I can squeeze out of it in the future if it ever starts struggling with newer games.
 
Last edited:
Soldato
Joined
13 Jun 2009
Posts
6,847
Inspired by @thebennyboy, mine seems happy at 1652 MHz @ 1000 mV. Since my "Cool N Quiet" profile uses 1000 mV for P7 anyway I guess I can bump up the clocks.

I don't really understand how P6 and P7 states work. If I set P6 and P7 to 1652 MHz @ 1000 mV, the card sits at 1630 MHz, fine. If I change P6 to 1602 MHz @ 950 mV, it sits at 1580 MHz...I would expect it to be somewhere between the P6 and P7 frequencies, no? Also, power usage doesn't seem to be solely dictated by voltage. I can run 1652 MHz @ 1000 mV and @ 950 mV and get the same performance and power usage, which is weird.
 
Last edited:
Associate
Joined
15 Oct 2010
Posts
1,461
Anyone had any issues with games just straight CTD with no warnings? Just to clarify, never had any issues with other GPUs.
 
Last edited:
Soldato
Joined
13 Jun 2009
Posts
6,847
Anyone had any issues with games just straight CTD with no warnings? Just to clarify, never had any issues with other GPUs.
Hmm not so far, sorry.

Cool N Quiet v2:
P6: 950 mV, 1627 MHz
P7: 1000 mV, 1652 MHz
HBM: 960 MHz
Power Limit: +50%
GPU Score: 7061 (+9.1% compared to stock)
Peak Sustained Core Clock: 1580 MHz
Peak Sustained Power Usage: 165 W

3% more power (+5 W) for 2.5% more performance. As soon as the sustained core clock approaches 1600 MHz (which seems to require the average of P6 & P7 to be 1650+ MHz), power usage shoots up considerably.
 
Associate
Joined
26 Apr 2017
Posts
1,252
Hmm not so far, sorry.

Cool N Quiet v2:
P6: 950 mV, 1627 MHz
P7: 1000 mV, 1652 MHz
HBM: 960 MHz
Power Limit: +50%
GPU Score: 7061 (+9.1% compared to stock)
Peak Sustained Core Clock: 1580 MHz
Peak Sustained Power Usage: 165 W

3% more power (+5 W) for 2.5% more performance. As soon as the sustained core clock approaches 1600 MHz (which seems to require the average of P6 & P7 to be 1650+ MHz), power usage shoots up considerably.

You can check gamer nexus when he remove the power limit for the vega 56 card and went beyond the 2070.
The power draw is not recommended tho :)
 
Associate
Joined
26 Apr 2017
Posts
1,252
Only programmes I have open are Steam, Uplay and Siege (one of the games that keeps crashing with no log or error).

windows have logs for events.
seems the game is likely the cause but you want to check windows logs to exclude your hardware/software as reasons.
 
Soldato
Joined
13 Jun 2009
Posts
6,847
Only programmes I have open are Steam, Uplay and Siege (one of the games that keeps crashing with no log or error).
Do you have monitoring software running like Rivatuner Statistics Server? That can cause such issues, especially if it is out of date. Otherwise, maybe try a DDU and driver reinstall?

You can check gamer nexus when he remove the power limit for the vega 56 card and went beyond the 2070.
The power draw is not recommended tho :)
Yeah I've seen that actually, pretty crazy. Personally I'm not interested in pure performance though, just looking for the best realistic performance with reasonable power consumption and temperatures. It does seem as though the process really holds the Vega architecture back though; it'd probably be a total beast on 7 nm!
 
Associate
Joined
15 Oct 2010
Posts
1,461
Do you have monitoring software running like Rivatuner Statistics Server? That can cause such issues, especially if it is out of date. Otherwise, maybe try a DDU and driver reinstall?


Yeah I've seen that actually, pretty crazy. Personally I'm not interested in pure performance though, just looking for the best realistic performance with reasonable power consumption and temperatures. It does seem as though the process really holds the Vega architecture back though; it'd probably be a total beast on 7 nm!

No, no monitoring software. Seems odd as it only happens on certain games, I ran DOOM and it was fine playing for a solid 20 minutes whereas the other games (Siege, POE, seem to crash within the first couple of minutes.
 
Associate
Joined
25 Feb 2012
Posts
348
Also, I got this yesterday after tweaking using some figures posted in this thread over the last few pages.
Time Spy
(7 564 - Graphics Score:7333,CPU Score:9 213)
https://www.3dmark.com/spy/5094733
Firestrike
(18523 - Graphics Score: 24234, Physics Score: 20050, Combined Score: 6428)
https://www.3dmark.com/fs/17117874

I have 16GB team group 8pack 3200 bdie RAM - with 3333 (FAST Preset Ryzen Dram Calc) low latency timing (recently achieved thanks- AGESA 1.0.0.6). My rzyen 1700 is OCed to 4.0GHz @ 1.4v

So far I have a gaming, a balanced and a coolnquiet profile.

These cards are great - really impressed.
 
Last edited:
Associate
Joined
26 Apr 2017
Posts
1,252
No, no monitoring software. Seems odd as it only happens on certain games, I ran DOOM and it was fine playing for a solid 20 minutes whereas the other games (Siege, POE, seem to crash within the first couple of minutes.

trouble shooting can be difficult.
remove any overclocks on cpu/gpu
try play.
check logs if any weird windows software of hardware error happens a lot.
run memory checks.
update any drivers, chipset, gpu other devices.
check bios version on mboard, if old, get the latest and flash it.

In essence you want to check if its a hardware or a software issue.
 
Associate
Joined
11 Jun 2013
Posts
1,087
Location
Nottingham
I'd start with running a benchmark like Heaven, it basically taxes only the GPU and it's a good way to check that the GPU is running stable, at the right clocks, not overheating etc. And if you are using the settings from the threads here on the forum you can also compare performance against the numbers of others.

Use a benchmark like Firestrike to also tax the CPU.

Once you can run the benchmark(s) without issues for a good amount of time you could try the games.
 
Associate
Joined
15 Oct 2010
Posts
1,461
trouble shooting can be difficult.
remove any overclocks on cpu/gpu
try play.
check logs if any weird windows software of hardware error happens a lot.
run memory checks.
update any drivers, chipset, gpu other devices.
check bios version on mboard, if old, get the latest and flash it.

In essence you want to check if its a hardware or a software issue.

Going through all the checks now to eliminate any hardware issues. Will report back later.

I'd start with running a benchmark like Heaven, it basically taxes only the GPU and it's a good way to check that the GPU is running stable, at the right clocks, not overheating etc. And if you are using the settings from the threads here on the forum you can also compare performance against the numbers of others.

Use a benchmark like Firestrike to also tax the CPU.

Once you can run the benchmark(s) without issues for a good amount of time you could try the games.

Ran Heaven for a little while earlier with some settings from above and the temps and everything looked stable and friendly, no issues with the benchmark in that regards. Will try firestrike next though.
 
Back
Top Bottom