• 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 64 Owners Thread

Took my rx vega 64 windforce back today & store replaced with a sapphire nitro+! The difference in cards is unbelieveable. I was getting 62fps on GTAv with the windforce, im now getting 90+fps with the nitro+ in 4k with settings on very high. I honestly think i had a faulty windforce card.

I wouldn't be surprised, It's always best to research the cards before buying, The Gigabyte is inferior to the other three imo, You can find review vid's on Youtube comparing them that come to the same opinion. Gigabyte is one to avoid for AMD cards, & Asus imo. Play it safe & always go for an AMD only brand like Sapphire or Powercolor, that's what I do now.
 
I wouldn't be surprised, It's always best to research the cards before buying, The Gigabyte is inferior to the other three imo, You can find review vid's on Youtube comparing them that come to the same opinion. Gigabyte is one to avoid for AMD cards, & Asus imo. Play it safe & always go for an AMD only brand like Sapphire or Powercolor, that's what I do now.

Could not agree more ;)
 
Anyone got any good ways to test for proper stability. I've got an undervolt/overclock on my Vega and it will run fine in most games for hours. Then for what ever reason I run the same game another day for a couple hours and I get a driver crash and bsod. Doesn't seem to matter what game it is or if I have chill on or not. It just randomly decides to crash when it's been fine for hours.
 
Im a bit gutted that my 2x 6+2 pin version of the nitro+ doesnt have led lights around the fans. Only have the sapphire logo on the side & nitro logo on top that light up. Would have looked nice with the fans lit.
 
@Illuminist Just to let you know I got Shadow of the Tomb Raider working with 2 cards in DX12.

Had to completely disable Crossfire fully, not just in the specific game profile. It then was running at around 60 FPS with most settings Max. It does look strange though as monitoring tools show 1 card as inactive, yet there is no way other single Vega could be running the game like that so I know the Async compute is working.
That's weird for me I just switch the game to dx 12 after deleting and disabling the crossfire profile But hey if that works for you and both cards tach lights are pumping then it's all good
 
Anyone got any good ways to test for proper stability. I've got an undervolt/overclock on my Vega and it will run fine in most games for hours. Then for what ever reason I run the same game another day for a couple hours and I get a driver crash and bsod. Doesn't seem to matter what game it is or if I have chill on or not. It just randomly decides to crash when it's been fine for hours.

Hi, I don't use Chill but I've found that the Relive Instant replay feature can cause instability when overclocked in some games, A few weeks back I booted up Dead Island Riptide Definitive Edition for the first time and was treated to a slideshow, I eventually worked out that Instant Replay was the cause and since then I've tested it with other games that had what I thought were occasional issues with my 24/7 overclock, For games like Strange Brigade & Civ 6 I was previously creating individual overclock profiles because like you one day they ran fine all day & then the next they'd crash again & again with the same clock speeds, At the time I had instant replay set up to save it's max of the last twenty minutes of gameplay whenever I told it to, I've since found that reducing the length to just a few minutes or turning it off altogether have solved what I thought were overclock issues.

I'm still using the 18.12.1 drivers but these are my 100% stable clocks with the Vega 64 Red Devil for anyone that's interested.

meYIuN1.png
 
That's weird for me I just switch the game to dx 12 after deleting and disabling the crossfire profile But hey if that works for you and both cards tach lights are pumping then it's all good

The game crashes for me unless I completely disable crossfire with the pop up message "dxgi_error_device_hung".
 
That's weird for me I just switch the game to dx 12 after deleting and disabling the crossfire profile But hey if that works for you and both cards tach lights are pumping then it's all good

The game crashes for me unless I completely disable crossfire with the pop up message "dxgi_error_device_hung".

I think it is definitely the game as it also causes problems on NVidia SLI.
 
Having a few probs with Firestrike extreme crashing , potentially on stock settings.

Ryzen 2700x , mem 3200 xmp at cl16, asrock taichi x470, Vega 64 LC now with EK block. Superflower 1kw PSU, Vega power connectors on separate rails.

Pbo is on and a voltage offset of -0.06250, otherwise nothing else changed.

Temps are good, hotspot is good etc and games etc seem stable. Thought i'd start with undervolting the Vega.

Timespy and gaming seemed fine.

But Firestrike Extreme keeps crashing on the Physics test. So put everything back to standard vega profile, but still lots of crashes.

Perhaps its not Vega settings as its probably CPU / Mem on the physics tests .. but see further below ...? . I've had no issues with gaming , BFV etc very stable once 1st patches out. Moved back from 18.12.3 to 18.12.2 but still same.

Relive disabled, free sync disabled, chill disabled.

Heres the odd bit

Timespy / Timespy extreme fine on Balanced and various undervolt settings. I did get an occasional issue so intended to start again with the UV and keep HBM at stock whilst slowly found stable settings. Played BFV for a few hours without issue ( i know its not the best stability test )

Tried Firestrike extreme , crashed on undervolt on physics test.

So put Vega back to balanced. Still crashed.

Put to power save , worked fine.

Put back to balanced after power save , worked fine.

Put to turbo, worked fine.

Restarted, put back to balanced, crashed.

Put to custom, kept stock settings bar Power Supply at either 25% or 50%, worked fine.

Any ideas, or am i barking up the wrong tree ?

ewY0gyX.jpg.png
 
Having a few probs with Firestrike extreme crashing , potentially on stock settings.

Ryzen 2700x , mem 3200 xmp at cl16, asrock taichi x470, Vega 64 LC now with EK block. Superflower 1kw PSU, Vega power connectors on separate rails.

Pbo is on and a voltage offset of -0.06250, otherwise nothing else changed.

Temps are good, hotspot is good etc and games etc seem stable. Thought i'd start with undervolting the Vega.

Timespy and gaming seemed fine.

But Firestrike Extreme keeps crashing on the Physics test. So put everything back to standard vega profile, but still lots of crashes.

Perhaps its not Vega settings as its probably CPU / Mem on the physics tests .. but see further below ...? . I've had no issues with gaming , BFV etc very stable once 1st patches out. Moved back from 18.12.3 to 18.12.2 but still same.

Relive disabled, free sync disabled, chill disabled.

Heres the odd bit

Timespy / Timespy extreme fine on Balanced and various undervolt settings. I did get an occasional issue so intended to start again with the UV and keep HBM at stock whilst slowly found stable settings. Played BFV for a few hours without issue ( i know its not the best stability test )

Tried Firestrike extreme , crashed on undervolt on physics test.

So put Vega back to balanced. Still crashed.

Put to power save , worked fine.

Put back to balanced after power save , worked fine.

Put to turbo, worked fine.

Restarted, put back to balanced, crashed.

Put to custom, kept stock settings bar Power Supply at either 25% or 50%, worked fine.

Any ideas, or am i barking up the wrong tree ?

ewY0gyX.jpg.png
Are you sure the memory is stable, also the cpu?
Have you tried a complete driver clean with DDU ?
What bios is the gpu using?
Did you change anything else recently?
 
The game crashes for me unless I completely disable crossfire with the pop up message "dxgi_error_device_hung".
Am i right in thinking you have a "free" copy of the game?
I'll bet that if you turn everything to max but turn off the shadows, that the game starts (with crossfire enabled in the main driver settings but removed from all profiles)
Just a hunch as mine would crash with DX12 and shadows turned on to anything other than low with that same error and also a "out of gpu memory error"
 
Are you sure the memory is stable, also the cpu?
Have you tried a complete driver clean with DDU ?
What bios is the gpu using?
Did you change anything else recently?

Not really changed much, its kind of a new system, Ryzen etc from early December, Vega have owned for a while, PSU a while.

No not sure if its stable, but odd that if its the CPU / Mem, why does higher power on the GPU stop the crash. Whats my best bet for eeking out any CPU / Mem instability . Mem Test for mem ? Anything that will hammer both together ?

Bios, its a standard Vega LC , Sapphire, are there bios updates for the LC ?

Cheers
 
Not really changed much, its kind of a new system, Ryzen etc from early December, Vega have owned for a while, PSU a while.

No not sure if its stable, but odd that if its the CPU / Mem, why does higher power on the GPU stop the crash. Whats my best bet for eeking out any CPU / Mem instability . Mem Test for mem ? Anything that will hammer both together ?

Bios, its a standard Vega LC , Sapphire, are there bios updates for the LC ?

Cheers

Have you tried increasing your cpu offset vcore to around stock or so, whilst experimenting with the same gpu profiles. Does the system crash in exactly the same part of the physics test, everytime?
 
Not really changed much, its kind of a new system, Ryzen etc from early December, Vega have owned for a while, PSU a while.

No not sure if its stable, but odd that if its the CPU / Mem, why does higher power on the GPU stop the crash. Whats my best bet for eeking out any CPU / Mem instability . Mem Test for mem ? Anything that will hammer both together ?

Bios, its a standard Vega LC , Sapphire, are there bios updates for the LC ?

Cheers
You could start by using ryzen master and checking it in there by turning on the mem and cpu check options then running a test, and also mem test yes , thats my go to, (the one you have to run multiple instances of in windows)
AIDA 64 will also do these things.
Yes there is also more recent versions of the LC bios that might help you, depends which version yours is (gpu-z will show you) then pop to techpowerup for a list of the vega bios's and use DDU and atiflash to achieve the update of bios.
If you set you Vega to +50% power , 1000mv for the HBM, and p6 1080 and p7 1100 does it still do this crashing?
 
Am i right in thinking you have a "free" copy of the game?
I'll bet that if you turn everything to max but turn off the shadows, that the game starts (with crossfire enabled in the main driver settings but removed from all profiles)
Just a hunch as mine would crash with DX12 and shadows turned on to anything other than low with that same error and also a "out of gpu memory error"

I have the game on steam. I'll have a further mess around with it and adjust the Shadows to see if it does anything.
 
Having a few probs with Firestrike extreme crashing , potentially on stock settings.

Ryzen 2700x , mem 3200 xmp at cl16, asrock taichi x470, Vega 64 LC now with EK block. Superflower 1kw PSU, Vega power connectors on separate rails.

Pbo is on and a voltage offset of -0.06250, otherwise nothing else changed.

Temps are good, hotspot is good etc and games etc seem stable. Thought i'd start with undervolting the Vega.

Timespy and gaming seemed fine.

But Firestrike Extreme keeps crashing on the Physics test. So put everything back to standard vega profile, but still lots of crashes.

Perhaps its not Vega settings as its probably CPU / Mem on the physics tests .. but see further below ...? . I've had no issues with gaming , BFV etc very stable once 1st patches out. Moved back from 18.12.3 to 18.12.2 but still same.

Relive disabled, free sync disabled, chill disabled.

Heres the odd bit

Timespy / Timespy extreme fine on Balanced and various undervolt settings. I did get an occasional issue so intended to start again with the UV and keep HBM at stock whilst slowly found stable settings. Played BFV for a few hours without issue ( i know its not the best stability test )

Tried Firestrike extreme , crashed on undervolt on physics test.

So put Vega back to balanced. Still crashed.

Put to power save , worked fine.

Put back to balanced after power save , worked fine.

Put to turbo, worked fine.

Restarted, put back to balanced, crashed.

Put to custom, kept stock settings bar Power Supply at either 25% or 50%, worked fine.

Any ideas, or am i barking up the wrong tree ?

ewY0gyX.jpg.png


Physics test on Firestrike crashing is the CPU 99.9999% of the time. The GPU does practically nothing for this test.
 
Have you tried increasing your cpu offset vcore to around stock or so, whilst experimenting with the same gpu profiles. Does the system crash in exactly the same part of the physics test, everytime?

You could start by using ryzen master and checking it in there by turning on the mem and cpu check options then running a test, and also mem test yes , thats my go to, (the one you have to run multiple instances of in windows)
AIDA 64 will also do these things.
Yes there is also more recent versions of the LC bios that might help you, depends which version yours is (gpu-z will show you) then pop to techpowerup for a list of the vega bios's and use DDU and atiflash to achieve the update of bios.
If you set you Vega to +50% power , 1000mv for the HBM, and p6 1080 and p7 1100 does it still do this crashing?

Physics test on Firestrike crashing is the CPU 99.9999% of the time. The GPU does practically nothing for this test.

Thanks i'll try all those suggestions.

I'm on the latest Bios as far as i can see for the Sapphire LC's 8774

It pretty much goes at the same point. It feels like its close to the end of the test and its switching out ready for the next one.
The screen goes black in the same way it does when switching, so not sure when its crashed or ended.
Occasionally its not dropped back to the app and alt tabbing doesnt help, i've had to turn the monitor on and off and thats brought the picture back.

Likely CPU / Mem etc , but wondered if its ending and its somehow the GPU resolution / clock change thats making it go . I'm going to turn off PBO and the offset and try again.

Does the HBM mv do anything on the LC, i thought i read it only took notice of the Frequency changes. On mine I can also now change any states P1 to p7, not just 6 and 7. Should i make i.e. 5 match p6 on MV etc ( p5 is 1562 @ 1100mv at the mo )

Thanks for all the help
 
Last edited:
So far haven't thrown any errors doing some stress testing with Aida, Ryzen Master and Occt without changing much from last night.

Aida 64 hammers the GPU more than most as far as i can tell, got the Gpu freq upto 1780 ( per hwinfo) which is the highest i've seen.

Did a few runs with PBO off / on via Ryzen master and all seemed well.

Havent done long enough on any of them. Also have Realbench to take a look at.

One thing i noticed was my Minimum processor state was 5% in the Power settings, and it should be at 90% on Ryzen balanced to stop core parking. I probably changed that early days , my bad. Will see if that helps later. Gym time.
 
Back
Top Bottom