The RX Vega 64 Owners Thread

[TheSwede]

you would have to be very unlucky to get 2 bad vega 64's that had run fine for months. I have had a bad HDMI cable before which gave me all kinds of trouble and had me looking at other things until I just changed it for the sake of it, and it cured it.
I have to say all though I would say 750w PSU is enough it does not mean it cant fail or have a cap leak. there have been faulty PSU's from good brands reported before, when you have a fault like this it can be a money pit, but also annoying it keep happening.

first thing I would do is swap out PSU and eliminate that first. make sure its a single rail PSU and high enough output. if you have a friend close who would be willing to try one of your cards for a while with a decent system it might save you expense. or even willing to swap out PSU's

Thank you for your quick reply Matt ! Much appreciated.

Well, I have a SuperFlower Leadex GOLD 850W Fully Modular "80 Plus Gold" that was bought from OcUK in end of Nov 2014, so a bit old but should be ok to try with that you think ? Not sure how to find out if it is a single rail PSU though, but I am guessing 850W is high enough output, or you referring to something else ?

 

[vapor matt]

Thank you for your quick reply Matt ! Much appreciated.

Well, I have a SuperFlower Leadex GOLD 850W Fully Modular "80 Plus Gold" that was bought from OcUK in end of Nov 2014, so a bit old but should be ok to try with that you think ? Not sure how to find out if it is a single rail PSU though, but I am guessing 850W is high enough output, or you referring to something else ?

That will be a single rail and leadex is a top brand, so give it a go and see how you get on. I would say it would be my first port of call the PSU. one of the most important things in any system and often overlooked because it seems to be working. and then we look at other things trying to find faults which are not there.

 

[eeii]

@TheSwede I know this is going to be painful but I would reinstall windows and go to windows 10.

 

[RyzenJunkie]

Hi All,

Got an issue. Cant get Freesync to work.

Sapphire Nitro+ Vega 64
LG 27UK600 Freesync Monitor

Got Freesync turned on on both the graphics card and the monitor (tried both standard freesync range and extended)

Loading the AMD Windmill Freesync test and the red horizontal bar only stops the screen tearing when i click vsync. If i click freesync it still does it.

Monitor is installed, graphics card fully installed.

Any ideas?

 

[RyzenJunkie]

Also just noticed , it keeps turning freesync off in the AMD Radeon settings?

 

[malachi]

Also just noticed , it keeps turning freesync off in the AMD Radeon settings?

Check your cables, are you using HDMI or DisplayPort?

 

[RyzenJunkie]

Using Displayport. Check as in , try another? Will have to look for one

 

[RyzenJunkie]

Tried 2 other DP cables and still the same

 

[RyzenJunkie]

Using Adrenalin 2019 Edition 19.4.1. Anybody else running the same as me and freesync works?

 

[RyzenJunkie]

Anyone else have any ideas?

 

[Goose]

I was wondering if someone might help me out with settings for Wattman for my custom water cooled Vega 64? I know this question must get asked a lot but it's a rather large thread to go hunting.

When under water should we still be undervolting? Also, i've seen mention of "locking the P states" but i'm not sure what that actually means and how to do it.

Any other tips would be greatly appreciated.

Thanks

 

[gpuerrilla]

@Goose definitely undervolt as it keeps your temps down, as the voltage provided is overkill in the default settings - it can run far lower even when clocks are increased.

You generally want to run the mem volts as your floor and keep both p6&7 volts same or higher.

See this excellent guide on what to absorb: https://forums.overclockers.co.uk/t...king-undervolting-information-guide.18793012/

 

[Goose]

@Goose definitely undervolt as it keeps your temps down, as the voltage provided is overkill in the default settings - it can run far lower even when clocks are increased.

You generally want to run the mem volts as your floor and keep both p6&7 volts same or higher.

See this excellent guide on what to absorb: https://forums.overclockers.co.uk/t...king-undervolting-information-guide.18793012/

Thanks.

Is the overdriventool not required any more since you lock pstates in Wattman?

Also, my card is a reference blower and i'm wondering if i should flash the LC bios since it's now under water.

Thanks

 

[iamcll]

@Goose definitely undervolt as it keeps your temps down, as the voltage provided is overkill in the default settings - it can run far lower even when clocks are increased.

You generally want to run the mem volts as your floor and keep both p6&7 volts same or higher.

See this excellent guide on what to absorb: https://forums.overclockers.co.uk/t...king-undervolting-information-guide.18793012/

Funny enough mine get's worse with a overclock-undervolt, Even on default with alphacool liquid it reaches 72c in 27c room... 20c room around 66c. ***** annoying

 

[gpuerrilla]

Thanks.

Is the overdriventool not required any more since you lock pstates in Wattman?

Also, my card is a reference blower and i'm wondering if i should flash the LC bios since it's now under water.

Thanks

Well wattman has got better over time but overdriventool can be simple enough and preferred by some. Depending on what your playing I know there are some guys getting better results from a stock BIOS and the new drivers if you look about.

 

[Dempsey22]

My crashing is back.

Checked graphs on AB after crash and the card was plotting along nice and steady core clock and mem clock, then it spikes up to 1675mhz which makes driver crash and game crash.

Anyone experiencing similar or know a fix?

This is even on power saver mode where such a high core clock isn't set.

 

[Grog]

what is your power limit set to?

When I had my PL set to 50% I would get those clock spikes leading to crashes. I turned it down to 20% and they don't seem to happen that often now.

 

[Dempsey22]

what is your power limit set to?

When I had my PL set to 50% I would get those clock spikes leading to crashes. I turned it down to 20% and they don't seem to happen that often now.

It was on Balanced mode, also tried Power Saver/Turbo/Custom. It can happen randomly. I checked again and it doesn't necessarily spike when it crashes. So its something else.

I tried MemTest over night, no errors reported. I'm gonna try clean install of Windows, if that doesnt help then I'm leaning towards a faulty PCIE Riser cable.

 

[Dempsey22]

Clean install of windows with newest AMD drivers.

Event 4101 AMD driver crashed and recovered.

Beginning to tear my hair out now.

I ran Furmark/Valley/Heaven/Superpositon and they all passed with flying colours and no crashes. That was at stock and with GPU OC cranked up.

I cant test in another system as it's a water cooled hardlined build.

 

[Illuminist]

Clean install of windows with newest AMD drivers.

Event 4101 AMD driver crashed and recovered.

Beginning to tear my hair out now.

I ran Furmark/Valley/Heaven/Superpositon and they all passed with flying colours and no crashes. That was at stock and with GPU OC cranked up.

I cant test in another system as it's a water cooled hardlined build.

I used to have this problem, i found that reducing the powerlimit helped, and locking the p states helped also, it happens on the radeon 7 as well, it is perfectly fine and then thinks to itself "ooh i have more headroom lets boost" and then crashes.
Lock the p states by setting the highest clock on wattman as the min and the max by right clicking on it. or dont clock the core so high.
I used to use 1650mhz and 1050mhz on the hbm with 1000mv for the hbm and 1050 1100 on the last 2 P states.
Hope that helps a little, also make sure you use DDU in safe mode at least twice then re install the drivers just to be sure first.