RX Vega Owners Thread

So I have been playing with bios edits in the extremely limited amount of free time I have had (seriously I worked the last 6 days straight and have another 3 long days to finish this project)...With 75% Power Limit and my Vcore set to 1.05V, my card will sustain ~1750 with 2 random drops in Super Position...GTA V keeps the card clocked @ 1750Mhz and I have to say with HBM @ 1100 the card flies!! I can keep my min FPS right @ my refresh rate of 80Hz...

Just for giggles I have played with PL of 105%, and the card will clock @ 1760Mhz and actually stays extremely cool, with HBM temps in the low 30s, rising to ~35C after an hour or so, and Core temps right around there...Hot Spot Max according to GPU-Z was a spike to 46C..If I keep the PL @ 75~80% it stays @ 42C for the Hot Spot.

Once I get back from my 6 week trip to Los VEGAs (heheh I could not help it) I am going to find the best FreeSync LCD I can...I am not sure I want 4K, but I think I will look into the ~100Hz IPS 34"s...The HP Omen has been on my radar as well since I got to play with one @ MicoCenter but the lowish refresh rate (75Hz, could probably OC to 80~85Hz like my current LCD) and the slowish VA panel have me a bit concerned.
 
For those on Windows 7, do you see the HBCC option under Radeon Settings? Works under windows 10 but missing in Windows 7.

I just went all through the Radeon settings and noticed that is no longer there, which is extremely odd..When I first installed my card, I was using the 178.2 with my 290...I removed the 290 and just used the same driver install...When I started playing with editing the BIOS of the card, I had to DDU in safe mode and reinstall (17.8.2) to get the increased Power Limits to work properly..I didn't pay much attention until you said something but it is definitely no longer showing up, where as before it was showing up but "disabled"...:mad::mad::mad:

Looks like now whenever I try to change anything from default settings, my system crashes when loading windows (including just power limit +50%). Might roll back the 17.8.x since it seems 17.9.1. is hot garbage.

I have stayed far far away from 17.9.1 after hearing all the horror stories. ..17.8.2 has been rock solid for me so I am staying here for now since I am not playing any super modern game at the moment.
 
6 weeks in Vegas? Nice. Hope after all the strippers and gambling you have some money left for the monitor.

Lol, I intend to have plenty! I am going out for my brother's wedding, then house/pet sitting for 2 weeks while they are on honeymoon, then hanging out with them for a few weeks. I actually play poker semi professionally, so it is a "working" vacation ;)..Am I the only one that has modified their Power Limit to enable an increase past 50%?
 
Any performance improvements noted on single GPU with 17.9.2? Usually AMD would indicate performance improvements with a new driver and here there are none.
 
Awful. How can we go from a flawless crossfire experience with my Fury X to this?


I'm a hair away from abandoning my freesync monitors and buying gsync. If they had a 38" gsync I'd do it right now.
Mismatched modeled GPUs, 56 and a 64. That may or may not have caused the poor scaling. Now in the past a 290 and 290x was excellent in CFX in Win 7 for me. Probably too early yet to determine how well Vega will scale with DX 11 in CFX.
 
Honestly it's pretty strange to mismatch GPUs. I'd be more interested in a matched analysis.
 
OK, so I downloaded the drivers for my 2x RX Vega 64s. To be honest, the situation is really bad.

I'm pretty sure the hardware is working. Dishonored got a nice boost from around 100 fps to over 200 fps 4K max settings, both GPUs in the 90% usage range. That's about the only clear win here.

Borderlands went from around 60 fps to around 90 fps 4K maxed, which is something, but the game was already comfortably running with 1 Vega.

GTA V, which traditionally had great CF support, was a shit-show. First off, the RivaTuner overlay was flickering wildly, not a great start. Fps was in the 40 - 90 range, but with lots of variance. With 1 card, I got about a solid 50 fps. With 2, I could get as high as 90 but it would frequently drop to as low as 40 and the whole experience felt choppy.

Far Cry Primal, I'm not sure performance was even better. Still around the 50 fps mark. Highs were a bit better, getting into the mid-60's but overall it was not meeting performance and was regularly in the 40's or 50's.

Everybody's Gone to the Rapture, performance was better at 1080p. It went from around 60 fps to 90 fps. Was able to bump from High to Very High and performance was still good. Going to 4K, perf was at the 30 fps mark (same as w/ 1 Vega) but now half the screen was covered in rainbow artifacts.

In DOOM, which doesn't even support CF, performance got worse. 1 Vega was getting a solid 60 fps 4K Ultra settings. With CF Vulkan, it dropped to 50 fps. With OpenGL it was much worse, at 40 fps.

Rise of the Tomb Raider did work, with DX12 mGPU. Performance went from 45 fps to around 70 fps. However, like GTA V, the perf was extremely variable, going from as high as 95 to as low as 55. This is about playable to me, but I would have liked to see 60+ fps solid.

Overall, I'm really disappointed. I know the driver is beta, but in some cases I would say it's almost worse than running on a single card. This is especially concerning as I sold my 2 RX 480s to go to 2 RX Vegas and I'm not sure it's even an improvement.

I'll give AMD some time to update the drivers, but this is not a good start.

EDIT: I'm sorry, the settings were wrong. Somehow Chill mode got enabled (what I bad default) and that messed up performance. With Chill off, performance is great. Will test again.
 
Last edited:
Chill is basically awful.

Any of those power saving functions should be off on a performance card from the start. The fact that driver updates sometimes enable it makes me frustrated too. Random enabling of Chill through driver updates frustrated me on my Fury X cards on rare occasion as well.

I don't get why AMD released Vega in this driver state. I'm truly disappointed. A couple friends and I are trying to mine with them - and they are not very stable, they are loud, they are hot, wattman settings constantly reset, Crossfire wasn't enabled. Eyefinity bezel correction doesn't work right, performance isn't consistent. Performance on PUBG is poor, even after the targeted driver update specifically for that game. 40-90fps at 1440p --- but way too much time at 40fps. My PC is acting wonky sometimes since installing Vega - tonight it just went to a blue background and seemed unresponsive immediately after a reboot. This machine was stable before Vega.
Heck for a week or two we dealt with this browser lockups for a few seconds at a time with Google Chrome!?!?!? Yes they fixed it, but you're going to tell me they didn't test, discover and fix the google chrome browser situation until 2-3 weeks after launch??? The whole MSRP fiasco still is a bit irksome - no matter who is at fault, as is the lack of cards, as is the mixed messaging (AMD is pro gaming, and we want to keep the cards out of miners hands.-- oh and by the way here's a miner driver for our cards). Finally when Raj basically confirmed the false mining numbers...it really blew my respect in AMD.

The gaming performance being lower than Nvidia 1080ti isn't what bothers me.

BUT...

The fact I'm beta testing on a product that was in development for so long (and that I overpaid above MSRP to be said beta tester) does bother me. Especially after having such a overall smooth experience with my AMD Vega 285 and AMD Furys. I didn't think AMD deserved the bad wrap the forum trolls gave them with recent cards. ( I had 285 and Fury X cards and they we're pretty much as problem free as any Nvidia card I ever owned in the last 15 + years). However, this release is just checking the boxes for everything the Nvidia fan boys constantly poke at.

At this point I feel I should have kept my Fury X pair and sat out this beta testing period.
I didn't think it would be like this.
I figured it would be pretty polished at release as the architecture changes seemed rather slight.
I don't enjoy playing the role of a $800 liquid cooled Vega 64 and 2x $500 Vega 56 beta tester ---- that's not what I thought I was getting.

I've been generally keeping my mouth shut, but AMD has to turn this around or they are going to lose their supporters. AMD - Stop the false promises, stop the mixed messaging, pick up the pieces and get us some reliable drivers pronto! You've had years to get the flagship in order ---- as it is --- it's sinking before it makes it out of the harbor.
 
Last edited:
Mismatched modeled GPUs, 56 and a 64. That may or may not have caused the poor scaling. Now in the past a 290 and 290x was excellent in CFX in Win 7 for me. Probably too early yet to determine how well Vega will scale with DX 11 in CFX.

Honestly it's pretty strange to mismatch GPUs. I'd be more interested in a matched analysis.

Given how close they are in performance, and that there are no major architectural differences (i.e., different amounts of RAM), they should be able to get Crossfire scaling back up to where it was on previous architectures.
 
OK, so now I am getting decent performance. Much better than before.

GTA V 4K Max settings (expect advanced) gets me around 75 fps and it is relatively solid (outside a few intensive areas where it is hovering around 60fps).

Far Cry Primal on 4K Very High settings, getting around 65 fps. In a few intensive areas it would drop as low as 55 fps, but this is definitely playable.

Rise of the Tomb Raider, 4K High settings. Getting around 70 fps. Like Far Cry, there were a few times it dipped into the high 50's but infrequent enough that it's fine.

So this is about the most I could hope for. Modern games on High/Very High and playable in 4K. Still disappointing I had to go through an emotional roller coaster due to Chill mode, but I'm happy I got it working.

One thing I wasn't too happy about, though, is that I got 2 game lock ups during level loading (on RotTR and Everyone's Gone to the Rapture). Was not able to Alt + F4 or even get into Task Manager, but I was able to logout/login to the computer to avoid a restart.
 
Last edited:
Given how close they are in performance, and that there are no major architectural differences (i.e., different amounts of RAM), they should be able to get Crossfire scaling back up to where it was on previous architectures.
I would think so as well, the extra shaders do not look like they add that much to most games.
 
PUBG just crashed on my Vega with a red screen.

Whole computer crashed.
 

Attachments

  • IMG_0720.JPG
    IMG_0720.JPG
    402.6 KB · Views: 36
PUBG just crashed on my Vega with a red screen.

Whole computer crashed.

I have had that happen over the generations with AMD cards. The red screen of death. I usually fix it with changing to a different driver.
 
I have had that happen over the generations with AMD cards. The red screen of death. I usually fix it with changing to a different driver.
Also a huge sign of a failing card as well....yea i agree not always but its exactly what happened on a card i had to rma before. I think we can all agree its always video card related
 
It's crashed three times since my red screen in Player Unknown Battlegrounds. One a complete machine lockup, but no red screen. Twice just the game crashed gracefully and let me send a bug report.
That's four crashes in PubG in just a couple hours of playtime. Anyone else have this problem? Driver version is x.9.1

Orgin client just crashed too.

It's not worth it when I have two 1080TI's mining in the room next door and had no bugs with those cards.

I'm downloading the newest driver x.9.2 right now. If the problem persists I'm probably selling my freesync monitors and selling the Vega. I'm a father of four. My game time is precious and limited. I don't want to be beta testing drivers and having buggy operation when I'd like to be playing with my friends.
 
Probably Wattman issues as it will crash if I set my DDR4 too fast, if I mess with the speed of my video cards, open a video while the video card is OC'd too high, etc. Before you get too frustrated, try resetting your Win 10 install and use the latest driver.
 
It's crashed three times since my red screen in Player Unknown Battlegrounds. One a complete machine lockup, but no red screen. Twice just the game crashed gracefully and let me send a bug report.
That's four crashes in PubG in just a couple hours of playtime. Anyone else have this problem? Driver version is x.9.1

Orgin client just crashed too.

It's not worth it when I have two 1080TI's mining in the room next door and had no bugs with those cards.

I'm downloading the newest driver x.9.2 right now. If the problem persists I'm probably selling my freesync monitors and selling the Vega. I'm a father of four. My game time is precious and limited. I don't want to be beta testing drivers and having buggy operation when I'd like to be playing with my friends.
Sorry. Cant say I have any helpful advice. So far mine has been flawless, no issues gaming DX9, DX11 or DX12 or Vulkan with DOOM. I don't have PubG but I did play on origin the other day and no issues either. Only had a slight issue with Forza 7 CTD but allowing it to run in the welcome screen for a few seconds longer before hitting start helped, beautiful game as far as graphics.
 
Probably Wattman issues as it will crash if I set my DDR4 too fast, if I mess with the speed of my video cards, open a video while the video card is OC'd too high, etc. Before you get too frustrated, try resetting your Win 10 install and use the latest driver.
Every thing is stock on the card, but I had moved from balanced to turbo. Each time it locked up wattman was saying it reset settings. I downloaded the newest drivers and didn't change a thing. We'll see if it's stable as I play some this afternoon.
 
Sorry. Cant say I have any helpful advice. So far mine has been flawless, no issues gaming DX9, DX11 or DX12 or Vulkan with DOOM. I don't have PubG but I did play on origin the other day and no issues either. Only had a slight issue with Forza 7 CTD but allowing it to run in the welcome screen for a few seconds longer before hitting start helped, beautiful game as far as graphics.
Nice signature. ;). Razor1 lover. Hah
 
I have Sapphire Vega 56. Pretty sure undervolting is broken for me. Tried on 17.9.1 and now 17.9.2 also. Not messing with power target. On stock, the average frequency in game is around ~1250 MHz. But adjusting voltage in any way leads to this dropping to ~1100 MHz and HBM to 700 MHz. It's almost like it defaults to some slower P state. Weird still, the GPU only power draw remains same as on stock! That makes no sense. Then increasing power target gets the clock speeds back up, but it's worse results than leaving the voltage adjustments untouched. No matter what I try, I'm simply getting the best results by leaving voltage at stock, which is obviously unoptimal.

My goal is not to overclock or anything, but actually just maximize perf/W at the stock TDP. I've done the same with RX 480 and 580 to great effect.

I have tried with both OverdriveNTool and Wattman, same results. (Before you ask about Watttool, I can't get that as I simply unable to register on Overclock.net. But pretty sure the results will be the same)

By the way, just OC HBM2 works. Stable at 930 MHz.

Any ideas what's going on? I've tried searching about this online, but can't find anything.
 
Every thing is stock on the card, but I had moved from balanced to turbo. Each time it locked up wattman was saying it reset settings. I downloaded the newest drivers and didn't change a thing. We'll see if it's stable as I play some this afternoon.

What power supply do ya have and when did ya get it? Had something similar with my gtx 1080ti gaming system, couldn't figure out the problem, the power supply even tested out fine, finally got fed up with it and started replacing components, first thing was the power supply and bham system starting working fine.
 
What power supply do ya have and when did ya get it? Had something similar with my gtx 1080ti gaming system, couldn't figure out the problem, the power supply even tested out fine, finally got fed up with it and started replacing components, first thing was the power supply and bham system starting working fine.
Reasonable suggestion


Corsair 1000x
Less than a year.

No crashes yet on x.9.2 in the last hour in two rounds of PUBG
 
ah cool, probably not that then, I had a 500 watt bronze left over from an old HP, before I replaced it with a 650 watts (EVGA g2), pushing around 450 watts, so it was stressing the PSU nicely.
 
Every thing is stock on the card, but I had moved from balanced to turbo. Each time it locked up wattman was saying it reset settings. I downloaded the newest drivers and didn't change a thing. We'll see if it's stable as I play some this afternoon.

I was eager to replace my RX 480 because it kept messing up with all types of system hangs, corruption on screen, etc. I figured it was bad and I'd RMA after I get VEGA 64. Guess what? Same sh*t happens when Wattman decides to do whatever it wants to do. Most of the time it is because I added a slight OC, or tried something that you'll suggested on the forums like undervolting.

If you owned Nvidia cards around the 9800 GTX era, Wattman is catching up to it in annoyances. I wish I could remember the exact text from those errors; Nvidia driver blah, blah, blah has recovered. What it should have said was "The Nvidia driver has recovered and I hope that you saved everything that you were working on as your PC is going to automatically restart in 5, 4, 3, 2, 1...."


I believe that mine did the same on Turbo, but I got a message that the card was overheating before Wattman took matters into its own hands. And before you ask, I called myself setting a much more aggressive fan curve, but Wattman said I was an idiot and the fan speed never got loud enough to be audible. User error or A.I. error? I'm no Wattman expert, so I'll eat the bullet on that one this time.

Also have you seen this? I get this issue on my RX 480 and Vega 64. Every once in a while I will see a line flash across the screen then soon after this follows. Using HDMI on a Samsung TV. It ONLY happens in Google Chrome while playing a video. Sometimes exiting the video makes the screen do this. I just play all videos in Firefox Nightly and I have had zero issues. To remedy the issue I logout of my Windows account and then immediately log back in and it is fixed. I don't even need to restart the PC.

Sd2cm2TxS_m2qhrspCQW6m9s3Dr09020EIdMkwDYZIKk4v7TAeH3Ccd3owGSlWY5CkSizF79TP-rQHFyb8oFV_5zn4PtNUbYDwytVAUyT72PAK3Wers_YgO4oiCQqPJJHgpXSeWnZtrbepOA-Ys00GCJijE5FNhku4G4_LvVdylpCBIT2k9EBJ7d6Z6uvDtWD5KMW3cpN6PZZoSU0WTmVtb4UTCYrbctKptv93jf5keGuZ8yIt7RRfPYe8A9IR6LMIf0hoG9mpu1bPtTC7cyle4YDjgzbRhIqdKK2hg1kXj8S2K1IF9O39Kk0-aDsGJ1RIr3xGDBUoczBvH-WLiNT9bZdxrOeG9ei2eUcr8ISPDixnMfUNr6bFHnO4UV06O6Xjlg_qPd2ZfBZm2iBflOFxC8PSivSj7ftENBoJKrnb56oNHWD6y8GC4zDZigG21h58kBEWTexWeCzAGmw_VoVOmNTyvlg_-9N992NJtH0NPKZRwjRhCnKbS9diLc1apV7WTuJ6PvnUa2PFGCigjjw2LQoSrkO2CnaYV0JOzYveirCWJrC_zc4WSAUIUf4DdqeYw13J41mz7VJ2pBf7_t_bR55iZOab93u6Er8Vk2lPoniSh8cHyeArzgc2tVze5I1hiChxDd0sg9rQ-GfsudndGFUjLeoUImOtg=w2739-h2054-no
 
cageymaru, I haven't seen that yet.

ughh.

The x.9.2 drivers let me complete about 8 rounds without an issue. I quit because I have to go do some other things - but that might have resolved my problem?!?! with that game anyway.
 
cageymaru, I haven't seen that yet.

ughh.

The x.9.2 drivers let me complete about 8 rounds without an issue. I quit because I have to go do some other things - but that might have resolved my problem?!?! with that game anyway.

I think that was the PUBG driver from AMD. I forget which improved that game's performance.
 
I was eager to replace my RX 480 because it kept messing up with all types of system hangs, corruption on screen, etc. I figured it was bad and I'd RMA after I get VEGA 64. Guess what? Same sh*t happens when Wattman decides to do whatever it wants to do. Most of the time it is because I added a slight OC, or tried something that you'll suggested on the forums like undervolting.

If you owned Nvidia cards around the 9800 GTX era, Wattman is catching up to it in annoyances. I wish I could remember the exact text from those errors; Nvidia driver blah, blah, blah has recovered. What it should have said was "The Nvidia driver has recovered and I hope that you saved everything that you were working on as your PC is going to automatically restart in 5, 4, 3, 2, 1...."


I believe that mine did the same on Turbo, but I got a message that the card was overheating before Wattman took matters into its own hands. And before you ask, I called myself setting a much more aggressive fan curve, but Wattman said I was an idiot and the fan speed never got loud enough to be audible. User error or A.I. error? I'm no Wattman expert, so I'll eat the bullet on that one this time.

Also have you seen this? I get this issue on my RX 480 and Vega 64. Every once in a while I will see a line flash across the screen then soon after this follows. Using HDMI on a Samsung TV. It ONLY happens in Google Chrome while playing a video. Sometimes exiting the video makes the screen do this. I just play all videos in Firefox Nightly and I have had zero issues. To remedy the issue I logout of my Windows account and then immediately log back in and it is fixed. I don't even need to restart the PC.

Sd2cm2TxS_m2qhrspCQW6m9s3Dr09020EIdMkwDYZIKk4v7TAeH3Ccd3owGSlWY5CkSizF79TP-rQHFyb8oFV_5zn4PtNUbYDwytVAUyT72PAK3Wers_YgO4oiCQqPJJHgpXSeWnZtrbepOA-Ys00GCJijE5FNhku4G4_LvVdylpCBIT2k9EBJ7d6Z6uvDtWD5KMW3cpN6PZZoSU0WTmVtb4UTCYrbctKptv93jf5keGuZ8yIt7RRfPYe8A9IR6LMIf0hoG9mpu1bPtTC7cyle4YDjgzbRhIqdKK2hg1kXj8S2K1IF9O39Kk0-aDsGJ1RIr3xGDBUoczBvH-WLiNT9bZdxrOeG9ei2eUcr8ISPDixnMfUNr6bFHnO4UV06O6Xjlg_qPd2ZfBZm2iBflOFxC8PSivSj7ftENBoJKrnb56oNHWD6y8GC4zDZigG21h58kBEWTexWeCzAGmw_VoVOmNTyvlg_-9N992NJtH0NPKZRwjRhCnKbS9diLc1apV7WTuJ6PvnUa2PFGCigjjw2LQoSrkO2CnaYV0JOzYveirCWJrC_zc4WSAUIUf4DdqeYw13J41mz7VJ2pBf7_t_bR55iZOab93u6Er8Vk2lPoniSh8cHyeArzgc2tVze5I1hiChxDd0sg9rQ-GfsudndGFUjLeoUImOtg=w2739-h2054-no
I did have that happen on my RX470, but not since I updated to my Vega 64 on the latest driver.

If you turn off Hardware Acceleration, it should stop. You can usually re-fullscreen, then it'll fix after a few seconds.

I believe it's a known issue they're working on.... maybe report it on the AMD site just in case?
 
.9.2 seems to have fixed my "renderer not found" errors in Overwatch, fwiw.

It also fixed a goofy flicker I would get when in full screen PUBG.
 
I have Sapphire Vega 56. Pretty sure undervolting is broken for me. Tried on 17.9.1 and now 17.9.2 also. Not messing with power target. On stock, the average frequency in game is around ~1250 MHz. But adjusting voltage in any way leads to this dropping to ~1100 MHz and HBM to 700 MHz. It's almost like it defaults to some slower P state. Weird still, the GPU only power draw remains same as on stock! That makes no sense. Then increasing power target gets the clock speeds back up, but it's worse results than leaving the voltage adjustments untouched. No matter what I try, I'm simply getting the best results by leaving voltage at stock, which is obviously unoptimal.

My goal is not to overclock or anything, but actually just maximize perf/W at the stock TDP. I've done the same with RX 480 and 580 to great effect.

I have tried with both OverdriveNTool and Wattman, same results. (Before you ask about Watttool, I can't get that as I simply unable to register on Overclock.net. But pretty sure the results will be the same)

By the way, just OC HBM2 works. Stable at 930 MHz.

Any ideas what's going on? I've tried searching about this online, but can't find anything.
When I tried different settings on my WC Vega64 the Power target at 0% kept the clocks in state6 and wouldn't touch state7, so it would only do 1660 and not the 1750 from state7. Vega is far too new and dynamic for the old ways, which I prefer. You may need to increase the power target to +20% at least and still lower voltage to get higher stable clocks even the stock clocks.
 
When I tried different settings on my WC Vega64 the Power target at 0% kept the clocks in state6 and wouldn't touch state7, so it would only do 1660 and not the 1750 from state7. Vega is far too new and dynamic for the old ways, which I prefer. You may need to increase the power target to +20% at least and still lower voltage to get higher stable clocks even the stock clocks.

I agree 100%..Set your state 6 to 1750Mhz (since these cards to not ever reach state 7) and set your Voltage to 1.05V with 50% Power Limit..You will see much better performance then stock with the same or a tiny increase in power usage. I am still humming along with my 80% PL and my Vcore @ 1.05V..Still seeing 1725-1750 depending on how stressful the load. I honestly couldn't be happier..
 
Back
Top