• 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 AMD Driver Thread

Figured out I had the minimal installer rather than the offline one so grabbed that and I'm back on 17.4.4. Will test if that's fixed anything tonight.



17.1.1 is pretty old by now but maybe Windows installed it? Just do a manual update check in Crimson.

ah cheers, thats part of my problem though as there is no manual update button any more and it stopped it's auto ones.

got the latest ones downloaded ill get them installed when im sober, is ddu for uninstalling driver remnants still a good choice?
 
I'm having weird issues that might be 17.7.2 related. I was on 17.4.4 but updated mainly to try out Enhanced Sync. Now I notice that in Overwatch three things have changed:

- My FPS drops (to 50-60 or even lower from 100+) in heavy fight scenes. Never used to happen - I'm very sensitive to drops like that.
- My in-game GPU temperature is 10-15 degrees higher for no apparent reason. Fan curves are the same (assuming latest Afterburner is actually setting and reporting correctly) and we're not exactly in a heatwave right now!
- My GPU isn't holding its maximum clock any more (used to be 1300 MHz solid, now hovers between 1280 and 1300 MHz).

The problem is I haven't played in around a month and there's been three changes since then: new AMD drivers, big Overwatch patch, and new Afterburner. I can see lots of posts on the Overwatch forums about recent performance problems in the latest patch but they could also be coincidence as others are likely to have updated drivers also. My first step was to revert to the 17.4.4 drivers but when I tried to install it, the installer just downloaded an "installer update" and installed 17.7.2 instead!

So basically: Anyone else experienced similar issues with 17.7.2? Anyone know how to force-install an old driver?
Tested again with 17.4.4 last night, all of these issues disappeared. Typical FPS is the same, temperatures are back down to mid-60s, and no crazy FPS drops during heavy fights. I found a bunch of people with the same issue on Reddit - some say they only downgraded to 17.7.1 and got the same improvements but I can't really be bothered to try right now so I'll stick with 17.4.4 until they fix it.

got the latest ones downloaded ill get them installed when im sober, is ddu for uninstalling driver remnants still a good choice?
I only bother with DDU when downgrading drivers or swapping GPUs. Upgrading drivers normally is fine.
 
Last edited:
Does anyone else have an issue with windows remembering where your monitors are with the new drivers. I have my screen stacked one on top of the other, but everytime i shut down my PC it reverts to putting them side by side and i have to manually go in and change them and apply new setting.

An other thing which may be unrelated is that AMD control panel and windows dont agree which screen is number 1 and 2 Window is correct but, AMD control panel switches them.

Both screens connected by DP.
 
Was playing it fine the other week on my R9 290.
Maybe it was just a problem with the Fury X then as I know someone else was having the same issue. After 16.11.5 the game would drop to 8 FPS, go back to normal frame rates for a second and then drop back down to 8 again - this was happening even on the intro and menu pages yet when I rolled back the drivers it was fine again. I'm getting the same behaviour with the Vega drivers - are there any other Vega users that could try The Evil Within to see if it's working?
 
Interesting comment there from a Vega FE user:

The reason performance falls off is core temp. As soon as vega gets over 60C performance falls off dramatically.

I'm currently running my Vega FE's at 1137MHz core clock and 1100MHz HBM2 clock. 55C core temp. 42.1MH/s.

Seems a pretty decent rate with improvements supposedly still to come, at that core clock speed I imagine power draw is way down as well.
 
Maybe it was just a problem with the Fury X then as I know someone else was having the same issue. After 16.11.5 the game would drop to 8 FPS, go back to normal frame rates for a second and then drop back down to 8 again - this was happening even on the intro and menu pages yet when I rolled back the drivers it was fine again. I'm getting the same behaviour with the Vega drivers - are there any other Vega users that could try The Evil Within to see if it's working?

Happens on the rx580 too.
Playing in window mode helps.
 
Just got a R9 290 Tri-x from members market and run my usual tests on it and even played a few games with no crashes, but when idle I keep getting crashes. Windows 10 locks up on a black screen mostly and I have to hold the power off button for 5 seconds in order to turn pc off. It is not a heat issue btw. It happens when 40-45C.

I uninstalled drivers, run DDU in safe mode and installed latest drivers, still same problem.

I get the feeling it is caused by Wattman which is applying high voltages on lower power states. By default voltage is on auto, but when toggle over to manual most states are 1250 with some being 1212, but the lowest one is 1250 which is obvious way to high and is likely why I get the lock ups I assume?

Can someone else with a R9 290 share with me their voltages for all the states so I can manually enter them please?

@AMDMatt - I know you are very busy right now, but any ideas?
 
This from an MSI R290 Gaming:

State 1: 1143
State 2: 1143
State 3: 1137
State 4: 1168
State 5: 1206
State 6: 1237
State 7: 1250

HTH.
Many thanks.

If anyone else can chip in, it would be helpful. As those numbers seem a little higher than some others I have seen online. Though, those were for a 390.

Seems a bit odd State 3 would be lower than State 2 and 1.
 
If the voltages are reading as you say then you should dump the bios using GPU-Z and then use Hawaii reader to make sure that the voltages haven't been hard wired. In the card's BIOS each DPM state should rely on a lookup table which sets the voltage for a given state based on the asic quality of the card (I believe). Therefore the default voltages are different for each card, apart from state 0 which is always 968mv I think (wattman doesn't let you change that).

For instance, here is my card which is running quite undervolted (-88mv under stock):
XUvkHW2.png

You can see in the GPU Freq table how my voltages go 65282, 65283, 65284, 68285 then 1150, 1156, 1162. The last three are not stock voltages, those are set by me and duplicated onto the limits tab as well. So make sure yours run up like the 65282 etc to make sure you have a stock bios first of all (those 65k numbers aren't voltages, they tell the card to use the lookup table for its voltages).

Secondly, regarding black screen on 290s etc, I think those issues were mostly due to voltage too low. My card did that at idle when I undervolted using MSI AB because I think it was applying the undervolt to all the states or something odd.

Edit: Oh I should add, if you do change voltages in your cards bios you MUST have uninstalled afterburner and drivers first, otherwise if those tools apply an overvolt on top of any changes that could cause issues so always DDU your drivers -> flash bios -> reinstall drivers if you need to go down this route.

Edit2: To answer your exact question also my voltages are:
S1: 1031
S2: 1037
S3: 1075
S4: 1118
S5: 1150 (hardwired)
S6: 1156 (hardwired)
S7: 1162 (hardwired) (1237 stock +13mv manufacturer offset = 1250mv)

Card is an MSI 290x gaming.
 
Last edited:
Many thanks.

If anyone else can chip in, it would be helpful. As those numbers seem a little higher than some others I have seen online. Though, those were for a 390.

Seems a bit odd State 3 would be lower than State 2 and 1.

Yeah, I noticed that, but I've never looked at them before, so I'm assuming they are correct. I've never had black screen issues, so YMMV.
 
Anyone got invited to that Driver Beta testing program yet?

Still waiting for the official package of 17.8.1 (for all cards), i dont wanna install the beta6.
 
If the voltages are reading as you say then you should dump the bios using GPU-Z and then use Hawaii reader to make sure that the voltages haven't been hard wired. In the card's BIOS each DPM state should rely on a lookup table which sets the voltage for a given state based on the asic quality of the card (I believe). Therefore the default voltages are different for each card, apart from state 0 which is always 968mv I think (wattman doesn't let you change that).

For instance, here is my card which is running quite undervolted (-88mv under stock):
XUvkHW2.png

You can see in the GPU Freq table how my voltages go 65282, 65283, 65284, 68285 then 1150, 1156, 1162. The last three are not stock voltages, those are set by me and duplicated onto the limits tab as well. So make sure yours run up like the 65282 etc to make sure you have a stock bios first of all (those 65k numbers aren't voltages, they tell the card to use the lookup table for its voltages).

Secondly, regarding black screen on 290s etc, I think those issues were mostly due to voltage too low. My card did that at idle when I undervolted using MSI AB because I think it was applying the undervolt to all the states or something odd.

Edit: Oh I should add, if you do change voltages in your cards bios you MUST have uninstalled afterburner and drivers first, otherwise if those tools apply an overvolt on top of any changes that could cause issues so always DDU your drivers -> flash bios -> reinstall drivers if you need to go down this route.

Edit2: To answer your exact question also my voltages are:
S1: 1031
S2: 1037
S3: 1075
S4: 1118
S5: 1150 (hardwired)
S6: 1156 (hardwired)
S7: 1162 (hardwired) (1237 stock +13mv manufacturer offset = 1250mv)

Card is an MSI 290x gaming.

Thanks for the detailed response. Since I manually input the voltage settings in wattman, had no issues, but I am not fully convinced yet.

I only use the AMD drivers, no afterburner or anything like that. I installed the hawaii reader. Here is how it looks like:

1.png

Is there a way of checking if it is the VRM? Maybe that is running hot somehow?



@Ultra_Extreme - Maybe you can chime in here instead of members market. Is this the default BIOS? Any info you can give to sort this out would be grateful.
 
Just got a R9 290 Tri-x from members market and run my usual tests on it and even played a few games with no crashes, but when idle I keep getting crashes. Windows 10 locks up on a black screen mostly and I have to hold the power off button for 5 seconds in order to turn pc off. It is not a heat issue btw. It happens when 40-45C.

I uninstalled drivers, run DDU in safe mode and installed latest drivers, still same problem.

I get the feeling it is caused by Wattman which is applying high voltages on lower power states. By default voltage is on auto, but when toggle over to manual most states are 1250 with some being 1212, but the lowest one is 1250 which is obvious way to high and is likely why I get the lock ups I assume?

Can someone else with a R9 290 share with me their voltages for all the states so I can manually enter them please?

@AMDMatt - I know you are very busy right now, but any ideas?


I have that card, under wattman mine are

0 N/A
1 1125
2 1112
3 1112
4 1150
5 1175 (can change this to 1150)
6 1212 (can change this to 1150)
7 1237 (can change this to 1150)

By altering the last 3 to 1150, power limit 50%, it knocks off about 5c and keeps the fans down, so its much nicer when gaming. The only issue when rebooting the PC i have to set them again as wattman does not keep settings (is that normal?)

*Update did not have a game profile in the gaming section once i did, it keeps the global settings,
 
Last edited:
I have that card, under wattman mine are

0 N/A
1 1125
2 1112
3 1112
4 1150
5 1175 (can change this to 1150)
6 1212 (can change this to 1150)
7 1237 (can change this to 1150)

By altering the last 3 to 1150, power limit 50%, it knocks off about 5c and keeps the fans down, so its much nicer when gaming. The only issue when rebooting the PC i have to set them again as wattman does not keep settings (is that normal?)

*Update did not have a game profile in the gaming section once i did, it keeps the global settings,

Thanks for the info.


@TNA That bios looks stock to me so no idea why it's default volts are high for all states. VRM temp can be seen on the gpu-z sensors tab.
All I see there as a temperature is GPU Temp. The rest display V (Vaults).

What should I be looking for, the name?
 
Hi M8, im afraid i have no info to give? I bought the card and i put it in my PC ive never altered the bios. Neither did i ever experience black idle hangs either. Im afraid all this voltage testing and overclocking stuff is stuff i did long in the past and im well out of date. The PC it was pulled from was basically on 24/7 and only in use hour or 2 a day at most, generally a file server. Never had any issues. I think those cards have a physically switchable bios dont they?
 
Back
Top Bottom