The Radeon VII Owners Thread

northun_munky · 14 Feb 2020 at 09:17

@Thomasck Got the core clocks working

Holding nice and steady at 1950 when set to 2000mhz. Having and issue with the HBM clocks though. Set them as you said to 1150 in both bottom boxes (DPM 2 and 3) but the mem clock in game are 800...

Tried changing it in Wattman with no joy. Am I missing something?

Jesaul · 14 Feb 2020 at 10:09

HBM memory is changed in driver UI. If it is 800 after that, dfu install again.

northun_munky · 14 Feb 2020 at 11:53

Jesaul said:
HBM memory is changed in driver UI. If it is 800 after that, dfu install again.

Ok, thanks.

Seem to have found an answer that works for me actually on another forum. Might have been mentioned here as well I suppose but read that changing the fan profile to 100% seemed to allow OC clocks to stick with wattman on recent drivers.

Working for me so far actually. Time will tell but may be sorted without having to resort to MPT.

@Chud_Jenny you tried this?

Jesaul · 14 Feb 2020 at 11:55

Yeah, I forgot, that even on water I set 100% to fan speed in MPT because of that reason

northun_munky · 14 Feb 2020 at 16:39

Spoke to early didn't I

Only seems to be working in one game, which happens to be the game I've been playing so only when I tried it on something else was I back to gimped clocks.

Will mess about with MPT again once I've finished playing this one game it works for. Damn, thought the fan fix was it!

Thomasck · 14 Feb 2020 at 18:48

northun_munky said:
@Thomasck Got the core clocks working Holding nice and steady at 1950 when set to 2000mhz. Having and issue with the HBM clocks though. Set them as you said to 1150 in both bottom boxes (DPM 2 and 3) but the mem clock in game are 800...

Tried changing it in Wattman with no joy. Am I missing something?

Setting the fans to 100% solves momentary, after a reboot, or crash, it goes messy. HBM OC with MPT does not work after 1100MHz, or something like that. I never mess to much with HBM clocks it does not yield too much performance anyways.

No driver will work for OC with Wattman flawlessly unless you use 19.7.5 (I heard about 19.7.11 too) or less. The fan trick was never 100% reliable here - just my opinion.

Kavalier · 16 Feb 2020 at 16:05

Hi guys,

So I am running the latest AMD Ryzen 2700x Chipset Driver (2.01.15.2138) and AMD Radeon Software (Version 20.2.1), but am still getting the occasional system hang, with the following Event IDs:

Event ID: 1000
Faulting application name: Radeonsoftware.exe, version: 10.1.2.1788, time stamp: 0x5e330423
Faulting module name: Radeonsoftware.exe, version: 10.1.2.1788, time stamp: 0x5e330423
Exception code: 0xc0000005
Fault offset: 0x0000000000405a07
Faulting process ID: 0x2f60
Faulting application start time: 0x01d5e4e1e8cd1739
Faulting application path: C:\Program Files\AMD\CNext\CNext\Radeonsoftware.exe
Faulting module path: C:\Program Files\AMD\CNext\CNext\Radeonsoftware.exe
Report ID: 71857567-05d9-4fab-bfe1-53afaf4d3bb3
Faulting package full name:
Faulting package-relative application ID:

Event ID: 1001
The computer has rebooted from a bugcheck. The bugcheck was: 0x00000116 (0xffff8b0f1eb7b2c0, 0xfffff8011fab0818, 0xffffffffc0000001, 0x0000000000000003). A dump was saved in: C:\WINDOWS\MEMORY.DMP. Report Id: 00000000-0000-0000-0000-000000000000.

No blue screens are displayed, just the system hangs and a reboot is followed.

The annoying thing is that this event is random and there is no pattern.

Can you please suggest what is the culprit? I do not believe that it has anything to do with the card?

Thomasck · 16 Feb 2020 at 23:00

@Kavalier is anything overclocked? when it happens, gaming/browsing?

Hades · 17 Feb 2020 at 09:17

@Kavalier Is your VII on the latest firmware? I know there weer a lot of stability issues on the early firmware versions.

It's probably unrelated but I had similar issues when my motherboard was dying. I thought it was the VII to start with but it's been completely stable since I replaced the mobo.

Kavalier · 17 Feb 2020 at 22:14

In terms of tuning, I just did a modest under volt and tweaked the fan-curve and power limit, since I have enough juice from power supply. @Thomasck

@Hades I am running the latest firmware. Also the motherboard is new. From the Event Logs, there are no system errors apart the occasional one from the AMD application. Also no blue screens or other other system related failures.
Also I did countless stress test, and no artifacts or driver failures were recorded. This is why I believe either it is a driver issue, which is strange or something is not 100% with the card. The tricky part is that since this event is rare, I do not know how to replicate it :s

Attached please find 2 snips for reference regarding the above.

northun_munky · 18 Feb 2020 at 15:58

Kavalier said:
In terms of tuning, I just did a modest under volt and tweaked the fan-curve and power limit, since I have enough juice from power supply. @Thomasck

@Hades I am running the latest firmware. Also the motherboard is new. From the Event Logs, there are no system errors apart the occasional one from the AMD application. Also no blue screens or other other system related failures.
Also I did countless stress test, and no artifacts or driver failures were recorded. This is why I believe either it is a driver issue, which is strange or something is not 100% with the card. The tricky part is that since this event is rare, I do not know how to replicate it :s

Attached please find 2 snips for reference regarding the above.

I appreciate the signs point to an issue with the card or drivers but your symptoms sound awfully like what happened to me when one of my ram sticks went bad. Worth checking if just to rule it out

Kavalier · 18 Feb 2020 at 19:51

Lets say something is wrong with one of my sticks, should the issue come out during a stress test focused on the RAM Sticks?

Other than that, from the issues that I highlighted, anyone else can provide some other insights please?

northun_munky · 18 Feb 2020 at 20:07

Kavalier said:
Lets say something is wrong with one of my sticks, should the issue come out during a stress test focused on the RAM Sticks?

Other than that, from the issues that I highlighted, anyone else can provide some other insights please?

Well, yes. Lets say the issue was a defective stick of ram, running memtest86 for example would either confirm, or negate the issue. Im not saying it is, but i went round the houses with a LOT of other things before testing the RAM. Doesnt take long and if nothing else, you can eliminate it as a cause. I was glad i did!

pingu666 · 20 Feb 2020 at 03:34

yeah run memtest86, its very good. sure its a hassle sometimes, but its worth it

Thomasck · 21 Feb 2020 at 17:01

Hi all, thanks for the time you've spent answering here, looks like the Bykski block is not doing a great contact somewhere, which has not been reported by any user so far.

Here's a screenshot of it running it with the stock cooler, no problem whatsoever sustaining 105-110C in the JT.

https://imgur.com/nU2UjXy

Now I gotta think what I'm going to do as I'm not gonna buy another gpu block.

MeddlE · 21 Feb 2020 at 17:41

Thomasck said:
Hi all, thanks for the time you've spent answering here, looks like the Bykski block is not doing a great contact somewhere, which has not been reported by any user so far.

Here's a screenshot of it running it with the stock cooler, no problem whatsoever sustaining 105-110C in the JT.

https://imgur.com/nU2UjXy

Now I gotta think what I'm going to do as I'm not gonna buy another gpu block.

Have you used enough thermal paste? They do require more than is usual.

Thomasck · 21 Feb 2020 at 21:40

@MeddlE yes I did. And when I put the stock cooler back I've used same amount. looks like I will need to get used to it and not push the card 2000mhz+

northun_munky · 21 Feb 2020 at 22:29

@Thomasck, long shot here but having looked back through you posts, there a link to imgur with a pic that shows the contact made by the block?

I've looked back through my pics when I fitted my ek block and it would appear that you haven't got a thermal pad covering the whole of the VRM's?

Forgive me as they may just have peeled off in your picture but heres what was covered in mine:

http://imgur.com/gallery/8glG2vZ

You may find that your missing some thermal pads?

Thomasck · 21 Feb 2020 at 23:09

@northun_munky Hi, they are all fine, I've fitted them to the block and they stuck in there

was easier that way. but then now, IDK, I'm not taking any futher action. This sample is nothing special in terms of ocing, I'll stick around 1950/1100/1100mV and that's it. This past year with the card I can say 30% of the time I was DDUing some driver, or something like that, and 70% enjoying it. I don't think this is a RMA case. When I was with the stock cooler the card was fine doing it's sustained 105C, but I could prove it's not a good one for oc, I set the rig in the window to get really cold air and after 1165mV it starts to glitch even with lower JT.

Kavalier · 22 Feb 2020 at 08:14

Hi guys,

Good Morning.

So I did some benchmarks this morning and no interruptions were noticed again.

The following is a link to one of the benchmarks.

UserBenchmarks:
CPU: AMD Ryzen 7 2700X - 84.5%
GPU: AMD Radeon-VII - 114.3%
SSD: WD Black SN750 NVMe PCIe M.2 500GB (2019) - 188.9%
SSD: Crucial M4 128GB - 61%
HDD: WD Green 1TB (2011) - 1,650.6%
RAM: Team Group Inc. TEAMGROUP-UD4-3200 2x8GB - 111.5%
MBD: Asus ROG STRIX X470-F GAMING