Support needed with 7900 XTX

[SpudMaster]

Should buy your parts from the rainforest and return if the new part doesn't fix. Jeff's a good friend to have like that

 

[Nothingdoing079]

One thing I've just spotted while playing around with things this evening is that my Chipset temps are sitting around 75C when not under load.

I'm therefore wondering whether the GPU when under load might be pumping more heat onto the chipset and over heating it, thus causing a shut down?

 

[wookiee87]

75c is way too hot idle for chipset, check the fan is spinning and clean if dusty. i'm on x670 but my chipset under load normally hits around 64c for both, i know x570 ran warm but thats why they had the fan on the chipset. this is mine idle in windows

 

[GhostDog1981]

Should buy your parts from the rainforest and return if the new part doesn't fix. Jeff's a good friend to have like that

That’s bad consumerism

 

[eeii]

Don’t know if you have tried this before. Can you reduce your max clock speed to 2500. Try a few games and let us know how you get on.

 

[Nothingdoing079]

Don’t know if you have tried this before. Can you reduce your max clock speed to 2500. Try a few games and let us know how you get on.

That is one of things I tried early on, and unfortunately still was getting crashes.

One thing I've noticed this morning when monitoring temps is that currently the GPU is sitting at 56C while idle, and the Hot Spot at 66C.

Interestingly when I then try playing a game it ramps up but doesn't exceed 75C.

Meanwhile the Chipset is holding at around 67C

 

[GhostDog1981]

That is one of things I tried early on, and unfortunately still was getting crashes.

One thing I've noticed this morning when monitoring temps is that currently the GPU is sitting at 56C while idle, and the Hot Spot at 66C.

Interestingly when I then try playing a game it ramps up but doesn't exceed 75C.

Meanwhile the Chipset is holding at around 67C

Random black screens is a known fault of recent 7000 cards. It was supposed to be fixed in the latest Adrenalin drivers. Have you tried the Fluid Motion Frames preview driver to see if that improves things?

The only other thing I can think of is the vram- if this is bad then random restarts etc make sense. But this is very hard to prove unfortunately (happy to be corrected).

Why don’t you post in the CS section of the forum and see if they’ll take another look at your card?

 

[eeii]

Random black screens is a known fault of recent 7000 cards. It was supposed to be fixed in the latest Adrenalin drivers. Have you tried the Fluid Motion Frames preview driver to see if that improves things?

The only other thing I can think of is the vram- if this is bad then random restarts etc make sense. But this is very hard to prove unfortunately (happy to be corrected).

Why don’t you post in the CS section of the forum and see if they’ll take another look at your card?

I’d also point them to this thread.

 

[wookiee87]

Meanwhile the Chipset is holding at around 67C

I hope thats under load, if not its still warm but its getting better, with regard to why your gpu is warm when idle all modern gpu's have a function called fan stop,
if the gpu drops below 50c the fans will turn off until the card is hit with a load, then when it goes past 50-55c area the fans will spin to dissipate the heat.

i'd have a quick look at case airflow and see if you can get a fan in the general gpu/chipset area and run it a bit quicker to allow fresh air to your board and gpu.

 

[Nothingdoing079]

Random black screens is a known fault of recent 7000 cards. It was supposed to be fixed in the latest Adrenalin drivers. Have you tried the Fluid Motion Frames preview driver to see if that improves things?

The only other thing I can think of is the vram- if this is bad then random restarts etc make sense. But this is very hard to prove unfortunately (happy to be corrected).

Why don’t you post in the CS section of the forum and see if they’ll take another look at your card?

Cheers I'll post something there and point them to this thread.

I'm also speaking to one of the Senior Technicians at Overclockers at the moment, and was also going to point them to the thread.

I hope thats under load, if not its still warm but its getting better, with regard to why your gpu is warm when idle all modern gpu's have a function called fan stop,
if the gpu drops below 50c the fans will turn off until the card is hit with a load, then when it goes past 50-55c area the fans will spin to dissipate the heat.

i'd have a quick look at case airflow and see if you can get a fan in the general gpu/chipset area and run it a bit quicker to allow fresh air to your board and gpu.

No that's while it's idle which is worrying me, as starting to wonder if that the cause of my troubles (though it doesn't really explain why I had the same crashes happen on the second pc)

I'm going add a couple new fans to my computer this evening to see if I can increase airflow to that area, and will also check if the fan is working correctly (there isn't any dust build up from what I can see)

 

[wookiee87]

No that's while it's idle which is worrying me, as starting to wonder if that the cause of my troubles (though it doesn't really explain why I had the same crashes happen on the second pc)

I'm going add a couple new fans to my computer this evening to see if I can increase airflow to that area, and will also check if the fan is working correctly (there isn't any dust build up from what I can see)

the second pc you ran the gpu in, is that a similar x570 system or a intel based rig?

i've included a screen shot of your board and which screws to take out to remove the chipset cover so you can access the fan and check its spinning, you could always set a manual fan curve on the gpu, given the location of the chipset fan it's effectivly blocked when a gpu is installed, under load it'll get hot but at least idle the gpu should bring down the chipset temp if on a manual fan curve.

 

[wookiee87]

Storage - 3xM.2, 2xSSD and 1xHDD

Just noticed this in your original post, Thats a hell of a lot of storage connected to the board, i wonder what pcie speed your gpu is running at, ideally you should have 1 m.2 installed in the top slot between the cpu and gpu, and then attach sata based drives otherwise you'll very quickly run out of pcie lanes.

Your cpu has 24 pcie lanes to use but 4 are reserved for the connection to the chipset, 16 lanes of the cpu go to the primary pcie slot where the gpu goes, the other 4 are for the m.2 slot between the cpu and gpu.

x570 motherboards have an additional 16 lanes, but bear in mind that some i/o functions do take a few, not to mention the sata ports and the remaing m.2 slots and lower pcie slots, its a real jungle to figure out what to run.

If it were me i'd take out 2 of the m.2's at the very least, check what speed the gpu is running at via hwinfo under load, it should read x16, any lower and your bandwith limited, go too low and it will cause major headaches and problems.

 

[Nothingdoing079]

the second pc you ran the gpu in, is that a similar x570 system or a intel based rig?

No the second pc was intel based and running a Z170 MB

I had checked the Chipset fan when I was troubleshooting the other day but didn't notice anything amiss at the time with how it was working when I removed the cover.

I will however be giving it a closer look, and playing with the curves to reduce GPU fan use in that area

f it were me i'd take out 2 of the m.2's at the very least, check what speed the gpu is running at via hwinfo under load, it should read x16, any lower and your bandwith limited, go too low and it will cause major headaches and problems.

I had a similar thought about the pcie connection the other day, and have already removed one of the m.2 which id had connected via an extender.

At the moment I now have 2 M.2 connected via the 2 M.2 spaces on the motherboard, 2 SSD connected via the SATA ports, and 1 HDD also connected via SATA.

Lane capacity confuses me somewhat, do you recommend reducing the drives further?

 

[wookiee87]

Lane capacity confuses me somewhat, do you recommend reducing the drives further?

I would yes the sata based drives should be ok but try with only 1 m.2, if still no joy take 1 sata drive out at a time and test a game(s)

 

[Nothingdoing079]

Running a sensor report via AIDA64 I am getting the following temps and voltages while idle

Sensor Properties:
			Sensor Type	Nuvoton NCT6798D (ISA A00h)
			GPU Sensor Type	PMLog (ADL2)
			Motherboard Name	Asus Prime X570-Pro / ROG Strix X570-E/F/I Gaming
			Chassis Intrusion Detected	No
		Temperatures:
			Motherboard	54 °C
			CPU	37 °C
			CPU Diode	41 °C
			CPU Package	41 °C
			Chipset	72 °C
			PCH Diode	72 °C
			GPU	56 °C
			GPU Hotspot	64 °C
			Samsung SSD 970 EVO Plus 2TB	66 °C / 65 °C
			Samsung SSD 850 EVO 250GB	46 °C
			CT1000MX500SSD1	45 °C
			Samsung SSD 970 EVO Plus 500GB	53 °C / 60 °C
			ST3000DM001-1ER166	41 °C
		Cooling Fans:
			CPU	651 RPM
			PCH	3355 RPM
			Chassis #1	564 RPM
			Chassis #2	399 RPM
			GPU	614 RPM (15%)
		Voltage Values:
			CPU Core	0.984 V
			CPU VID	1.000 V
			+3.3 V	3.296 V
			+5 V	5.020 V
			+12 V	12.076 V
			+3.3 V Standby	3.408 V
			VBAT Battery	3.376 V
			CPU VDD	0.981 V
			CPU VDDNB	0.988 V
			GPU Core	0.585 V
		Current Values:
			CPU VDD	8.31 A
			CPU VDDNB	7.22 A
		Power Values:
			CPU Package	38.24 W
			CPU VDD	8.16 W
			CPU VDDNB	7.13 W
			GPU	90.00 W
			Debug Info F	65535 65535 65535 65535 65535 65535
			Debug Info T	00 00 00
			Debug Info V	36 00 62 62 62 FF FF (00) / FF FF FF FF FF FF FF FF / FF FF FF FF FF FF FF FF
			Debug Info I	FF D42B

 

[Nothingdoing079]

I would yes the sata based drives should be ok but try with only 1 m.2

Cheers, will give it a go

 

[wookiee87]

That 2tb Evo plus is a bit warm, that a SSD or a m.2?

 

[Nothingdoing079]

That 2tb Evo plus is a bit warm, that a SSD or a m.2?

Its a M.2 which is located directly underneath the 7900 XTX.

Have also looked at the PCI-E Controllers in use according to AIDA64, and have the following

North Bridge

PCI-E 2.0x4 port #1 - In use @ x4 (Samsung NVMe SSD Controller)
PCI-E 2.0 x8 port #0 In Use @ x4 (AMD 500-Series Chipset - Switch Upstream in PCIe --> AMD 500-Series Chipset - PCIe GPP Bridge --> Intel I211 Gigabit Network Connection, AMD 500-Series Chipset - PCIe GPP Bridge --> Samsung NVMe SSD Controller)
PCI-E 2.0 x16 port #0 In Use @ x16 (AMD Navi 10 - PCI Express Upstream Switch Port --> AMD Navi 10 - PCI Express Downstream Switch Port --> AMD Navi 31 - High Definition Audio Controller, Sapphire RX 7900 XTX Video Adapter)

South Bridge
PCI-E 2.0 x4 port #1 In Use @ x4 (Samsung NVMe SSD Controller)
PCI-E 2.0 x1 port #5 In Use @ x1 (Intel I211 Gigabit Network Connection)
PCI-E 2.0 x16 port #0 In Use @ x16 (AMD K19.2 - PCIe Dummy Function, AMD K19.2 - USB 3.1 xHCI Controller)
PCI-E 2.0 x16 port #0 In Use @ x16 (AMD K19.2 FCH - SATA AHCI Controller)
PCI-E 2.0 x16 port #0 In Use @ x16 (AMD K19.2 FCH - SATA AHCI Controller)

 

[wookiee87]

Its a M.2 which is located directly underneath the 7900 XTX.

It's around 20c hotter than it needs to be tbh, m.2's should idle around the 43-45c mark and then under load with a heatsink sit in the mid 60-70c region, there are a fair few controllers on the board that will take 1pcie lane, so I would say you have too many drives especially m.2's installed.

Also pcie gen 2 is pretty slow your drives and GPU are capable of gen 4 speeds.

 

[Nothingdoing079]

It's around 20c hotter than it needs to be tbh, m.2's should idle around the 43-45c mark and then under load with a heatsink sit in the mid 60-70c region, there are a fair few controllers on the board that will take 1pcie lane, so I would say you have too many drives especially m.2's installed.

Also pcie gen 2 is pretty slow your drives and GPU are capable of gen 4 speeds.

Interesting, I took those readings while at Idle, so will try under load when I have a chance to see if that changes things to Gen 4.

Its a shame if I can't use the 2 M.2, as currently have 1 M.2 as my Boot Drive, and then the second as a storage drive (previously it was 500GB but had filled it).