New build crashing in games (but not stress tests)

[siluro818]

Have you tried disabling Resizable BAR btw? This can cause instability on certain configs in games.
It is just strange to me that an unstable CPU can keep running stress tests, but only crash during play...
This is exactly the situation you would get with ReBAR problems though.

 

[goshsowitty]

Essentially assume I’m a 5 year old

I’ve only just learnt about resizable BAR now and, honestly, I’m looking for any logical explanation for the behaviour I’m seeing and a technology that targets games specifically does start to make this whole thing make more sense.

I’ve not enabled this so I’m unsure if I should expect to find it enabled by default but I will look into it when I can and disable it and give it another try.

Thank you!

 

[freeagent]

If vdimm is 1.35v I would try 1.375, them 1.4v. Shouldn’t need 1.1v soc more like 1v.. but try giving the sticks some more juice

 

[goshsowitty]

Have you tried disabling Resizable BAR btw? This can cause instability on certain configs in games.
It is just strange to me that an unstable CPU can keep running stress tests, but only crash during play...
This is exactly the situation you would get with ReBAR problems though.

Yeah so it is already disabled. Which is actually annoying. I thought we might be on to something.

 

[chaoshusky]

If this is tl;dr: Update your BIOS. UPDATE BIOS. BIOS BIOS BIOS. lol

Indeed Ryzen can be a pain with RAM if it's not "Ryzen Optimized" or you don't use a calculator. Unless the 6700XT has nasty power spike issues i doubt its the PSU either, but that behaviour could be several things, including a faulty motherboard or GPU but... Have you tried updating your motherboard's BIOS? Quite often they are sitting around for a while and still contain a fairly old BIOS. Usually a rule of thumb is if you don't have any problems, don't update the BIOS.. But you are having problems! So do it! I just got a new board, a z790 PG Sonic.. Got it to post fine-ish but it locked up a couple of times booting Windows, wasn't quite applying the correct XMP settings for the Corsair Dominator DDR5 RAM etc.. But the board was on the original BIOS! I updated and all is well, no longer has to mess around training and starts instantly. So, not sure if anyone else suggested it, but update your BIOS! Fairly sure that board can handle the wattage of a 5600X too.. Hmm.. Could even be you were sent a rare occurrence of a faulty second PSU. Shame you don't have any other machines to try the components in, but i repeat..if it no likey memory a new BIOS could fix that for ya and end up training the RAM correctly. If all else fails, swap the kit for one that is "Ryzen Optimized" then you won't have to dick around with a calculator tool!

Edit: Also, keep ReBAR ("Smart Memory") enabled! Apps, drivers etc mitigate the feature to soft disabled if a game, application etc is known to be buggy with it or not like it, so unless you have really serious hardware incompatibilities, especially for an AMD (eww) system, keep it on! I can't believe that cost you a grand either, prices are so crazy still!

 

[mechtech]

hmmm weird - not much to add to others have already mentioned.
Have you tried setting ram to non-xmp settings, as in jedec defaults?

Your psu, in theory should be more than enough, but with a 5600X and 6700XT total system may hit 350W+, maybe it's just a dud psu? memory testing doesnt use much power, relative to gaming.

You could try running the donut (furmark) along with cpu stress test at the same time to pull a lot of power, just keep an eye on temps, the donut can bring the pain to gpus. It should rule out the psu.

As for games is it every game or specific ones?

Only other thing might be temps? CPU and GPU ok while gaming? You can use Techpowerups GPUZ to check gpu temps while gaming, or just running furmark and compare with techpowerup review temps.

For P95 you can set it to fit everything(tests) in cpu cache you can do this to kind of differentiate between cpu and ram.

good luck

 

[Colddecked]

If vdimm is 1.35v I would try 1.375, them 1.4v. Shouldn’t need 1.1v soc more like 1v.. but try giving the sticks some more juice

Actually with Samsung stuff other than b die he may want to try going the other direction and lower voltage by .01. My Samsung c die hated more voltage and refused to go over 1.35 at all. Worked fine at xmp timings @1.32v.

 

[chaoshusky]

I think things are getting off-track here messing with voltages.. If you have to do that, something isn't right, the kit isn't Ryzen optimized and Ryzen is known to be picky, plus its a B550 board. BIOS update, if that doesn't fix it, swap the kit and ideally go for 3600MHz to give you a nice performance boost on Ryzen. My new board and RAM kit didn't like each other as much as they were supposed to, but i got them running fine.. Then updated the BIOS as i say and it just works now. Next: overclocking! But yeah, increasing the voltage to get them to run stock? That's an indication things aren't quite working as they should, though its hard to measure if there's a vast difference in voltage supplied vs voltage set by XMP or manually, which requires a DMM.

In some cases it could even be the CPU's IMC that's bad! And giving it more voltage than it should have without changing other things, sure fire way to upset your IMC too! (Memory controller)

 

[Deleted member 185158]

After 10 years it was time to build a new gaming PC on a relatively small budget but something that would last hopefully another 10 years (at a push), so I bought and built this over Christmas and New Year:

PCPartPicker Part List

Type Item Price

CPU	AMD Ryzen 5 5600X 3.7 GHz 6-Core Processor	£160.99 @ Amazon UK
Motherboard	Gigabyte B550 AORUS ELITE AX V2 ATX AM4 Motherboard	£140.99 @ Technextday
Memory	Corsair Vengeance LPX 32 GB (2 x 16 GB) DDR4-3200 CL16 Memory	£87.49 @ Amazon UK
Storage	Crucial P5 Plus 1 TB M.2-2280 PCIe 4.0 X4 NVME Solid State Drive	£86.31 @ CCL Computers
Video Card	PowerColor Fighter Radeon RX 6700 XT 12 GB Video Card	£414.11 @ Amazon UK
Case	Lian Li LANCOOL II-W ATX Mid Tower Case	£109.95 @ Amazon UK
Power Supply	Corsair TX650M Gold 650 W 80+ Gold Certified Semi-modular ATX Power Supply	£79.98 @ Amazon UK
Prices include shipping, taxes, rebates, and discounts
Total	£1079.82
Generated by PCPartPicker 2023-01-05 00:04 GMT+0000

Return said memory kit.

Memory

Corsair Vengeance LPX 32 GB (2 x 16 GB) DDR4-3200 CL16 Memory

£87.49 @ Amazon UK

It's Garbage. Consider it Value ram.

Return and pay a little more for some quality IC's.
This set is 10 bucks more and you'll have a much better experience with it.

https://www.amazon.co.uk/G-SKILL-F4-3200C16D-16GTZR-Trident-PC4-25600-Channel/dp/B01MTDEYHU/ref=sr_1_4?crid=29NNV1UYEHUFG&keywords=F4-3200C16D-32GTZA&qid=1672989028&sprefix=f4-3200c16d-32gtza%2Caps%2C168&sr=8-4

 

[A Computer Guy]

Actually with Samsung stuff other than b die he may want to try going the other direction and lower voltage by .01. My Samsung c die hated more voltage and refused to go over 1.35 at all. Worked fine at xmp timings @1.32v.

That sounds similar to an experience I had with my CMW32GX4M2C3200C16 kit (I forgot the version number) but I had to drop voltage to 1.33v to get DDR4-3800 to work.

 

[goshsowitty]

But get that memtesting down asap, make sure TM5 shows @anta777Extreme1 or @1usmusv3

Ran TM5 for hours using the @anta777Extreme1 profile and I think it did 3-4 cycles without errors or crashing.

I may fiddle with voltages and stuff. I'm not seeing this as a solution, necessarily, but for validation that the RAM is/isn't working as expected.

To address some other comments:

• The PSU has already been swapped and the issue persists.
• The BIOS has already been updated to the latest version and the issue persisted before/after the BIOS update.
• Done various shenanigans with chipset and display drivers too. Tried both recommended chipset drivers from Gigabyte and latest from AMD. Did a DDU for display drivers and installed driver-only (no Adrenaline software).
• I agree the PSU wattage should be fine by most accounts, at least in theory. Monitoring I've done suggests usage is way under 650W.
• Temps are fine. I ran Furmark and HeavyLoad simultaneously for an hour this morning where CPU, GPU and RAM were all maxed and temperatures were stable and hovered around 80 degrees.
• The issue does not arise in all games. Most games from a few years ago are fine. Fortnite (which is fairly graphically intensive these days) is even fine if we turn down the settings a little.

Does anyone have any thoughts as to how this is manifesting itself from a software perspective? For example the dump files are pointing towards Authenticamd.sys is this appearing in the dumps caused by the crash or could it be causing the crash?

I'm leaning towards it being a hardware issue and that's simply how it manifests itself in the crash reporting.

Based on further Googling around similar WHEA-logger errors and BSODs related to Authenticamd.sys it would seem to suggest people may have had some success with replacing or re-seating the CPU. People are bad at coming back to forum threads and posting a conclusion so it's not 100% clear but I'm going to turn my attention to the CPU for now and see if anything makes a difference.

To everyone who has replied so far, I am extremely grateful. I've had far more input here and a little on the AMD forum compared to other places I've posted like Reddit where I've had hardly any replies. Firm believer that forums like this are the cornerstone of the internet and you're all doing a great job and being a huge help. Thank you!

 

[tabascosauz]

@goshsowitty

Are you still within return window when you can go back to the store and quickly swap for a replacement 5600X? That can quickly rule out two possible culprits at this time (CPU and RAM).

TM5 is pretty determinative, but if you want extra confirmation you can run some 1usmusv3 (faster, you can run more cycles), or edit the anta777 config file to do 10-20 cycles (full overnight testing).

Like I said, except for very few errant exceptions the three RAM stability domains for Ryzen are clear and distinct from one another. If the Corsair kit is testing stable across 1usmusv3 and anta777, it's not the problem.

I haven't seen much of the authenticamd.sys bsod, but in terms of Bus/Interconnect there's not much you can do aside from play with the voltages. I don't think I've ever seen reseating CPU suggested or worked as a genuine solution to WHEAs. Again, tweaking is worth it if you're running into some pushback trying to make 1800-2000MHz Fabric work, but at bare spec 1600 all this troubleshooting is honestly not worth the time, take it back.

VSOC: try whatever works, up to 1.2V
VDDG_CCD: can stay 0.9-1.0
VDDG_IOD: set 0.05V below VSOC
VDDP: no need to exceed 0.9-0.95V

 

[A Computer Guy]

@goshsowitty

Are you still within return window when you can go back to the store and quickly swap for a replacement 5600X? That can quickly rule out two possible culprits at this time (CPU and RAM).

TM5 is pretty determinative, but if you want extra confirmation you can run some 1usmusv3 (faster, you can run more cycles), or edit the anta777 config file to do 10-20 cycles (full overnight testing).

Like I said, except for very few errant exceptions the three RAM stability domains for Ryzen are clear and distinct from one another. If the Corsair kit is testing stable across 1usmusv3 and anta777, it's not the problem.

I haven't seen much of the authenticamd.sys bsod, but in terms of Bus/Interconnect there's not much you can do aside from play with the voltages. I don't think I've ever seen reseating CPU suggested or worked as a genuine solution to WHEAs. Again, tweaking is worth it if you're running into some pushback trying to make 1800-2000MHz Fabric work, but at bare spec 1600 all this troubleshooting is honestly not worth the time, take it back.

VSOC: try whatever works, up to 1.2V
VDDG_CCD: can stay 0.9-1.0
VDDG_IOD: set 0.05V below VSOC
VDDP: no need to exceed 0.9-0.95V

in another thread you suggested "Also try bumping 1.8V PLL, up to a max of 2.0V. One of the tools to support Fabric stability."
Would that possibly help OP here? I didn't have a chance to try it yet on my thread.

• Done various shenanigans with chipset and display drivers too. Tried both recommended chipset drivers from Gigabyte and latest from AMD. Did a DDU for display drivers and installed driver-only (no Adrenaline software).

Have you tried using older WHQL display drivers? Just throwing that out there cause sometimes it seems the newest driver can be more buggy from time to time.

 

[goshsowitty]

Are you still within return window when you can go back to the store and quickly swap for a replacement 5600X? That can quickly rule out two possible culprits at this time (CPU and RAM).

Not quickly because purchased online, although I can probably just buy some new RAM and sort out the return of the Corsair stuff later just to make things a little quicker.

But after the recent judgement of the RAM I decided on originally I'm really hesitant and worried I might make a bad choice again.

@ShrimpBrime you recommended the G.SKILL RAM earlier. This isn't quite equivalent in terms of value. I have 2 x 16GB which cost me £84.99. You suggested 2 x 8GB for not much less. Can anyone recommend anything that doesn't completely make you baulk but roughly equivalent value?

Essentially I just want something that works and may be more reliable/compatible. I'm not interested in overclocking and RGB on RAM isn't a priority for me.

So if there's a compromise to be had based on those parameters I'll just buy a set and we'll know once and for all if it's the RAM or not. That being said, I'm pretty sceptical at this point.

 

[Deleted member 185158]

Not quickly because purchased online, although I can probably just buy some new RAM and sort out the return of the Corsair stuff later just to make things a little quicker.

But after the recent judgement of the RAM I decided on originally I'm really hesitant and worried I might make a bad choice again.

@ShrimpBrime you recommended the G.SKILL RAM earlier. This isn't quite equivalent in terms of value. I have 2 x 16GB which cost me £84.99. You suggested 2 x 8GB for not much less. Can anyone recommend anything that doesn't completely make you baulk but roughly equivalent value?

Essentially I just want something that works and may be more reliable/compatible. I'm not interested in overclocking and RGB on RAM isn't a priority for me.

So if there's a compromise to be had based on those parameters I'll just buy a set and we'll know once and for all if it's the RAM or not. That being said, I'm pretty sceptical at this point.

I'd love to suggest top tier kits, but that's a 300$ adventure I'm not sure you're willing to take.

A lot of ddr4 is RGB because it's in the middle of the RGB era of computing. So most kits flashy flash.

The 16gb suggestion is because less memory would be less strain for your AMD cpu.

Many forget XMP profiles on memory kits are mostly an Intel thing. A lot of kits where designed around the z170 and up platforms. AMD conforms to XMP profile by using the terminology DOCP, which memory kits are not sold as DOCP, but may be labeled "AMD compatible".

I just know from some 10 to 15 thousand posts at LTT Forums that Corsair LPX and most AMD systems just don't play well together.

Looking back, 2 things where replaced. CPU and memory kits. I'd start with a better kit first. If the issue where to remain, RMA the cpu.

 

[qubit]

I had highly intermittent problems like this with my PC for years where it might BSOD on the desktop, doing nothing, but ran just fine when stressed and on memtest. Turned out to be a bad RAM stick once the problem finally got worse and could be properly troubleshot. Replacing it fixed it. You seem to have ruled out the memory though, so it might be the PSU potentially, or perhaps mobo. Good luck.

 

[goshsowitty]

I'd love to suggest top tier kits, but that's a 300$ adventure I'm not sure you're willing to take.

A lot of ddr4 is RGB because it's in the middle of the RGB era of computing. So most kits flashy flash.

The 16gb suggestion is because less memory would be less strain for your AMD cpu.

Many forget XMP profiles on memory kits are mostly an Intel thing. A lot of kits where designed around the z170 and up platforms. AMD conforms to XMP profile by using the terminology DOCP, which memory kits are not sold as DOCP, but may be labeled "AMD compatible".

I just know from some 10 to 15 thousand posts at LTT Forums that Corsair LPX and most AMD systems just don't play well together.

Looking back, 2 things where replaced. CPU and memory kits. I'd start with a better kit first. If the issue where to remain, RMA the cpu.

How do we feel about Kingston Fury? It's on the AMD QVL, a good price, 3600MHz rather than 3200MHz.

Seems like a pretty good deal, plus I can get it quickly which satisfies my urgency to put this to bed once and for all!

https://www.amazon.co.uk/dp/B097HPN1S3?th=1

 

[Deleted member 185158]

How do we feel about Kingston Fury? It's on the AMD QVL, a good price, 3600MHz rather than 3200MHz.

Seems like a pretty good deal, plus I can get it quickly which satisfies my urgency to put this to bed once and for all!

https://www.amazon.co.uk/dp/B097HPN1S3?th=1

You've got nothing to loose at this point.

 

[OkieDan]

I'm not convinced it's RAM in fact I doubt it is if you can't get memtest to fail on XMP profile and you still cant run stable using a JEDEC profile, especially one stick at a time... but I suppose it's always handy to have multiple kits on hand for troubleshooting.

 

[phanbuey]

Ran TM5 for hours using the @anta777Extreme1 profile and I think it did 3-4 cycles without errors or crashing.

I may fiddle with voltages and stuff. I'm not seeing this as a solution, necessarily, but for validation that the RAM is/isn't working as expected.

To address some other comments:

• The PSU has already been swapped and the issue persists.
• The BIOS has already been updated to the latest version and the issue persisted before/after the BIOS update.
• Done various shenanigans with chipset and display drivers too. Tried both recommended chipset drivers from Gigabyte and latest from AMD. Did a DDU for display drivers and installed driver-only (no Adrenaline software).
• I agree the PSU wattage should be fine by most accounts, at least in theory. Monitoring I've done suggests usage is way under 650W.
• Temps are fine. I ran Furmark and HeavyLoad simultaneously for an hour this morning where CPU, GPU and RAM were all maxed and temperatures were stable and hovered around 80 degrees.
• The issue does not arise in all games. Most games from a few years ago are fine. Fortnite (which is fairly graphically intensive these days) is even fine if we turn down the settings a little.

Does anyone have any thoughts as to how this is manifesting itself from a software perspective? For example the dump files are pointing towards Authenticamd.sys is this appearing in the dumps caused by the crash or could it be causing the crash?

I'm leaning towards it being a hardware issue and that's simply how it manifests itself in the crash reporting.

Based on further Googling around similar WHEA-logger errors and BSODs related to Authenticamd.sys it would seem to suggest people may have had some success with replacing or re-seating the CPU. People are bad at coming back to forum threads and posting a conclusion so it's not 100% clear but I'm going to turn my attention to the CPU for now and see if anything makes a difference.

To everyone who has replied so far, I am extremely grateful. I've had far more input here and a little on the AMD forum compared to other places I've posted like Reddit where I've had hardly any replies. Firm believer that forums like this are the cornerstone of the internet and you're all doing a great job and being a huge help. Thank you!

Have you tried raising the voltage curve? It sounds like you're crashing when the cpu is transitioning in power states, one of the states/cores is just barely unstable and the power savings is too aggressive. This is actually not that uncommon.

 

[Deleted member 185158]

I'm not convinced it's RAM in fact I doubt it is if you can't get memtest to fail on XMP profile and you still cant run stable using a JEDEC profile, especially one stick at a time... but I suppose it's always handy to have multiple kits on hand for troubleshooting.

Very likely the memory.

I will example helping someone with a link.
Many many threads helping people with AMD and certain memory kits.
There can be many settings to change to get a stable system.
It's easier to suggest a better memory kit most of the time. As you see 3 pages deep, so far no resolution.

OP can follow this link, this person had purchased 2 sticks of memory at different times.
However the symptoms of AMD memory issues always seem to follow 1 of 2 paths.
First path is memory kits supposed compatibility, yes some user tweaking may be needed, see above 2 pages.....
Second path generally at a much smaller percentage, CPU IMC is burnt. That's forever whea errors (can go deeper, but not rn)

RAM not being able to sustain stock speeds

Hey guys! I am new to this forum, but there has been something, that has been bothering me about my PC: Some time ago I bought a second stick of RAM (which is closely matched, but certainly not the same as the other one) for my desktop, so I set them to use the slower timings and speed that were ...

linustechtips.com

No I don't participate any longer over there. Spending my time elsewhere these days.

 

[OkieDan]

Very likely the memory.

I will example helping someone with a link.
Many many threads helping people with AMD and certain memory kits.
There can be many settings to change to get a stable system.
It's easier to suggest a better memory kit most of the time. As you see 3 pages deep, so far no resolution.

OP can follow this link, this person had purchased 2 sticks of memory at different times.
However the symptoms of AMD memory issues always seem to follow 1 of 2 paths.
First path is memory kits supposed compatibility, yes some user tweaking may be needed, see above 2 pages.....
Second path generally at a much smaller percentage, CPU IMC is burnt. That's forever whea errors (can go deeper, but not rn)

RAM not being able to sustain stock speeds

Hey guys! I am new to this forum, but there has been something, that has been bothering me about my PC: Some time ago I bought a second stick of RAM (which is closely matched, but certainly not the same as the other one) for my desktop, so I set them to use the slower timings and speed that were ...

linustechtips.com

No I don't participate any longer over there. Spending my time elsewhere these days.

I still have my doubts, but I do think another kit will be helpful for troubleshooting by either fixing the issue or ruling out the RAM. Hopefully OP will keep us updated regardless. I'm glad I didn't have issues with the 4x16GB from two LPX kits I ran with my old 5600x.

 

[Haytham]

With all respect to all the members and their responses, but this problem happened to me before, and over the course of two months I tried everything, and at the end I replaced the processor and the problem ended, knowing that the new processor works fine on the same hardware.

 

[eidairaman1]

With all respect to all the members and their responses, but this problem happened to me before, and over the course of two months I tried everything, and at the end I replaced the processor and the problem ended, knowing that the new processor works fine on the same hardware.

Not everyones case though

For 23 years experience the issues of a ram kit causing problems is still the most common, and Corsair has been the biggest villain on Ryzen that I do not suggest them at all.

 

[goshsowitty]

With all respect to all the members and their responses, but this problem happened to me before, and over the course of two months I tried everything, and at the end I replaced the processor and the problem ended, knowing that the new processor works fine on the same hardware.

I also believe it is the CPU at this point but we will find out tomorrow.

But I feel like all of this process is required. It’s a difficult situation going back to a retailer asking them to exchange or return something that may or may not be faulty.

At least by the time we’ve excluded the RAM - if we do - then fingers are most likely pointing to the CPU as the culprit.

I guess that’s when we’ll find out it’s the motherboard, anyway