NVIDIA GeForce 536.67 WHQL Drivers Released

[Deleted member 232534]

After issuing command "nvidia-smi -i 0 -lmc 810,10501" the idle memory is up from ~100mhz to ~200mhz and it might lowered the DPC a bit, nvidia driver is not the one causing the highest latency.

The idle power consumption is up 4watt. From ~15watt to ~19watt.

I have the power saving options turned on.

Improve DPC Latency Spikes For Ada Lovelace-based GPUs in R536.67 Driver

Improve DPC Latency Spikes For Ada Lovelace-based GPUs in R536.67 Driver NVIDIA driver R536.67+ has improved the DPC Latency Spikes issues. But some...

forums.guru3d.com

 

[Mussels]

Those values look reasonable considering you've got monitoring software running, which does poll the GPU
anything that alters the display will cause a spike, like enabling or disabling another display, changing to fullscreen exclusive etc etc

That link mentions

'When I open the web browser. That maybe produces a higher spike of DPC Latency.'

Because they're opening a browser that's using 3D acceleration, it's changing the 3D profile of the GPU and locking it to a higher state


Seems like Nvidia lowered the VRAM clocks to get idle power consumption down, but on some setups it's too low?
This shouldn't cause issues anywhere except an idle 2D desktop, since any load whatsoever would keep the VRAM clocks higher

 

[Klemc]

I have brighness variances (flashings) in games loading screens since this update...

//

Only on my Asus OLED 27" 240

 

[Carlyle2020hs]

I run a 13900ks without ecores with Win10 and no dpc problems until this update.

Now the tables have turned and the internet praises this update and i curse its devs.

Not even a partial activation of ecores does help.
I has has to be all of them before i can understand what the rest is praising.

I'm at a loss on what´s going on.
msi is active with undefined priority for my 4090.

 

[Mussels]

I run a 13900ks without ecores with Win10 and no dpc problems until this update.

Now the tables have turned and the internet praises this update and i curse its devs.


View attachment 306106

Not even a partial activation of ecores does help.
I has has to be all of them before i can understand what the rest is praising.

I'm at a loss on what´s going on.
msi is active with undefined priority for my 4090.

you'll need to roll back a version and do before/after clean installs to be sure nothing else got in the way, windows updates for example love to sneak in and screw things up


mine looks nothing like yours, do be careful not to crop things so much that its out of context... if i was unfamilar with latencymon i'd have zero idea what you were showing or why

 

[Tomorrow]

that ultimate plan is not a good idea, it's hidden for a reason.
You're breaking the cores being able to sleep, which harms their boost performance.

You should avoid things like that in the future, as they break a lot more than they fix - they often come paired with lots of other tweaks that only help reduce the damage tweaks like that power plan cause in the first place.

As someone running even higher IF speeds than you, it does sound like the traditional infinity fabric dropout - which component drops depends on the board and CPU, but it can be CPU, PCI-E lanes, USB or SATA.

Your VDIMM is really high, which adds the risk of simple RAM instability being the cause too. B-die doesnt like heat at all, and can get unstable as low as 40c with such high voltages - it only takes a few correctable errors to make a stutter, or the high power draw to cause an IF dropout
I'm pretty sure you need to raise other voltages at the same time as VDIMM, but 1.60v is so high i'm not sure it's even safe to do so.

first google result for B-die on AM4 confirmed that idea

There are also people with similar RAM that's hynix, which behaves entirely differently
Bought Viper Steel 4400MHz c19 and turns out to be HYNIX die. Have anyone experienced the same thing? : overclocking (reddit.com)

Your problems not GPU driver related, it's that you're running some extremely uncommon and not recommended tweaks in hardware and software.

As for boosting im running BIOS level per core undervolt so my CPU actually boosts better than stock, consumes less power and runs cooler. For this test i obviously ran at stock settings.
IF is stable at 1867 for me. I have tested stability before with y-cruncher. 1900 wont boot and 1933 boots but throws WHEA errors so i know 1867 is max stable.
I have very good airflow inside my case so high VDIMM is not an issue.
An my sticks use B-Die. I know later revisions introduced Hynix too but that's not my concern.

Nope. Clearly Nvidia driver fault. I closed all background apps, reset BIOS to optimized defaults and selected balanced power plan.
None of those made a difference. With a few minutes of running and starting up a game the DPC latency spiked to 3400+ like before.
Win11 install is fresh (from yesterday) so that, too is not an issue. I even used 536.40 instead of 536.67 in case the newest one had a bug or something.
Any more excuses for Nvidia's poor drivers?

 

[Mussels]

DDR4 2133?

And yes, if you've got similar hardware and the same software and only you have the problem that's a pretty big clue. The problem is either unique to your hardware, or your software.

Either something's not right at a hardware level, from overheating to a faulty device (Any attached device can be responsible here, from a USB device to a PCI-E riser to a network card. ANYTHING can cause the PCI-E bus or OS to have a delay, and have it show as the nvidia driver since it's waiting on external events/data to continue its job.


You can bark up the wrong tree, or you can do a default generic install of the OS on a clean drive and see how that performs, and if the issue still exists you need to investigate your hardware - hows your chipset temps? VRMs? GPU's VRAM temps? Use a thermal probe and don't rely on software measurements if something doesnt math up right.

 

[Tomorrow]

DDR4 2133?

And yes, if you've got similar hardware and the same software and only you have the problem that's a pretty big clue. The problem is either unique to your hardware, or your software.

Either something's not right at a hardware level, from overheating to a faulty device (Any attached device can be responsible here, from a USB device to a PCI-E riser to a network card. ANYTHING can cause the PCI-E bus or OS to have a delay, and have it show as the nvidia driver since it's waiting on external events/data to continue its job.


You can bark up the wrong tree, or you can do a default generic install of the OS on a clean drive and see how that performs, and if the issue still exists you need to investigate your hardware - hows your chipset temps? VRMs? GPU's VRAM temps? Use a thermal probe and don't rely on software measurements if something doesnt math up right.

Optimized defaults sets RAM to JEDEC speeds. Not my choice.
Clearly it's hardware. Nvidia hardware. Cant wait to upgrade my GPU at the start of next year. Hopefully that will end this saga once and for all.
As it stands i doubt i can do anything to fix this without a GPU swap. I do have an old GTX 950 from the Maxwell era laying around. I could try and see if putting that in, instead of 2080 Ti changes something.

 

[Chomiq]

I have brighness variances (flashings) in games loading screens since this update...

//

Only on my Asus OLED 27" 240

Not a driver issue, google "OLED GSync/VRR near-black gamma shift and flickering"

 

[Mussels]

Not a driver issue, google "OLED GSync/VRR near-black gamma shift and flickering"

Was this in one of the changelogs? I've definitely seen it mentioned as an OLED specific issue

 

[Klemc]

Looks like OLED only, Since it doesn't on my VA 4K 60hz panel...

 

[Chomiq]

Was this in one of the changelogs? I've definitely seen it mentioned as an OLED specific issue

What changelogs? It's WOLED tech issue, LG bypasses it by raising gamma on their gaming OLED displays.

 

[Mussels]

What changelogs? It's WOLED tech issue, LG bypasses it by raising gamma on their gaming OLED displays.

Nvidias changelogs? This is an nvidia driver thread

 

[Chomiq]

Nvidias changelogs? This is an nvidia driver thread

Maybe if you would actually follow the thread you'd understand why this was brought up...

 

[R-T-B]

Not a driver issue, google "OLED GSync/VRR near-black gamma shift and flickering"

It happens on all emissive techs that implement gsync to some degree. Fortunately at 120hz I don't really need gsync. Got enough steps in there for plain ol' vsync.