Hi EVGA Team & Community!
 
I have a question about the 3080 Ti FTW3 Ultra performance:
I just posted on official bios thread that i had to re-flash both bios version to original ones from the Bios Thread and upgrade them but this time without fastboot enabled under my bios.
Everything went great and i am really happy for that cause i was thinking that my card is broken or whatever :) But then after upgrading the new bios/last bios from july 2021 on the normal mode
to this version: VBIOS Version: 94.02.71.80.74
https://www.techpowerup.com/vgabios/235802/evga-rtx3080ti-12288-210712 my issue started again. What is my issue? Well the card is getting to hot and somehow i see stuttering but only in game (i only have fortnite installed or at least i only play this game. I could test some others if needed.
 
Since english is not my spoken language i will try to explain this with screenshots i have done.
https://ibb.co/cyBCksy px1 start - seem all normal
https://ibb.co/r2pdmQs what you can see on this screenshot is a game running in the background...
IMPORTANT:
All the stutterings and unstable FPS or as you can see the memory clock changing from 95xx to half ... this is happening while i'm IN GAME...
as SOON i get PX1 on the screen to do screenshots - as you can see the HW Monitor shows me everything stable.
https://ibb.co/L1pDd2Y ^ same as 2nd Link, from HW Monitoring from PX1 but just more or other sensores. You can compare them you will see it match - there is a clear picture when is wwhat happening when "In-Game vs. Notingame" but the game was running in the background.
https://ibb.co/LC1mM1k GPU 10min Stresstest or Test with OCCT
 
This was working under my OC bios from the first day i recieved my card until, as written, few days ago when i found i never tested the normal bios, went under the normal bios and upgraded to the latest version.
 
What are other differences between NORMAL vs. OC Bios? I always read "only fan curve and normal has 0% spin option" but if so why would my PC handle the one same game so unstable? It has to be something which tells the card that if the Game somehow is maybe not so demanding, or like i run fortnite under performance mode that the card should not clock to it's max performance.
Ps:
Yes i reinstalled the drivers / used DDU latest version from Guru3D / i also runed HYDRA latest version today and OC my system again,
i dont have any other issues, my bios as written is on the latest version, no other monitoring tools but i tested GPU 10mins as you can see on screenshot to see if there will be errors but non found.
 
find my PC Specs or Hardware below:
 
Driver Easy TXT Report
-------------------------------------------------------------------------
Overview
    Machine                ASUS
    Operating System        Microsoft Windows 10 Home (64-bit)
    Memory                32 GB
    Processor            AMD Ryzen 9 5900X 12-Core Processor            
    Motherboard            ASUSTeK COMPUTER INC. - ROG CROSSHAIR VIII DARK HERO
    Video Cards            NVIDIA GeForce RTX 3080 Ti
    Monitors            ROG PG259QN
    Hard Disk            WDS500G1X0E-00AFY0 (500.0 GB)
    Network Card            Intel(R) I211 Gigabit Network Connection

CPU

    Specification            AMD Ryzen 9 5900X 12-Core Processor            
    Family                15
    Revision            B0
    Instructions            MMX (+), SSE, SSE2, SSE3, SSSE3, SSE4.1, SSE4.2, SSE4A, x86-64, NX, VMX, AES, AVX, AVX2, FMA3


Clocks
    Core Speed            3699.1 MHz
    Multiplier            x 37
    Bus Speed            100 MHz
    QPILink                --
^ CPU OC done with latest Hydra from 1usmus.
^ my RAM is tuned manual also done with Ryzen DRAM Calculator from Yuri and timing from Yuri approved and no errors when checking my Ram.
I have also CMOS bios and run only XMP or DOCP ram without to change anything manual and left everything else on Auto within the bios
but the issue was the same.

Cache
    L1 Data                32 KBytes
    L1 Inst                32 KBytes
    Level 2                512 KBytes
    Level3                32768 KBytes

Motherboard
-------------------------------------------------------------------------


Motherboard
    Manufacturer            ASUSTeK COMPUTER INC.
    Model                ROG CROSSHAIR VIII DARK HERO

BIOS on the Motherboard <- latest from the Asus Dark Hero web-page.

    Core Speed            American Megatrends Inc.
    Version                4006
    Date                03/07/2022

Memory
General
    Type                DDR4
    Size                32768 MBytes
    Channel#            --
    DC Model            --
    NB Frequency            --


Timings
    DRAM Frequency            1899.6 MHz
    FSB:DRAM            --
    CAS# Latency(CL)        14 clocks
    RAS# to CAS# Delay(tRCD)    15 clocks
    RAS# Precharge(tRP)        16 clocks
    Cycle Time(rRAS)        34 clocks
    Bank Cycle Time(tRC)        --
    Command Rate(CR)        1 clocks




Graphics
GPU
    Name                NVIDIA GeForce RTX 3080 Ti
Memory
    Size                12288 MBytes
    Type                15






 

Sajin
If you go back to the older vbios the issue goes away?

Thanks Sajin for the fast Question. I just updated the post with screenshots and tried to explain.
I just re-flashed the normal bios on the card with the first one from Bios Thread and did not upgrade it yet. I will do a test and will post here an update.
 
Ps: comparing to normal bios, when gaming under the OC bios i see under PX1 / Hardware Monitoring everything maxed out. There is nothing like "card getting used only as much as needed"
Under Normal bios i also tried to lock frequencies and under nvidia pannel i tried to put maximal performance but still did not worked out. PX1 locked frequencies but performance was still bad.
This also could be re-produced after re-flashing all to original and upgrading to last versions without fastboot enabled.
 
Yes, i have not tested with the first bios. Will do soon and update the thread.
Thanks
Dali

Hi Sajin,
 
The game itself can be played but i see stuttering when i change direction for example and that happeneds really fast. I also note a huge lag compared to oc mode because the whole moving feels laggy ingame. Below the screenshots made from hardware monitoring and also CPU vs GPU idle temp before i started any application... there are 24c difference between in idle mode. Of course i dont know how much would temp be under OC mode if the fans were off (haha) but it's still really hot.
And the strange point is: If you take a look from one PX1 screenshot while i was in-game just before - you will not note any big temp. difference.
Yesterday (please EVGA Team / Community and GPU Gods forgive me...) i installed GeForce Expirience after did clean new driver install with DDU and played the game with DX12 with all settings hit on maximum... And with everything set to maximum the card performed... i still had to turned off after that effects / so much shadows and so on BUT it was not so unstable with memory clock and mhz clock and so on.
 
Is there any information avaiable about thresholds and how which bios on this card is handling the loads?
 
https://ibb.co/yf5F03n - HW - Monitor gpu usage / gpu voltage / framerate.

https://ibb.co/6tcMB07 - HW - Monitor gpu usage / gpu voltage / framerate.

https://ibb.co/gyG2WDS - 58c with fans activated (under load) ... if there is no load the card will go also that hot or even hit up 60C without to activate fans.

https://ibb.co/WDnqdgs - PX1 HW Monitor in-game pt. 1

https://ibb.co/SsDkG3h - idle temp CPU vs GPU

https://ibb.co/Yfc6qMS - PX1 HW monitor in-game pt. 2
 
Normal Bios was flashed back to the version: VBIOS Version: 94.02.71.00.AF
 
I noted a difference, but it's a small one. I dont use msi afterburner but yesterday i also installed it and saw that the powerboard limit and also same for geforce expirience tool, can be extended for max. "12%". PX1 allowes +13.
 
I dont know... still does not feels right. I dont know if the card also is not performing because the temps from the beggining are very high and GPU Memory is between 60 and 62degrees? but it's a totally different behavior on the same hardware.
 
Thanks
Dali

For now. Let me see what i can find on my steam account.. i believe i should have Immortals Fenyx Rising if i‘m not wrong.

Ps: feel free to tell me if there is no issue at all and this behavior on the gpu is very normal? I mean - only because i see something for the first time doesnt means that this is an issue…

Will update this thread as soon i have done some more tests.

Ps: just to answer the question about the cpu temps: my cpu stays around 47/50c and hits oe has peaks sometimes around 60/60+ but only peakwize.

i also never seen memory clock jumping around like that or that temp stays where is on idle…
4K is something i would need to fake i guess since i only have 1 1920x1080 monitor (pg259qn) running at 360Hz with very good DP cable and GSync activated. I dont have other monitors at home.

But i can try with dx11 and everything on max with and without dlss and dx12 with ray tracing ti see if the card will perform better.

Until done i can only imagine that on normal bios the card is somehow „pre-checking“ how much resources it takes to run the game but again - memory clock shoudnt be jumping even the clock speed i can agree can jump. Memory usage is ok if goes up and down.

Hi rjbarker,
 
Yes i did but i would never run 1T. I have 2T gear down disabled and power down disabled.
https://ibb.co/7V7ZwW1
I've attached a screenshot with my RAM timings.

Sorry for feedback delay. i'm, to be honest, still shocked but it  turned out that my issue has been resolved with uninstalling the PX1. Before uninstalling Px1 i tried following steps:
as suggested back on "no docp / xmp on" even me and my oh-great performance were bleeding inside... but did not helped.
bios: load optimal defaults, only xmp on, all stock - did not helped
got my profile in bios back, reloaded, safe mode without network, ddu 1.0.4.9 version (last one) clean deinstall...after that safe boot again - latest nvidia drivers on - nothing happened.
PX1 removed - downloaded .exe again from the evga website - re-installed - still same. the sensors were just flickering.
I decided to remove / deinstall (over windows app manager - just normally deinstalling obvious) the PX1, no clean up, not registry cleaning, nothing... installed msi afternburner and see the screenshots:
 
https://ibb.co/0nSDyQF there is one straight long line on the middle-left side where the cpu goes 4925/4950 for a veery long period of time... the up and downs are not concirning since the other threads are taking over so all fine. my FPS counter feels like smashed to that 360fps counter
 
https://ibb.co/VB3RBvr  and here is the exact same FTW3 Ultra 3080Ti.... no hickups. Smooth, great performance.
and here is the https://www.3dmark.com/3dm/74496608 Firestrike - Score "40472" (did some people complained that ftw3 3080 ti ultra cannot reach more than 400W? I reached in basic firestrike today up to 445-446W...
 
BUT... i'm still MORE than interessted to find out what did wrong with px1 and what changed after the PX1 was removed? That's really strange.

PX1 strikes again. 

Sajin
PX1 strikes again. 

oh i looked now on google and been found lot of reporting for this kind of issue.

But i might know then what is the reason for the issue and why i did not had it before.

Since i only, as written, play fortnite on my desktop, i, for the first time at all, tried the OSD feature since i read fortnite been added.

Once enabled - things worked broken even after reinstalling px1 and even clean reinstall of new nvidia drivers in safe mode.

Maybe there is someone from evga interested at this thread from evga px1 team.

Could be that i am totally wrong but since enabling osd i run into this issue, not before. not with any other cards and i had 2080ti, 3070 ftw3, 3070ti ftw3 ultra, 3080 10gb no lhr and now 3080ti and i was always using the original evga tool but this was my osd try.

But there are also few other things not working in px1 correct which needs to be fixed.

Sajin
Yeah, when I said px1 strikes again I was making a comment to your reply that px1 was causing your issue. This isn’t the first time px1 has caused issues for someone.

there is the MSI with fantastic afterburner but buggy dragon center which made me give up on msi. Then i see so many people loving gigabyte... i absolute could throw-up when i take a look how modern and well-working is their software... That's really sad for 2020 and later...
 
And we have EVGA and i love these cards but we again have PX1 and that's evga own tool not working as it should...
 
I will of course stay with EVGA since i had so many other cards from so many different vendors but i cannot understand how can a company made such a great hardware and let their own highend hardware to be managed by some buggy tools... and even worse - it works better if you do management with a third party application. And oldest posts are back from 2017/2018.... It cannot be that complicated.
Also changing from main window to an another under setttings even i dont have to change any setting in PX1, i get the option to click on apply. Why should i apply the same configuration?
 
But thanks for this input! it was very helpful!

Chesty Puller
*EVGA RTX 3080 Ti FTW3 ULTRA GAMING FIX* (workaround)
I found if I just disable G-sync all stuttering and refresh rate varying stops and I get over 2k fps framerates that I then enable "Max Framerate" setting to a comfortable 200 fps to keep my card from bursting into flames. I really am not a fan of all this variable framerate crap they are trying to tell me is supposed to "work so well" because it just doesn't.
 
Before: monitor refresh rate set to 160fps 4k 1440 and I was getting horrible performance lags, stuttering, and varying refresh rates to as low as 48 fps
 
After: now have steady refresh rate stability at solid 160fps and doesn't budge a bit and no stuttering or performance issues.
 
This is how my setup will run it until they fix the issues with whatever is causing G-sync to be all over the place.

G-Sync is VRR. It is not supposed to be fixed refresh rate. It really does work well preventing screen tearing and increasing smoothness BELOW your native refresh rate.
 

HeavyHemi

Chesty Puller
*EVGA RTX 3080 Ti FTW3 ULTRA GAMING FIX* (workaround)
I found if I just disable G-sync all stuttering and refresh rate varying stops and I get over 2k fps framerates that I then enable "Max Framerate" setting to a comfortable 200 fps to keep my card from bursting into flames. I really am not a fan of all this variable framerate crap they are trying to tell me is supposed to "work so well" because it just doesn't.
 
Before: monitor refresh rate set to 160fps 4k 1440 and I was getting horrible performance lags, stuttering, and varying refresh rates to as low as 48 fps
 
After: now have steady refresh rate stability at solid 160fps and doesn't budge a bit and no stuttering or performance issues.
 
This is how my setup will run it until they fix the issues with whatever is causing G-sync to be all over the place.

G-Sync is VRR. It is not supposed to be fixed refresh rate. It really does work well preventing screen tearing and increasing smoothness BELOW your native refresh rate.
 

And I will clarify and state again, I am not a fan of this VRR technology when it causes performance degradation like I have been experiencing. I would rather experience tearing, of which I am having practically none without the use of VRR. What I am attempting to say is that it seems to be that there may be some issue with the algorithm used in the application of VRR. Without the use of VRR, I am experiencing better performance because of the glitching it is causing when it is used. Until they find a better way of implementing this without the "jerkiness and stuttering" it causes, I will not be using it and I urge others to try this and judge it for themselves. At this point currently, I am having a hard time believing that VRR is actually working as it was intended to be. This so called "smoothness" that you speak of, I wish it was, but it just isn't there for me. When the problem has been evaluated and rectified, I will revisit whether or not it is in fact working as intended.