[GPU] @ DEV Team - This specific Zone Rendering is a perfect zone to troubleshoot FPS/Rendering Issues [idle GPU?]

[hyperf0x]

Video @ https://www.reddit.com/r/newworldgame/comments/pyzqja/gpu_dev_team_this_specific_zone_rendering_is_a/

(Look at top left corner GPU monitoring)

I have no clue why this does happen, in technical terms it doesn’t even make sense and it might be reason why some MOSFET’s are going to card heaven by putting sudden random voltage load in the coils by looking at certain angles multiple times randomly.

I don’t know why this is happening at this specific frame and possible thousands of other angles ingame. But it makes no sense to my rendering knowledge.

PROBLEM: Looking at this specific frame causes the card to go dormant/undervolted from 1.1 to 0.7. Everything starts dropping, FPS drop from 120+ are forced down to 30-35. The Clock drops, the Wattage consumption drops and the Fans shutdown.

Now the weird thing is… This image rendering has absolutely no sense why it would drop or force the card to “chill” when it’s actually a more complex scene than looking at the ground/sky or even around!

So here’s the logical Paradox:

• Looking at the sky, or ground which has the SAME scene complexity for the GPU, has the GPU running at high clock speed which makes it deliver a steady 140+ FPS. Looking pretty much anywhere that delivers with MORE render complexity also has high/steady FPS’s because it forces the clock speed to go up to render.

The Paradox - Yet all the time I look at this particular angle I lose a good 300% FPS which in normal technical terms emphasizes that 1) Either the GPU is under stress by the rendering complexity and the FPS’s are dropped because the GPU cannot keep up with the rendering, but that means the clock speed and voltage are at their maximum OR 2) The FPS’s drop because the card goes in idle state, the voltage drops, the clock drops and therefore the FPS’s drop.

Now the thing is… I don’t know why this specific angle is forcing the card to go idle (or lose FPS’s without the clock @ max (??) ) when the scene complexity is the same as looking down on Y axis or anywhere else where there is small rendering to be done.

Can someone else try to reproduce my exact location found in the video and confirm this also? I am pretty sure this is embedded in the engine/DirectX call…

[hyperf0x]

On my investigation this seems to go on par and on-pair with this issue:

https://forums.evga.com/Fixing-EVGA39s-7-Figure-Problem-with-FTW3-30-Series-cards-m3217284.aspx

Where it is figuratively a card’s issue. But it kind of makes sense why there are so many cards (included to this game) having issues. Not only to this game, same goes with HALO and LOL. But I am pretty sure that the voltage spikes, mainly from the Undervolted → 1.1v stage at high clocks are causing some components of the card which were not properly tested/Quality Controled to degrade quickly and brick the cards.

I don’t really want to test it but it seems these cards are having weird voltage regulators when going from low/idle voltages to high voltages - someone was able to capture some but they were so fast in voltage spike that standard capture devices had a hard time catching it. These are repicable ingame - which aren’t “optimal” but cause no harm - except in some of the cards, I suppose due to manufacturing, that aren’t handling these voltage spikes.

And this games seems to be providing a lot of situations where the card loses voltage and then spikes up to 1.1v which is the firmware stock and somehow degrading the components along the way.

My 3070 started coil whining after playing this game, and I assume it’s due to the constant voltage change, putting a more-than usual stress on the coils and degrading a bit of the soldering/glueing.

Regardless, this would undoubtedly be a 100% issue of the hardware of the card. The game just simply provides more events for this to happen - but some cards would just be a time bomb anyway. And, I assume, this is due to the manufacturing of the voltage controllers or some cards who shouldn’t’ve had passed Quality Control in the first place. Then again this is all speculation, that makes some sense.

Like someone mentions on the EVGA forum:
The issue with replicating this problem is I think you need games or benchmarks that vary significantly in performance, i.e pushing the card up and down the V/F curve (3DMark etc tends not to do this). Problem is this actually does happen a lot in games when you transition through cut scenes, enter menus etc, but often can take 1h+ for the issue to arise.

And this game has a very floppy performance and provides a lot of these odd voltage spikes, mainly that skybox angle I’ve provided that forces the card to go undervolted somehow.

[PirateLemon]

This should get more attention. Regardless who people blame, getting to the root of this issue is the main thing to do.

Have a like!

[ItsMySecondDay]

This needs attention, also able to reproduce this in multiple places across the map, Including this one.

[landdon]

The problem is, even if it’s hardware related, you’re not going to get card replaced just because of a game. I think this is a deal breaker for me. I’m not frying my card for a game.

[hyperf0x]

Nah it’s definitely not hardware related. It seems it’s just dorky programming.