Jump to content

RTX 3070 / 3080 / 3090


hobbyist

Recommended Posts

@DasFrodo Ah, I didn't know that - I thought the capacitor type was the issue. Any reference to that uodated info? I'd like to read / view it.

 

Total speculation on my part... But could they be covering this up a little? If it was the capacitor, then my guess is that replacing that physically would not be a simple swap - different pins etc, so would require redesign, new tooling etc = very expensive. Jayz did say, in that first video report, that this could be worked around using a driver update to reduce max boost clocks. So I wonder if 'its just a driver issue' might be an excuse?

 

Link to comment
  • Customer
31 minutes ago, Mike A said:

@DasFrodo Ah, I didn't know that - I thought the capacitor type was the issue. Any reference to that uodated info? I'd like to read / view it.

 

Total speculation on my part... But could they be covering this up a little? If it was the capacitor, then my guess is that replacing that physically would not be a simple swap - different pins etc, so would require redesign, new tooling etc = very expensive. Jayz did say, in that first video report, that this could be worked around using a driver update to reduce max boost clocks. So I wonder if 'its just a driver issue' might be an excuse?

 

 

Really does just look like faulty driver issue affecting Windows machines. Hardware Unboxed done a summing up here

 

 

Link to comment

@hobbyistThanks - that's useful info and sounds like a sensible analysis from the Hardware Unboxed guy. I'm not a great Jayz fan - but that was the only video I'd previously seen on the issue! Moral of the story: don't always believe what you read on the interwebs : )

Link to comment
20 hours ago, hobbyist said:

 

Yes the capacitors look like a red herring and the latest driver update resolved crashing issues.

Well if the community was right the capacitors could still have been the issue. Apparently those cards crashed because they clocked sliiiightly into instability, so if the newest driver just lowered the max GPU clock it's basically a bandaid.

Visit my Portfolio on Artstation.com & Follow me on Instagram

Link to comment
  • Customer
17 minutes ago, DasFrodo said:

Well if the community was right the capacitors could still have been the issue. Apparently those cards crashed because they clocked sliiiightly into instability, so if the newest driver just lowered the max GPU clock it's basically a bandaid.

 

I have seen some videos where the 30 series is described as one of the worst power to performance cards that leaves no room for overclocking. 

I don't think it's as straightforward as lowering max GPU clock, but this isn't really something I'm knowledgeable on. I'm practically just repeating what I've seen in small sample of videos...even at that I might be miscommunicating it 😄.

 

JayTwoCents goes over it from around 8min 40sec 

 

 

My main interest in the cards are how they render with the different 3rd parties. Maybe overclocking is a thing in the 3D rendering world? I've got a FE 1080Ti and don't overclock that. 

 

There's probably a gap in the YouTube market just for just focusing on rendering different scenes with render engines on different rigs. Hmmmm 🤔

Link to comment

I would never overclock a gpu for rendering, there is so little speed to be gained, many cards are already 99% up against their thermal and power limits at stock. I always laugh when I see "Super OC edition" gpus released, where they bump the clock speed from 1720MHz to 1760MHz as if it matters.

 

Re: The POSCAP issue, well it sort of is and isn't a driver problem. Yes they have fixed the problem in a driver update, but that is a software fix to a hardware problem as far as I have read. If my car engine explodes from overheating, they could still issue a software fix to prevent it from being allowed to rev high enough to break. 

Link to comment
16 hours ago, imashination said:

I would never overclock a gpu for rendering, there is so little speed to be gained, many cards are already 99% up against their thermal and power limits at stock. I always laugh when I see "Super OC edition" gpus released, where they bump the clock speed from 1720MHz to 1760MHz as if it matters.

 

Re: The POSCAP issue, well it sort of is and isn't a driver problem. Yes they have fixed the problem in a driver update, but that is a software fix to a hardware problem as far as I have read. If my car engine explodes from overheating, they could still issue a software fix to prevent it from being allowed to rev high enough to break. 

 

Your point is logically sound. However, according to the evaluation by Hardware Unboxed, the new drivers resulted in such a nominal down-throttling that it was almost irrelevant. And early reports are encouraging that it did resolve the crashes.

 

To be cautious I wouldn't buy one in the next week or two (even if I could) but I think the issue will die out and the card will be a big success.

 

 

Link to comment

Igor, the Guy who found out about the POSCAP issue already analyzed the new driver version.

 

https://www.igorslab.de/en/wonder-how-invidia-the-crashes-of-the-force-rtx-3080-andrtx-3090-will-be-removed-and-still-will-be-removed-even-from-the-power-supplies-analysis/

 

if you understand german  ( I dont think the subtitles are to good) here is the video which is much more detailed:

 

 

best regards

Jops

Link to comment
×
×
  • Create New...

Copyright Core 4D © 2023 Powered by Invision Community