• Competitor rules

    Please remember that any mention of competitors, hinting at competitors or offering to provide details of competitors will result in an account suspension. The full rules can be found under the 'Terms and Rules' link in the bottom right corner of your screen. Just don't mention competitors in any way, shape or form and you'll be OK.

AMD Navi 23 ‘NVIDIA Killer’ GPU Rumored to Support Hardware Ray Tracing, Coming Next Year

Status
Not open for further replies.
g67575 said:
Are you sure? This article: https://www.tomsguide.com/uk/news/nvidia-rtx-3080-crashes-caused-by-capacitors-says-evga

suggests that the crashing problem is caused by the type of capacitors used, and that the RTX 3080 FE isnt affected.
Click to expand...

No one except Nvidia will ever truly know, it is probably a combination of things suggested by everyone so far.

The one thing we can say for sure, the launch was rushed, botched and the cards were not ready to be sold to the public in a desperate bid to beat AMD to market.
 
Tiny_D said:
No one except Nvidia will ever truly know, it is probably a combination of things suggested by everyone so far.

The one thing we can say for sure, the launch was rushed, botched and the cards were not ready to be sold to the public in a desperate bid to beat AMD to market.
Click to expand...

wasn’t just amd they wanted to beat though I think consoles played a huge part also if the leaks are true that a Xbox series x is around the 2080/s level it’s not very good for optics that a console is as good as there high end line:)
 
Th0nt said:
FE still crashed, they didnt know what was causing it. New driver stopped a lot of the crashes. I believe there are still people with CTD etc. but it could be down to poor PSU or other system weaknesses.
Click to expand...
Have been looking around.. new driver increased tdp by 10w and shaved off 30mhz from core clk for increased stability.. that seems to have solved the issue for many folks
 
Nablaoperator said:
Have been looking around.. new driver increased tdp by 10w and shaved off 30mhz from core clk for increased stability.. that seems to have solved the issue for many folks
Click to expand...

Like I said "stopped a lot of the crashes". People are still having them. From an experienced AMD user, just like the 5700XT black screens, you can try defend it, but when general population can have under-specced parts, poor PSU, bad cables or daisy chain pin delivery from the same rail - owners will still have the issues no matter what "solved the issue for many folks" you say.
 
Th0nt said:
Like I said "stopped a lot of the crashes". People are still having them. From an experienced AMD user, just like the 5700XT black screens, you can try defend it, but when general population can have under-specced parts, poor PSU, bad cables or daisy chain pin delivery from the same rail - owners will still have the issues no matter what "solved the issue for many folks" you say.
Click to expand...

Yup there are many factors.. but with a driver level change of that sort it seems more like chip level issues... rather than caps.. that's what I was trying to imply. Just a speculation.. rest is outside nvidias control anyways
 
If its just a case of insufficient power being supplied, NV should offer refunds / working replacements for the FE 3080s where crashes are still occuring, and modify the recommended TDP. Customers will (rightly) be expecting that the company with a better reputation (Nvidia) would have the money, resources and time to design and test their graphics cards, before they are released.
 
g67575 said:
If its just a case of insufficient power being supplied, NV should offer refunds for the FE addition where crashes are still occuring, and modify the recommended TDP. Customers will (rightly) be expecting that the company with a better reputation (Nvidia) would have the money, resources and time to design and test their graphics cards, before they are released.
Click to expand...

there are to many moving parts for even a company like nvidia to get it right. They spent a long time getting everything ready for launch and did all they could to get the design right at least. But not just the card themselves is the issues you have a bios that AIB edit for there won cards then you have drivers controlling all aspects of the card only takes 1 line of code to be wrong and the whole system can fail.
 
No acceptable excuse for the FE Rtx 3080s crashing though. They are designed and built by Nvidia. They can test GPUs and drivers before launch until they are ready. Delay a launch a bit if needed, or dont set a date until ready, preferably. They can certainly afford to test thoroughly with all the profit they make from GPUs.

Better to reduce GPU clocks to achieve stability, than increase them and give a false impression. For the high price, I'd be pretty disappointed if crashes occured, and I had a sufficient power supply installed.
 
Last edited:
Troezar said:
I don't see there being a 3080Ti. There's no gap to place it in. Who's buying a card only 10% better than a 3080?
Click to expand...

Well that's the point though, it would knee cap the sales of the 3090 surely unless people are then going to stump up for, say 5% extra performance of the 3090 if the 3080ti was 5% better than the 3080. I suppose they might make the 3080ti with only slightly more VRAM and up it to 16GB which would leave the 3090 with the extra still VRAM of 24GB.
 
g67575 said:
No acceptable excuse for the FE Rtx 3080s crashing though. They are designed and built by Nvidia. They can test GPUs and drivers before launch until they are ready. Delay a launch a bit if needed, or dont set a date until ready, preferably.

Better to reduce GPU clocks to achieve stability, than increase them and give a false impression.
Click to expand...

that really a old way of thinking nvidia or any other major brand when a release date is set it’s very costly to change it. In this allways connected world we live in it’s easier to send out a rushed product and then fix later. Nvidia did there job got the small amount of cards out beat amd new gpu and also released ahead of consoles.
Took them very few days to drop a driver that improved stability for a lot of people so now the cards are moser stable. Much cheaper then delaying the launch changing all the adds they brought etc
 
If a new driver fixes the issues for most, thats alright. But, the rest should get refunds or replacements if the product (FE) simply isnt functioning as intended.

No need to defend NV they are doing just fine... No one forced Nvidia to set a release date, its just their usual company policy I suppose. Its all part of getting ppl invested well in advance (and stock investors too) but in my view, doesnt benefit customers.

To be fair, AMD too is guilty of hyping stuff up before release and endlessly announcing minor snippets of info, or 'special events' before a product release.
 
Last edited:
TrM said:
Took them very few days to drop a driver that improved stability for a lot of people so now the cards are moser stable. Much cheaper then delaying the launch changing all the adds they brought etc
Click to expand...
g67575 said:
If a new driver fixes the issues for most, thats alright.
Click to expand...

Hang on, it's not a little driver fix to fix a bug or some bad code to stop crashes. They have increased the TDP and lowered the boost clocks post-release for stability. They have altered the product. This isn't a normal 'modern' process post release and if AMD did it, they would have been slaughtered in the press. It basically invalidates all the stock reviews as an aside.
 
A boost clock is type of built in overclock, and could be described as extra performance, in addition to the standard clock rate, so arguably it can be adjusted as needed. Adjusting the TDP by 10w seems unlikely to cause issues, especially if a user meets the required PSU spec, which is often a but higher than is strictly neccesary.

For larger adjustments in tdp/ clock rates, id be less accepting.
 
Nutella33 said:
Hang on, it's not a little driver fix to fix a bug or some bad code to stop crashes. They have increased the TDP and lowered the boost clocks post-release for stability. They have altered the product. This isn't a normal 'modern' process post release and if AMD did it, they would have been slaughtered in the press. It basically invalidates all the stock reviews as an aside.
Click to expand...
A lot of reviewers won't bother to back and retest, so when Navi is released it will be compared to the release day nvidia performance and not the dialled back version.

g67575 said:
A boost clock is type of built in overclock, and could be described as extra performance, in addition to the standard clock rate, so arguably it can be adjusted as needed. Adjusting the TDP by 10w seems unlikely to cause issues, especially if a user meets the required PSU spec, which is often a but higher than is strictly neccesary.

For larger adjustments in tdp/ clock rates, id be less accepting.
Click to expand...

As my ventus 3080 only had a locked 320w tdp to begin with the extra 10 has now effectively limited it to a 310w card. Never had no issues with crashing but now I'm losing performance as I was already stuck at the power limit before the new driver so can't even overclock to offset the clock changes as I'm even more power limited now.
 
Last edited:
Status
Not open for further replies.
Back
Top Bottom