Why would someone have to defend the purchase? Surely they bought 'X' GPU knowing that there could well be something faster coming. If I had got any of the 3 series, I would just enjoy it. Bit of a daft questionaire in truth!
-
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
- Thread starter Gregster
- Start date
- Status
Soldato
- Joined
- 21 Jul 2005
- Posts
- 21,107
- Location
- Officially least sunny location -Ronskistats
Yes this was part of my point. I defend nvidia on the hand like what happened to AMD with the 5700XT, many users blamed the card when in some cases it was likely the PSU or bungle of build because we seen in the driver thread people were owning up to daisy chaining leads or swapping PSU and the issue went away.
However the FE cards are crashing and the driver has not fixed everything. So there is a deeper issue at hand, but something the company can weather now through criticism just like AMD had to. This is why we said its a GPU industry thing not really specific to one brand. The bad press AMD got is now suddenly on the other foot before people dismiss it again.
I've subsequently revised my opinion after other input. Limited run 3090 then pitch the 3080Ti at £1,000 and somewhere above the 3080. As you say new buyers coming in who don't want to pay £1,400+ for GPU.
Associate
- Joined
- 19 Jun 2017
- Posts
- 1,029
Fair point. PC building is so accessible now, any idiot can build one, and often do
Sometimes i feel i should have been like those Apple dolts.. just order that latest iPhone, showoff to friends and be happy in that little bubble..
but no hell no i had to pick up PC Gaming as a hobby
oh yeah, I'm aware of that. That's why I find it odd to just see it handwaved as a 'driver fix'. It goes well beyond the accepted norm for a driver fix and more like a post-release firmware flash without the bad press that that would generate.
The stock boost was altered, not the factory overclock boost. Boost clocks are a part of the stock performance of the item and how they are marketed and sold. Look at the issues AMD ran into with the boost behaviour of the 3000 series, inc threatened lawsuits over advertising/sale information.
Whether it causes issues or not to current owners is irrelevant. It's a post-release alteration of the power scheme of the board, something that features heavily in reviews and indirectly affects other performance metrics due to how integral boost behaviours work on modern chips. If Intel sold you a SC boost 5Ghz 100TDP Processor and then released a stability patch that dropped your boost to 4.9Ghz and raised your TDP to 110 two weeks after launch, you have been mis-sold and it is not the modern 'normal' thing to happen or standard practice, or as simple a thing as 'a driver fix'.
Associate
- Joined
- 19 Jun 2017
- Posts
- 1,029
Does any one know how the rx 5700 juggles between int/full precision/partial precision ops.. we might be able to hypothetically compare it with the 3080?
here is the thing though they haven’t done either they changed there boost Algorithm. Not changed anything fundamentally about the base of the card. They still run at 1510mhz and 1710:MHz There is no place that stats you sould be boosting to 2.2ghz or even 2ghz they changed a feature of the card to make all brands of the card behaviour better and more stable. Boosting is not a guaranteed to be anything. Amd have done the same things on there CPU’s and gpu’s in the past and they will continue to do the same.
you mention amd pacifically but the matter of the truth if is it was amd all the youtubers and everyone In the press would have just jumped to drivers and this issue wouldn’t of come to light. Your perception and basis for your thought are all based on a idea that was brought forward by untested and unverified idea from ther best guess.
Soldato
- Joined
- 17 Aug 2003
- Posts
- 20,160
- Location
- Woburn Sand Dunes
Wellll...There's what people should do, and there's reality
so what's the performance penalty?
Associate
- Joined
- 19 Jun 2017
- Posts
- 1,029
Are you serious? its 30 MHz
My entire point is that the reviews should be redone to find that out? If there was no penalty, the question becomes why were they clocked that high initially. But we know from the Zotac underclock faff that the small changes do actually affect performance if only minimally.
and a change to TDP, which effects boosting. Factory overclocked cards sometimes only have 30-50mhz on the boost and people pay extra for them. Zotac had to underclock their 3080 cards by a similar amount for stability and their product segmentation and they got a small performance hit. Also the whole deal with AMD's boost performance issues on the 3000 series was over 25-50mhz boost steps. Handwaving it as a simple bios fix is trying to sweep it under the rug.
I'm not claiming they should be boosting higher? The point is they changed the stock features of the card, two weeks after release/reviews.
If you read my post you'd see I mentioned this. They got threatened with lawsuits despite the boost behaviour being a bios bug and later BIOSes fixed the boost behaviour. This is not what has happened with this release though, where metrics have purposefully being altered with no apparent future plan to return to the release settings.
I can't really work out your point here as its a bit mangled.
Last edited:
Mine is the wait for a good water blockMy biggest concern is RDNA2 will be really good and I'll want to upgrade my 5700XT
Yes, all true. I don't buy in to just RTX, I am buying 2nd gen RT support with dedicated AI hardware. I just can't beleive AMD can touch the performace offered by Nvidia, at least not with their 1st gen. I hope I am wrong here and AMD come forward with the goods. If they do I will cancel my 3080 order.
I was able to get my EK WB before the 5700xt - and I had both on pre-orderMine is the wait for a good water block
Hope that's the case. I've been enjoying the Aquacomputer blocks for the 3080/90 series with the inbuild LCD screen so I'm hoping they are on the ball with Navi and produce something similar.I was able to get my EK WB before the 5700xt - and I had both on pre-order
Soldato
- Joined
- 17 Aug 2003
- Posts
- 20,160
- Location
- Woburn Sand Dunes
We dont need an entire set of reviews to be redone to find that out. We need one person with a card to start with. It's all very well you saying reviews are invalidated but if the difference is naff all (and we already know overclocking the 3080 nets very little as it is) then why care? I assumed by the way you are describing it that there was a significant negative affect on performance but that would appear to not be the case. Some people in the ampere thread are even saying their cards boost higher now so?
In humbug napkin math world the difference is negligible. In the real world it's 88%
You'd love it if I went back to the Nvidia thread, love it
Last edited:
Same reason people care about the crap Zotac basic boards reducing performance by 1-5%? Or why people would care if that potential drop was inflicted on your card even if it was perfectly stable like in @Joxeon 's case? 1-5% can be the difference between a 'good' AIB and a 'bad' AIB and people pay extra money for that. Such a drop across the release may be 'naff all' in the scheme of things but it then becomes clear that Nvidia pumped the numbers at release at the cost of stability and then retroactively baselined performance, which if it happened, it should be known. The point is people buy off reviews. In the same way we know overclocking is 'naff all' on those cards but some people still consider it when purchasing whichever variant they want?
- Status