Associate
- Joined
- 19 Jun 2017
- Posts
- 1,029
-
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
Caporegime
- Joined
- 18 Oct 2002
- Posts
- 29,852
Soldato
From calculation on what we do know (IE what AMD said) the GPU should be faster than a 3080. Especially the 80CU one, at the rumoured clock speeds of 2ghz+ stock.
And remember, those numbers (80 CU and 2ghz+) were not pulled out of thin air. TSMC leaked the 80CU part (they have a hole in their bucket) and AMD did a presentation a year ago in which their scientist talked about "Multi ghz GPUs". IE, multi usually means more than one
So there are some loose facts there. From that point the speculators can start coming up with numbers. Which are the numbers that are going around everywhere. So either AMD are trying to silence those rumours (which would make sense but lose them sales) or they are playing a game.
Nobody would bolt their load 20 days before launch. Nobody, not even Jen.
Associate
- Joined
- 19 Jun 2017
- Posts
- 1,029
Associate
- Joined
- 19 Jun 2017
- Posts
- 1,029
And no one said anything about 80CU during the reveal. But you can go ahead with that assumption in your calculations
Soldato
There are plenty of sources out there making those claims by doing the maths. Which isn't my strong point, and TBH? I would never try doing that any way. It has always been quite easy to estimate the TFLOPS of a GPU at a given speed given the full spec.
I'm not saying I believe it. However, I know that certain mathematical equations don't lie.
TBH? if AMD wanted to clarify anything with that "thing" yesterday then all they have done is cause more confusion. Their fans were bitterly disappointed (come on, admit it) and it hardly got any Nvdia fans thinking did it? "Oh hey look it's about as fast as a 3080 so I will not order a 3080 and wait for AMD!". Said no Nvidia fan ever.
No company in their right mind would bolt their load 20 days before launch is what I am saying. Why? because look at the reaction it has caused. Especially when they have been so incredibly tight lipped so far, with the only leaks coming out of TSMC as per usual.
I just said it wasn't me who did the maths. Read what I said.
The 80 CU thing has been bouncing around for ages. I didn't say anything about the "reveal" if you could even call it that. They revealed, absolutely bugger all apart from some very loose words and three towers with numbers at the bottom.
Caporegime
LLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLadies and gentlemen, my names Paul, and in this Redgamingtech.com video.......................
After re watching AMds stream I noticed something interesting
the big Navi benchmarks were done using a 5950x which is a cpu that no one can buy today and it's the fastest gaming cpu in the world
which means that big navis numbers in that benchmark are inflated versus what we can achieve today when benchmarking competing gpus
What this means is that we can't compare them against a rtx3080 for example until we can also re test the 3080 using a 5950x, it's not fair to run the 3080 with a 10900k/3950x and then attribute the difference in performance all to big Navi
I certainly hope AMD doesn't use the 5950x to benchmark big Navi on October 28 or it totally invalidates their claims but I belive that's exactly what they will do
the big Navi benchmarks were done using a 5950x which is a cpu that no one can buy today and it's the fastest gaming cpu in the world
which means that big navis numbers in that benchmark are inflated versus what we can achieve today when benchmarking competing gpus
What this means is that we can't compare them against a rtx3080 for example until we can also re test the 3080 using a 5950x, it's not fair to run the 3080 with a 10900k/3950x and then attribute the difference in performance all to big Navi
I certainly hope AMD doesn't use the 5950x to benchmark big Navi on October 28 or it totally invalidates their claims but I belive that's exactly what they will do
It was at 4k. From hardware unboxed 14 game average were they tested AMD vs intel they found that the CPU did not make a difference at 4k.
Unless these games have a lot computations to run the GPU will be the bottleneck at 4k.
There are still outliers at 4k, games like the far cry series and flight simulator 2020 are still super cpu bound at 4k - all it takes is for a couple of these games in your benchmark suite to cause a problem with the results.
And you can be sure amd is going to show 1080p and 1440p numbers on October 28
AMD's top Zen 3 CPU is only a bit faster than Intel's top Comet Lake CPU. And the benchmark is in 4K, reducing the chance of CPU bottleneck.
And the games are not especially CPU intensive games, such as Total War Warhammer I/II. Instead they picked GPU intensive first/third person shooters.
The preview was meant to give us some ballpark figures, to give us an idea of what the GPU is capable of, without a CPU bottleneck.
The mean (average) FPS for the three benchmark results was 74 FPS, in my view, quite respectable. They threw in a challenging benchmark too, with Borderlands 3 on the highest 'Badass' setting.
I see there's a random twitter account (maybe others too) claiming it wasn't 'Big Navi'. Why give this any credability?
And the games are not especially CPU intensive games, such as Total War Warhammer I/II. Instead they picked GPU intensive first/third person shooters.
The preview was meant to give us some ballpark figures, to give us an idea of what the GPU is capable of, without a CPU bottleneck.
The mean (average) FPS for the three benchmark results was 74 FPS, in my view, quite respectable. They threw in a challenging benchmark too, with Borderlands 3 on the highest 'Badass' setting.
I see there's a random twitter account (maybe others too) claiming it wasn't 'Big Navi'. Why give this any credability?
Last edited:
AMD have always been truthful with their benchmark results. Too many people looking for drama tbh.
3080 performance ±10%* is good enough for me, but I'm more curious about software innovations & pricing & of course the most important - availability.
* edit: forgot to add %
3080 performance ±10%* is good enough for me, but I'm more curious about software innovations & pricing & of course the most important - availability.
* edit: forgot to add %
Last edited:
Soldato
He was annoyed about the same thing most people were who'd been waiting for the announcement. Not the performance (it looks fantastic on that front and not that there were price hikes (we'd been expecting them) but more at the way the sku's had been positioned. A 5600X instead of just a 5600 bumped their bottom CPU up to nearly £300 (replacing the 3600 which was probably the most popular PC enthusiast CPU on the market for less than £200. Then there was the 5800x pushing 8 core Ryzen 7's a lot higher than the previous gen's point of entry. NAAF's big gripe with any of the manufacturer's have generally been when they take the biscuit on price. Hence why he was positive about the 3080 FE launch and MSRP but less so when they've forced a shortage to keep high margins through the AIB's. With the Radeon reveal he may have fallen into the trap of a lot of people thinking that was the best they had. Logic dictates they would save something big for the 28th. Otherwise you might as well have done the full launch alongside Zen 3.
Soldato
Not in the part I watched. He was absolutely devastated, because all of the unfounded stuff he had been talking about was crushed. IE, it's no 3090 competitor. From what they showed it barely competes with the 3080. And of course there was no DLSS or RT in there.
And that Red Gaming Tech bloke talks poo as well. Just had a flick through that video and noted he said "AMD may have something as fast as a 3080 or are hiding something faster". Funny, that's not what he was saying last week when he too was getting all excited and claiming that AMD could actually beat the 3090.
I don't understand it all myself. Back when I grew up if you talked crap you got busted and called out as a Billy BSer. These days people actually get paid to talk crap. I love how he conveniently takes what AMD did yesterday and manipulates it so that he doesn't look like a top belmet. Which of course he is.
I can't believe people actually encourage these people. "Yeah man, sit and talk crap all day and we will watch your ads so you get paid for it".
Same here, I'm building a second PC and will be using my 2080 Super in that. So wanting something to replace it.
Option A) Potentially* over-egged expectations
Option B) Nvidia trolling
*I say potentially because it’s probably premature to call that one.
A card that matches a 3080 but with an extra 6gb would do me pretty nicely and give me a good excuse to jump over to AMD, especially if it’s a little cheaper. A card that beats the 3080 would also be great, but it has to be at a realistic enough price for me to care. They could bring out a 3090 stomper for £1500 and it would probably mean very little to me!
I’m mostly just happy that we aren’t looking at only having an XX70 equivalent this time.
Do you have proof that the games shown by AMd are outliers?
Last edited:
Paul (RGT) clearly has some valid sources... for example he called the Radeon VII even when AIBs were saying it wasn’t a thing, various bits about epyc and zen 2 (eg the first I saw to leak the correct release date for zen 2), he called out the name “infinity cache” a couple of weeks before AMD trademarked it, he had pictures of the GPU before anyone else I’m aware of....
I think you are being premature to call him full of crap when we have precisely zero hard evidence and only individual conjecture about what AMD did or didn’t show.
In 3 weeks, that’s when you can confidently declare him and the others to be full of it.
Last edited:
Soldato
Come on man, this is the new age con stuff. Just take streaming in general. Every kid out there wants to stream, yet in reality who wants to watch someone else play a video game? Yeah like you wont catch me with an hour or tops two after work and kids in bed hopping onto twitch to watch Grim play like a sac on BF5 or something. Its whats selling the PC components im afraid, just look at nvidia suite of software with 'broadcast' and noise cancelling to talk to your mates while your mums hoovering round your scummy bedroom!
- Status