That's why I put it in quotation marks... it's not especially low, but it's low compared to 3080's 760gbps, which will no doubt be the point of comparison for most people.
-
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
That's why I put it in quotation marks... it's not especially low, but it's low compared to 3080's 760gbps, which will no doubt be the point of comparison for most people.
The memory throughput is just an irrelevant detail. What matters the most is the end performance or FPS on screen, TDP in check and affordable and fair pricing.
I have no idea about Chiphell but I hope it is right not because I'm an AMD fan, I'm a fan of competition that drives innovation and performance gains at fair prices, Intel shafted us all for years because they could, Nvidia doing the same, AMD righting the ship's both CPU and GPU even if you hate AMD for what ever personal reasons you cant deny this is just great for us as consumers and should be giving AMD a high five.
I think the shown performance is the RX 6800 XT (or Navi 21 XL with disabled shaders and lower specs) which is right under the RTX 3080 on Chiphell's graph.
AMD been tight lipped about cpu and gpu so none really knows outside amd.
While there are some info we can say the 6800xt goes vs the 3080 and beats the 3070 any version so
its the nvidiakiller in the $500 price range that if you have 1440p, 1080p and even 4k will then buy.
Nvidia make no sense to buy then...
Yea Lisa su and team are bringing the performance uplift this generation on a level people cant understand how good it is. when I see people complain about a $50 buck addition to 19% ipc lift and 28% gaming improvement, I am like you do understand Intel would have added $500?
People are funny.
Yes the Nvidia killer at the $500 price bracket.
The majority of users buy that price range for the added superior amd performance
It seems to be predicated on a 50% (probably taking the 50% perf/watt efficiency gain literally) increase in efficiency from RDNA1 to RDNA2?
I can't believe those GPU 'average' frequencies are true, they are surely too high.
Soldato
- Joined
- 21 Jul 2005
- Posts
- 21,107
- Location
- Officially least sunny location -Ronskistats
Is this the hype train?
You mean the 2000-2300 mhz is too high?
If they've moved to plus variant of the node max frequency increase is 15% - average clock at around 2.3GHz is probably not possible above ~150mm2 die size (ignoring overclocking, etc.)
Also a weird metric - I assume this corresponds with AMD's Game Frequency notation - otherwise in benchmarking they'd be more varied card to card.
Could not agree more with this only difference is I am HUGE AMD fan and nearly all my systems have been AMD, both GPU and CPU, with a few exception. If rumours are correct and AMD deliver on performance as we suspect then they have killed both Intel's and NVIDIA's price gouging crowns. Without competition we get crazy prices for silly performance increase. However this competition theory does not seem to apply to phone prices, still trying to figure that one out.
There was an image like this that was released prior the 3080 being announced. Does anyone remember who or where on the forums it was posted. I want to compare it to what we know now the 3080 and 3090 are released.
These are coming from Chiphell user Polymorph and usually he puts out such graph 1-2 month prior to a product launch and from previous graphs these seems to be quite accurate like the one prior to Nvidia 3000 series launch. (Not the Wccftech one as he updated the graph some days later)
This was the last graph about Nvidia 3000 series.
The AMD topic from him is this one :
https://www.chiphell.com/forum.php?...&extra=page=1&filter=typeid&typeid=223&page=1
PS: Obviously he cannot calculate with later deleted SKUs like 3070ti. (If that was existed/true)
This was the last graph about Nvidia 3000 series.
The AMD topic from him is this one :
https://www.chiphell.com/forum.php?...&extra=page=1&filter=typeid&typeid=223&page=1
PS: Obviously he cannot calculate with later deleted SKUs like 3070ti. (If that was existed/true)
Last edited:
So basically trying to simulate synthetic performance from half an idea of the true specs. Looks like his AMD ones have misunderstood the way the 50% perf/watt improvement works and would have to wind them down 15-20% to get closer to true performance (depending a bit on final clock speeds).
Still if you check back his Nvidia 3000,2000,1000 series graphs and the fact that this guy leaked the first original pic of the 3080 GPU chip as well, most of the time this guy is quite close to what we really get after launch. Of course sometimes not the first graph he puts out like we saw at the 3000 series.
Obviously it still means nothing, as nobody really sees behind the curtain. Maybe some aspects are revealed for some, like this guy with the 3080 GPU chip, but not the whole image for sure.
At least his graphs are not made with red/green glasses on.
There is definitely some information on there which no one else was talking about before and/or couldn't have been guessed at which he got right but there is other information on there which looks like it was guessed at.
I agree.
If this true then its game over for NVIDIA>
But surely this is not true given the benches AMD showed in their presentation?
If AMD beat NVIDIA... i honestly dont know how NVIDIA respond. They for the first time will behind on VRAM, performance, longevity and on an inferior process..
I don't think AMD will touch 3090 but I am hopeful I am wrong.
- Status