Monday, January 29th 2024
Top AMD RDNA4 Part Could Offer RX 7900 XTX Performance at Half its Price and Lower Power
We've known since way back in August 2023, that AMD is rumored to be retreating from the enthusiast graphics segment with its next-generation RDNA 4 graphics architecture, which means that we likely won't see successors to the RX 7900 series squaring off against the upper end of NVIDIA's fastest GeForce RTX "Blackwell" series. What we'll get instead is a product stack closely resembling that of the RX 5000 series RDNA, with its top part providing a highly competitive price-performance mix around the $400-mark. A more recent report by Moore's Law is Dead sheds more light on this part.
Apparently, the top Radeon RX SKU based on the next-gen RDNA4 graphics architecture will offer performance comparable to that of the current RX 7900 XTX, but at less than half its price (around the $400 mark). It is also expected to achieve this performance target using a smaller, simpler silicon, with significantly lower board cost, leading up to its price. What's more, there could be energy efficiency gains made from the switch to a newer 4 nm-class foundry node and the RDNA4 architecture itself; which could achieve its performance target using fewer numbers of compute units than the RX 7900 XTX with its 96.
When it came out, the RX 5700 XT offered an interesting performance proposition, beating the RTX 2070, and forcing NVIDIA to refresh its product stack with the RTX 20-series SUPER, and the resulting RTX 2070 SUPER. Things could go down slightly differently with RDNA4. Back in 2019, ray tracing was a novelty, and AMD could surprise NVIDIA in the performance segment even without it. There is no such advantage now, ray tracing is relevant; and so AMD could count on timing its launch before the Q4-2024 debut of the RTX 50-series "Blackwell."
Sources:
Moore's Law is Dead (YouTube), Tweaktown
Apparently, the top Radeon RX SKU based on the next-gen RDNA4 graphics architecture will offer performance comparable to that of the current RX 7900 XTX, but at less than half its price (around the $400 mark). It is also expected to achieve this performance target using a smaller, simpler silicon, with significantly lower board cost, leading up to its price. What's more, there could be energy efficiency gains made from the switch to a newer 4 nm-class foundry node and the RDNA4 architecture itself; which could achieve its performance target using fewer numbers of compute units than the RX 7900 XTX with its 96.
292 Comments on Top AMD RDNA4 Part Could Offer RX 7900 XTX Performance at Half its Price and Lower Power
www.techpowerup.com/gpu-specs/radeon-780m.c4020
Even faster if you have tight and fast RAM
When MS released DX12 it killed all SLI/Crossfire for gaming.
There is 1-2 benchmark games that utilize it and that is it.
DX12 makes it so you can mix and match AMD and Nvidia gpus in a multi-gpu config...
but it also puts all of the effort to make it work on the game programmers.
SLI and Xfire worked because AMD and Nvidia supported it.
So, hydrid crossfire was impossible in 2005.About Radeon RX 8700 XT.
Rumour specification state a 256-bit memory bus with old GDDR6 and using the old TSMC 5nm+ process (labeled 4N).
Performance around RTX 4070 Ti - RTX 4070Ti S.
Navi 48 ~ 300-350 mm2, 256-bit, GDDR6, 5nm, RTX 4070 Ti.
About Radeon RX 8600 XT.
Navi 44 ~ 200-210 mm2, 128-bit, GDDR6, 5nm, Radeon RX 7700 XT.
Navi 44 a new version of Navi 10, Navi 23, Navi 33.
www.notebookcheck.net/RDNA-4-Navi-48-and-Navi-44-GPUs-leak-with-details-of-performance-memory-spec-and-more.802603.0.html
www.digitaltrends.com/computing/amd-rdna-4-news-release-date-price-rumors/
Or is your pedantness limited to others?
I actually considered returning my 5800X3D to try and score an AM5 APU. (Couldn't. Platform cost was too high)
Why? So I could remove the 6500XT I use for AFMFing my Vega 10.
To me, the Navi 24 is just a primitive less-featured version of what I'd expect out of an AM5 APU.Hmmm? "Crossfire" is dead.
HOWEVER there's more than just "AMD MGPU" to multi-GPU.
If InfintyFabric could be packetized over PCIe, "Hybrid Crossfire"-like functionality could exist again.
No, this functionality is not coming back, unless DX12's successor changes who does the work.
DX12 supports multi gpu and even vendor mixing but the enablement is on the game dev side, so it doesn't happen.
DX11 Nvidia and AMD supported making multi gpu work, DX12 game devs do, and only the super sponsored titles and benchmarks ever had multi gpu support.
For all intents and purposes multi gpu for gaming is dead.
For workstations/servers we have NVlink and IF.
Navi 48: 64 Compute Units, 256-bit GDDR6 693 GB/s, 240mm², PCIe 5.0 ~ RX 7800 XT
Navi 44: 32 Compute Units, 128-bit GDDR6 288 GB/s, 130mm², PCIe 5.0 ~ RX 7600
www.guru3d.com/story/more-amd-rdna-4-gpu-lineup-info-surfaces-navi-48-and-navi-44-details/
*Manufacturing process, die size, and clues regarding architectural alterations: Examining AMD’s RDNA 4 Changes in LLVM – Chips and Cheese
Different architectures. Check 4080 performance and available bandwidth, for example.
The changes implemented by AMD will enhance processing efficiency and improve code management at runtime.
RX 7900 XT 20GB: Pixel Rate 459.6 GPixel/s; Texture Rate 804.4 GTexel/s; Bandwidth 800.0 GB/s; L0 Cache 64 KB per WGP; L1 Cache 256 KB per Array; L2 Cache 6 MB; L3 Cache 80 MB
RTX 4080 16GB: Pixel Rate 280.6 GPixel/s; Texture Rate 761.5 GTexel/s; Bandwidth 716.8 GB/s; L1 Cache 128 KB (per SM); L2 Cache 64 MB
The thing that AMD must do is to copy nvidia and enable something like Radeon Boost as the default review setting :D
The theoreticals don't mean anything because RDNA 3 has obvious scaling problems, it performs far worse than it should.
Also slightly OT but since you're going to be predictably defending your purchasing choices again, Hardware Unboxed just dropped this video proving what we were telling you all along:
They'll need to crush the price to performance and make it markedly more efficient to even tempt consumers I think.
Something around the performance of a 7900GRE for 30% less is very promising. I'd sure be happy with a $€£ 399 GRE. It's going to handle 1440p high-refresh and probably 4K60 which is arguably something that price point has never attained.
I enabled Hyper-RX for the first time on my 7800XT last night and gave CP2077 another stab at path-tracing. I don't think it's usable (60-70fps with FMF is more like 30fps in terms of input lag) and the reflections in motion at that framerate are pretty noisy, just as they are on my RTX system, but the point is that AMD's tech of Boost, FMF, Anti-lag - all combined in one easy to set tickbox in the driver was a pretty seamless experience that made path tracing a better experience on the £500 7800XT than it is on the £500 4060Ti 16GB. (yes, I know the 4060Ti 16GB has since had a price cut, but it's the closest comparison point to my £480 7800XT from Nvidia right now.)
The only thing left is for CDPR to update FSR in CP2077 to a newer, better version because FSR2.1 suffers with ghosting behind vehicles pretty badly.
I only want them to fix the video playback power consumption. My 7800 XT eats more playing a film than my entire bedroom HTPC while playing a game, which is ridiculous.