Thursday, September 26th 2024
NVIDIA GeForce RTX 5090 and RTX 5080 Specifications Surface, Showing Larger SKU Segmentation
Thanks to the renowned NVIDIA hardware leaker kopite7Kimi on X, we are getting information about the final versions of NVIDIA's first upcoming wave of GeForce RTX 50 series "Blackwell" graphics cards. The two leaked GPUs are the GeForce RTX 5090 and RTX 5080, which now feature a more significant gap between xx80 and xx90 SKUs. For starters, we have the highest-end GeForce RTX 5090. NVIDIA has decided to use the GB202-300-A1 die and enabled 21,760 FP32 CUDA cores on this top-end model. Accompanying the massive 170 SM GPU configuration, the RTX 5090 has 32 GB of GDDR7 memory on a 512-bit bus, with each GDDR7 die running at 28 Gbps. This translates to 1,568 GB/s memory bandwidth. All of this is confined to a 600 W TGP.
When it comes to the GeForce RTX 5080, NVIDIA has decided to further separate its xx80 and xx90 SKUs. The RTX 5080 has 10,752 FP32 CUDA cores paired with 16 GB of GDDR7 memory on a 256-bit bus. With GDDR7 running at 28 Gbps, the memory bandwidth is also halved at 784 GB/s. This SKU uses a GB203-400-A1 die, which is designed to run within a 400 W TGP power envelope. For reference, the RTX 4090 has 68% more CUDA cores than the RTX 4080. The rumored RTX 5090 has around 102% more CUDA cores than the rumored RTX 5080, which means that NVIDIA is separating its top SKUs even more. We are curious to see at what price point NVIDIA places its upcoming GPUs so that we can compare generational updates and the difference between xx80 and xx90 models and their widened gaps.
Sources:
kopite7kimi (RTX 5090), kopite7kimi (RTX 5080)
When it comes to the GeForce RTX 5080, NVIDIA has decided to further separate its xx80 and xx90 SKUs. The RTX 5080 has 10,752 FP32 CUDA cores paired with 16 GB of GDDR7 memory on a 256-bit bus. With GDDR7 running at 28 Gbps, the memory bandwidth is also halved at 784 GB/s. This SKU uses a GB203-400-A1 die, which is designed to run within a 400 W TGP power envelope. For reference, the RTX 4090 has 68% more CUDA cores than the RTX 4080. The rumored RTX 5090 has around 102% more CUDA cores than the rumored RTX 5080, which means that NVIDIA is separating its top SKUs even more. We are curious to see at what price point NVIDIA places its upcoming GPUs so that we can compare generational updates and the difference between xx80 and xx90 models and their widened gaps.
183 Comments on NVIDIA GeForce RTX 5090 and RTX 5080 Specifications Surface, Showing Larger SKU Segmentation
Assuming these rumours are accurate. I'd be surprised to see an xx90 with full memory bus/die. 4090 was quite cut down. I don't think they're going to jump straight to 600 W from 450 W.
My guess 500 W.
20,000 cores.
Also I bet 21k cores of 5090 is for most defective dies available now. Later we will see higher grade dies with slightly more enabled cores.
BTW - they should also differentiate supply voltages between these two SKU imho. 24V would fit better for 600W thermal envelope.
So if you want a high end card, it will the same sxxt again.
The x90 variant will only be worth it.
The 4080 already runs nearly 10k shaders, at 320W.
The 4090 runs 16k right now, at 450W.
And then there's also a 4080S with 10240 with the same 320W.
So what's Blackwell's 5080 then? A massively OC'd 4080S? And why is there such a disparity between the power gap relative to shader gap between these two generations? Are these some special shaders that want much more juice on a 5080 than they do on a 5090? We haven't seen a big frequency gap between same stack GPUs on Nvidia lately, so what's that power doing there, all VRAM... ? And if that's true, we come back to the earlier point, what's that 5080 doing with 400W on the same 16GB as its Ada sibling?
I don't think there's much chance of a full die/full bus 5090.
Maybe 5090 Ti/Titan for 22k full die/full bus.
Yes, 5080 Ti somewhere in 15k range.
Heavily considering a couple of those now...
I know there's litterally zero chance of that happening but I wish they could leave the interconnection (NVLink) on the 5080 to bring back SLI
BTW
2 times No - You dont need different mobo or new PSU design - just minor changes in PSU - thats all.