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.
111 Comments on NVIDIA GeForce RTX 5090 and RTX 5080 Specifications Surface, Showing Larger SKU Segmentation
Can see 5080 being priced the same as 4080 @$1200 and likely $2-2.5k for the "mightyyyyyyyyyy" 5090 with such specs, 1.5TB/s vRAM though o_O
AMD better pull there finger out and give me an worthy upgrade to my 6800 for $300 almost 4 years since it's release and there is nothing without spending $500+ for a meaningful upgrade, stagnant :mad:
If this leak is correct then it's almost certain RTX 5080 is indeed an overclocked 4080. Which is foul. Especially if for >1000 USD we get <50% performance uplift (which is a given, thus it's foul).
I don't believe NVIDIA will stop shitposting but a GPU like 4070TS but for <500 USD and with >=16 GB VRAM would be much appreciated. Almost zero chance it happens before RTX 7000 series but wishing doesn't hurt.He still got a point. GPU prices are horribly high, and there's no way they become reasonable. Anti-monopoly departments and other organisations will have hard time proving NV guilty before the 2020s end. And even if they prove there's still no good answer. No good solution.
But ultra fast progress is what we want - isnt it ? Just be more patient and you can find used GPU at affordable price I'm sure.
If the rumour is true about the power, it also says to me whatever they got ready isnt ready for prime time if it requires another jacking up the power.
Progress is currently excruciatingly slow: previously, $300 GPUs made fun of $600 GPUs of the last gen (GTX 1060 VS GTX 980; HD 7870 VS HD 6970). Today, $300 GPUs match or slightly defeat yesterday's $300 GPUs. You can account inflation and DLSS all you please, it never denies the fact we need a full-scale price war but we can't get one because there's literally not a single company to unload grass for NVIDIA to touch. Notice I'm not even saying, "enough grass."
5080 will be a very expensive piece of rubbish. 5090 will be even more out of reach. 5070 downwards will just be a smidge better than their predecessors and won't cost noticeably less. We'll probably see some good news on the lowest end where Intel and AMD still have their saying but something more than casual gaming will be a millionaires' thing for a long while.