- Particulars of Nvidia’s quickest ever workstation video card have been noticed on-line
- There are RTX Pro 6000 Blackwell and RTX Pro 6000 X Blackwell flavors
- The non-X model has 24,064 CUDA cores and 96GB of GDDR7 ECC reminiscence
On-line leaks seem to reveal particulars of Nvidia’s upcoming RTX Pro 6000 Blackwell GPU, which guarantees to be the corporate’s most highly effective workstation card to date.
Data noticed on Leadtek’s web site, and backed by NBD transport knowledge, level to the card rocking 24,064 CUDA cores and 96GB of GDDR7 reminiscence.
These specs are considerably greater than each the present RTX 6000 Ada, which launched over two years in the past, and the anticipated RTX 5090 gaming GPU, which reportedly options 2,304 fewer cores.
NBD’s transport knowledge reveals two variations of the workstation GPU – the RTX Pro 6000 Blackwell and the RTX Pro 6000 X Blackwell – being despatched to India for “testing functions”. We’re undecided fairly what the X signifies, or how totally different it would be from the opposite model. It’s notable nonetheless, that Nvidia appears to be figuring out its new playing cards with a Pro label.
The info on one in every of these playing cards (the non-X model) was posted on X by Harukaze, exhibiting that the GPU runs on a 512-bit reminiscence interface, and GDDR7 helps ECC (error correction), making it a good selection for servers and methods the place knowledge accuracy is important.
Energy hungry beast
Energy necessities are notably excessive for the brand new GPU, with a 600W Whole Graphics Energy delivered via a 16-pin PCIe 5.0 connector. This places it among the many most power-hungry GPUs in Nvidia’s lineup – over twice that required by RTX 6000 Ada – however that shouldn’t come as an excessive amount of of a shock. Not like earlier workstation playing cards, this one will possible skip the standard blower-style cooler and take its design cues from the RTX 5090, with a dual-flow-through cooling system.
Writing in regards to the leak, Tom’s {Hardware} notes, “Workstation GPUs beneath Nvidia’s RTX lineup are tailored for skilled purposes. Nonetheless, in most situations, 96GB of VRAM is overkill except you are in search of coaching or inferencing AI regionally. Whilst you would not sometimes use these GPUs for gaming, the added reminiscence can be useful in duties concerned with sport improvement, content material creation, ProViz, and computer-aided design.”
The GPU is anticipated to be formally launched at Nvidia’s GPU Expertise Convention (GTC) in March 2025. Pricing hasn’t been confirmed, however it would possible fall in the identical high-end vary as its predecessor, which may be had for between $6,000 and $8,000.
You may also like
Source link
#RTX #Pro #Blackwell #GPU #Nvidias #expensive #video #card
