News Overview
- NVIDIA’s upcoming RTX Pro 6000 Blackwell GPU has been spotted, featuring 24,064 CUDA cores, 96GB of GDDR7 memory, and a 600W total graphics power (TGP).
- This new workstation GPU boasts 11% more cores than the GeForce RTX 5090, indicating a significant performance leap.
- The RTX Pro 6000 Blackwell is expected to arrive in two variants: the standard RTX Pro 6000 and the RTX Pro 6000 X, with the latter likely featuring a fully enabled GB202 die.
Original article link: NVIDIA RTX Pro 6000 Blackwell GPU spotted with 24,064 CUDA cores, 96GB GDDR7, and 600W — 11% more cores than RTX 5090
In-Depth Analysis
-
CUDA Cores and SMs:
The RTX Pro 6000 Blackwell is reported to feature 188 streaming multiprocessors (SMs), totaling 24,064 CUDA cores. This configuration represents a 97.9% enabled GB202 die, surpassing the GeForce RTX 5090’s 21,760 CUDA cores by 11%. -
Memory Configuration:
Equipped with 96GB of GDDR7 memory on a 512-bit interface, the GPU utilizes thirty-two 24Gb (3GB) memory chips in clamshell mode. This substantial memory capacity is particularly beneficial for professional applications such as digital content creation, professional visualization, and AI workloads. -
Power Consumption:
The RTX Pro 6000 Blackwell has a reported TGP of 600W, which is double that of its predecessor, the RTX 6000 Ada. This increase in power consumption aligns with the enhanced performance capabilities of the new GPU. -
Variants:
NVIDIA plans to release two versions of the GPU: the RTX Pro 6000 and the RTX Pro 6000 X. The “X” variant is expected to feature a fully enabled GB202 die with all 192 SMs, potentially offering even higher performance.
Commentary
The introduction of the RTX Pro 6000 Blackwell series signifies NVIDIA’s commitment to advancing workstation GPU performance. The substantial increase in CUDA cores and memory capacity positions this GPU as a formidable tool for professionals engaged in high-demand tasks such as 3D rendering, complex simulations, and AI model training. However, the notable rise in power consumption to 600W may necessitate enhanced cooling solutions and robust power supplies, potentially impacting workstation design considerations. Additionally, while the 96GB of GDDR7 memory offers significant advantages for data-intensive applications, it may be excessive for standard professional workloads, suggesting that NVIDIA is targeting specialized markets with this release. Pricing is anticipated to be in line with previous high-end workstation GPUs, likely ranging between $6,000 and $8,000, reflecting the advanced capabilities and targeted professional audience of the RTX Pro 6000 Blackwell series.