With up to 92 billion transistors, Blackwell is the most powerful consumer GPU ever created. The Blackwell Streaming Multiprocessor (SM) has been updated with increased processing throughput and better integration with Tensor Cores to optimize neural shader performance. Blackwell has been enhanced with several hardware and software innovations to improve shader execution reordering. The sorting logic is now twice as efficient, increasing the speed and accuracy of sorting and improving the performance of neural shaders.
The new Blackwell Tensor Cores are built with massive AI processing power to support fast processing of FP4 high-precision models. With FP4, Blackwell Tensor Cores can not only process models faster, but also use less graphics memory.
To support the frame pacing requirements of next-generation DLSS multi-frame generation, the Blackwell architecture is built with enhanced hardware flip metering capabilities to provide the speed and accuracy needed for a smooth, high-quality experience. .
Blackwell also features brand new RT Cores designed to ray trace large amounts of detailed geometry. RT Cores feature twice the ray triangle intersection rate of previous generations and enhanced compression designed to reduce memory footprint. This allows Blackwell GPUs to ray trace geometry at a level never before possible.
Blackwell is also powered by PCIe Gen5 and DisplayPort 2.1b UHBR20 to drive displays up to 8K 165Hz.
For GeForce RTX 50 Series laptops, new Max-Q technologies, including advanced power gating, low-latency sleep, and fast frequency switching, extend battery life by up to 40% compared to previous generations.
And to supply all this processing power, Blackwell is equipped with GDDR7, the world’s fastest memory with speeds of up to 30Gbps. Blackwell GPUs with G7 memory can provide up to 1.8TB/s of memory bandwidth.
