Unverified85% confidenceFactExact time
sm_103是NVIDIA Blackwell架构的CUDA计算能力标识符
1
Sources
85%
Confidence
Long-term
Relevance
5/27/2026
First Seen
Sources
SGLang v0.5.12.post1 Released: DeepSeek V4 Stability Fixes and Blackwell Adaptation
twitterlmsysorg5/27/2026
Related Entities
Related Claims
UnverifiedThe carry-less multiplication update in CUDA 13.3 can be seen as part of NVIDIA's roadmap for completing the confidential computing capabilities of the H100/H200.72% similarUnverifiedUnsloth has hand-written customized CUDA/Triton compute kernels for mainstream GPU architectures and rewritten the backpropagation algorithm.67% similarUnverifiedNVIDIA designed a dedicated Tensor Core instruction mma.kind::mxf4 for NVFP4 in the Blackwell architecture (SM 100/103, such as the B100 and B200 series), and adopted an FP8 scaling factor mechanism with one scale per group of 16 elements.67% similarUnverifiedA CUDA stream is an asynchronous execution queue on the GPU; operations in different streams can execute concurrently and are a key mechanism for overlapping computation and data transfer.67% similarUnverifiedThe GPGPU concept entered the mainstream in 2007 with the release of NVIDIA CUDA; CUDA is limited to NVIDIA hardware, whereas Vulkan compute shaders can run cross-platform.66% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/569API
curl https://kongchang.com/api/v1/knowledge/claims/569MCP
get_claim(id=569)