Popular repositories Loading

1. vllm-sm70 vllm-sm70 Public archive

   vLLM for nVidia Volta (sm_70)

   Dockerfile 9 4

2. superl8 superl8 Public

   DP4A FlashAttention-2 and GEMM for Volta GPUs. 46 TOP/s INT8 on CMP 100-210 where tensor cores are firmware-disabled. The .superl8 format loads weights at memory speed.

   Python 2

3. ComfyUI-superl8 ComfyUI-superl8 Public

   INT8 quantized diffusion nodes for ComfyUI on CMP 100-210 / Volta GPUs. Per-layer SQNR gating, GGUF loading, multi-GPU. DiTs that fit in 12 GB.

   Python 2

4. superl8-serve superl8-serve Public

   OpenAI-compatible INT8 inference server for CMP 100-210 and Volta GPUs. DP4A kernels, native GGUF loading, speculative decode. Runs Qwen, Gemma, DeepSeek on cards with disabled tensor cores.

   Python 2

5. gv100-fecs-limiter gv100-fecs-limiter Public

   GV100/CMP 100-210 GPU firmware research: VBIOS decode, FECS speed-select analysis, BAR0 register probes, CUDA tensor-path benchmarks

   C 2

6. vrm_automation vrm_automation Public

   Python 1