Instruction-level Tensor Core micro-kernel engineering on SM75 (RTX 2070), including PTX manual emission, shared-memory staging, and hardware-aware auto-tuning.
performance-engineering gpu cuda high-performance-computing microkernel auto-tuning ptx tensor-cores deep-learning-infrastructure sm75
-
Updated
Jun 26, 2026 - Cuda