Predicated Execution VS Conditional Execution 07-01-2026 07-02-2026 blog 17 minutes read (About 2609 words)where VS cond Accelerated Computing, CUDA, PyTorch, TensorRT, GPU, AOTInductor, TorchInductor, JAX, XLA, TPU Read More
PyTorch Custom Operation 05-10-2026 05-10-2026 blog 23 minutes read (About 3501 words)Implementing PyTorch Custom Operations In C++ and CUDA Using torch.library CPP, Python, CUDA, PyTorch, AOTInductor Read More
How Is FARS, The Fully Automated Research System? 04-22-2026 04-22-2026 blog 5 minutes read (About 740 words)The AI Just Tried To Fool People Artificial Intelligence, CUDA, Research, CUDA Graphs, Mixture-of-Experts Read More
Page Table for Page-Locked Host Memory 04-12-2026 04-12-2026 blog 17 minutes read (About 2541 words)Page Table GPU Memory Overhead and Sharing Page-Locked Host Memory Across Processes CUDA, NVIDIA, Computer Architecture, GPU, Memory Management Read More
CUDA_LAUNCH_BLOCKING=1 03-20-2026 03-20-2026 blog 4 minutes read (About 599 words)Debugging CUDA Applications CUDA, Debug Read More
CUDA Shared Memory Bank Conflict-Free Vectorized Access 02-13-2026 02-13-2026 blog 14 minutes read (About 2060 words)Instruction-Level Phase Based Bank Conflict-Free Execution CUDA, NVIDIA, Parallel Computing, GPU Read More
CUDA Rendezvous Stream 01-26-2026 01-26-2026 blog 11 minutes read (About 1690 words)Simplifying Synchronization Complexities Using CUDA Rendezvous Streams CUDA, NVIDIA, Parallel Computing, GPU Read More
PyTorch CUDA Graph Capture 01-12-2026 01-12-2026 blog 23 minutes read (About 3454 words)Using PyTorch CUDA Graph APIs CUDA, PyTorch, CUDA Graph, Perfetto Read More
NVIDIA NVML GPU Statistics 12-25-2025 12-25-2025 blog 15 minutes read (About 2214 words)Mimicking nvidia-smi dmon Using NVIDIA NVML CPP, CUDA, NVIDIA, GPU, NVML Read More
NVIDIA Tensor Core TN Layout MMA Instruction 12-06-2025 12-06-2025 blog 16 minutes read (About 2389 words)GEMM Layout, History, Performance, and Implementation CPP, CUDA, NVIDIA, CUTLASS, CuTe, MMA, GEMM, Tensor Core Read More
Benchmarking NVIDIA Tensor Core MMA Instruction Peak Performances 11-26-2025 11-26-2025 blog 11 minutes read (About 1646 words)Reproducing NVIDIA Advertised GPU AI Peak Performances Using CUTLASS and CuTe CPP, CUDA, NVIDIA, CUTLASS, CuTe, MMA, GEMM, Tensor Core Read More
Nsight Streamer 11-04-2025 11-04-2025 blog 3 minutes read (About 515 words)Nsight Systems and Nsight Compute GUIs In a Web Browser CUDA, NVIDIA, Nsight Compute, Nsight Systems, Nsight Streamer Read More