cloudhan

@Wongboo It works, but you need to carefully configure it though...

This should work: `.clangd` file: ``` CompileFlags: Add: - -std=c++14 - --cuda-path=C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.1 - --cuda-gpu-arch=sm_61 - -LC:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.1\lib\x64 - -IC:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.1\include...

It is possible to embed SIMT programming model in circle and mimic what [ISPC](https://ispc.github.io/) has achieved?

> It will take some imagination to figure out the inter-function bindings that will allow us to treat normal functions as SIMT functions. Hmm, What about treat the entry point...

I also experiance this problem.

For recording purpose, the perf difference with initial try ``` Latency(ms) Latency_P50 Latency_P75 Latency_P90 Latency_P95 Latency_P99 Throughput(QPS) model graph_optimization_level intra_op_num_threads batch_size sequence_length test_cases test_timesuse_gpu 113.03 113.01 113.15 113.26 113.38 113.53...

This PR is split into 2, the following #13116 the enabling and testing for it.

Use clangd as C++ language server gives you a uniform editing experience span C/C++ and CUDA and HIP code. clang-tidy and clang-format is also integrated. All you need is the...

nondiscard is great. Better to keep it on and fix them from the code side.

@justinchuby Just found a new common annoying case: > Floating point literal has suffix 'f', which is not uppercase (fix available)clang-tidy(readability-uppercase-literal-suffix) I think this should also be disabled.