Skip to content

Make cuda::stream_ref constructible on device - #4243

Merged
miscco merged 3 commits into
NVIDIA:mainfrom
miscco:fix_stream_ref_par_n
Mar 24, 2025
Merged

miscco merged 3 commits into
NVIDIA:mainfrom
miscco:fix_stream_ref_par_n

Conversation

@miscco

@miscco miscco commented Mar 24, 2025

Copy link
Copy Markdown
Contributor

There is no reason we should not be able to construct it or extract the pointer out of it on device.

Use that to drop the cudaStream_t constructors for thrust execution policies

There is no reason we should not be able to construct it or extract the pointer out of it on device.
@miscco
miscco requested review from a team as code owners March 24, 2025 08:53
@miscco
miscco requested review from fbusato and gevtushenko March 24, 2025 08:53
@github-project-automation github-project-automation Bot moved this to Todo in CCCL Mar 24, 2025
@cccl-authenticator-app cccl-authenticator-app Bot moved this from Todo to In Review in CCCL Mar 24, 2025
@miscco

miscco commented Mar 24, 2025

Copy link
Copy Markdown
Contributor Author

This avoids a ton of breakage for rapids

/home/coder/cuml/cpp/src/hdbscan/detail/../condensed_hierarchy.cu(212): error: more than one instance of overloaded function "thrust::cuda_cub::par_t::on" matches the argument list:
            function "thrust::cuda_cub::par_t::on(const cudaStream_t &) const" (declared at line 146 of /home/coder/cuml/cpp/build/conda/cuda-12.8/release/_deps/cccl-src/lib/cmake/thrust/../../../thrust/thrust/system/cuda/detail/par.h)
            function "thrust::cuda_cub::par_t::on(const cuda::__4::stream_ref &) const" (declared at line 151 of /home/coder/cuml/cpp/build/conda/cuda-12.8/release/_deps/cccl-src/lib/cmake/thrust/../../../thrust/thrust/system/cuda/detail/par.h)
            argument types are: (rmm::cuda_stream_view)
            object type is: const thrust::cuda_cub::par_t
      thrust::minmax_element(thrust::cuda::par.on(stream), parents_ptr, parents_ptr + n_edges);

@miscco
miscco requested a review from a team as a code owner March 24, 2025 10:24
@miscco
miscco force-pushed the fix_stream_ref_par_n branch from 693c8fc to d15f302 Compare March 24, 2025 10:40
@github-actions

Copy link
Copy Markdown
Contributor
ðŸŸĐ CI finished in 2h 14m: Pass: 100%/162 | Total: 1d 23h | Avg: 17m 35s | Max: 1h 07m | Hits: 92%/252726
  • ðŸŸĐ cub: Pass: 100%/45 | Total: 1d 05h | Avg: 39m 03s | Max: 55m 36s | Hits: 94%/53780

    ðŸŸĐ cpu
      ðŸŸĐ amd64              Pass: 100%/43  | Total:  1d 03h | Avg: 38m 45s | Max: 55m 36s | Hits:  94%/51336 
      ðŸŸĐ arm64              Pass: 100%/2   | Total:  1h 30m | Avg: 45m 26s | Max: 46m 31s | Hits:  96%/2444  
    ðŸŸĐ ctk
      ðŸŸĐ 12.0               Pass: 100%/5   | Total:  3h 26m | Avg: 41m 14s | Max: 43m 07s | Hits:  90%/5940  
      ðŸŸĐ 12.6               Pass: 100%/2   | Total:  1h 33m | Avg: 46m 43s | Max: 48m 15s | Hits:  89%/2260  
      ðŸŸĐ 12.8               Pass: 100%/38  | Total:  1d 00h | Avg: 38m 22s | Max: 55m 36s | Hits:  95%/45580 
    ðŸŸĐ cudacxx
      ðŸŸĐ ClangCUDA18        Pass: 100%/2   | Total:  1h 50m | Avg: 55m 06s | Max: 55m 36s | Hits:  99%/2108  
      ðŸŸĐ nvcc12.0           Pass: 100%/5   | Total:  3h 26m | Avg: 41m 14s | Max: 43m 07s | Hits:  90%/5940  
      ðŸŸĐ nvcc12.6           Pass: 100%/2   | Total:  1h 33m | Avg: 46m 43s | Max: 48m 15s | Hits:  89%/2260  
      ðŸŸĐ nvcc12.8           Pass: 100%/36  | Total: 22h 27m | Avg: 37m 26s | Max: 55m 03s | Hits:  95%/43472 
    ðŸŸĐ cudacxx_family
      ðŸŸĐ ClangCUDA          Pass: 100%/2   | Total:  1h 50m | Avg: 55m 06s | Max: 55m 36s | Hits:  99%/2108  
      ðŸŸĐ nvcc               Pass: 100%/43  | Total:  1d 03h | Avg: 38m 18s | Max: 55m 03s | Hits:  94%/51672 
    ðŸŸĐ cxx
      ðŸŸĐ Clang14            Pass: 100%/4   | Total:  2h 38m | Avg: 39m 30s | Max: 41m 23s | Hits:  93%/4896  
      ðŸŸĐ Clang15            Pass: 100%/2   | Total:  1h 14m | Avg: 37m 04s | Max: 37m 27s | Hits:  96%/2444  
      ðŸŸĐ Clang16            Pass: 100%/2   | Total:  1h 23m | Avg: 41m 58s | Max: 42m 24s | Hits:  93%/2444  
      ðŸŸĐ Clang17            Pass: 100%/2   | Total:  1h 16m | Avg: 38m 07s | Max: 38m 18s | Hits:  94%/2444  
      ðŸŸĐ Clang18            Pass: 100%/7   | Total:  4h 41m | Avg: 40m 15s | Max: 55m 36s | Hits:  98%/8218  
      ðŸŸĐ GCC7               Pass: 100%/2   | Total:  1h 20m | Avg: 40m 04s | Max: 42m 42s | Hits:  92%/2448  
      ðŸŸĐ GCC8               Pass: 100%/1   | Total: 38m 05s | Avg: 38m 05s | Max: 38m 05s | Hits:  92%/1224  
      ðŸŸĐ GCC9               Pass: 100%/2   | Total:  1h 16m | Avg: 38m 29s | Max: 39m 26s | Hits:  92%/2448  
      ðŸŸĐ GCC10              Pass: 100%/2   | Total:  1h 15m | Avg: 37m 41s | Max: 37m 56s | Hits:  94%/2448  
      ðŸŸĐ GCC11              Pass: 100%/2   | Total:  1h 15m | Avg: 37m 48s | Max: 37m 52s | Hits:  95%/2444  
      ðŸŸĐ GCC12              Pass: 100%/2   | Total:  1h 21m | Avg: 40m 43s | Max: 41m 48s | Hits:  93%/2444  
      ðŸŸĐ GCC13              Pass: 100%/11  | Total:  5h 56m | Avg: 32m 26s | Max: 50m 47s | Hits:  96%/13442 
      ðŸŸĐ MSVC14.29          Pass: 100%/2   | Total:  1h 35m | Avg: 47m 49s | Max: 52m 31s | Hits:  90%/2088  
      ðŸŸĐ MSVC14.42          Pass: 100%/2   | Total:  1h 49m | Avg: 54m 54s | Max: 55m 03s | Hits:  90%/2088  
      ðŸŸĐ NVHPC25.1          Pass: 100%/2   | Total:  1h 33m | Avg: 46m 43s | Max: 48m 15s | Hits:  89%/2260  
    ðŸŸĐ cxx_family
      ðŸŸĐ Clang              Pass: 100%/17  | Total: 11h 14m | Avg: 39m 39s | Max: 55m 36s | Hits:  96%/20446 
      ðŸŸĐ GCC                Pass: 100%/22  | Total: 13h 04m | Avg: 35m 39s | Max: 50m 47s | Hits:  95%/26898 
      ðŸŸĐ MSVC               Pass: 100%/4   | Total:  3h 25m | Avg: 51m 21s | Max: 55m 03s | Hits:  90%/4176  
      ðŸŸĐ NVHPC              Pass: 100%/2   | Total:  1h 33m | Avg: 46m 43s | Max: 48m 15s | Hits:  89%/2260  
    ðŸŸĐ gpu
      ðŸŸĐ h100               Pass: 100%/3   | Total:  1h 09m | Avg: 23m 18s | Max: 30m 16s | Hits:  98%/3666  
      ðŸŸĐ rtx2080            Pass: 100%/34  | Total:  1d 00h | Avg: 42m 34s | Max: 55m 36s | Hits:  93%/40338 
      ðŸŸĐ rtxa6000           Pass: 100%/8   | Total:  4h 00m | Avg: 30m 02s | Max: 40m 24s | Hits:  98%/9776  
    ðŸŸĐ jobs
      ðŸŸĐ Build              Pass: 100%/37  | Total:  1d 01h | Avg: 41m 37s | Max: 55m 36s | Hits:  93%/44004 
      ðŸŸĐ DeviceLaunch       Pass: 100%/1   | Total: 31m 27s | Avg: 31m 27s | Max: 31m 27s | Hits:  99%/1222  
      ðŸŸĐ GraphCapture       Pass: 100%/1   | Total: 21m 45s | Avg: 21m 45s | Max: 21m 45s | Hits:  99%/1222  
      ðŸŸĐ HostLaunch         Pass: 100%/3   | Total:  1h 29m | Avg: 29m 51s | Max: 30m 16s | Hits:  99%/3666  
      ðŸŸĐ TestGPU            Pass: 100%/3   | Total:  1h 15m | Avg: 25m 02s | Max: 27m 09s | Hits:  99%/3666  
    ðŸŸĐ sm
      ðŸŸĐ 90                 Pass: 100%/3   | Total:  1h 09m | Avg: 23m 18s | Max: 30m 16s | Hits:  98%/3666  
      ðŸŸĐ 90;90a;100         Pass: 100%/1   | Total: 50m 47s | Avg: 50m 47s | Max: 50m 47s | Hits:  94%/1222  
    ðŸŸĐ std
      ðŸŸĐ 17                 Pass: 100%/20  | Total: 13h 58m | Avg: 41m 55s | Max: 55m 03s | Hits:  93%/23662 
      ðŸŸĐ 20                 Pass: 100%/25  | Total: 15h 19m | Avg: 36m 45s | Max: 55m 36s | Hits:  96%/30118 
    
  • ðŸŸĐ thrust: Pass: 100%/45 | Total: 7h 08m | Avg: 9m 31s | Max: 35m 53s | Hits: 98%/79911

    ðŸŸĐ cmake_options
      ðŸŸĐ -DTHRUST_DISPATCH_TYPE=Force32bit Pass: 100%/2   | Total: 18m 13s | Avg:  9m 06s | Max: 11m 56s | Hits:  99%/3554  
    ðŸŸĐ cpu
      ðŸŸĐ amd64              Pass: 100%/43  | Total:  6h 58m | Avg:  9m 43s | Max: 35m 53s | Hits:  98%/76358 
      ðŸŸĐ arm64              Pass: 100%/2   | Total:  9m 57s | Avg:  4m 58s | Max:  5m 20s | Hits:  99%/3553  
    ðŸŸĐ ctk
      ðŸŸĐ 12.0               Pass: 100%/5   | Total: 45m 23s | Avg:  9m 04s | Max: 25m 19s | Hits:  98%/8876  
      ðŸŸĐ 12.6               Pass: 100%/2   | Total: 45m 00s | Avg: 22m 30s | Max: 23m 21s | Hits:  92%/3552  
      ðŸŸĐ 12.8               Pass: 100%/38  | Total:  5h 37m | Avg:  8m 53s | Max: 35m 53s | Hits:  99%/67483 
    ðŸŸĐ cudacxx
      ðŸŸĐ ClangCUDA18        Pass: 100%/2   | Total: 10m 08s | Avg:  5m 04s | Max:  5m 10s | Hits: 100%/3552  
      ðŸŸĐ nvcc12.0           Pass: 100%/5   | Total: 45m 23s | Avg:  9m 04s | Max: 25m 19s | Hits:  98%/8876  
      ðŸŸĐ nvcc12.6           Pass: 100%/2   | Total: 45m 00s | Avg: 22m 30s | Max: 23m 21s | Hits:  92%/3552  
      ðŸŸĐ nvcc12.8           Pass: 100%/36  | Total:  5h 27m | Avg:  9m 06s | Max: 35m 53s | Hits:  99%/63931 
    ðŸŸĐ cudacxx_family
      ðŸŸĐ ClangCUDA          Pass: 100%/2   | Total: 10m 08s | Avg:  5m 04s | Max:  5m 10s | Hits: 100%/3552  
      ðŸŸĐ nvcc               Pass: 100%/43  | Total:  6h 58m | Avg:  9m 43s | Max: 35m 53s | Hits:  98%/76359 
    ðŸŸĐ cxx
      ðŸŸĐ Clang14            Pass: 100%/4   | Total: 20m 42s | Avg:  5m 10s | Max:  5m 34s | Hits: 100%/7104  
      ðŸŸĐ Clang15            Pass: 100%/2   | Total: 10m 57s | Avg:  5m 28s | Max:  5m 34s | Hits: 100%/3552  
      ðŸŸĐ Clang16            Pass: 100%/2   | Total: 11m 21s | Avg:  5m 40s | Max:  6m 00s | Hits: 100%/3552  
      ðŸŸĐ Clang17            Pass: 100%/2   | Total: 10m 43s | Avg:  5m 21s | Max:  5m 29s | Hits: 100%/3552  
      ðŸŸĐ Clang18            Pass: 100%/7   | Total: 43m 50s | Avg:  6m 15s | Max: 11m 05s | Hits: 100%/12432 
      ðŸŸĐ GCC7               Pass: 100%/2   | Total: 10m 41s | Avg:  5m 20s | Max:  5m 21s | Hits:  99%/3554  
      ðŸŸĐ GCC8               Pass: 100%/1   | Total:  5m 27s | Avg:  5m 27s | Max:  5m 27s | Hits:  99%/1777  
      ðŸŸĐ GCC9               Pass: 100%/2   | Total: 11m 17s | Avg:  5m 38s | Max:  6m 13s | Hits:  99%/3554  
      ðŸŸĐ GCC10              Pass: 100%/2   | Total: 11m 06s | Avg:  5m 33s | Max:  5m 36s | Hits:  99%/3554  
      ðŸŸĐ GCC11              Pass: 100%/2   | Total: 12m 18s | Avg:  6m 09s | Max:  6m 19s | Hits:  99%/3554  
      ðŸŸĐ GCC12              Pass: 100%/2   | Total: 12m 47s | Avg:  6m 23s | Max:  6m 29s | Hits:  99%/3554  
      ðŸŸĐ GCC13              Pass: 100%/10  | Total:  1h 18m | Avg:  7m 53s | Max: 12m 09s | Hits:  99%/17770 
      ðŸŸĐ MSVC14.29          Pass: 100%/2   | Total: 53m 55s | Avg: 26m 57s | Max: 28m 36s | Hits:  91%/3540  
      ðŸŸĐ MSVC14.42          Pass: 100%/3   | Total:  1h 29m | Avg: 29m 48s | Max: 35m 53s | Hits:  93%/5310  
      ðŸŸĐ NVHPC25.1          Pass: 100%/2   | Total: 45m 00s | Avg: 22m 30s | Max: 23m 21s | Hits:  92%/3552  
    ðŸŸĐ cxx_family
      ðŸŸĐ Clang              Pass: 100%/17  | Total:  1h 37m | Avg:  5m 44s | Max: 11m 05s | Hits: 100%/30192 
      ðŸŸĐ GCC                Pass: 100%/21  | Total:  2h 22m | Avg:  6m 47s | Max: 12m 09s | Hits:  99%/37317 
      ðŸŸĐ MSVC               Pass: 100%/5   | Total:  2h 23m | Avg: 28m 40s | Max: 35m 53s | Hits:  92%/8850  
      ðŸŸĐ NVHPC              Pass: 100%/2   | Total: 45m 00s | Avg: 22m 30s | Max: 23m 21s | Hits:  92%/3552  
    ðŸŸĐ gpu
      ðŸŸĐ h100               Pass: 100%/2   | Total: 16m 13s | Avg:  8m 06s | Max: 11m 38s | Hits:  99%/3554  
      ðŸŸĐ rtx2080            Pass: 100%/33  | Total:  4h 43m | Avg:  8m 34s | Max: 29m 12s | Hits:  98%/58604 
      ðŸŸĐ rtx4090            Pass: 100%/10  | Total:  2h 09m | Avg: 12m 54s | Max: 35m 53s | Hits:  98%/17753 
    ðŸŸĐ jobs
      ðŸŸĐ Build              Pass: 100%/38  | Total:  5h 41m | Avg:  8m 59s | Max: 35m 53s | Hits:  98%/67481 
      ðŸŸĐ TestCPU            Pass: 100%/3   | Total: 39m 55s | Avg: 13m 18s | Max: 24m 21s | Hits:  99%/5323  
      ðŸŸĐ TestGPU            Pass: 100%/4   | Total: 46m 48s | Avg: 11m 42s | Max: 12m 09s | Hits:  99%/7107  
    ðŸŸĐ sm
      ðŸŸĐ 90                 Pass: 100%/2   | Total: 16m 13s | Avg:  8m 06s | Max: 11m 38s | Hits:  99%/3554  
      ðŸŸĐ 90;90a;100         Pass: 100%/1   | Total:  6m 08s | Avg:  6m 08s | Max:  6m 08s | Hits:  99%/1777  
    ðŸŸĐ std
      ðŸŸĐ 17                 Pass: 100%/20  | Total:  3h 13m | Avg:  9m 41s | Max: 29m 12s | Hits:  98%/35511 
      ðŸŸĐ 20                 Pass: 100%/23  | Total:  3h 36m | Avg:  9m 24s | Max: 35m 53s | Hits:  99%/40846 
    
  • ðŸŸĐ libcudacxx: Pass: 100%/43 | Total: 7h 18m | Avg: 10m 11s | Max: 35m 22s | Hits: 86%/106863

    ðŸŸĐ cpu
      ðŸŸĐ amd64              Pass: 100%/41  | Total:  7h 10m | Avg: 10m 29s | Max: 35m 22s | Hits:  85%/101018
      ðŸŸĐ arm64              Pass: 100%/2   | Total:  7m 53s | Avg:  3m 56s | Max:  4m 02s | Hits:  98%/5845  
    ðŸŸĐ ctk
      ðŸŸĐ 12.0               Pass: 100%/5   | Total:  1h 08m | Avg: 13m 43s | Max: 20m 13s | Hits:  74%/14234 
      ðŸŸĐ 12.6               Pass: 100%/2   | Total: 47m 50s | Avg: 23m 55s | Max: 35m 22s | Hits:  59%/5792  
      ðŸŸĐ 12.8               Pass: 100%/36  | Total:  5h 21m | Avg:  8m 56s | Max: 22m 56s | Hits:  90%/86837 
    ðŸŸĐ cudacxx
      ðŸŸĐ ClangCUDA18        Pass: 100%/2   | Total: 43m 49s | Avg: 21m 54s | Max: 22m 42s | Hits:  28%/5806  
      ðŸŸĐ nvcc12.0           Pass: 100%/5   | Total:  1h 08m | Avg: 13m 43s | Max: 20m 13s | Hits:  74%/14234 
      ðŸŸĐ nvcc12.6           Pass: 100%/2   | Total: 47m 50s | Avg: 23m 55s | Max: 35m 22s | Hits:  59%/5792  
      ðŸŸĐ nvcc12.8           Pass: 100%/34  | Total:  4h 37m | Avg:  8m 10s | Max: 22m 56s | Hits:  94%/81031 
    ðŸŸĐ cudacxx_family
      ðŸŸĐ ClangCUDA          Pass: 100%/2   | Total: 43m 49s | Avg: 21m 54s | Max: 22m 42s | Hits:  28%/5806  
      ðŸŸĐ nvcc               Pass: 100%/41  | Total:  6h 34m | Avg:  9m 37s | Max: 35m 22s | Hits:  89%/101057
    ðŸŸĐ cxx
      ðŸŸĐ Clang14            Pass: 100%/4   | Total: 20m 33s | Avg:  5m 08s | Max:  6m 37s | Hits:  96%/11580 
      ðŸŸĐ Clang15            Pass: 100%/2   | Total: 25m 36s | Avg: 12m 48s | Max: 21m 01s | Hits:  74%/5802  
      ðŸŸĐ Clang16            Pass: 100%/2   | Total:  9m 03s | Avg:  4m 31s | Max:  4m 34s | Hits:  98%/5802  
      ðŸŸĐ Clang17            Pass: 100%/2   | Total: 12m 04s | Avg:  6m 02s | Max:  7m 32s | Hits:  94%/5802  
      ðŸŸĐ Clang18            Pass: 100%/6   | Total:  1h 06m | Avg: 11m 03s | Max: 22m 42s | Hits:  70%/14530 
      ðŸŸĐ GCC7               Pass: 100%/2   | Total: 22m 47s | Avg: 11m 23s | Max: 18m 37s | Hits:  70%/5739  
      ðŸŸĐ GCC8               Pass: 100%/1   | Total:  4m 04s | Avg:  4m 04s | Max:  4m 04s | Hits:  99%/2880  
      ðŸŸĐ GCC9               Pass: 100%/2   | Total: 37m 34s | Avg: 18m 47s | Max: 18m 53s | Hits:  47%/5751  
      ðŸŸĐ GCC10              Pass: 100%/2   | Total:  8m 36s | Avg:  4m 18s | Max:  4m 20s | Hits:  98%/5808  
      ðŸŸĐ GCC11              Pass: 100%/2   | Total:  8m 09s | Avg:  4m 04s | Max:  4m 12s | Hits:  98%/5804  
      ðŸŸĐ GCC12              Pass: 100%/2   | Total:  8m 22s | Avg:  4m 11s | Max:  4m 16s | Hits:  98%/5804  
      ðŸŸĐ GCC13              Pass: 100%/10  | Total:  1h 22m | Avg:  8m 17s | Max: 17m 30s | Hits:  97%/14791 
      ðŸŸĐ MSVC14.29          Pass: 100%/2   | Total: 43m 09s | Avg: 21m 34s | Max: 22m 56s | Hits:  98%/5453  
      ðŸŸĐ MSVC14.42          Pass: 100%/2   | Total: 41m 11s | Avg: 20m 35s | Max: 21m 17s | Hits:  98%/5525  
      ðŸŸĐ NVHPC25.1          Pass: 100%/2   | Total: 47m 50s | Avg: 23m 55s | Max: 35m 22s | Hits:  59%/5792  
    ðŸŸĐ cxx_family
      ðŸŸĐ Clang              Pass: 100%/16  | Total:  2h 13m | Avg:  8m 21s | Max: 22m 42s | Hits:  84%/43516 
      ðŸŸĐ GCC                Pass: 100%/21  | Total:  2h 52m | Avg:  8m 12s | Max: 18m 53s | Hits:  88%/46577 
      ðŸŸĐ MSVC               Pass: 100%/4   | Total:  1h 24m | Avg: 21m 05s | Max: 22m 56s | Hits:  98%/10978 
      ðŸŸĐ NVHPC              Pass: 100%/2   | Total: 47m 50s | Avg: 23m 55s | Max: 35m 22s | Hits:  59%/5792  
    ðŸŸĐ gpu
      ðŸŸĐ h100               Pass: 100%/2   | Total: 17m 04s | Avg:  8m 32s | Max: 12m 51s | Hits:  99%/3012  
      ðŸŸĐ rtx2080            Pass: 100%/41  | Total:  7h 01m | Avg: 10m 16s | Max: 35m 22s | Hits:  86%/103851
    ðŸŸĐ jobs
      ðŸŸĐ Build              Pass: 100%/37  | Total:  6h 10m | Avg: 10m 00s | Max: 35m 22s | Hits:  86%/106823
      ðŸŸĐ NVRTC              Pass: 100%/2   | Total: 33m 50s | Avg: 16m 55s | Max: 17m 30s | Hits:  90%/40    
      ðŸŸĐ Test               Pass: 100%/3   | Total: 32m 10s | Avg: 10m 43s | Max: 12m 51s
      ðŸŸĐ VerifyCodegen      Pass: 100%/1   | Total:  2m 11s | Avg:  2m 11s | Max:  2m 11s
    ðŸŸĐ sm
      ðŸŸĐ 75                 Pass: 100%/2   | Total: 33m 50s | Avg: 16m 55s | Max: 17m 30s | Hits:  90%/40    
      ðŸŸĐ 90                 Pass: 100%/2   | Total: 17m 04s | Avg:  8m 32s | Max: 12m 51s | Hits:  99%/3012  
      ðŸŸĐ 90;90a;100         Pass: 100%/1   | Total:  7m 15s | Avg:  7m 15s | Max:  7m 15s | Hits:  90%/3012  
    ðŸŸĐ std
      ðŸŸĐ 17                 Pass: 100%/21  | Total:  3h 46m | Avg: 10m 46s | Max: 22m 56s | Hits:  85%/57127 
      ðŸŸĐ 20                 Pass: 100%/21  | Total:  3h 29m | Avg:  9m 58s | Max: 35m 22s | Hits:  87%/49736 
    
  • ðŸŸĐ cudax: Pass: 100%/22 | Total: 2h 03m | Avg: 5m 36s | Max: 16m 03s | Hits: 99%/11852

    ðŸŸĐ cpu
      ðŸŸĐ amd64              Pass: 100%/18  | Total:  1h 51m | Avg:  6m 12s | Max: 16m 03s | Hits:  99%/9512  
      ðŸŸĐ arm64              Pass: 100%/4   | Total: 11m 41s | Avg:  2m 55s | Max:  2m 58s | Hits:  99%/2340  
    ðŸŸĐ ctk
      ðŸŸĐ 12.0               Pass: 100%/1   | Total:  8m 54s | Avg:  8m 54s | Max:  8m 54s | Hits:  95%/282   
      ðŸŸĐ 12.6               Pass: 100%/2   | Total: 11m 33s | Avg:  5m 46s | Max:  5m 50s | Hits:  96%/754   
      ðŸŸĐ 12.8               Pass: 100%/19  | Total:  1h 42m | Avg:  5m 25s | Max: 16m 03s | Hits:  99%/10816 
    ðŸŸĐ cudacxx
      ðŸŸĐ nvcc12.0           Pass: 100%/1   | Total:  8m 54s | Avg:  8m 54s | Max:  8m 54s | Hits:  95%/282   
      ðŸŸĐ nvcc12.6           Pass: 100%/2   | Total: 11m 33s | Avg:  5m 46s | Max:  5m 50s | Hits:  96%/754   
      ðŸŸĐ nvcc12.8           Pass: 100%/19  | Total:  1h 42m | Avg:  5m 25s | Max: 16m 03s | Hits:  99%/10816 
    ðŸŸĐ cudacxx_family
      ðŸŸĐ nvcc               Pass: 100%/22  | Total:  2h 03m | Avg:  5m 36s | Max: 16m 03s | Hits:  99%/11852 
    ðŸŸĐ cxx
      ðŸŸĐ Clang14            Pass: 100%/1   | Total:  3m 28s | Avg:  3m 28s | Max:  3m 28s | Hits: 100%/587   
      ðŸŸĐ Clang15            Pass: 100%/1   | Total:  3m 39s | Avg:  3m 39s | Max:  3m 39s | Hits: 100%/585   
      ðŸŸĐ Clang16            Pass: 100%/1   | Total:  3m 33s | Avg:  3m 33s | Max:  3m 33s | Hits: 100%/585   
      ðŸŸĐ Clang17            Pass: 100%/1   | Total:  3m 36s | Avg:  3m 36s | Max:  3m 36s | Hits: 100%/585   
      ðŸŸĐ Clang18            Pass: 100%/4   | Total: 25m 27s | Avg:  6m 21s | Max: 16m 03s | Hits: 100%/2340  
      ðŸŸĐ GCC10              Pass: 100%/1   | Total:  3m 15s | Avg:  3m 15s | Max:  3m 15s | Hits:  99%/587   
      ðŸŸĐ GCC11              Pass: 100%/1   | Total:  3m 38s | Avg:  3m 38s | Max:  3m 38s | Hits:  99%/585   
      ðŸŸĐ GCC12              Pass: 100%/2   | Total: 17m 17s | Avg:  8m 38s | Max: 13m 44s | Hits:  99%/1170  
      ðŸŸĐ GCC13              Pass: 100%/6   | Total: 29m 00s | Avg:  4m 50s | Max: 13m 49s | Hits:  99%/3510  
      ðŸŸĐ MSVC14.39          Pass: 100%/1   | Total:  8m 54s | Avg:  8m 54s | Max:  8m 54s | Hits:  95%/282   
      ðŸŸĐ MSVC14.42          Pass: 100%/1   | Total: 10m 03s | Avg: 10m 03s | Max: 10m 03s | Hits:  95%/282   
      ðŸŸĐ NVHPC25.1          Pass: 100%/2   | Total: 11m 33s | Avg:  5m 46s | Max:  5m 50s | Hits:  96%/754   
    ðŸŸĐ cxx_family
      ðŸŸĐ Clang              Pass: 100%/8   | Total: 39m 43s | Avg:  4m 57s | Max: 16m 03s | Hits: 100%/4682  
      ðŸŸĐ GCC                Pass: 100%/10  | Total: 53m 10s | Avg:  5m 19s | Max: 13m 49s | Hits:  99%/5852  
      ðŸŸĐ MSVC               Pass: 100%/2   | Total: 18m 57s | Avg:  9m 28s | Max: 10m 03s | Hits:  95%/564   
      ðŸŸĐ NVHPC              Pass: 100%/2   | Total: 11m 33s | Avg:  5m 46s | Max:  5m 50s | Hits:  96%/754   
    ðŸŸĐ gpu
      ðŸŸĐ h100               Pass: 100%/2   | Total: 16m 59s | Avg:  8m 29s | Max: 13m 49s | Hits:  99%/1170  
      ðŸŸĐ rtx2080            Pass: 100%/20  | Total:  1h 46m | Avg:  5m 19s | Max: 16m 03s | Hits:  99%/10682 
    ðŸŸĐ jobs
      ðŸŸĐ Build              Pass: 100%/19  | Total:  1h 19m | Avg:  4m 11s | Max: 10m 03s | Hits:  99%/10097 
      ðŸŸĐ Test               Pass: 100%/3   | Total: 43m 36s | Avg: 14m 32s | Max: 16m 03s | Hits:  99%/1755  
    ðŸŸĐ sm
      ðŸŸĐ 90                 Pass: 100%/3   | Total: 20m 01s | Avg:  6m 40s | Max: 13m 49s | Hits:  99%/1755  
      ðŸŸĐ 90a                Pass: 100%/1   | Total:  3m 04s | Avg:  3m 04s | Max:  3m 04s | Hits:  99%/585   
    ðŸŸĐ std
      ðŸŸĐ 17                 Pass: 100%/4   | Total: 14m 43s | Avg:  3m 40s | Max:  5m 50s | Hits:  99%/2132  
      ðŸŸĐ 20                 Pass: 100%/18  | Total:  1h 48m | Avg:  6m 02s | Max: 16m 03s | Hits:  99%/9720  
    
  • ðŸŸĐ stdpar: Pass: 100%/4 | Total: 16m 19s | Avg: 4m 04s | Max: 5m 04s

    ðŸŸĐ cpu
      ðŸŸĐ amd64              Pass: 100%/2   | Total:  9m 54s | Avg:  4m 57s | Max:  5m 04s
      ðŸŸĐ arm64              Pass: 100%/2   | Total:  6m 25s | Avg:  3m 12s | Max:  3m 16s
    ðŸŸĐ ctk
      ðŸŸĐ 12.6               Pass: 100%/4   | Total: 16m 19s | Avg:  4m 04s | Max:  5m 04s
    ðŸŸĐ cudacxx
      ðŸŸĐ nvcc12.6           Pass: 100%/4   | Total: 16m 19s | Avg:  4m 04s | Max:  5m 04s
    ðŸŸĐ cudacxx_family
      ðŸŸĐ nvcc               Pass: 100%/4   | Total: 16m 19s | Avg:  4m 04s | Max:  5m 04s
    ðŸŸĐ cxx
      ðŸŸĐ NVHPC25.1          Pass: 100%/4   | Total: 16m 19s | Avg:  4m 04s | Max:  5m 04s
    ðŸŸĐ cxx_family
      ðŸŸĐ NVHPC              Pass: 100%/4   | Total: 16m 19s | Avg:  4m 04s | Max:  5m 04s
    ðŸŸĐ gpu
      ðŸŸĐ rtx2080            Pass: 100%/4   | Total: 16m 19s | Avg:  4m 04s | Max:  5m 04s
    ðŸŸĐ jobs
      ðŸŸĐ Build              Pass: 100%/4   | Total: 16m 19s | Avg:  4m 04s | Max:  5m 04s
    ðŸŸĐ std
      ðŸŸĐ 17                 Pass: 100%/2   | Total:  8m 20s | Avg:  4m 10s | Max:  5m 04s
      ðŸŸĐ 20                 Pass: 100%/2   | Total:  7m 59s | Avg:  3m 59s | Max:  4m 50s
    
  • ðŸŸĐ cccl_c_parallel: Pass: 100%/2 | Total: 19m 20s | Avg: 9m 40s | Max: 16m 51s | Hits: 98%/320

    ðŸŸĐ cpu
      ðŸŸĐ amd64              Pass: 100%/2   | Total: 19m 20s | Avg:  9m 40s | Max: 16m 51s | Hits:  98%/320   
    ðŸŸĐ ctk
      ðŸŸĐ 12.8               Pass: 100%/2   | Total: 19m 20s | Avg:  9m 40s | Max: 16m 51s | Hits:  98%/320   
    ðŸŸĐ cudacxx
      ðŸŸĐ nvcc12.8           Pass: 100%/2   | Total: 19m 20s | Avg:  9m 40s | Max: 16m 51s | Hits:  98%/320   
    ðŸŸĐ cudacxx_family
      ðŸŸĐ nvcc               Pass: 100%/2   | Total: 19m 20s | Avg:  9m 40s | Max: 16m 51s | Hits:  98%/320   
    ðŸŸĐ cxx
      ðŸŸĐ GCC13              Pass: 100%/2   | Total: 19m 20s | Avg:  9m 40s | Max: 16m 51s | Hits:  98%/320   
    ðŸŸĐ cxx_family
      ðŸŸĐ GCC                Pass: 100%/2   | Total: 19m 20s | Avg:  9m 40s | Max: 16m 51s | Hits:  98%/320   
    ðŸŸĐ gpu
      ðŸŸĐ rtx2080            Pass: 100%/2   | Total: 19m 20s | Avg:  9m 40s | Max: 16m 51s | Hits:  98%/320   
    ðŸŸĐ jobs
      ðŸŸĐ Build              Pass: 100%/1   | Total:  2m 29s | Avg:  2m 29s | Max:  2m 29s | Hits:  98%/160   
      ðŸŸĐ Test               Pass: 100%/1   | Total: 16m 51s | Avg: 16m 51s | Max: 16m 51s | Hits:  98%/160   
    
  • ðŸŸĐ python: Pass: 100%/1 | Total: 1h 07m | Avg: 1h 07m | Max: 1h 07m

    ðŸŸĐ cpu
      ðŸŸĐ amd64              Pass: 100%/1   | Total:  1h 07m | Avg:  1h 07m | Max:  1h 07m
    ðŸŸĐ ctk
      ðŸŸĐ 12.8               Pass: 100%/1   | Total:  1h 07m | Avg:  1h 07m | Max:  1h 07m
    ðŸŸĐ cudacxx
      ðŸŸĐ nvcc12.8           Pass: 100%/1   | Total:  1h 07m | Avg:  1h 07m | Max:  1h 07m
    ðŸŸĐ cudacxx_family
      ðŸŸĐ nvcc               Pass: 100%/1   | Total:  1h 07m | Avg:  1h 07m | Max:  1h 07m
    ðŸŸĐ cxx
      ðŸŸĐ GCC13              Pass: 100%/1   | Total:  1h 07m | Avg:  1h 07m | Max:  1h 07m
    ðŸŸĐ cxx_family
      ðŸŸĐ GCC                Pass: 100%/1   | Total:  1h 07m | Avg:  1h 07m | Max:  1h 07m
    ðŸŸĐ gpu
      ðŸŸĐ rtx2080            Pass: 100%/1   | Total:  1h 07m | Avg:  1h 07m | Max:  1h 07m
    ðŸŸĐ jobs
      ðŸŸĐ Test               Pass: 100%/1   | Total:  1h 07m | Avg:  1h 07m | Max:  1h 07m
    

👃 Inspect Changes

Modifications in project?

Project
CCCL Infrastructure
+/- libcu++
+/- CUB
+/- Thrust
CUDA Experimental
stdpar
python
CCCL C Parallel Library
Catch2Helper

Modifications in project or dependencies?

Project
CCCL Infrastructure
+/- libcu++
+/- CUB
+/- Thrust
+/- CUDA Experimental
+/- stdpar
+/- python
+/- CCCL C Parallel Library
+/- Catch2Helper

🏃‍ Runner counts (total jobs: 162)

# Runner
113 linux-amd64-cpu16
15 windows-amd64-cpu16
12 linux-arm64-cpu16
8 linux-amd64-gpu-rtx2080-latest-1
6 linux-amd64-gpu-rtxa6000-latest-1
5 linux-amd64-gpu-h100-latest-1
3 linux-amd64-gpu-rtx4090-latest-1

@miscco
miscco enabled auto-merge (squash) March 24, 2025 19:43
@miscco
miscco merged commit 8b4d386 into NVIDIA:main Mar 24, 2025
@github-project-automation github-project-automation Bot moved this from In Review to Done in CCCL Mar 24, 2025
@miscco
miscco deleted the fix_stream_ref_par_n branch March 25, 2025 06:50
caugonnet pushed a commit to caugonnet/cccl that referenced this pull request Mar 25, 2025
* Make `cuda::stream_ref` constructible on device

There is no reason we should not be able to construct it or extract the pointer out of it on device.

* Pass by value

* Make nvbench work
caugonnet added a commit that referenced this pull request Mar 26, 2025
* Enable NVHPC in CUDASTF CI

* Use nvtx_range in the POTRI example

* no need to include nvtx3 anymore here

* Do not use nvtx3 directly

* fix compilation

* WIP: try to pass the type of the execution place to the parallel_for_scope

* pass the exec_place type to parallel_for_scope

* remove a device annotation for a host only lambda

* is_shape_of_v is not working

* experiment to see if we can avoid generating device code in parallel_for with a host place

* fixes for a previous conflict merge

* disable an invalid test

* better typing for ctx.parallel_for with a partition

* use ::std and more types for parallel_for

* Add transform c parallel implementation (#4048)

* Enable setting _CUB_HAS_TRANSFORM_UBLKCP as flag

* Stale comment

* Update make_kernel_user_binary_operator to accept lhs/rhs types

* Add transform c.parallel implementation

* Add tests for c.parallel transform

* Use launcher factory to query ptx version

* Refactor how we get iterator names

* Add comment explaining why we need -default-device

* Address remaining review feedback

* Undo change to make_iterator_kernel_arg

* Restore the comment

* Restore comment

---------

Co-authored-by: Ashwin Srinath <shwina@users.noreply.github.com>

* Drop duplicated system header blocks (#4245)

Those were accidentally duplicated

* Exclude sm101 from RDC testing. (#4247)

* Make `cuda::stream_ref` constructible on device (#4243)

* Make `cuda::stream_ref` constructible on device

There is no reason we should not be able to construct it or extract the pointer out of it on device.

* Pass by value

* Make nvbench work

* Fix logic in test_segmented_reduce (#4198)

* Fix logic in test_segmented_reduce, also test over different types of offsets

To resolve gh-4197, use `cupy.cumsum` to accumulate over random partition
sizes to form correct offsets sequence.

Add assertions to verify that `offsets` is a non-decreasing sequence, and
that its last element equals the size of the input array.

Perform the test for several plausible offset data types.

* Changes per PR review comments

1. Use `cupy.random` to draw random sample on GPU, rather than on CPU
   followed by a transfer
2. Use `cp.empty` to allocate output, rather than `cp.zeros`

* Add new `WarpReduce` overloadings (#3884)

Co-authored-by: Bernhard Manfred Gruber <bernhardmgruber@gmail.com>

* Fix #4250 (#4251)

* Refactor fp masks (#4246)

* Implement `views::all` (#4244)

* Implement `views::all`

* [cudax] incorporate P3557 (constexpr completion signatures) into Âĩstdex (#3841)

* incorporate P3557 (constexpr completion signatures) into ustdex

* remove the need for nvcc-specific workarounds

* review feedback

* tweaks to ustdex for clangd

* fix typo

* use concepts portability macros for C++17 support

* Add fixed size segmented reduce (#3969)

* Adds fixed size segmented reduce

* remove unnecessary changes

* revert changes

* Properly deal with parallel_for overloads with grids, add an is_host() for exec_place

* Remove test doing illegal things, and which is not really useful anymore

---------

Co-authored-by: Ashwin Srinath <3190405+shwina@users.noreply.github.com>
Co-authored-by: Ashwin Srinath <shwina@users.noreply.github.com>
Co-authored-by: Michael Schellenberger Costa <miscco@nvidia.com>
Co-authored-by: Allison Piper <alliepiper16@gmail.com>
Co-authored-by: Oleksandr Pavlyk <21087696+oleksandr-pavlyk@users.noreply.github.com>
Co-authored-by: Federico Busato <50413820+fbusato@users.noreply.github.com>
Co-authored-by: Bernhard Manfred Gruber <bernhardmgruber@gmail.com>
Co-authored-by: David Bayer <48736217+davebayer@users.noreply.github.com>
Co-authored-by: Eric Niebler <eniebler@nvidia.com>
Co-authored-by: Srinivas Yadav <43375352+srinivasyadav18@users.noreply.github.com>
xwang233 added a commit to xwang233/pytorch that referenced this pull request Mar 30, 2025
davebayer pushed a commit to davebayer/cccl that referenced this pull request Apr 7, 2025
* Make `cuda::stream_ref` constructible on device

There is no reason we should not be able to construct it or extract the pointer out of it on device.

* Pass by value

* Make nvbench work
davebayer added a commit to davebayer/cccl that referenced this pull request Apr 7, 2025
* Enable NVHPC in CUDASTF CI

* Use nvtx_range in the POTRI example

* no need to include nvtx3 anymore here

* Do not use nvtx3 directly

* fix compilation

* WIP: try to pass the type of the execution place to the parallel_for_scope

* pass the exec_place type to parallel_for_scope

* remove a device annotation for a host only lambda

* is_shape_of_v is not working

* experiment to see if we can avoid generating device code in parallel_for with a host place

* fixes for a previous conflict merge

* disable an invalid test

* better typing for ctx.parallel_for with a partition

* use ::std and more types for parallel_for

* Add transform c parallel implementation (NVIDIA#4048)

* Enable setting _CUB_HAS_TRANSFORM_UBLKCP as flag

* Stale comment

* Update make_kernel_user_binary_operator to accept lhs/rhs types

* Add transform c.parallel implementation

* Add tests for c.parallel transform

* Use launcher factory to query ptx version

* Refactor how we get iterator names

* Add comment explaining why we need -default-device

* Address remaining review feedback

* Undo change to make_iterator_kernel_arg

* Restore the comment

* Restore comment

---------

Co-authored-by: Ashwin Srinath <shwina@users.noreply.github.com>

* Drop duplicated system header blocks (NVIDIA#4245)

Those were accidentally duplicated

* Exclude sm101 from RDC testing. (NVIDIA#4247)

* Make `cuda::stream_ref` constructible on device (NVIDIA#4243)

* Make `cuda::stream_ref` constructible on device

There is no reason we should not be able to construct it or extract the pointer out of it on device.

* Pass by value

* Make nvbench work

* Fix logic in test_segmented_reduce (NVIDIA#4198)

* Fix logic in test_segmented_reduce, also test over different types of offsets

To resolve NVIDIAgh-4197, use `cupy.cumsum` to accumulate over random partition
sizes to form correct offsets sequence.

Add assertions to verify that `offsets` is a non-decreasing sequence, and
that its last element equals the size of the input array.

Perform the test for several plausible offset data types.

* Changes per PR review comments

1. Use `cupy.random` to draw random sample on GPU, rather than on CPU
   followed by a transfer
2. Use `cp.empty` to allocate output, rather than `cp.zeros`

* Add new `WarpReduce` overloadings (NVIDIA#3884)

Co-authored-by: Bernhard Manfred Gruber <bernhardmgruber@gmail.com>

* Fix NVIDIA#4250 (NVIDIA#4251)

* Refactor fp masks (NVIDIA#4246)

* Implement `views::all` (NVIDIA#4244)

* Implement `views::all`

* [cudax] incorporate P3557 (constexpr completion signatures) into Âĩstdex (NVIDIA#3841)

* incorporate P3557 (constexpr completion signatures) into ustdex

* remove the need for nvcc-specific workarounds

* review feedback

* tweaks to ustdex for clangd

* fix typo

* use concepts portability macros for C++17 support

* Add fixed size segmented reduce (NVIDIA#3969)

* Adds fixed size segmented reduce

* remove unnecessary changes

* revert changes

* Properly deal with parallel_for overloads with grids, add an is_host() for exec_place

* Remove test doing illegal things, and which is not really useful anymore

---------

Co-authored-by: Ashwin Srinath <3190405+shwina@users.noreply.github.com>
Co-authored-by: Ashwin Srinath <shwina@users.noreply.github.com>
Co-authored-by: Michael Schellenberger Costa <miscco@nvidia.com>
Co-authored-by: Allison Piper <alliepiper16@gmail.com>
Co-authored-by: Oleksandr Pavlyk <21087696+oleksandr-pavlyk@users.noreply.github.com>
Co-authored-by: Federico Busato <50413820+fbusato@users.noreply.github.com>
Co-authored-by: Bernhard Manfred Gruber <bernhardmgruber@gmail.com>
Co-authored-by: David Bayer <48736217+davebayer@users.noreply.github.com>
Co-authored-by: Eric Niebler <eniebler@nvidia.com>
Co-authored-by: Srinivas Yadav <43375352+srinivasyadav18@users.noreply.github.com>
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

Archived in project

Development

Successfully merging this pull request may close these issues.

5 participants