Skip to content

Fix issue where calling merge_sort on custom types was failing - #4367

Merged
NaderAlAwar merged 1 commit into
NVIDIA:mainfrom
NaderAlAwar:merge-sort-custom-types-fix
Apr 8, 2025
Merged

NaderAlAwar merged 1 commit into
NVIDIA:mainfrom
NaderAlAwar:merge-sort-custom-types-fix

Conversation

@NaderAlAwar

Copy link
Copy Markdown
Contributor

Description

closes #4240

Checklist

  • New or existing tests cover these changes.
  • The documentation is up to date with these changes.

â€Ķpatching merge sort. This was causing the merge sort custom types test to fail
@NaderAlAwar
NaderAlAwar requested review from a team as code owners April 7, 2025 21:54
@github-project-automation github-project-automation Bot moved this to Todo in CCCL Apr 7, 2025
@cccl-authenticator-app cccl-authenticator-app Bot moved this from Todo to In Review in CCCL Apr 7, 2025
@NaderAlAwar
NaderAlAwar requested a review from shwina April 7, 2025 21:54
@github-actions

github-actions Bot commented Apr 7, 2025

Copy link
Copy Markdown
Contributor
ðŸŸĐ CI finished in 1h 34m: Pass: 100%/101 | Total: 2d 23h | Avg: 42m 37s | Max: 1h 31m | Hits: 78%/140297
  • ðŸŸĐ cub: Pass: 100%/47 | Total: 1d 21h | Avg: 58m 23s | Max: 1h 23m | Hits: 72%/56506

    ðŸŸĐ cpu
      ðŸŸĐ amd64              Pass: 100%/45  | Total:  1d 19h | Avg: 58m 02s | Max:  1h 23m | Hits:  73%/54050 
      ðŸŸĐ arm64              Pass: 100%/2   | Total:  2h 12m | Avg:  1h 06m | Max:  1h 07m | Hits:  66%/2456  
    ðŸŸĐ ctk
      ðŸŸĐ 12.0               Pass: 100%/5   | Total:  5h 35m | Avg:  1h 07m | Max:  1h 14m | Hits:  67%/5970  
      ðŸŸĐ 12.8               Pass: 100%/42  | Total:  1d 16h | Avg: 57m 21s | Max:  1h 23m | Hits:  73%/50536 
    ðŸŸĐ cudacxx
      ðŸŸĐ ClangCUDA19        Pass: 100%/2   | Total:  1h 50m | Avg: 55m 20s | Max: 55m 50s | Hits:  72%/2120  
      ðŸŸĐ nvcc12.0           Pass: 100%/5   | Total:  5h 35m | Avg:  1h 07m | Max:  1h 14m | Hits:  67%/5970  
      ðŸŸĐ nvcc12.8           Pass: 100%/40  | Total:  1d 14h | Avg: 57m 27s | Max:  1h 23m | Hits:  73%/48416 
    ðŸŸĐ cudacxx_family
      ðŸŸĐ ClangCUDA          Pass: 100%/2   | Total:  1h 50m | Avg: 55m 20s | Max: 55m 50s | Hits:  72%/2120  
      ðŸŸĐ nvcc               Pass: 100%/45  | Total:  1d 19h | Avg: 58m 31s | Max:  1h 23m | Hits:  72%/54386 
    ðŸŸĐ cxx
      ðŸŸĐ Clang14            Pass: 100%/4   | Total:  4h 14m | Avg:  1h 03m | Max:  1h 05m | Hits:  66%/4920  
      ðŸŸĐ Clang15            Pass: 100%/2   | Total:  2h 10m | Avg:  1h 05m | Max:  1h 06m | Hits:  66%/2456  
      ðŸŸĐ Clang16            Pass: 100%/2   | Total:  2h 01m | Avg:  1h 00m | Max:  1h 01m | Hits:  66%/2456  
      ðŸŸĐ Clang17            Pass: 100%/2   | Total:  2h 07m | Avg:  1h 03m | Max:  1h 07m | Hits:  66%/2456  
      ðŸŸĐ Clang18            Pass: 100%/2   | Total:  2h 05m | Avg:  1h 02m | Max:  1h 04m | Hits:  66%/2456  
      ðŸŸĐ Clang19            Pass: 100%/7   | Total:  5h 44m | Avg: 49m 10s | Max:  1h 05m | Hits:  78%/8260  
      ðŸŸĐ GCC7               Pass: 100%/2   | Total:  2h 09m | Avg:  1h 04m | Max:  1h 07m | Hits:  66%/2460  
      ðŸŸĐ GCC8               Pass: 100%/1   | Total:  1h 03m | Avg:  1h 03m | Max:  1h 03m | Hits:  66%/1230  
      ðŸŸĐ GCC9               Pass: 100%/2   | Total:  2h 14m | Avg:  1h 07m | Max:  1h 08m | Hits:  66%/2460  
      ðŸŸĐ GCC10              Pass: 100%/2   | Total:  2h 04m | Avg:  1h 02m | Max:  1h 02m | Hits:  66%/2460  
      ðŸŸĐ GCC11              Pass: 100%/2   | Total:  2h 11m | Avg:  1h 05m | Max:  1h 08m | Hits:  66%/2456  
      ðŸŸĐ GCC12              Pass: 100%/2   | Total:  2h 12m | Avg:  1h 06m | Max:  1h 08m | Hits:  66%/2456  
      ðŸŸĐ GCC13              Pass: 100%/11  | Total:  7h 19m | Avg: 39m 59s | Max:  1h 07m | Hits:  84%/13508 
      ðŸŸĐ MSVC14.29          Pass: 100%/2   | Total:  2h 37m | Avg:  1h 18m | Max:  1h 23m | Hits:  71%/2100  
      ðŸŸĐ MSVC14.42          Pass: 100%/2   | Total:  2h 42m | Avg:  1h 21m | Max:  1h 21m | Hits:  71%/2100  
      ðŸŸĐ NVHPC25.3          Pass: 100%/2   | Total:  2h 44m | Avg:  1h 22m | Max:  1h 22m | Hits:  66%/2272  
    ðŸŸĐ cxx_family
      ðŸŸĐ Clang              Pass: 100%/19  | Total: 18h 24m | Avg: 58m 09s | Max:  1h 07m | Hits:  70%/23004 
      ðŸŸĐ GCC                Pass: 100%/22  | Total: 19h 14m | Avg: 52m 29s | Max:  1h 08m | Hits:  75%/27030 
      ðŸŸĐ MSVC               Pass: 100%/4   | Total:  5h 19m | Avg:  1h 19m | Max:  1h 23m | Hits:  71%/4200  
      ðŸŸĐ NVHPC              Pass: 100%/2   | Total:  2h 44m | Avg:  1h 22m | Max:  1h 22m | Hits:  66%/2272  
    ðŸŸĐ gpu
      ðŸŸĐ h100               Pass: 100%/3   | Total:  1h 16m | Avg: 25m 35s | Max: 27m 48s | Hits:  88%/3684  
      ðŸŸĐ rtx2080            Pass: 100%/36  | Total:  1d 15h | Avg:  1h 06m | Max:  1h 23m | Hits:  67%/42998 
      ðŸŸĐ rtxa6000           Pass: 100%/8   | Total:  4h 29m | Avg: 33m 42s | Max:  1h 05m | Hits:  91%/9824  
    ðŸŸĐ jobs
      ðŸŸĐ Build              Pass: 100%/39  | Total:  1d 18h | Avg:  1h 05m | Max:  1h 23m | Hits:  67%/46682 
      ðŸŸĐ DeviceLaunch       Pass: 100%/1   | Total: 27m 17s | Avg: 27m 17s | Max: 27m 17s | Hits:  99%/1228  
      ðŸŸĐ GraphCapture       Pass: 100%/1   | Total: 20m 16s | Avg: 20m 16s | Max: 20m 16s | Hits:  99%/1228  
      ðŸŸĐ HostLaunch         Pass: 100%/3   | Total:  1h 19m | Avg: 26m 38s | Max: 27m 04s | Hits:  99%/3684  
      ðŸŸĐ TestGPU            Pass: 100%/3   | Total:  1h 07m | Avg: 22m 37s | Max: 23m 00s | Hits:  99%/3684  
    ðŸŸĐ sm
      ðŸŸĐ 90                 Pass: 100%/3   | Total:  1h 16m | Avg: 25m 35s | Max: 27m 48s | Hits:  88%/3684  
      ðŸŸĐ 90;90a;100         Pass: 100%/1   | Total:  1h 07m | Avg:  1h 07m | Max:  1h 07m | Hits:  66%/1228  
    ðŸŸĐ std
      ðŸŸĐ 17                 Pass: 100%/21  | Total: 23h 32m | Avg:  1h 07m | Max:  1h 23m | Hits:  67%/25010 
      ðŸŸĐ 20                 Pass: 100%/26  | Total: 22h 11m | Avg: 51m 13s | Max:  1h 21m | Hits:  77%/31496 
    
  • ðŸŸĐ thrust: Pass: 100%/47 | Total: 23h 46m | Avg: 30m 20s | Max: 1h 07m | Hits: 81%/83463

    ðŸŸĐ cmake_options
      ðŸŸĐ -DTHRUST_DISPATCH_TYPE=Force32bit Pass: 100%/2   | Total: 36m 31s | Avg: 18m 15s | Max: 25m 04s | Hits:  89%/3554  
    ðŸŸĐ cpu
      ðŸŸĐ amd64              Pass: 100%/45  | Total: 22h 52m | Avg: 30m 30s | Max:  1h 07m | Hits:  81%/79910 
      ðŸŸĐ arm64              Pass: 100%/2   | Total: 53m 12s | Avg: 26m 36s | Max: 28m 15s | Hits:  79%/3553  
    ðŸŸĐ ctk
      ðŸŸĐ 12.0               Pass: 100%/5   | Total:  2h 52m | Avg: 34m 31s | Max: 52m 45s | Hits:  78%/8876  
      ðŸŸĐ 12.8               Pass: 100%/42  | Total: 20h 53m | Avg: 29m 50s | Max:  1h 07m | Hits:  82%/74587 
    ðŸŸĐ cudacxx
      ðŸŸĐ ClangCUDA19        Pass: 100%/2   | Total: 49m 18s | Avg: 24m 39s | Max: 25m 14s | Hits:  79%/3552  
      ðŸŸĐ nvcc12.0           Pass: 100%/5   | Total:  2h 52m | Avg: 34m 31s | Max: 52m 45s | Hits:  78%/8876  
      ðŸŸĐ nvcc12.8           Pass: 100%/40  | Total: 20h 04m | Avg: 30m 06s | Max:  1h 07m | Hits:  82%/71035 
    ðŸŸĐ cudacxx_family
      ðŸŸĐ ClangCUDA          Pass: 100%/2   | Total: 49m 18s | Avg: 24m 39s | Max: 25m 14s | Hits:  79%/3552  
      ðŸŸĐ nvcc               Pass: 100%/45  | Total: 22h 56m | Avg: 30m 35s | Max:  1h 07m | Hits:  81%/79911 
    ðŸŸĐ cxx
      ðŸŸĐ Clang14            Pass: 100%/4   | Total:  1h 58m | Avg: 29m 42s | Max: 33m 11s | Hits:  79%/7104  
      ðŸŸĐ Clang15            Pass: 100%/2   | Total: 56m 46s | Avg: 28m 23s | Max: 28m 35s | Hits:  79%/3552  
      ðŸŸĐ Clang16            Pass: 100%/2   | Total:  1h 01m | Avg: 30m 59s | Max: 31m 41s | Hits:  79%/3552  
      ðŸŸĐ Clang17            Pass: 100%/2   | Total: 57m 05s | Avg: 28m 32s | Max: 28m 47s | Hits:  79%/3552  
      ðŸŸĐ Clang18            Pass: 100%/2   | Total: 56m 31s | Avg: 28m 15s | Max: 28m 36s | Hits:  79%/3552  
      ðŸŸĐ Clang19            Pass: 100%/7   | Total:  2h 28m | Avg: 21m 14s | Max: 28m 17s | Hits:  85%/12432 
      ðŸŸĐ GCC7               Pass: 100%/2   | Total:  1h 01m | Avg: 30m 38s | Max: 30m 53s | Hits:  79%/3554  
      ðŸŸĐ GCC8               Pass: 100%/1   | Total: 30m 33s | Avg: 30m 33s | Max: 30m 33s | Hits:  79%/1777  
      ðŸŸĐ GCC9               Pass: 100%/2   | Total:  1h 05m | Avg: 32m 46s | Max: 34m 05s | Hits:  79%/3554  
      ðŸŸĐ GCC10              Pass: 100%/2   | Total:  1h 00m | Avg: 30m 29s | Max: 31m 45s | Hits:  79%/3554  
      ðŸŸĐ GCC11              Pass: 100%/2   | Total: 59m 17s | Avg: 29m 38s | Max: 29m 54s | Hits:  79%/3554  
      ðŸŸĐ GCC12              Pass: 100%/2   | Total:  1h 01m | Avg: 30m 59s | Max: 31m 03s | Hits:  79%/3554  
      ðŸŸĐ GCC13              Pass: 100%/10  | Total:  3h 28m | Avg: 20m 52s | Max: 33m 58s | Hits:  87%/17770 
      ðŸŸĐ MSVC14.29          Pass: 100%/2   | Total:  1h 46m | Avg: 53m 22s | Max: 54m 00s | Hits:  74%/3540  
      ðŸŸĐ MSVC14.42          Pass: 100%/3   | Total:  2h 23m | Avg: 47m 49s | Max: 57m 34s | Hits:  80%/5310  
      ðŸŸĐ NVHPC25.3          Pass: 100%/2   | Total:  2h 07m | Avg:  1h 03m | Max:  1h 07m | Hits:  74%/3552  
    ðŸŸĐ cxx_family
      ðŸŸĐ Clang              Pass: 100%/19  | Total:  8h 19m | Avg: 26m 18s | Max: 33m 11s | Hits:  81%/33744 
      ðŸŸĐ GCC                Pass: 100%/21  | Total:  9h 08m | Avg: 26m 06s | Max: 34m 05s | Hits:  83%/37317 
      ðŸŸĐ MSVC               Pass: 100%/5   | Total:  4h 10m | Avg: 50m 02s | Max: 57m 34s | Hits:  78%/8850  
      ðŸŸĐ NVHPC              Pass: 100%/2   | Total:  2h 07m | Avg:  1h 03m | Max:  1h 07m | Hits:  74%/3552  
    ðŸŸĐ gpu
      ðŸŸĐ h100               Pass: 100%/2   | Total: 27m 10s | Avg: 13m 35s | Max: 16m 15s | Hits:  89%/3554  
      ðŸŸĐ rtx2080            Pass: 100%/35  | Total: 19h 41m | Avg: 33m 44s | Max:  1h 07m | Hits:  78%/62156 
      ðŸŸĐ rtx4090            Pass: 100%/10  | Total:  3h 37m | Avg: 21m 46s | Max: 57m 20s | Hits:  90%/17753 
    ðŸŸĐ jobs
      ðŸŸĐ Build              Pass: 100%/40  | Total: 22h 18m | Avg: 33m 27s | Max:  1h 07m | Hits:  78%/71033 
      ðŸŸĐ TestCPU            Pass: 100%/3   | Total: 44m 04s | Avg: 14m 41s | Max: 28m 35s | Hits:  99%/5323  
      ðŸŸĐ TestGPU            Pass: 100%/4   | Total: 43m 57s | Avg: 10m 59s | Max: 11m 31s | Hits:  99%/7107  
    ðŸŸĐ sm
      ðŸŸĐ 90                 Pass: 100%/2   | Total: 27m 10s | Avg: 13m 35s | Max: 16m 15s | Hits:  89%/3554  
      ðŸŸĐ 90;90a;100         Pass: 100%/1   | Total: 33m 58s | Avg: 33m 58s | Max: 33m 58s | Hits:  79%/1777  
    ðŸŸĐ std
      ðŸŸĐ 17                 Pass: 100%/21  | Total: 12h 25m | Avg: 35m 29s | Max:  1h 07m | Hits:  78%/37287 
      ðŸŸĐ 20                 Pass: 100%/24  | Total: 10h 44m | Avg: 26m 50s | Max:  1h 00m | Hits:  83%/42622 
    
  • ðŸŸĐ stdpar: Pass: 100%/4 | Total: 19m 47s | Avg: 4m 56s | Max: 5m 51s

    ðŸŸĐ cpu
      ðŸŸĐ amd64              Pass: 100%/2   | Total: 11m 16s | Avg:  5m 38s | Max:  5m 51s
      ðŸŸĐ arm64              Pass: 100%/2   | Total:  8m 31s | Avg:  4m 15s | Max:  4m 19s
    ðŸŸĐ ctk
      ðŸŸĐ 12.8               Pass: 100%/4   | Total: 19m 47s | Avg:  4m 56s | Max:  5m 51s
    ðŸŸĐ cudacxx
      ðŸŸĐ nvcc12.8           Pass: 100%/4   | Total: 19m 47s | Avg:  4m 56s | Max:  5m 51s
    ðŸŸĐ cudacxx_family
      ðŸŸĐ nvcc               Pass: 100%/4   | Total: 19m 47s | Avg:  4m 56s | Max:  5m 51s
    ðŸŸĐ cxx
      ðŸŸĐ NVHPC25.3          Pass: 100%/4   | Total: 19m 47s | Avg:  4m 56s | Max:  5m 51s
    ðŸŸĐ cxx_family
      ðŸŸĐ NVHPC              Pass: 100%/4   | Total: 19m 47s | Avg:  4m 56s | Max:  5m 51s
    ðŸŸĐ gpu
      ðŸŸĐ rtx2080            Pass: 100%/4   | Total: 19m 47s | Avg:  4m 56s | Max:  5m 51s
    ðŸŸĐ jobs
      ðŸŸĐ Build              Pass: 100%/4   | Total: 19m 47s | Avg:  4m 56s | Max:  5m 51s
    ðŸŸĐ std
      ðŸŸĐ 17                 Pass: 100%/2   | Total:  9m 37s | Avg:  4m 48s | Max:  5m 25s
      ðŸŸĐ 20                 Pass: 100%/2   | Total: 10m 10s | Avg:  5m 05s | Max:  5m 51s
    
  • ðŸŸĐ cccl_c_parallel: Pass: 100%/2 | Total: 23m 38s | Avg: 11m 49s | Max: 21m 17s | Hits: 97%/328

    ðŸŸĐ cpu
      ðŸŸĐ amd64              Pass: 100%/2   | Total: 23m 38s | Avg: 11m 49s | Max: 21m 17s | Hits:  97%/328   
    ðŸŸĐ ctk
      ðŸŸĐ 12.8               Pass: 100%/2   | Total: 23m 38s | Avg: 11m 49s | Max: 21m 17s | Hits:  97%/328   
    ðŸŸĐ cudacxx
      ðŸŸĐ nvcc12.8           Pass: 100%/2   | Total: 23m 38s | Avg: 11m 49s | Max: 21m 17s | Hits:  97%/328   
    ðŸŸĐ cudacxx_family
      ðŸŸĐ nvcc               Pass: 100%/2   | Total: 23m 38s | Avg: 11m 49s | Max: 21m 17s | Hits:  97%/328   
    ðŸŸĐ cxx
      ðŸŸĐ GCC13              Pass: 100%/2   | Total: 23m 38s | Avg: 11m 49s | Max: 21m 17s | Hits:  97%/328   
    ðŸŸĐ cxx_family
      ðŸŸĐ GCC                Pass: 100%/2   | Total: 23m 38s | Avg: 11m 49s | Max: 21m 17s | Hits:  97%/328   
    ðŸŸĐ gpu
      ðŸŸĐ rtx2080            Pass: 100%/2   | Total: 23m 38s | Avg: 11m 49s | Max: 21m 17s | Hits:  97%/328   
    ðŸŸĐ jobs
      ðŸŸĐ Build              Pass: 100%/1   | Total:  2m 21s | Avg:  2m 21s | Max:  2m 21s | Hits:  95%/164   
      ðŸŸĐ Test               Pass: 100%/1   | Total: 21m 17s | Avg: 21m 17s | Max: 21m 17s | Hits:  98%/164   
    
  • ðŸŸĐ python: Pass: 100%/1 | Total: 1h 31m | Avg: 1h 31m | Max: 1h 31m

    ðŸŸĐ cpu
      ðŸŸĐ amd64              Pass: 100%/1   | Total:  1h 31m | Avg:  1h 31m | Max:  1h 31m
    ðŸŸĐ ctk
      ðŸŸĐ 12.8               Pass: 100%/1   | Total:  1h 31m | Avg:  1h 31m | Max:  1h 31m
    ðŸŸĐ cudacxx
      ðŸŸĐ nvcc12.8           Pass: 100%/1   | Total:  1h 31m | Avg:  1h 31m | Max:  1h 31m
    ðŸŸĐ cudacxx_family
      ðŸŸĐ nvcc               Pass: 100%/1   | Total:  1h 31m | Avg:  1h 31m | Max:  1h 31m
    ðŸŸĐ cxx
      ðŸŸĐ GCC13              Pass: 100%/1   | Total:  1h 31m | Avg:  1h 31m | Max:  1h 31m
    ðŸŸĐ cxx_family
      ðŸŸĐ GCC                Pass: 100%/1   | Total:  1h 31m | Avg:  1h 31m | Max:  1h 31m
    ðŸŸĐ gpu
      ðŸŸĐ rtx2080            Pass: 100%/1   | Total:  1h 31m | Avg:  1h 31m | Max:  1h 31m
    ðŸŸĐ jobs
      ðŸŸĐ Test               Pass: 100%/1   | Total:  1h 31m | Avg:  1h 31m | Max:  1h 31m
    

👃 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: 101)

# Runner
72 linux-amd64-cpu16
9 windows-amd64-cpu16
6 linux-arm64-cpu16
6 linux-amd64-gpu-rtxa6000-latest-1
3 linux-amd64-gpu-h100-latest-1
3 linux-amd64-gpu-rtx4090-latest-1
2 linux-amd64-gpu-rtx2080-latest-1

@oleksandr-pavlyk oleksandr-pavlyk left a comment

Copy link
Copy Markdown
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Just curious how you discovered this? Perhaps we should run cccl/c/parallel test suite from under compute-sanitizer.

@NaderAlAwar

Copy link
Copy Markdown
Contributor Author

Just curious how you discovered this?

@shwina reported the issue. Coincidentally I noticed that we were using the wrong key and value sizes while working on radix_sort, so I checked the existing algorithms to see where else we were doing that. Fixing it in merge_sort fixed this issue.

Perhaps we should run cccl/c/parallel test suite from under compute-sanitizer

I would definitely support that

@NaderAlAwar
NaderAlAwar merged commit 86b2470 into NVIDIA:main Apr 8, 2025
@github-project-automation github-project-automation Bot moved this from In Review to Done in CCCL Apr 8, 2025
davebayer pushed a commit to davebayer/cccl that referenced this pull request Apr 11, 2025
â€Ķpatching merge sort. This was causing the merge sort custom types test to fail for larger sizes (NVIDIA#4367)
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.

[BUG]: cccl.c merge_sort with custom types falis for larger inputs

4 participants