Task06 Александр Исаенков

[PlayingPeano]

Локальный вывод

$ ./main_merge_sort 1
Found 3 GPUs in 0.0845624 sec (OpenCL: 0.0448216 sec, Vulkan: 0.0395478 sec)
Available devices:
  Device #0: API: Vulkan. iGPU. AMD Radeon Vega 8 Graphics (RADV RAVEN). Free memory: 1786/2970 Mb.
  Device #1: API: OpenCL. CPU. cpu-haswell-AMD Ryzen 5 3500U with Radeon Vega Mobile Gfx. AuthenticAMD. Total memory: 5146 Mb.
  Device #2: API: Vulkan. CPU. llvmpipe (LLVM 21.1.6, 256 bits). Free memory: 6862/6862 Mb.
Using device #1: API: OpenCL. CPU. cpu-haswell-AMD Ryzen 5 3500U with Radeon Vega Mobile Gfx. AuthenticAMD. Total memory: 5146 Mb.
Using OpenCL API...
n=100000000 values in range [1; 2147483646]
sorting on CPU...
CPU std::sort finished in 35.2513 sec
CPU std::sort effective RAM bandwidth: 0.0211356 GB/s (2.83677 uint millions/s)
Kernels compilation done in 0.093194 seconds
GPU merge-sort times (in seconds) - 10 values (min=18.2102 10%=18.4008 median=20.9557 90%=21.5827 max=21.5827)
GPU merge-sort median effective VRAM bandwidth: 0.035554 GB/s (4.77198 uint millions/s)

[GPUcourseBOT]

✅ Результаты тестирования PR #1048

Логи тестирования (нажмите чтобы развернуть)

=== СТАТУС: Успешно выполнены программы: main_merge_sort ===
=== main_merge_sort stdout (exit code: -11 (segfault после выполнения)) ===
Found 1 GPUs in 8.65016 sec (CUDA: 0.117925 sec, OpenCL: 1.56522 sec, Vulkan: 6.96696 sec)
Available devices:
Device #0: API: CUDA+OpenCL+Vulkan. GPU. Tesla T4 (CUDA 12020). Free memory: 14822/14930 Mb.
Using device #0: API: CUDA+OpenCL+Vulkan. GPU. Tesla T4 (CUDA 12020). Free memory: 14822/14930 Mb.
Using OpenCL API...
n=100000000 values in range [1; 2147483646]
sorting on CPU...
CPU std::sort finished in 11.564 sec
CPU std::sort effective RAM bandwidth: 0.0644289 GB/s (8.6475 uint millions/s)
Kernels compilation done in 4.6272 seconds
GPU merge-sort times (in seconds) - 10 values (min=0.270342 10%=0.270975 median=0.271458 90%=4.99486 max=4.99486)
GPU merge-sort median effective VRAM bandwidth: 2.74466 GB/s (368.381 uint millions/s)

Посмотреть полные логи

[GPUcourseBOT]

✅ Результаты тестирования PR #1048

Логи тестирования (нажмите чтобы развернуть)

=== СТАТУС: Успешно выполнены программы: main_merge_sort ===
=== main_merge_sort stdout (exit code: -11 (segfault после выполнения)) ===
Found 1 GPUs in 0.288912 sec (CUDA: 0.122584 sec, OpenCL: 0.0374053 sec, Vulkan: 0.128866 sec)
Available devices:
Device #0: API: CUDA+OpenCL+Vulkan. GPU. Tesla T4 (CUDA 12020). Free memory: 14822/14930 Mb.
Using device #0: API: CUDA+OpenCL+Vulkan. GPU. Tesla T4 (CUDA 12020). Free memory: 14822/14930 Mb.
Using OpenCL API...
n=100000000 values in range [1; 2147483646]
sorting on CPU...
CPU std::sort finished in 11.3193 sec
CPU std::sort effective RAM bandwidth: 0.0658215 GB/s (8.83442 uint millions/s)
Kernels compilation done in 0.0603096 seconds
GPU merge-sort times (in seconds) - 10 values (min=0.270826 10%=0.271254 median=0.271912 90%=0.415427 max=0.415427)
GPU merge-sort median effective VRAM bandwidth: 2.74007 GB/s (367.766 uint millions/s)

Посмотреть полные логи

[GPUcourseBOT]

✅ Результаты тестирования PR #1048

Логи тестирования (нажмите чтобы развернуть)

=== СТАТУС: Успешно выполнены программы: main_merge_sort ===
=== main_merge_sort stdout (exit code: -11 (segfault после выполнения)) ===
Found 1 GPUs in 0.315448 sec (CUDA: 0.120847 sec, OpenCL: 0.0374985 sec, Vulkan: 0.157041 sec)
Available devices:
Device #0: API: CUDA+OpenCL+Vulkan. GPU. Tesla T4 (CUDA 12020). Free memory: 14822/14930 Mb.
Using device #0: API: CUDA+OpenCL+Vulkan. GPU. Tesla T4 (CUDA 12020). Free memory: 14822/14930 Mb.
Using OpenCL API...
n=100000000 values in range [1; 2147483646]
sorting on CPU...
CPU std::sort finished in 11.4936 sec
CPU std::sort effective RAM bandwidth: 0.0648236 GB/s (8.70048 uint millions/s)
Kernels compilation done in 0.0645857 seconds
GPU merge-sort times (in seconds) - 10 values (min=0.270352 10%=0.270777 median=0.272053 90%=0.436599 max=0.436599)
GPU merge-sort median effective VRAM bandwidth: 2.73865 GB/s (367.575 uint millions/s)

Посмотреть полные логи

[GPUcourseBOT]

✅ Результаты тестирования PR #1048

Логи тестирования (нажмите чтобы развернуть)

=== СТАТУС: Успешно выполнены программы: main_merge_sort ===
=== main_merge_sort stdout (exit code: -11 (segfault после выполнения)) ===
Found 1 GPUs in 0.296224 sec (CUDA: 0.122169 sec, OpenCL: 0.0395973 sec, Vulkan: 0.134392 sec)
Available devices:
Device #0: API: CUDA+OpenCL+Vulkan. GPU. Tesla T4 (CUDA 12020). Free memory: 14822/14930 Mb.
Using device #0: API: CUDA+OpenCL+Vulkan. GPU. Tesla T4 (CUDA 12020). Free memory: 14822/14930 Mb.
Using OpenCL API...
n=100000000 values in range [1; 2147483646]
sorting on CPU...
CPU std::sort finished in 11.5124 sec
CPU std::sort effective RAM bandwidth: 0.0647178 GB/s (8.68627 uint millions/s)
Kernels compilation done in 0.0577272 seconds
GPU merge-sort times (in seconds) - 10 values (min=0.270201 10%=0.270948 median=0.272898 90%=0.410731 max=0.410731)
GPU merge-sort median effective VRAM bandwidth: 2.73018 GB/s (366.438 uint millions/s)

Посмотреть полные логи