Skip to content

Instantly share code, notes, and snippets.

@genkuroki

genkuroki/Octavian M=K=N=1024.ipynb

Last active May 29, 2021 12:09
Show Gist options
  • Save genkuroki/6123aef79488bc20b52047656fc6f015 to your computer and use it in GitHub Desktop.
Save genkuroki/6123aef79488bc20b52047656fc6f015 to your computer and use it in GitHub Desktop.
Download ZIP
Octavian!
Display the source blob
Display the rendered blob
Raw
Octavian M=K=N=1024.ipynb
Loading
Sorry, something went wrong. Reload?
Sorry, we cannot display this file.
Sorry, this file is invalid so it cannot be displayed.
Display the source blob
Display the rendered blob
Raw
Octavian M=K=N=128.ipynb
Loading
Sorry, something went wrong. Reload?
Sorry, we cannot display this file.
Sorry, this file is invalid so it cannot be displayed.
Display the source blob
Display the rendered blob
Raw
Octavian M=K=N=32.ipynb
Loading
Sorry, something went wrong. Reload?
Sorry, we cannot display this file.
Sorry, this file is invalid so it cannot be displayed.
Display the source blob
Display the rendered blob
Raw
Octavian M=K=N=512.ipynb
{
"cells": [
{
"metadata": {},
"cell_type": "markdown",
"source": "https://discourse.julialang.org/t/intel-c-c-compiler-performance-versus-julia/61929/18"
},
{
"metadata": {
"trusted": true
},
"cell_type": "code",
"source": "versioninfo()",
"execution_count": 1,
"outputs": [
{
"output_type": "stream",
"text": "Julia Version 1.7.0-DEV.1129\nCommit 9117b4d6d6 (2021-05-20 16:42 UTC)\nPlatform Info:\n OS: Windows (x86_64-w64-mingw32)\n CPU: Intel(R) Core(TM) i7-10750H CPU @ 2.60GHz\n WORD_SIZE: 64\n LIBM: libopenlibm\n LLVM: libLLVM-11.0.1 (ORCJIT, skylake)\nEnvironment:\n JULIA_NUM_THREADS = 12\n JULIA_PYTHONCALL_EXE = C:\\Users\\genkuroki\\.julia\\conda\\3\\python.exe\n",
"name": "stdout"
}
]
},
{
"metadata": {
"trusted": true
},
"cell_type": "code",
"source": "using LinearAlgebra\nusing BLASBenchmarksCPU\nusing Octavian\nusing BenchmarkHistograms",
"execution_count": 2,
"outputs": []
},
{
"metadata": {
"trusted": true
},
"cell_type": "code",
"source": "M = K = N = 512\nA = rand(M, K)\nB = rand(K, N)\nC1 = @time(A * B)\nC0 = similar(C1);",
"execution_count": 3,
"outputs": [
{
"output_type": "stream",
"text": " 0.607834 seconds (2.53 M allocations: 136.127 MiB, 6.48% gc time, 99.52% compilation time)\n",
"name": "stdout"
}
]
},
{
"metadata": {
"trusted": true
},
"cell_type": "code",
"source": "@benchmark mul!($C0, $A, $B)",
"execution_count": 4,
"outputs": [
{
"output_type": "execute_result",
"execution_count": 4,
"data": {
"text/plain": "samples: 2404; evals/sample: 1; memory estimate: 0 bytes; allocs estimate: 0\nns\n\n (1.36e6 - 1.64e6 ] \u001b[32m▋\u001b[39m40\n (1.64e6 - 1.93e6 ] \u001b[32m█▉\u001b[39m124\n (1.93e6 - 2.21e6 ] \u001b[32m██████████████████████████████ \u001b[39m2053\n (2.21e6 - 2.5e6 ] \u001b[32m██▎\u001b[39m148\n (2.5e6 - 2.78e6 ] \u001b[32m▌\u001b[39m28\n (2.78e6 - 3.06e6 ] \u001b[32m▏\u001b[39m2\n (3.06e6 - 3.35e6 ] \u001b[32m▏\u001b[39m1\n (3.35e6 - 3.63e6 ] \u001b[32m▏\u001b[39m1\n (3.63e6 - 3.92e6 ] \u001b[32m▏\u001b[39m3\n (3.92e6 - 4.2e6 ] \u001b[32m \u001b[39m0\n (4.2e6 - 4.48e6 ] \u001b[32m▏\u001b[39m1\n (4.48e6 - 4.77e6 ] \u001b[32m \u001b[39m0\n (4.77e6 - 1.145e7] \u001b[32m▏\u001b[39m3\n\n Counts\n\nmin: 1.359 ms (0.00% GC); mean: 2.065 ms (0.00% GC); median: 2.046 ms (0.00% GC); max: 11.455 ms (0.00% GC)."
},
"metadata": {}
}
]
},
{
"metadata": {
"trusted": true
},
"cell_type": "code",
"source": "# MKL dgemm\n@benchmark gemmmkl!($C0, $A, $B)",
"execution_count": 5,
"outputs": [
{
"output_type": "execute_result",
"execution_count": 5,
"data": {
"text/plain": "samples: 2801; evals/sample: 1; memory estimate: 0 bytes; allocs estimate: 0\nns\n\n (1.03e6 - 1.19e6 ] \u001b[32m▎\u001b[39m10\n (1.19e6 - 1.34e6 ] \u001b[32m▋\u001b[39m35\n (1.34e6 - 1.5e6 ] \u001b[32m█▏\u001b[39m63\n (1.5e6 - 1.65e6 ] \u001b[32m█▉\u001b[39m108\n (1.65e6 - 1.81e6 ] \u001b[32m██████████████████████████████ \u001b[39m1745\n (1.81e6 - 1.96e6 ] \u001b[32m████████████▊\u001b[39m737\n (1.96e6 - 2.12e6 ] \u001b[32m█▌\u001b[39m83\n (2.12e6 - 2.27e6 ] \u001b[32m▎\u001b[39m11\n (2.27e6 - 2.42e6 ] \u001b[32m▏\u001b[39m5\n (2.42e6 - 2.58e6 ] \u001b[32m \u001b[39m0\n (2.58e6 - 2.73e6 ] \u001b[32m \u001b[39m0\n (2.73e6 - 2.89e6 ] \u001b[32m▏\u001b[39m1\n (2.89e6 - 2.869e7] \u001b[32m▏\u001b[39m3\n\n Counts\n\nmin: 1.035 ms (0.00% GC); mean: 1.771 ms (0.00% GC); median: 1.757 ms (0.00% GC); max: 28.691 ms (0.00% GC)."
},
"metadata": {}
}
]
},
{
"metadata": {
"trusted": true
},
"cell_type": "code",
"source": "# MKL dgemm_direct\n@benchmark gemmmkl_direct!($C0, $A, $B)",
"execution_count": 6,
"outputs": [
{
"output_type": "execute_result",
"execution_count": 6,
"data": {
"text/plain": "samples: 2736; evals/sample: 1; memory estimate: 0 bytes; allocs estimate: 0\nns\n\n (1.07e6 - 1.39e6] \u001b[32m█\u001b[39m67\n (1.39e6 - 1.71e6] \u001b[32m█████▎\u001b[39m376\n (1.71e6 - 2.03e6] \u001b[32m██████████████████████████████ \u001b[39m2187\n (2.03e6 - 2.35e6] \u001b[32m█▎\u001b[39m90\n (2.35e6 - 2.67e6] \u001b[32m▏\u001b[39m1\n (2.67e6 - 2.99e6] \u001b[32m▏\u001b[39m1\n (2.99e6 - 3.31e6] \u001b[32m▏\u001b[39m4\n (3.31e6 - 3.63e6] \u001b[32m▏\u001b[39m3\n (3.63e6 - 3.95e6] \u001b[32m▏\u001b[39m1\n (3.95e6 - 4.27e6] \u001b[32m▏\u001b[39m1\n (4.27e6 - 4.59e6] \u001b[32m▏\u001b[39m2\n (4.59e6 - 4.91e6] \u001b[32m \u001b[39m0\n (4.91e6 - 5.23e6] \u001b[32m▏\u001b[39m3\n\n Counts\n\nmin: 1.074 ms (0.00% GC); mean: 1.811 ms (0.00% GC); median: 1.803 ms (0.00% GC); max: 5.230 ms (0.00% GC)."
},
"metadata": {}
}
]
},
{
"metadata": {
"trusted": true
},
"cell_type": "code",
"source": "# Octavian.jl\n@benchmark matmul!($C0, $A, $B)",
"execution_count": 7,
"outputs": [
{
"output_type": "execute_result",
"execution_count": 7,
"data": {
"text/plain": "samples: 2801; evals/sample: 1; memory estimate: 0 bytes; allocs estimate: 0\nns\n\n (1.05e6 - 1.16e6 ] \u001b[32m▍\u001b[39m11\n (1.16e6 - 1.27e6 ] \u001b[32m▍\u001b[39m16\n (1.27e6 - 1.38e6 ] \u001b[32m█▎\u001b[39m50\n (1.38e6 - 1.5e6 ] \u001b[32m█▍\u001b[39m58\n (1.5e6 - 1.61e6 ] \u001b[32m█\u001b[39m39\n (1.61e6 - 1.72e6 ] \u001b[32m█████████████▊\u001b[39m587\n (1.72e6 - 1.83e6 ] \u001b[32m██████████████████████████████ \u001b[39m1290\n (1.83e6 - 1.94e6 ] \u001b[32m██████████████▉\u001b[39m637\n (1.94e6 - 2.05e6 ] \u001b[32m██▎\u001b[39m95\n (2.05e6 - 2.17e6 ] \u001b[32m▎\u001b[39m9\n (2.17e6 - 2.28e6 ] \u001b[32m▏\u001b[39m5\n (2.28e6 - 2.39e6 ] \u001b[32m▏\u001b[39m1\n (2.39e6 - 2.999e7] \u001b[32m▏\u001b[39m3\n\n Counts\n\nmin: 1.049 ms (0.00% GC); mean: 1.773 ms (0.00% GC); median: 1.774 ms (0.00% GC); max: 29.993 ms (0.00% GC)."
},
"metadata": {}
}
]
},
{
"metadata": {
"trusted": true
},
"cell_type": "code",
"source": "",
"execution_count": null,
"outputs": []
}
],
"metadata": {
"@webio": {
"lastKernelId": null,
"lastCommId": null
},
"kernelspec": {
"name": "julia-1.7-depwarn-o3",
"display_name": "Julia 1.7.0-DEV depwarn -O3",
"language": "julia"
},
"language_info": {
"file_extension": ".jl",
"name": "julia",
"mimetype": "application/julia",
"version": "1.7.0"
},
"toc": {
"nav_menu": {},
"number_sections": true,
"sideBar": true,
"skip_h1_title": false,
"base_numbering": 1,
"title_cell": "Table of Contents",
"title_sidebar": "Contents",
"toc_cell": false,
"toc_position": {},
"toc_section_display": true,
"toc_window_display": false
},
"gist": {
"id": "6123aef79488bc20b52047656fc6f015",
"data": {
"description": "Octavian!",
"public": true
}
},
"_draft": {
"nbviewer_url": "https://gist.github.com/6123aef79488bc20b52047656fc6f015"
}
},
"nbformat": 4,
"nbformat_minor": 5
}
Display the source blob
Display the rendered blob
Raw
Octavian M=K=N=8.ipynb
Loading
Sorry, something went wrong. Reload?
Sorry, we cannot display this file.
Sorry, this file is invalid so it cannot be displayed.
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment