Hi, Intel x86 simd sort library can be used via C++ template. If I wanted to make this library available to the maximum of Julia users (a bit like numpy did) what would be the best steps to follow? Should I use BinaryBuilder.jl? Thank you

BinaryBuilder would be the way to do this. that said, as an alternative, it shouldn’t be too hard to copy the algorithm to Julia

We are clearly not getting the same results. Sorry for that. What I don’t understand is that in anotger thread that I linked above you are showing numpy outperforming julia. So it seems that you have inconsistent results.

[image] gitboy16: So it seems that you have inconsistent results. Did you notice the different CPU (that was AVX2, this is AVX512)?

Which makes no sense at all :rofl: Anyway let’s put that to bed, I think we are both wasting our time on this one.

[image] gitboy16: If you don’t have a intel processor Is the library really specific to Intel processors, not x86 with AVX2 or AVX512 in general?

[image] Benny: s the library really specific to Intel processors I don’t see anything suggesting that in the README of the project , and again , it’d be hard to artificially cripple performance on non-Intel CPUs in open source code as that’d be in plain sight. Their code uses intrinsics for the…

[image] gitboy16: Which makes no sense at all :rofl: In the original post, it was explained quite clearly why AVX2 performance was bad in Julia [image] Numpy.sort vs Julia sort This is because I was optimizing sorting while developing on an apple machine without AVX2. If we had cross-plat…

[image] giordano: [image] Benny: s the library really specific to Intel processors I don’t see anything suggesting zen4 issue : performance on amd 7950x ... · Issue #6 · intel/x86-simd-sort · GitHub And OpenJDK adapted the code ( OpenJDK Merges Intel’s x86-simd-sort For Speeding Up…

As requested here is my code: /* test.cpp */ #include "src/x86simdsort-static-incl.h" // g++ test.cpp -mavx512f -mavx512dq -mavx512vl -O3 -o libsort.dll -shared -I./x86-simd-sort -fopenmp -DXSS_USE_OPENMP extern "C" void qsort_float(float *arr, size_t size) { x86simdsortStatic::qsort(arr, si…

[image] ImreSamu: zen4 issue Wonder if this is better with Google’s Highway , another manually vectorized C++ library? They seem to more explicitly take processors into account and its vqsort showed some advantage in a benchmark by ipnsort ’s author a couple years ago. I think any serious bench…

[image] ImreSamu: [image] giordano: [image] Benny: s the library really specific to Intel processors I don’t see anything suggesting zen4 issue : performance on amd 7950x … · Issue #6 · intel/x86-simd-sort · GitHub That’s a relevant issue, but a performance trap in a single inst…

Bring Intel x86 simd sort library to Julia

General Usage

jling August 20, 2025, 12:36am 22

In the original post, it was explained quite clearly why AVX2 performance was bad in Julia

Topic		Replies	Views
Why is Julia faster than C++ for quicksort? Performance performance , quicksort	15	2133	August 15, 2023
Squeeze out the last 10% of performance for a sorting function? Performance sort	26	3598	July 18, 2021
We love sorting -- latest news in sorting Performance	1	537	December 3, 2024
In as few lines as possible describe why you love julia New to Julia	113	13076	May 21, 2021
Ironic observation about `sort` and `sortperm` speed for "small integers" vs R Performance sort , sortperm , r	32	4835	February 4, 2018

Bring Intel x86 simd sort library to Julia

Related topics