Is there a reference-level implementation of BLAS in C/C++?

Question

The netlib BLAS implementation is an excellent reference, being mostly un-optimized and well documented (e.g. zgemm). However, it is in Fortran 77, making it somewhat inaccessible to those with a more modern programming education. Is there a reference-level implementation of BLAS, like netlib, in C/C++?

score 10 · Accepted Answer · answered Oct 27 '13 at 17:31

10

Have you looked at GNU Scientific Library's implementation? I find the source code to be sufficiently readable and the routines are well documented.

answered Oct 27 '13 at 17:31

Juan M. Bello-Rivas

3,994
16
24

Looks good to me. The documentation is a bit lacking, but the variable names are chosen well enough that I think its clear. I'll probably prepend the opening comment of the netlib BLAS routines. What exactly do you take issue with? Do you have an alternative? – Max Hutchinson Oct 28 '13 at 10:21

score 6 · Answer 2 · answered Oct 28 '13 at 01:33

6

A notable, C language implementation of BLAS is ATLAS. Among useful features:

Algebra routines implemented both as straightforward C as well as highly-optimized assembler assisted versions for multiple architectures and variants.
The build system features an "auto-tuner" which compiles multiple variants of the ATLAS library to establish which one will be the fastest on the given machine.

http://math-atlas.sourceforge.net/

answered Oct 28 '13 at 01:33

oakad

161
1

I looked at ATLAS but missed this. The path to the reference implementation is "src/blas/reference", with "ref" inserted between the type character and routine name and with character arguments appended. – Max Hutchinson Oct 28 '13 at 03:04

score 3 · Answer 3 · answered Oct 27 '13 at 17:09

3

Netlib also produces CLAPACK, which includes BLAS, but it is just the fortran code run through f2c and is therefore a bit clunky (e.g. zgemm).

answered Oct 27 '13 at 17:09

Max Hutchinson

3,051
16
29

score 1 · Answer 4 · edited Apr 01 '18 at 22:35

1

For a high-performance implementation that is not only among the highest performing (better than 85% of peak on 60 cores of the Intel Xeon Phi), but is also imho the most beautifully written, have a look at BLIS:

https://github.com/flame/blis

edited Apr 01 '18 at 22:35

Juan M. Bello-Rivas

3,994
16
24

answered Apr 11 '14 at 14:37

Robert van de Geijn

215
1
3

This does not seem like it claims to be a reference-level implementation – mabraham Apr 17 '14 at 20:54
We like to thin of it as a high-performance reference implementation. – Robert van de Geijn Apr 01 '18 at 13:55

score 0 · Answer 5 · answered Oct 27 '13 at 18:50

0

I have implementations of some bits of BLAS/LAPACK in RNP and RNP2.

answered Oct 27 '13 at 18:50

Victor Liu

4,480
18
28

This is interesting, but it is definitely post-BLAS and less straight forward than GSL and the reference implementation in ATLAS. – Max Hutchinson Oct 28 '13 at 03:13

score -1 · Answer 6 · answered Apr 01 '18 at 13:58

We are currently working on a Massive Open Online Course, "LAFF-On High-Performance Computing" that uses dgemm as the example that leads one through different levels of parallelization: instruction level, OpenMP, MPI.

This is not a reference implementation for the BLAS, but it is a reference for how to code the BLAS (for performance). To be kept informed, visit www.ulaff.net

Is there a reference-level implementation of BLAS in C/C++?

6 Answers6