feat: add C and Fortran implementation for blas/base/zaxpy

[gururaj1512]

Resolves a part of #2039.

Description

What is the purpose of this pull request?

This pull request:

• Adds C and fortran implementation for blas/base/zaxpy

Related Issues

Does this pull request have any related issues?

This pull request:

• Resolves a part of [RFC]: Add BLAS bindings and implementations for linear algebra (tracking issue) #2039.

Questions

Any questions for reviewers of this pull request?

No.

Other

Any other information relevant to this pull request? This may include screenshots, references, and/or implementation notes.

No.

Checklist

Please ensure the following tasks are completed before submitting this pull request.

• Read, understood, and followed the contributing guidelines.

---

@stdlib-js/reviewers

[gururaj1512]

@kgryte, @aman-095, how can i resove below error
65 | intrinsic dcabs1 | 1 Error: ‘dcabs1’ declared INTRINSIC at (1) does not exist

[kgryte]

@gururaj1512 Yes, this is a known issue. Namely, our tooling needs to be updated to handle Fortran dependencies. Until then, adding Fortran implementations for packages, such as zaxpy, is blocked.

[gururaj1512]

@kgryte, Can't we solve it by

function dcabs1(z) result(res)
    implicit none
    complex(kind=kind(0.0d0)), intent(in) :: z
    real(kind=kind(0.0d0)) :: res
    
    res = abs(real(z)) + abs(aimag(z))
end function dcabs1

[kgryte]

@gururaj1512 That could be a quick fix, but it would not solve the larger issue for other packages which also rely on Fortran functions defined in other packages. In those instances, we need to the resolution of package Fortran files and compiling together. We can already do this for C, but we've yet to update to tooling to appropriately handle Fortran source files. In this case, we've added a dcabs package, but are currently unable to use it within other BLAS routines.

[kgryte]

I'm fine with doing something similar to your suggestion as a quick fix for this package with a FIXME indicating that this needs to be patched once we've fixed project tooling.

[gururaj1512]

@kgryte, Should I proceed with quick fix with a FIXME tag for now to keep this package functional or we can proceed once the Fortran tooling integration is improved. Let me know if there's anything specific you'd like me to include in the comment or implementation.

[kgryte]

@gururaj1512 Sure, feel free to proceed. As for comment, "FIXME: replace with blas/base/dcabs1 once library manifest tooling is fixed".

[gururaj1512]

Thank you for clarifying!

[gururaj1512]

@kgryte, Are there any commands to test fortran implementation before pushing the code.

[kgryte]

The way we test all native implementations is via JavaScript. Namely, you need to compile the native add-on against the desired set of source files. Locally, you can tweak the manifest.json file such that, based on your OS, project tooling resolves the correct set of source files. Note, however, that if you are on Windows, you likely won't have a Fortran compiler installed. And I am not sure how well, say, GFortran works under WSL (Linux subsystem for Windows).

[kgryte]

If you want to know which source files are being resolved during compilation of the native add-on, you can set the DEBUG environment variable to *. E.g.,

DEBUG=* make install-node-addons NODE_ADDONS_PATTERN="blas/base/zaxpy"

[gururaj1512]

The way we test all native implementations is via JavaScript. Namely, you need to compile the native add-on against the desired set of source files. Locally, you can tweak the manifest.json file such that, based on your OS, project tooling resolves the correct set of source files. Note, however, that if you are on Windows, you likely won't have a Fortran compiler installed. And I am not sure how well, say, GFortran works under WSL (Linux subsystem for Windows).

GFortran is not working well, at least on my machine. I am trying to set up Linux on my machine to move forward.

[stdlib-bot]

Coverage Report

Package	Statements	Branches	Functions	Lines
blas/base/zaxpy	$471/471$ $+100.00\mathrm{\%}$	$19/19$ $+100.00\mathrm{\%}$	$4/4$ $+100.00\mathrm{\%}$	$471/471$ $+100.00\mathrm{\%}$

The above coverage report was generated for the changes in this PR.

[gururaj1512]

@kgryte, Instead of

function dcabs1(z) result(res)
    implicit none
    complex(kind=kind(0.0d0)), intent(in) :: z
    real(kind=kind(0.0d0)) :: res
    
    res = abs(real(z)) + abs(aimag(z))
end function dcabs1

I have used

! Intrinsic functions:
intrinsic abs
if ( ( abs(real(za)) + abs(aimag(za)) ) <= 0.0D0 ) then
    return
end if

Which is working fine. PR ready for review