Where communities thrive

Join over 1.5M+ people
Join over 100K+ communities
Free without limits
Create your own community

Explore more communities

libmir/public

People

Repo info

Activity

Jul 04 06:01

tastyminerals commented #403
Jul 03 07:09

9il commented #403
Jul 03 05:16

tastyminerals starred libmir/mir
Jul 03 05:15

tastyminerals closed #403
Jul 03 05:15

tastyminerals commented #403
Jul 03 05:13

tastyminerals commented #403
Jul 03 05:12

tastyminerals commented #403
Jul 03 04:03

9il commented #403
Jul 03 01:25

mridsole starred libmir/mir
Jul 02 18:29

tastyminerals edited #403
Jul 02 18:29

tastyminerals opened #403
May 29 23:33

miguelmartin75 starred libmir/mir
Apr 26 05:56

atsushiKojima starred libmir/mir
Mar 31 12:30

9il on master

Update README.md (compare)
Mar 30 06:00

9il on v3.2.1

(compare)
Mar 30 06:00

9il on master

fix meson (compare)
Mar 30 05:54

9il on master

update readme (compare)
Mar 30 05:53

9il on master

add meson & update ci (compare)
Jan 27 17:48

zweimach starred libmir/mir
Dec 27 2019 17:00

chapman2014 starred libmir/mir

Ilya Yaroshenko

@9il

proposed new Mir logo

Feedback or alternatives are welcome. I am looking for a soft logo instead of current one

Nicholas Wilson

@thewilsonator

IDK, i like the current one, it looks distinctive. That one looks ...bland.

Ilya Yaroshenko

@9il

We need a bland one :-(

or somtihing soft

Maybe we can change it back after a while

Matt Ranger

@VHRanger

hey guys. I think for the long run (3-5 years) it would be in the best interest for libmir to have an interface with apache arrow. Apache arrow is poised to replace pandas in that timeframe

see here: http://wesmckinney.com/blog/apache-arrow-pandas-internals/

and interfacing efficiently with it would potentially be a killer app. I'm interested in contributing here, just throwing this out there

Ilya Yaroshenko

@9il

Hey Matt @VHRanger . Thank you for letting us know. You can count on our help if you need new mir-algorithm features for Apache Arrow. Also, we can create a repo in libmir org and include you to the organizatoin if you want Apache Arrow be part of the libmir.

Mathias L. Baumann

@Marenz

Greetings. I try to build a simple project using lubeck, mir-glas and mir-cpuid. I am getting errors and I am not really sure how to deal with them: https://gist.github.com/Marenz/a33645a8ab6e6b8b1d4fa5234c124f6a

I am getting the same result if I write ldmd2

and even if I write --compiler=dmd it's using my system ldc (a very old version) and fails

Mathias L. Baumann

@Marenz

ah.. it seems mir-glas' dub file is hard coding quite a lot of things

Mathias L. Baumann

@Marenz

This is all a bit confusing really

lubeck says it requires blas, cblas and lapack as static libraries.. which confused me as I thought it would be a pure D library

After linking with those, I stumbled over kaleidicassociates/lubeck#3

Today I discovered that mir-glas aims to provide .. static c libraries

so I figured I Can use mir-glas in-place

mir-glas claims to provide blas and glas. For some reason lubeck still wants cblas though.

So, linking with cblas, it finally compiled

however, I again stumbled over the already mentioned bug, this time the program at least exits with code 255 and says something slightly different

I just updated the ticket accordingly

Mathias L. Baumann

@Marenz

though I don't even know why cblas is required, my assumption was that mir-glas would provide all.

Ilya Yaroshenko

@9il

mir-glas is not used in Lubeck

lubeck uses only common cblas / lapack libs. Glas does not provide full cblas

The bug in lubeck code

or even in https://github.com/libmir/mir-blas

Mathias L. Baumann

@Marenz

hmm alright. But I suppose I don't actually need lubeck and just use mir-glas directly

btw, I don't know when opBinary was implemented, but it is soooo nice to be able to write m[] = m1 + m2 (it wasn't possible last time I used it)

me gusta

Mathias L. Baumann

@Marenz

I assume there is no libary that offers something like a lazy gemm abstraction so I could write things like m[] = m1.mtimes(m2) + m3;

It seems not trivial to write something like that myself as I discovered

at least if you want to avoid temporary allocations

Ilya Yaroshenko

@9il

m[] = m1.mtimes(m2) + m3; should be possible with lubeck

I will fix the mtimes bug
Thank you for the report!

Mathias L. Baumann

@Marenz

Yeah, it is with lubeck, but as I understood, lubeck doesn't use mir-glas, right?

Ilya Yaroshenko

@9il

Yes, lubeck does not

Mathias L. Baumann

@Marenz

another question: what exactly does the universal() function do? It doesn't take the slice as a ref-argument so I am wondering if it modifies the original slice in any way or if it just returns a new slice that has a different "view" on the same data? Or would it be a mistake to use the original slice after I called universal() on it?

I mean, a Slice is a struct referencing some external data. Is that data modified by universal()? the docu isn't very clear here

My use case is when using gemm from mir-glas. It requires the SliceKind.universal but all I have is contiguous, so my solution so far was to just do gemm(1.0f, m1.universal, m2.universal, 0.0f, tmp_result.universal);

but i am not very faithful in my solution

especially when using m1 afterwards

Ilya Yaroshenko

@9il

Hi @Marenz , universal as well as other contents of topology does not modify the data. universal just returns a slice of Universal kind with the same strides and shapes. The difference is only in kind.

Mathias L. Baumann

@Marenz

thank you for that info :)

I still haven't figured out all those details

Mind if I create PRs that add what you just told me?

for the docu

Ilya Yaroshenko

@9il

Sure, hope you will improve English