On Wed, 2014-09-17 at 09:30 -0700, Andrei Alexandrescu via Digitalmars-d
wrote:
[…]
> Nothing wrong with starting small and moving later to a larger pond when the need arises. Let's just get this started without worrying too much about minutia. -- Andrei

OK so to the minutiae :-) which repository is this directory in?

-- 
Russel. ============================================================================= Dr Russel Winder      t: +44 20 7585 2200   voip: sip:russel.winder@ekiga.net 41 Buckmaster Road    m: +44 7770 465 077   xmpp: russel@winder.org.uk London SW11 1EN, UK   w: www.russel.org.uk  skype: russel_winder

On Wednesday, 17 September 2014 at 14:45:59 UTC, Andrei
Alexandrescu wrote:
> On 9/16/14, 11:59 PM, bearophile wrote:
>> Andrei Alexandrescu:
>>
>>> https://issues.dlang.org/show_bug.cgi?id=13487
>>
>> If the upload conditions and site are sufficiently good I am willing to
>> offer some implementations in D and to keep them updated. I suggest to
>> add two D versions for some benchmarks, one that shows short high level
>> code, and one that shows longer hairier fast code.
>>
>> In some cases I'd even like to show a third "safe" version (that tries
>> to be more correct), but most Shootout/ComputerGame benchmarks are not
>> very fit for this (you can see some examples of this on Rosettacode).
>
> I'm glad folks like Alexander, Russel, and yourself are interested in chipping in. Thanks!
>
> Let's do it all on the github repo. Add a directory e.g. tools/benchmarks/ and organize things in there.
>
>
> Thanks,
>
> Andrei

Would this be limited to the Computer Language Benchmarks Game?
This paper has gotten a lot of attention among economists:

http://www.econ.upenn.edu/~jesusfv/comparison_languages.pdf

They mention D, but only to say "D, which generates code usually
roughly of the same speed as C++, is less popular". Demonstrating
that you can write code that is competitive with C++ but in an
elegant language would be powerful.

Another site is

http://quant-econ.net/

which has gotten a lot of attention. It uses Python. I'd like to
write D versions of the programs to demonstrate that you can have
a language that's both nicer than Python and much, much faster.

On Wed, Sep 17, 2014 at 06:45:21PM +0100, Russel Winder via Digitalmars-d wrote:
> On Wed, 2014-09-17 at 09:30 -0700, Andrei Alexandrescu via Digitalmars-d
> wrote:
> […]
> > Nothing wrong with starting small and moving later to a larger pond when the need arises. Let's just get this started without worrying too much about minutia. -- Andrei
> 
> OK so to the minutiae :-) which repository is this directory in?
[...]

I believe Andrei was referring to:

	https://github.com/D-Programming-Language/tools


T

-- 
Too many people have open minds but closed eyes.

On Wednesday, 17 September 2014 at 14:59:48 UTC, Andrei Alexandrescu wrote:
> Awesome. Suggestion in order to leverage crowdsourcing: first focus on setting up the test bed such that adding benchmarks is easy. Then you and others can add a bunch of benchmarks.

On a somewhat related note, I've been working on a CI system to keep tabs on the compile-time/run-time performance, memory usage and file size for our compilers. It's strictly geared towards executing the same test case on different compiler configurations, though, so it doesn't really overlap with what is proposed here.

Right now, its continually building DMD/GDC/LDC from Git and measuring some 40 mostly small benchmarks, but I need to improve the web UI a lot before it is ready for public consumption. Just thought I would mention it here to avoid scope creep in what Peter Alexander (and others) might be working on.

Cheers,
David

On Wednesday, 17 September 2014 at 18:30:37 UTC, David Nadlinger wrote:
> On Wednesday, 17 September 2014 at 14:59:48 UTC, Andrei Alexandrescu wrote:
>> Awesome. Suggestion in order to leverage crowdsourcing: first focus on setting up the test bed such that adding benchmarks is easy. Then you and others can add a bunch of benchmarks.
>
> On a somewhat related note, I've been working on a CI system to keep tabs on the compile-time/run-time performance, memory usage and file size for our compilers. It's strictly geared towards executing the same test case on different compiler configurations, though, so it doesn't really overlap with what is proposed here.
>
> Right now, its continually building DMD/GDC/LDC from Git and measuring some 40 mostly small benchmarks, but I need to improve the web UI a lot before it is ready for public consumption. Just thought I would mention it here to avoid scope creep in what Peter Alexander (and others) might be working on.

That sounds great. I'm not planning anything grand with this. I'm just going to get the already exiting benchmark framework working with dmd, ldc, and gdc; and put it on github so people can contribute implementations.

I imagine what you have could probably be extended to do comparisons with other languages, but I think there's still value in getting these benchmarks working because they are so well known and respected.

On Wed, 2014-09-17 at 17:50 +0000, bachmeier via Digitalmars-d wrote: […]

> which has gotten a lot of attention. It uses Python. I'd like to write D versions of the programs to demonstrate that you can have a language that's both nicer than Python and much, much faster.

You can no longer bet on that and expect to win. Python + Numba is likely as fast as C/C++/D/….

OK so quants are currently using Python + NumPy which C/C++/D can generally comfortably beat. But you need to get you returns in quick before I and others get the quants using Python + Numba.

-- 
Russel. ============================================================================= Dr Russel Winder      t: +44 20 7585 2200   voip: sip:russel.winder@ekiga.net 41 Buckmaster Road    m: +44 7770 465 077   xmpp: russel@winder.org.uk London SW11 1EN, UK   w: www.russel.org.uk  skype: russel_winder

On Wednesday, 17 September 2014 at 18:58:20 UTC, Peter Alexander wrote:
> I imagine what you have could probably be extended to do comparisons with other languages,

Not really. Well, I could always add some sort of "reference" functionality to compare the results to equivalent C++/… implementations, but my focus definitely is on tracking compiler development over time.

> but I think there's still value in getting these benchmarks working because they are so well known and respected.

Definitely! I'll be adding them to my test suite as well, but my audience is a different one.

David

On Wednesday, 17 September 2014 at 18:30:37 UTC, David Nadlinger wrote:
-snip-
> On a somewhat related note, I've been working on a CI system to keep tabs on the compile-time/run-time performance, memory usage and file size for our compilers.


Maybe you've seen Emery Berger's work on Stabilizer?

http://plasma.cs.umass.edu/emery/stabilizer

I'm not smart enough to understand how that could be applied to D compilers ;)

On 9/17/14, 10:45 AM, Russel Winder via Digitalmars-d wrote:
> On Wed, 2014-09-17 at 09:30 -0700, Andrei Alexandrescu via Digitalmars-d
> wrote:
> […]
>> Nothing wrong with starting small and moving later to a larger pond when
>> the need arises. Let's just get this started without worrying too much
>> about minutia. -- Andrei
>
> OK so to the minutiae :-) which repository is this directory in?
>

Just put it in tools/benchmarks/ for now and stop worrying about it. We can move it later. -- Andrei

On Tuesday, 16 September 2014 at 05:27:15 UTC, Isaac Gouy wrote:
> On Tuesday, 16 September 2014 at 02:20:18 UTC, Freddy wrote:
>
>> What The D community do wrong in the first place?
>
> Nothing. There are just too many language implementations. It takes more time than I choose to donate. Been there; done that.

Hi Isaac,

Off-topic question: I've been wondering, how do you magically appear here every time the Shootout is mentioned? Someone contacted you, or HTTP referrer logs?