Re: Toy computational "benchmark" in Ada (new blog post)

[David Trudgett <dktrudgett@gmail.com>]

Il giorno venerdì 7 giugno 2019 10:14:18 UTC+10, Paul Rubin ha scritto:
> 
> I managed to build this with GNAT 6.3.0 on an Intel i5-3570S (3.4 GHz)
> with 16GB of memory.  Results:
> 
>     SINGLE-THREADED V2
>     Data points    :  1024000000
>     Sum of squares :  1.01064747680946E+10
>     Duration (ms)  :  2456.406644100
> 
>     MULTI-THREADED V2
>     Data points    :  1024000000
>     Worker tasks   :  4
>     Sum of squares :  1.01064749025115E+10
>     Duration (ms)  :  689.120666560
> 
Those times are amazingly slow. Perhaps you would like to post the disassembly of the inner loop to see what the issue is?


> I don't know what those ms durations measure.

This means that you did not read either the article or the source code.


>  Actual elapsed times were
> 2 min 5 sec for the single threaded version and 36.484s for the parallel
> version.  The reported usermode cpu times were 2m3.2s for the single
> threaded version and 2m18s (across the 4 cores) for the parallel version.
> 
> I'll see if I can code and time a C++ version.

Righto.

Cheers,
David