Prev: freeBSD and compilers
Next: proof of existing quine
From: aminer on 7 May 2010 04:07

Hello all,



I have used my Parallel Sort Library and benchmarked an Object Pascal
program that sorts a dynamic array of 10 millions of strings.

Please look at the benchmarks in following page:

http://pages.videotron.com/aminer/parallelsort/parallelsort.htm


On four cores four threads, here is the results:

Parallel Quicksort gives 5.31x speed and 2.877 seconds
Parallel Heapsort gave me 4.72x and 7.452 seconds

Note: Parallel Quicksort is much faster in pratice than parallel
heapsort
or parallel mergesort.





Sincerely
Amine Moulay Ramdane.

From: aminer on 8 May 2010 05:10

On May 8, 3:03 am, "Skybuck Flying" <IntoTheFut...(a)hotmail.com> wrote:
> "aminer" <ami...(a)videotron.ca> wrote in message
>
> news:b25f4f21-b361-4565-bd04-516b476f0f15(a)t15g2000vbo.googlegroups.com...
>
> Skybuck wrote:
> > For example: 14.307 (1 core) / 2.877 (4 cores) = 4.972x speed up for
> > parallel quicksort.
>
> > While you state 5.31x speed up ?!?

Sorry Skybuck, it was a typo , and it is 15.307 seconds
for one core/one thread, i have put again the right numbers,
look again:

http://pages.videotron.com/aminer/parallelsort/parallelsort.htm




Sincerely,
Amine.



On May 7, 4:07 am, aminer <ami...(a)videotron.ca> wrote:
> Hello all,
>
> I have used my Parallel Sort Library and benchmarked an Object Pascal
> program that sorts a dynamic array of 10 millions of strings.
>
> Please look at the benchmarks in following page:
>
> http://pages.videotron.com/aminer/parallelsort/parallelsort.htm
>
> On four cores four threads, here is the results:
>
> Parallel Quicksort gives 5.31x speed and 2.877 seconds
> Parallel Heapsort gave me 4.72x  and 7.452 seconds
>
> Note: Parallel Quicksort is much faster in pratice than parallel
> heapsort
>          or parallel mergesort.
>
> Sincerely
> Amine Moulay Ramdane.

 | 
Pages: 1
Prev: freeBSD and compilers
Next: proof of existing quine