Path: csiph.com!v102.xanadu-bbs.net!xanadu-bbs.net!feeder.erje.net!eu.feeder.erje.net!eternal-september.org!feeder.eternal-september.org!news.eternal-september.org!.POSTED!not-for-mail
From: Ramine <ramine@1.1>
Newsgroups: comp.programming,comp.programming.threads
Subject: Re: FPC 64-bit  2.7.1
Date: Sun, 29 Jun 2014 14:30:10 -0700
Organization: A noiseless patient Spider
Lines: 97
Message-ID: <loplt2$8tr$1@dont-email.me>
References: <loph0e$506$1@dont-email.me>
Mime-Version: 1.0
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Content-Transfer-Encoding: 7bit
Injection-Date: Sun, 29 Jun 2014 18:28:50 +0000 (UTC)
Injection-Info: mx05.eternal-september.org; posting-host="518d70aeead21e0a1a15f1c91bb07e09"; logging-data="9147"; mail-complaints-to="abuse@eternal-september.org";	posting-account="U2FsdGVkX19qN7h1IlxTPCqGEyCGh4sJ"
User-Agent: Mozilla/5.0 (Windows NT 6.0; WOW64; rv:24.0) Gecko/20100101 Thunderbird/24.6.0
In-Reply-To: <loph0e$506$1@dont-email.me>
Cancel-Lock: sha1:iQl/1YMZgkk6zt3dOe9GwPiZLD8=
Xref: csiph.com comp.programming:4650 comp.programming.threads:2547


Hello,


FreePascal 2.7.1 64-bit is much faster than Delphi XE4 64-bit

Here is the results on my Quadcore Q6600:


For Delphi XE4 64-bit we have:

Mininum running time = 2.00 seconds
Composite Score MFlops:   310.22
FFT             Mflops:   410.41    (N=1024)
SOR             Mflops:   446.88    (100 x 100)
MonteCarlo:     Mflops:   110.83
Sparse matmult  Mflops:  482.42    (N=1000, nz=5000)
LU              Mflops:  1065.42    (M=100, N=100)

For FreePascal 2.7.1 we have:

Mininum running time = 2.00 seconds
Composite Score MFlops:   578.95
FFT             Mflops:   393.07    (N=1024)
SOR             Mflops:   717.27    (100 x 100)
MonteCarlo:     Mflops:   113.03
Sparse matmult  Mflops:   814.43    (N=1000, nz=5000)
LU              Mflops:   856.94    (M=100, N=100)



Thank you,
Amine Moulay Ramdane.



On 6/29/2014 1:06 PM, Ramine wrote:
>
> Hello,
>
> I have benchmarked FPC 2.7.1 64-bit against the newest version
> of tdm-gcc 64-bit and the FreePascal scimark2 benchmark is much better:
>
>
> Here is the scimark2 results on my Quadcore Q6600:
>
>
> FPC 2.7.1 64-bit gives:
>
> Mininum running time = 2.00 seconds
> Composite Score MFlops:   578.95
> FFT             Mflops:   393.07    (N=1024)
> SOR             Mflops:   717.27    (100 x 100)
> MonteCarlo:     Mflops:   113.03
> Sparse matmult  Mflops:   814.43    (N=1000, nz=5000)
> LU              Mflops:   856.94    (M=100, N=100)
>
>
> And tdm-gcc 64-bit gives:
>
> Using       2.00 seconds min time per kenel.
> Composite Score:          754.95
> FFT             Mflops:   664.22    (N=1024)
> SOR             Mflops:   720.90    (100 x 100)
> MonteCarlo:     Mflops:   223.05
> Sparse matmult  Mflops:   985.50    (N=1000, nz=5000)
> LU              Mflops:  1181.08    (M=100, N=100)
>
>
> So  if you have noticed on LU there a 38% difference in performance
> and on Sparse matmult there is a 21% difference in performance
> and on SOR it's the same performance
> and on FFT there is 69% difference in performance.
> and on the overall Composite Score there is only 30% difference in
> performance.
>
>
> So i think that FPC 2.7.1 64-bit is much better and i think
> FPC will get much better on performance in the future.
>
> You can download FPC 2.7.1 64-bit from here:
>
> ftp://ftp.freepascal.org/pub/fpc/snapshot/trunk/
>
>
> FPC 2.7.1 64-bit is much better on scimark benchmark than Delphi XE 64-bit
>
>
> Thank you,
> Amine Moulay Ramdane.
>
>
>
>
>
>