I propose to renato copy this post in harbour website
so it will be intresting for all harbour newbbies
2009 Speed Comparision

2009/3/19 Viktor Szakáts <harbour...@syenar.hu>:
> Additional information, ranks by executable size in KB, plus
> some conclusions:
> st-c52bli        271
> st-hb-msvc       566
> mt-hb-msvc       586
> st-hb-bcc        629
> mt-hb-bcc        652
> st-hb-pocc       669
> mt-hb-pocc       695
> st-hb-ow         760
> mt-hb-ow         807
> st-hb-icc        811
> mt-hb-icc        843
> st-xhb-xcc       962
> st-hb-mingw432  1076
> st-hb-mingw433  1100
> mt-hb-mingw432  1136
> mt-hb-mingw433  1160
> mt-xpp           252 (shared)
> For me the clear winner in all technical aspects is MSVC,
> creating almost the quickest executables (probably LTCG
> can enhance it a bit), and the smallest ones. It's also
> gratis, supports x64, unfortunately it's not free software
> and not multiplatform at all, but very-well supported by
> 3rd parties. Next one is MinGW, which is free, open source
> and multiplatform, also very well supported, gives the best speed
> amongst free tools (and the smallest MT overhead), but it
> generates huge executables on Windows. ICC stays being
> a curiosity because it's payware, Open Watcom performs
> unfortunately poorly in all aspects, I'm not sure what benefit
> can it do us for now, maybe in the future, or maybe it's possible
> to tweak performance. BCC and POCC have similar middle-ground
> performance, both closed-source and gratis, both with bugs
> and largely missing support from 3rd parties, uncertain future.
> POCC supports ARM and x64, uses standard COFF libs and
> all editions are gratis (not just some old versions with weird
> licensing), POCC can also generate proper .dlls, so it clearly
> wins over BCC.
> So my overall rank goes like this:
> - MSVC
> - MinGW (top rank, if being free software is important)
> - POCC
> Maybe owatcom, if it can be made faster, but otherwise,
> it's unfortunately useless. I see no reason to deal with
> BCC at all, since POCC is a better alternative.
> Most of these isn't news, what's new is that POCC is a very
> good replacement for BCC (if MSVC and MinGW isn't an option),
> and the unfortunate results of Open Watcom.
> [ Notice that Open Watcom's .dll support in Harbour is broken.
> Anyone to help fixing it? ]
> Rgds,
> Viktor
> On Thu, Mar 19, 2009 at 2:14 PM, Viktor Szakáts <harbour...@syenar.hu>
> wrote:
>>
>> Hi Marek,
>> Yes, I've run tests ST and MT with pocc, ow, bcc, mingw and
>> icc for comparison,
>> also included previous Clipper, xpp and xhb tests.
>> Get them here:
>> http://www.syenar.hu/harbour/winbench.zip
>> Machine:
>> P4 2.6 HT 3GB RAM. No other apps running.
>> Command lines:
>> hbmk2 speedtst.prg -gc3 -mt -ospeedtstmt
>> hbmk2 speedtst.prg -gc3
>> Ranks based on "Total app time" [sec]:
>> st-hb-icc      ..27.39
>> st-hb-msvc     ..29.88
>> st-hb-mingw432 ..40.08
>> st-hb-mingw433 ..41.91
>> st-hb-bcc      ..42.81
>> st-hb-pocc     ..42.86
>> mt-hb-icc      ..48.22
>> mt-hb-msvc     ..49.66
>> st-hb-ow       ..53.56
>> mt-hb-mingw433 ..56.94
>> mt-hb-mingw432 ..57.02
>> mt-hb-bcc      ..64.47
>> mt-hb-pocc     ..69.09
>> mt-hb-ow       ..77.03
>> mt-xpp         .108.45
>> st-xhb-xcc     .116.11
>> st-c52bli      .259.30
>> Rgds,
>> Viktor
>> 2009/3/19 Horodyski Marek (PZUZ) <m.horody...@pzuzycie.com.pl>
>>>
>>>
>>>
>>> ________________________________
>>> From: Viktor Szakáts [mailto:harbour...@syenar.hu]
>>> Sent: Wednesday, March 18, 2009 2:39 PM
>>> To: Harbour Project Main Developer List.
>>> Subject: Re: [Harbour] Speedtest results
>>>
>>> Here is mine, just for fun :) [ BTW, Chen, you'd wonder how
>>> much faster Harbour would be with MSVC or MinGW. ]
>>>
>>>
>>> [...]
>>>
>>>
>>> --- exe size: 586KB (static)
>>> 03/18/09 11:10:46 Windows XP 5.1.2600 Service Pack 3
>>> Harbour 1.1.0dev (Rev. 10642) Microsoft Visual C++ 15.0.26569 (32-bit)
>>> THREADS: 0
>>> N_LOOPS: 1000000
>>> [ T000: empty loop overhead ]...................................0.05
>>> ====================================================================
>>> [ T001: x := L_C ]..............................................0.03
>>>
>>>
>>> [...]
>>>
>>>  [ T056: f_prv( c ) ]............................................0.39
>>> ====================================================================
>>> [ total application time: ]....................................29.81
>>> [ total real time: ]...........................................29.98
>>>
>>> Viktor,
>>> can you present too test with OW on this computer ?
>>>
>>> Regards,
>>> Marek Horodyski
>>>
>>> _______________________________________________
>>> Harbour mailing list
>>> Harbour@harbour-project.org
>>> http://lists.harbour-project.org/mailman/listinfo/harbour
>>>
>>
>
>
> _______________________________________________
> Harbour mailing list
> Harbour@harbour-project.org
> http://lists.harbour-project.org/mailman/listinfo/harbour
>
>



-- 
Massimo Belgrano
_______________________________________________
Harbour mailing list
Harbour@harbour-project.org
http://lists.harbour-project.org/mailman/listinfo/harbour

Reply via email to