Ho fiara, nu era necesar sa sari asa in sus, observa ca nu m-am legat de
o distributie si inceteaza sa mai citesti cu alte organe in afara de
ochi & creier.
Ãn data de Mi, 09-02-2005 la 21:09 +0200, Alexandru N. Barloiu a scris:
> one of my points este flexibilitatea. nu ai adus nici un argument
> contra. in schimb vii cu un argument, ce e drept semipertinent, cum ca
> "majoritatea" fac runtime detect. sunt de acord ca face asta mplayer
> ACUM. dar de cand o face? then again cate din alea "majoritatea" o fac.
> doua? trei?
Toate cele pentru care performanta este critica, nu flag-urile pe care
le pune utilizatorul la compilare. Codere/decodere video (still sau
stream), codere/decodere audio, jocuri.
Daca nu ma crezi pe mine pe cuvant uite citate din oameni care imi place
sa cred ca stiu ce spun:
<citat>
De la:
Dave Jones <[EMAIL PROTECTED]>
RÄspuns-la::
Development discussions related to
Fedora Core <fedora-devel-
[EMAIL PROTECTED]>
La:
Development discussions related to
Fedora Core <fedora-devel-
[EMAIL PROTECTED]>
Cc:
'Matthew Miller' <[EMAIL PROTECTED]>
Subiect:
Re: RFC: Optimizing for 386
Data:
Wed, 19 Jan 2005 11:42:19 -0500
(18:42 EET)
Mailer:
Mutt/1.4.1i
On Wed, Jan 19, 2005 at 10:26:51AM -0600, Joseph D. Wagner wrote:
> > They're optimized for pentium4, but use the i386 instruction set.
> There are instructions which have come out since the 386, like MMX,
that
> could improve the performance of programs. In this case, graphics
programs.
Many graphics programs test for features such as mmx/3dnow/sse at
runtime
and run alternative hand-written assembly routines when present.
This way the program runs on any x86 platform regardless of featureset
without needing to be compiled for a specific variant.
> Why should my graphics programs suffer because some fool is running a
20
> year old computer?
You forgot to attach the benchmarks showing the improvements.
</citat>
<citat>
De la:
Alan Cox <[EMAIL PROTECTED]>
RÄspuns-la::
Development discussions related to
Fedora Core <fedora-devel-
[EMAIL PROTECTED]>
La:
Development discussions related to
Fedora Core <fedora-devel-
[EMAIL PROTECTED]>
Cc:
'Matthew Miller' <[EMAIL PROTECTED]>
Subiect:
Re: RFC: Optimizing for 386
Data:
Wed, 19 Jan 2005 19:12:28 -0500
(Jo, 02:12 EET)
Mailer:
Mutt/1.4.1i
On Wed, Jan 19, 2005 at 10:26:51AM -0600, Joseph D. Wagner wrote:
> > They're optimized for pentium4, but use the i386 instruction set.
>
> There are instructions which have come out since the 386, like MMX,
>that could improve the performance of programs. In this case,
>graphics programs. Why should my graphics programs suffer because
>some fool is running a 20 year old computer?
MMX using apps all check for MMX v 3Dnow v SSE already and have too. The
instructions in question are not compiler generated as they are
specialist
</citat>
> te rog ia in calcul ca nu spun ca este un boost uluitor dar este
> pertinent sa spui ca functie de arhitectura exista un boost, fie el cat
> de mic. ia in calcul ca posibilitatea de a seta CFLAGS/CXXFLAGS/LDFLAGS
> samd este doar un bonus (sa zicem) de la avantaj mai mare si anume
> flagul USE. daca nu taiai "cra"-ul de mai sus vedeai si alte flaguri pe
> langa cele in discutie.
<citat>
De la:
William Cohen <[EMAIL PROTECTED]>
RÄspuns-la::
Development discussions related to
Fedora Core <fedora-devel-
[EMAIL PROTECTED]>
La:
Development discussions related to
Fedora Core <fedora-devel-
[EMAIL PROTECTED]>
Subiect:
Re: RFC: Optimizing for 386
Data:
Wed, 19 Jan 2005 12:28:33 -0500
(19:28 EET)
Mailer:
Mozilla/5.0 (X11; U; Linux i686; en-
US; rv:1.4.3) Gecko/20050104 Red
Hat/1.4.3-3.0.7
Benchmarking is important in this case. Use of the new instructions is
only going to help in very specific cases where the code is CPU bound
and the operations fit the style of parallel processing supported by
the
new instructions. In some cases the performance could be hurt due to
the
overheads of packing and unpacking the data for the MMX/SSE/3DNOW!
register, besides on P4 processors the SSE hardware works at only half
the speed as the integer units, so the gains are not going to be as big
on P4.
Blindly compiling all the sources to use the latest instruction set
isn't
going to magically improve the performance of the code. There are other
lots of other things besides processor instructions affecting
performance like the memory hierarchy, interprocessor communciation,
and
algorithms.
</citat>
> in fond la asta se reduce tot flame-ul referitor
> la gentoo. flexibilitate vs timp pentru compilare. vin si intreb: e
> chiar asa de rau ca anumite persoane prefera gentoo pentru motivele mai
> sus numite?
Dupa cum spuneam si la inceput, in postul meu nu se gasea cuvantul
"gentoo" decat in subiect (care a fost preluat de reply).
Nu am nici o problema cu gentoo, cui ii place sa-l foloseasca in pace.
Ah, si dupa cum spunea Dave Jones ai uitat sa vii cu benchmark-urile.
> una peste alta ce incerci sa imi explici defapt?
Ca soaftele pentru care performanta este critica pana la marginea de 1%
in genere nu stau dupa CFLAGS, LDFLAGS, USE & shit.
Si mai era un point dar nu am mai gasit mesajul respectiv dar care suna
de genul: "Ce frumos e cand unii pretind ca stiu mai multe despre flag-
urile compilatoarelor decat cei care au facut compilatorul respectiv..."
EOT pentru mine, keep barking daca te face sa te simti mai bine.
--
Cioby
---
Detalii despre listele noastre de mail: http://www.lug.ro/
