Re: int a = a (Was: Bart's Language)

Liste des GroupesRevenir à cl c 
Sujet : Re: int a = a (Was: Bart's Language)
De : david.brown (at) *nospam* hesbynett.no (David Brown)
Groupes : comp.lang.c
Date : 19. Mar 2025, 11:40:32
Autres entêtes
Organisation : A noiseless patient Spider
Message-ID : <vre6v0$lb74$1@dont-email.me>
References : 1 2 3 4 5 6 7
User-Agent : Mozilla/5.0 (X11; Linux x86_64; rv:102.0) Gecko/20100101 Thunderbird/102.11.0
On 18/03/2025 23:27, Janis Papanagnou wrote:
On 18.03.2025 20:51, David Brown wrote:
[...]  A brief check suggests that gcc will generate code as it
would for "int a = 0;", but it is certainly possible for a compiler to
avoid any kind of initialisation here and let the register or stack slot
used for "a" stay as it was.  That would be a pretty minor efficiency
improvement,
 
but optimised code is mostly the sum of lots of tiny improvements.
 Interesting view. I've learned that such Peephole Optimizations were
not what contribute to optimizations most. It's rather transformations
of structure of various forms that is what "mostly" matters. - That's
what I've learned many decades ago, of course. - So I'm curious where
you've got that view from. (Some reference, maybe? Or was that just a
personal opinion?)
 
Some optimisations have big effects, certainly - good register allocation and lifetime analysis, and optimisations that move code around (loop transformations, inlining, etc.) are the big factors. However, in modern compilers there are lots of minor optimisations that only apply in a few cases and only help a few percent in those cases. Each does little on its own, but in sum the results can be significant.
But you are right that it is not really fair to say that optimisation is "mostly" the sum of tiny improvements - it's a small number of big and important transforms, and /then/ the sum of a large number of small ones.
One complicating factor about these small optimisations is that the observable effect on code speed is highly dependent on the rest of the code and the type of processor involved.  A peephole optimisation that removes an extra register-to-register move will save a cycle on a microcontroller, but on an x86 system such a move might be merged in the register renaming hardware of the cpu's prefetch queues and thus disappear entirely.  Reducing instruction cycles matters a lot on microcontrollers, while on a big processor they might not make any difference if the cpu is waiting for memory.

Date Sujet#  Auteur
18 Mar 25 * Bart's Language57bart
18 Mar 25 `* Re: Bart's Language56Waldek Hebisch
18 Mar 25  `* Re: Bart's Language55bart
18 Mar 25   `* Re: Bart's Language54Waldek Hebisch
18 Mar 25    +* Re: Bart's Language46bart
18 Mar 25    i+* Re: Bart's Language36David Brown
18 Mar 25    ii`* int a = a (Was: Bart's Language)35Kenny McCormack
18 Mar 25    ii +* Re: int a = a (Was: Bart's Language)25Janis Papanagnou
18 Mar 25    ii i+- Re: int a = a (Was: Bart's Language)1Kaz Kylheku
19 Mar 25    ii i`* Re: int a = a (Was: Bart's Language)23David Brown
19 Mar 25    ii i +- Re: int a = a (Was: Bart's Language)1Kaz Kylheku
19 Mar 25    ii i +* Re: int a = a14Keith Thompson
20 Mar 25    ii i i+* Re: int a = a12Tim Rentsch
20 Mar 25    ii i ii`* Re: int a = a11Keith Thompson
20 Mar 25    ii i ii +* Re: int a = a8David Brown
20 Mar 25    ii i ii i`* Re: int a = a7Keith Thompson
21 Mar 25    ii i ii i `* Re: int a = a6David Brown
21 Mar 25    ii i ii i  `* Re: int a = a5Keith Thompson
21 Mar 25    ii i ii i   +- Re: int a = a1David Brown
22 Mar 25    ii i ii i   `* Re: int a = a3Tim Rentsch
22 Mar 25    ii i ii i    `* Re: int a = a2Keith Thompson
28 Apr 25    ii i ii i     `- Re: int a = a1Tim Rentsch
29 Apr 25    ii i ii `* Re: int a = a2Tim Rentsch
29 Apr 25    ii i ii  `- Re: int a = a1Keith Thompson
20 Mar 25    ii i i`- Re: int a = a1David Brown
19 Mar 25    ii i +* Re: int a = a (Was: Bart's Language)5Chris M. Thomasson
20 Mar 25    ii i i`* Re: int a = a (Was: Bart's Language)4David Brown
20 Mar 25    ii i i `* Re: int a = a (Was: Bart's Language)3bart
20 Mar 25    ii i i  `* Re: int a = a (Was: Bart's Language)2David Brown
20 Mar 25    ii i i   `- Re: int a = a (Was: Bart's Language)1wij
20 Mar 25    ii i `* Re: int a = a (Was: Bart's Language)2Tim Rentsch
20 Mar 25    ii i  `- Re: int a = a (Was: Bart's Language)1David Brown
18 Mar 25    ii +* Re: int a = a (Was: Bart's Language)3David Brown
18 Mar 25    ii i`* Re: int a = a (Was: Bart's Language)2Janis Papanagnou
19 Mar 25    ii i `- Re: int a = a (Was: Bart's Language)1David Brown
19 Mar 25    ii `* Re: int a = a (Was: Bart's Language)6Tim Rentsch
19 Mar 25    ii  +* Re: int a = a2Keith Thompson
27 Apr 25    ii  i`- Re: int a = a1Tim Rentsch
19 Mar 25    ii  +- Re: int a = a (Was: Bart's Language)1David Brown
19 Mar 25    ii  `* Re: int a = a (Was: Bart's Language)2Rosario19
20 Mar 25    ii   `- Re: int a = a (Was: Bart's Language)1Tim Rentsch
20 Mar 25    i`* Re: Bart's Language9Waldek Hebisch
21 Mar 25    i `* Re: Bart's Language8Keith Thompson
22 Mar 25    i  +* Re: Bart's Language5Waldek Hebisch
22 Mar 25    i  i`* Re: Bart's Language4James Kuyper
22 Mar 25    i  i +* Re: Bart's Language2Waldek Hebisch
23 Mar 25    i  i i`- Re: Bart's Language1James Kuyper
23 Mar 25    i  i `- By definition... (Was: Bart's Language)1Kenny McCormack
27 Apr 25    i  `* Re: Bart's Language2Tim Rentsch
27 Apr 25    i   `- Re: Bart's Language1Keith Thompson
18 Mar 25    `* Re: Bart's Language7bart
20 Mar 25     `* Re: Bart's Language6Waldek Hebisch
21 Mar 25      +* Re: Bart's Language4Kaz Kylheku
21 Mar 25      i`* Re: Bart's Language3bart
21 Mar 25      i `* Re: Bart's Language2Kaz Kylheku
22 Mar 25      i  `- Re: Bart's Language1Tim Rentsch
21 Mar 25      `- Re: Bart's Language1bart

Haut de la page

Les messages affichés proviennent d'usenet.

NewsPortal