View previous topic :: View next topic |
Author |
Message |
wilburpan l33t
Joined: 21 Jan 2003 Posts: 977
|
Posted: Fri Mar 18, 2005 12:10 pm Post subject: |
|
|
That's really funny.
However, I think it's a sad commentary on the limitations of forum communication when you have to plaster "Danger: Humor Ahead" warnings all over a post like this. _________________ I'm only hanging out in OTW until I get rid of this stupid l33t ranking.....Crap. That didn't work. |
|
Back to top |
|
|
discomfitor l33t
Joined: 21 Feb 2003 Posts: 927 Location: None
|
Posted: Fri Mar 18, 2005 2:31 pm Post subject: |
|
|
ciaranm wrote: | No, see, the difference is that those guys understand what they're writing, and they're quite clear that they're pushing out experimental code to be used on test systems by people who have a clue. Big difference between that and randomly applying patches that have cool-sounding names... |
As far as "experimental code" goes, I'm pretty sure that Con Kolivas was the one who tuned ingosched (O(1)) in the 2.5 development series. Prior to his work, it was an interactivity mess that nobody wanted to touch. Once he was finished with it, O(1) actually began to work, and this led to the release of 2.6.
So really, if you're saying that his staircase scheduler is experimental, you need to say that his work on O(1) is experimental as well. If that's the case, then thousands of unwitting linux users have become ricers because they don't know any better. _________________ There is no substitute for experience.
Imperfection indicates a lack of effort. |
|
Back to top |
|
|
ciaranm Retired Dev
Joined: 19 Jul 2003 Posts: 1719 Location: In Hiding
|
Posted: Fri Mar 18, 2005 2:42 pm Post subject: |
|
|
wilburpan wrote: | However, I think it's a sad commentary on the limitations of forum communication when you have to plaster "Danger: Humor Ahead" warnings all over a post like this. |
Come on, we're not exactly dealing with the smartest cookies in the pot here...
Darckness wrote: | ciaranm wrote: | No, see, the difference is that those guys understand what they're writing, and they're quite clear that they're pushing out experimental code to be used on test systems by people who have a clue. Big difference between that and randomly applying patches that have cool-sounding names... |
As far as "experimental code" goes, I'm pretty sure that Con Kolivas was the one who tuned ingosched (O(1)) in the 2.5 development series. Prior to his work, it was an interactivity mess that nobody wanted to touch. Once he was finished with it, O(1) actually began to work, and this led to the release of 2.6.
So really, if you're saying that his staircase scheduler is experimental, you need to say that his work on O(1) is experimental as well. If that's the case, then thousands of unwitting linux users have become ricers because they don't know any better. |
Not at all. "Began as experimental" doesn't always mean "is still experimental". |
|
Back to top |
|
|
discomfitor l33t
Joined: 21 Feb 2003 Posts: 927 Location: None
|
Posted: Fri Mar 18, 2005 3:16 pm Post subject: |
|
|
ciaranm wrote: | Not at all. "Began as experimental" doesn't always mean "is still experimental". |
Exactly .
I'm not trying to start an argument with you, merely pointing out that I believe ck has moved past the "experimental" stage. If you'll check its mailing lists, you'll see that there aren't really any bugs referred to in recent messages about recent versions of staircase or mapped watermark or the like. Anyway, that's most likely enough on ck for now. I like it, you don't. Somehow I doubt an argument here will change that . _________________ There is no substitute for experience.
Imperfection indicates a lack of effort. |
|
Back to top |
|
|
cokey Advocate
Joined: 23 Apr 2004 Posts: 3355
|
Posted: Fri Mar 18, 2005 8:00 pm Post subject: |
|
|
darckness, you forgot to add about setting CTARGET in the ricing tutorial _________________ https://otw20.com/ OTW20 The new place for off the wall chat |
|
Back to top |
|
|
discomfitor l33t
Joined: 21 Feb 2003 Posts: 927 Location: None
|
Posted: Fri Mar 18, 2005 8:06 pm Post subject: |
|
|
cokehabit: Whoops...another thing I missed. I'm all over it now.
EDIT: Up now. _________________ There is no substitute for experience.
Imperfection indicates a lack of effort. |
|
Back to top |
|
|
rhill Retired Dev
Joined: 22 Oct 2004 Posts: 1629 Location: sk.ca
|
Posted: Fri Mar 18, 2005 8:33 pm Post subject: |
|
|
pure gold. _________________ by design, by neglect
for a fact or just for effect |
|
Back to top |
|
|
ciaranm Retired Dev
Joined: 19 Jul 2003 Posts: 1719 Location: In Hiding
|
Posted: Fri Mar 18, 2005 8:39 pm Post subject: |
|
|
I'm almost certainly going to regret mentioning that you can do this... But you're forgetting to mention that you should symlink package.unmask to package.mask so that you automatically get all the top sekrit super fast experimental ebuilds enabled. |
|
Back to top |
|
|
ravloony n00b
Joined: 04 Feb 2005 Posts: 54 Location: France
|
Posted: Fri Mar 18, 2005 8:55 pm Post subject: |
|
|
Anyway, the whole kernel thing is overrated, the best way to get a fast kernel is to not bloat it with stuff you don't need, so it stays small. Oh and use modules, I reckon.
Funny post though. When I started with Gentoo, I actually looked around to see what I could do to CFLAGS et al. Luckily I was talked out of it on #gentoo by someone. But I'm sure there are a lot of people who aren't talked out of it. Makes me cringe really.
seeya
rav _________________ No sig yet, sig ebuild up soon |
|
Back to top |
|
|
thebigslide l33t
Joined: 23 Dec 2004 Posts: 792 Location: under a car or on top of a keyboard
|
Posted: Fri Mar 18, 2005 9:05 pm Post subject: |
|
|
[offtopic]
ciaranm wrote: | yaneurabeya wrote: | I know I was just joking (well, partially; mostly thinking about quicksort under some cases ). That would of course be the holy grail of computer science if someone figured out that . There are specialized algorithms though which under certain cases do function in log(n) time though of course . |
Eh? No, a log(N) sorting algorithm is impossible (except in special cases where you know that, say, only the first item is out of place). Doing a guaranteed time n*log(N) sort is easy, and doesn't involve anything even remotely related to qsort. Going beyond log-linear in the general case is impossible. |
Code: | const long LENGTH=1000;
randomize();
long unsortedList[LENGTH];
long listData[length];
int i;
for (i=0;i<LENGTH;i++) unsortedList[i]=ipart(999*rand());
for (i=0;i<LENGTH;i++) listData[i]=0;
for (i=0;i<LENGTH;i++) listData[unsortedList[i]]++;
for (i=0;i<LENGTH;i++) if(listData[i]) {
cout << i << endl;
listData[i--]--;
} |
This is close, works with strings by calling recursively.
[/offtopic] |
|
Back to top |
|
|
dcrook Tux's lil' helper
Joined: 22 Jul 2004 Posts: 83
|
Posted: Fri Mar 18, 2005 9:06 pm Post subject: |
|
|
ciaranm wrote: | I'm almost certainly going to regret mentioning that you can do this... But you're forgetting to mention that you should symlink package.unmask to package.mask so that you automatically get all the top sekrit super fast experimental ebuilds enabled. |
Hey, thanks for the tip Ciaranm, thats like Gento Ricer gold there, bling-bling. I'm going to add that right away and crank up my cron job is that it syncs and emerges everthing every 30 minutes.
Oh yea, I'm having some problems with my raid 0 array running reiser 4 can I like call you at home and ask you about that? |
|
Back to top |
|
|
discomfitor l33t
Joined: 21 Feb 2003 Posts: 927 Location: None
|
Posted: Fri Mar 18, 2005 10:51 pm Post subject: |
|
|
ciaranm wrote: | I'm almost certainly going to regret mentioning that you can do this... But you're forgetting to mention that you should symlink package.unmask to package.mask so that you automatically get all the top sekrit super fast experimental ebuilds enabled. |
ciaranm: That's spectacular. Thanks! Updating...
EDIT: Updated, complete with boldfaced warning to not do this in the "real" section.
EDIT #2: I'm assuming that you meant /usr/portage/profiles/package.mask to /etc/portage/package.unmask since that would unmask all of the broken ebuilds, thus creating more ricer speed? Maybe I'm thinking about this the wrong way though... _________________ There is no substitute for experience.
Imperfection indicates a lack of effort. |
|
Back to top |
|
|
feld Guru
Joined: 29 Aug 2004 Posts: 593 Location: WI, USA
|
Posted: Sat Mar 19, 2005 5:17 am Post subject: |
|
|
rofl....
funny darckness...
but don't you have something better to do? go watch some porn lol
-Feld _________________ < bmg505> I think the first line in reiserfsck is
if (random(65535)< 65500) { hose(partition); for (i=0;i<100000000;i++) print_crap(); } |
|
Back to top |
|
|
LucaSpiller Apprentice
Joined: 10 Sep 2004 Posts: 188 Location: Censorship Land (aka England)
|
Posted: Sat Mar 19, 2005 8:09 am Post subject: |
|
|
Fantastic.
Has there ever been a thread on the 'fastest, most insanly optimized systems'? _________________ :: Luca :: Mac Fag :: Original Macbook, 2g RAM :: Closet Linux user (seasoned with salt and pepper) :: C2D E4400 @ 2ghz, 4g RAM (only 3.2g detected under 64bit...), Nvidia 9600GSO :: |
|
Back to top |
|
|
discomfitor l33t
Joined: 21 Feb 2003 Posts: 927 Location: None
|
Posted: Sat Mar 19, 2005 8:32 am Post subject: |
|
|
feld: As no doubt many have noticed, ricing is my life. _________________ There is no substitute for experience.
Imperfection indicates a lack of effort. |
|
Back to top |
|
|
DoctorPizza n00b
Joined: 19 Mar 2005 Posts: 2
|
Posted: Sat Mar 19, 2005 10:09 am Post subject: |
|
|
ciaranm wrote: | yaneurabeya wrote: | I know I was just joking (well, partially; mostly thinking about quicksort under some cases ). That would of course be the holy grail of computer science if someone figured out that . There are specialized algorithms though which under certain cases do function in log(n) time though of course . |
Eh? No, a log(N) sorting algorithm is impossible (except in special cases where you know that, say, only the first item is out of place). Doing a guaranteed time n*log(N) sort is easy, and doesn't involve anything even remotely related to qsort. Going beyond log-linear in the general case is impossible. |
*Comparison* sorts that are better than O(N log N) (where the base of the logarithm is the number of elements you can simultaneously compare), but non-comparison sorts (radix sorts, bin sorts, "spaghetti sorts") are O(N). |
|
Back to top |
|
|
pilla Bodhisattva
Joined: 07 Aug 2002 Posts: 7729 Location: Underworld
|
Posted: Sat Mar 19, 2005 1:02 pm Post subject: |
|
|
DoctorPizza wrote: | ciaranm wrote: | yaneurabeya wrote: | I know I was just joking (well, partially; mostly thinking about quicksort under some cases ). That would of course be the holy grail of computer science if someone figured out that . There are specialized algorithms though which under certain cases do function in log(n) time though of course . |
Eh? No, a log(N) sorting algorithm is impossible (except in special cases where you know that, say, only the first item is out of place). Doing a guaranteed time n*log(N) sort is easy, and doesn't involve anything even remotely related to qsort. Going beyond log-linear in the general case is impossible. |
*Comparison* sorts that are better than O(N log N) (where the base of the logarithm is the number of elements you can simultaneously compare), but non-comparison sorts (radix sorts, bin sorts, "spaghetti sorts") are O(N). |
do you mean O(N^2)? _________________ "I'm just very selective about the reality I choose to accept." -- Calvin |
|
Back to top |
|
|
Earthwings Bodhisattva
Joined: 14 Apr 2003 Posts: 7753 Location: Germany
|
Posted: Sat Mar 19, 2005 1:32 pm Post subject: |
|
|
pilla wrote: | DoctorPizza wrote: | ciaranm wrote: | yaneurabeya wrote: | I know I was just joking (well, partially; mostly thinking about quicksort under some cases ). That would of course be the holy grail of computer science if someone figured out that . There are specialized algorithms though which under certain cases do function in log(n) time though of course . |
Eh? No, a log(N) sorting algorithm is impossible (except in special cases where you know that, say, only the first item is out of place). Doing a guaranteed time n*log(N) sort is easy, and doesn't involve anything even remotely related to qsort. Going beyond log-linear in the general case is impossible. |
*Comparison* sorts that are better than O(N log N) (where the base of the logarithm is the number of elements you can simultaneously compare), but non-comparison sorts (radix sorts, bin sorts, "spaghetti sorts") are O(N). |
do you mean O(N^2)? |
I think he means O(N), but that's only half truth. See http://en.wikipedia.org/wiki/Sort_algorithm |
|
Back to top |
|
|
ciaranm Retired Dev
Joined: 19 Jul 2003 Posts: 1719 Location: In Hiding
|
Posted: Sat Mar 19, 2005 2:23 pm Post subject: |
|
|
Yes, you can do O(N) sorts of integers using radix sort. However, this assumes that you already know the range of integers in the list, and that said range is small. So it's not a general sort. Still useful sometimes, of course. |
|
Back to top |
|
|
pilla Bodhisattva
Joined: 07 Aug 2002 Posts: 7729 Location: Underworld
|
Posted: Sat Mar 19, 2005 2:27 pm Post subject: |
|
|
Yes, you can even sort a vector in O(1) if it is known to be sorted _________________ "I'm just very selective about the reality I choose to accept." -- Calvin |
|
Back to top |
|
|
ebrostig Bodhisattva
Joined: 20 Jul 2002 Posts: 3152 Location: Orlando, Fl
|
Posted: Sat Mar 19, 2005 4:50 pm Post subject: |
|
|
Meeehhh ricers.... amateurs... pure virgins in the world of -O's....
You had me scared for a minute there, that is when I saw -O4. "How the heck did he get hold of my own optimization code for gcc???" was my first thought... Then i understood he had not and that he was trying a joke! Puuhhhh....
Erik _________________ 'Yes, Firefox is indeed greater than women. Can women block pops up for you? No. Can Firefox show you naked women? Yes.' |
|
Back to top |
|
|
pilla Bodhisattva
Joined: 07 Aug 2002 Posts: 7729 Location: Underworld
|
Posted: Sat Mar 19, 2005 5:20 pm Post subject: |
|
|
I remember when I was a child and we didn't care about optimization (after all, it was BASIC). _________________ "I'm just very selective about the reality I choose to accept." -- Calvin |
|
Back to top |
|
|
cokey Advocate
Joined: 23 Apr 2004 Posts: 3355
|
Posted: Sat Mar 19, 2005 6:48 pm Post subject: |
|
|
pilla wrote: | I remember when I was a child and we didn't care about optimization (after all, it was BASIC). |
no no no, when you were a child pilla, it was COBOL _________________ https://otw20.com/ OTW20 The new place for off the wall chat |
|
Back to top |
|
|
superstoned Guru
Joined: 17 Dec 2004 Posts: 432
|
Posted: Sat Mar 19, 2005 7:02 pm Post subject: |
|
|
Darckness wrote: | ciaranm wrote: | Not at all. "Began as experimental" doesn't always mean "is still experimental". |
Exactly .
I'm not trying to start an argument with you, merely pointing out that I believe ck has moved past the "experimental" stage. If you'll check its mailing lists, you'll see that there aren't really any bugs referred to in recent messages about recent versions of staircase or mapped watermark or the like. Anyway, that's most likely enough on ck for now. I like it, you don't. Somehow I doubt an argument here will change that . |
Exactly. Con's cpuscheduler is as stable as mainline's, if not more (as ingo's scheduler is much more complex in design). the reason it doesn't end up in mainline is that there is only one scheduler, and that's ingo's. he doesn't want another scheduler, and his works fine.
Now Con is busy getting the ability to choose from multiple cpu schedulers at runtime into the kernel, this will make inclusion of staircase more likely. but now (partly thanx to the testing in -ck) timesliced-cfq from jens axboe is stable (and will end up in Suse 9.3) there aren't many (if any) experimental things in -ck, especially not in the -ck for a stable kernel (as con only adds bugfixes to a 'stable' ck).
anyway, everyone is free to join testing -ck and help getting the kernel more responsive for the desktop, see you on the mailinglist |
|
Back to top |
|
|
Bob P Advocate
Joined: 20 Oct 2004 Posts: 3355 Location: Jackass! Development Labs
|
Posted: Sat Mar 19, 2005 7:07 pm Post subject: |
|
|
ebrostig wrote: | You had me scared for a minute there, that is when I saw -O4. "How the heck did he get hold of my own optimization code for gcc???" |
its no secret, Erik. many of us know that you have to edit the GCC ebuild to remove the flag substitution that reverts everything to -O2. we all know that -O4 code executes much faster than -Os, and will make a Pentium 60 run as fast as a un-optimized Pentium 4. _________________ .
Stage 1/3 | Jackass! | Rockhopper! | Thanks | Google Sucks |
|
Back to top |
|
|
|