Gentoo Forums
Gentoo Forums
Gentoo Forums
Quick Search: in
Acovea-4.0.0 : Try out my ebuilds (and scripts)
View unanswered posts
View posts from last 24 hours

Goto page Previous  1, 2, 3 ... 6, 7, 8 ... 14, 15, 16  Next  
Reply to topic    Gentoo Forums Forum Index Gentoo Chat
View previous topic :: View next topic  
Author Message
Hypnos
Advocate
Advocate


Joined: 18 Jul 2002
Posts: 2887
Location: Omnipresent

PostPosted: Tue Apr 27, 2004 12:15 am    Post subject: Reply with quote

Daagar wrote:
Hypnos wrote:
Depends -- would you be happy with a 0-3% performance improvement for that time invested?


Heheh... for me personally, sure. I'm twisted like that. However, as the previous poster had found, there are instances where the performance goes _backwards_. I guess the question is whether the perofrmance gains will in general outweigh the reverse for an average gentoo'ers system (based on the assumption that most gentoo'ers are in a workstation envrionment, and not doing 24/7 number crunching). .

That's the art of it -- finding a composite benchmark that represents the whole system code base, except smaller (but not so small that size-based optimizations are overly weighted).

Having read the benchmarking HOWTO, I think the thing to try is to isolate specific operations done by the desktop (see glib, gdk) and maximize each individually, and hope that the composite will be a good approximation to the combined action.
_________________
Personal overlay | Simple backup scheme
Back to top
View user's profile Send private message
Daagar
Tux's lil' helper
Tux's lil' helper


Joined: 14 Mar 2003
Posts: 78

PostPosted: Tue Apr 27, 2004 12:20 am    Post subject: Reply with quote

By all means, I agree with that. It appears, however, that the current benchmarks we have are not the appropriate benchmarks for general use. I simply don't want people getting the wrong impression and running acovea + emerge -e world expecting miracles (which was my initial impression until getting poisson's feedback), especially considering the time involved in that process (especially on older/slower machines, which are going to be the people most interested in eeking out every % of performance).
Back to top
View user's profile Send private message
Hypnos
Advocate
Advocate


Joined: 18 Jul 2002
Posts: 2887
Location: Omnipresent

PostPosted: Tue Apr 27, 2004 12:33 am    Post subject: Reply with quote

Daagar wrote:
By all means, I agree with that. It appears, however, that the current benchmarks we have are not the appropriate benchmarks for general use. I simply don't want people getting the wrong impression and running acovea + emerge -e world expecting miracles (which was my initial impression until getting poisson's feedback), especially considering the time involved in that process (especially on older/slower machines, which are going to be the people most interested in eeking out every % of performance).

Well, I'm skeptical about the new benchmarks hurting, if at all; but, you are correct that the jury is out on how much they help, if at all. My Quake and UT framerates are up after recompiling Xorg with the Acovea flags (~5 fps in Quake), and my executables are built faster and are smaller and so load quicker, so I'm pleased ;)

Perhaps we can think harder about a desktop-centric benchmark if and when Scott Ladd adds "make" support to Acovea -- e.g., glib+gdk.
_________________
Personal overlay | Simple backup scheme
Back to top
View user's profile Send private message
Daagar
Tux's lil' helper
Tux's lil' helper


Joined: 14 Mar 2003
Posts: 78

PostPosted: Tue Apr 27, 2004 12:37 am    Post subject: Reply with quote

That's encouraging to hear then! Re-reading the past posts, it appears _most_ people feel that their system improved after an acovea analysis. Granted, this could be the placebo effect, but no one so far has complained about perceived loss off performance. Good enough for me, YMMV. Heh.
Back to top
View user's profile Send private message
nmcsween
Guru
Guru


Joined: 12 Nov 2003
Posts: 381

PostPosted: Tue Apr 27, 2004 1:38 am    Post subject: Reply with quote

I just finished running acovea 4 times and got 4 different results in a old PIII 800MHz I also didn't touch, move, look at the computer while it was running the tests. Seems acovea should be run multiple times to get an idea of what flags will optimize
Back to top
View user's profile Send private message
Daagar
Tux's lil' helper
Tux's lil' helper


Joined: 14 Mar 2003
Posts: 78

PostPosted: Tue Apr 27, 2004 2:00 am    Post subject: Reply with quote

Care to post your Hypnos-parsed output for both runs? As it is using GA to arrive at a solution, I'm not sure this is necessarily 'wrong' that it gives inconsistent results. With a greater number of generations, they both might(?) eventually converge on the same solution. It could be random luck... if the algorithm started down a 'bad' genetic path, it may have wasted some iterations on it. A 3rd run might conclude which options are 'right', but that's a lot of extra cpu crunching :)
Back to top
View user's profile Send private message
wilburpan
l33t
l33t


Joined: 21 Jan 2003
Posts: 977

PostPosted: Tue Apr 27, 2004 2:38 am    Post subject: Reply with quote

Hypnos wrote:
I am pleased -- my speed boost is not placebo: Q3A is playable now at max settings, and UT2004 is more responsive....Overall, it's comparable to the boost you get moving from 2.4->2.6.

And later:
Hypnos wrote:
Depends -- would you be happy with a 0-3% performance improvement for that time invested?

Um, which is it for you? I have a hard time saying as to whether my recompiled system is much faster. It seems so, but it may also be placebo.

I can say that my CFLAGS settings generated by Acovea are really stable. Some packages that previously been sensitive to CFLAGS settings compiled without a hitch.
_________________
I'm only hanging out in OTW until I get rid of this stupid l33t ranking.....Crap. That didn't work.
Back to top
View user's profile Send private message
Hypnos
Advocate
Advocate


Joined: 18 Jul 2002
Posts: 2887
Location: Omnipresent

PostPosted: Tue Apr 27, 2004 3:09 am    Post subject: Reply with quote

wilburpan wrote:
Hypnos wrote:
I am pleased -- my speed boost is not placebo: Q3A is playable now at max settings, and UT2004 is more responsive....Overall, it's comparable to the boost you get moving from 2.4->2.6.

And later:
Hypnos wrote:
Depends -- would you be happy with a 0-3% performance improvement for that time invested?

Um, which is it for you? I have a hard time saying as to whether my recompiled system is much faster. It seems so, but it may also be placebo.

Well, the Quake improvement is real -- ~5fps out 50 (so that's more like 10%). The other stuff is harder to separate from placebo -- definitely feels snappier; Mozilla loads up in ~3.5 seconds for 2nd run forward on my laptop, which is pretty good.

Quote:
I can say that my CFLAGS settings generated by Acovea are really stable. Some packages that previously been sensitive to CFLAGS settings compiled without a hitch.

True, and they compile somewhat faster.
_________________
Personal overlay | Simple backup scheme
Back to top
View user's profile Send private message
Daagar
Tux's lil' helper
Tux's lil' helper


Joined: 14 Mar 2003
Posts: 78

PostPosted: Wed Apr 28, 2004 3:11 pm    Post subject: Reply with quote

Yet another question. Why do most people continue to start with -O2, and then use all the 'Yes' options? According to the output, -O2 usually includes some flags that turn up as 'No' according to acovea. I believe wilburpan brought this up a few pages back, but there was little comment on it. It appears that if you wish to trust acovea, you should really start with -O1, and then add all the 'Yes' options (which will turn some O1 flags off, and turn _some_ O2/O3 flags on, depending on system).


Also.... after reading more on the official Acovea site - while Scott may be adding makefile support in the future, making use of it in any sane way will be torturous. As he says in his example, something that takes 1min to compile and 1min to run will take 5.5days to evolve optimized settings. Extrapolate that to something like X.org, or even glibc, and you have quite a problem. Expanding the set of benchmarks already in acovea will likely be the shortterm solution, and even then you are tacking on 3-15hours/benchmark. It sounds as if the options we are getting from acovea now are about as optimal as we are likely to get in the near future (barring massive changes to how acovea itself works).
Back to top
View user's profile Send private message
Hypnos
Advocate
Advocate


Joined: 18 Jul 2002
Posts: 2887
Location: Omnipresent

PostPosted: Wed Apr 28, 2004 7:19 pm    Post subject: Reply with quote

Daagar wrote:
Yet another question. Why do most people continue to start with -O2, and then use all the 'Yes' options? According to the output, -O2 usually includes some flags that turn up as 'No' according to acovea. I believe wilburpan brought this up a few pages back, but there was little comment on it. It appears that if you wish to trust acovea, you should really start with -O1, and then add all the 'Yes' options (which will turn some O1 flags off, and turn _some_ O2/O3 flags on, depending on system).

Well, for me, nearly all of the O2 flags are yes/maybe, so it's just most expedient to run with that , and tweak as necessary (-O1 overrides, etc.).
_________________
Personal overlay | Simple backup scheme
Back to top
View user's profile Send private message
ed0n
l33t
l33t


Joined: 23 Apr 2003
Posts: 638
Location: Prishtine/Kosove

PostPosted: Wed Apr 28, 2004 7:46 pm    Post subject: Reply with quote

I am still at evo bench, I was wondering some things.
Do you get noticable speed after compiling with acovea generated flags?
Sorry if someone asked and answered this question before, I was lazy to read all the thread.
Back to top
View user's profile Send private message
Vagabond
Apprentice
Apprentice


Joined: 19 Jan 2003
Posts: 192

PostPosted: Wed Apr 28, 2004 11:18 pm    Post subject: Reply with quote

how exactly are ypou supposed to deal with the (! -O1) flags? I'm using O1 with the safe Cflags acovea recommended, but I didn't know what to do with the not -O1 flags or whatever....

Vag
Back to top
View user's profile Send private message
Hypnos
Advocate
Advocate


Joined: 18 Jul 2002
Posts: 2887
Location: Omnipresent

PostPosted: Thu Apr 29, 2004 1:03 am    Post subject: Reply with quote

Vagabond wrote:
how exactly are ypou supposed to deal with the (! -O1) flags? I'm using O1 with the safe Cflags acovea recommended, but I didn't know what to do with the not -O1 flags or whatever....

Use them to override stuff implied by -O1.
_________________
Personal overlay | Simple backup scheme
Back to top
View user's profile Send private message
wilburpan
l33t
l33t


Joined: 21 Jan 2003
Posts: 977

PostPosted: Fri Apr 30, 2004 4:37 am    Post subject: Reply with quote

Hypnos wrote:
Daagar wrote:
Yet another question. Why do most people continue to start with -O2, and then use all the 'Yes' options? According to the output, -O2 usually includes some flags that turn up as 'No' according to acovea. I believe wilburpan brought this up a few pages back, but there was little comment on it....

Well, for me, nearly all of the O2 flags are yes/maybe, so it's just most expedient to run with that, and tweak as necessary (-O1 overrides, etc.).

Just to clarify, I wound up using this approach as well. With the results of my run, most of the -O2 flags were "yes/maybe", and only 4 were "no", so I set my CFLAGS to include -O2 and all the rest of the "yes" options. After reading the info on the acovea site, I think that the "no" options certainly won't improve performance, but they won't necessarily degrade system performance, either.

They key thing for me is that this set of CFLAGS is the most stable I've ever had, and programs actually seem to compile faster, although I haven't done any actual timing to be sure.
_________________
I'm only hanging out in OTW until I get rid of this stupid l33t ranking.....Crap. That didn't work.
Back to top
View user's profile Send private message
Hypnos
Advocate
Advocate


Joined: 18 Jul 2002
Posts: 2887
Location: Omnipresent

PostPosted: Fri Apr 30, 2004 4:59 am    Post subject: Reply with quote

wilburpan wrote:
They key thing for me is that this set of CFLAGS is the most stable I've ever had, and programs actually seem to compile faster, although I haven't done any actual timing to be sure.

I concur.
_________________
Personal overlay | Simple backup scheme
Back to top
View user's profile Send private message
Andrej
n00b
n00b


Joined: 04 Oct 2003
Posts: 60
Location: Latvia

PostPosted: Fri Apr 30, 2004 5:28 am    Post subject: Reply with quote

Why the perl script does not work for me? Only writes
Code:
 Score |  So?  | Switch (annotation)
-------------------------------------
Anybody can tell me what flags must i use? Here it:
Code:

alma


optimistic options:
                   -fforce-mem (1.844)
              -fschedule-insns (2.555)
                 -falign-loops (1.335)
            -finline-functions (1.233)
                      -ftracer (1.284)
                -malign-double (1.132)
   -funsafe-math-optimizations (2.81)

pessimistic options:
             -fno-thread-jumps (-1.055)
            -fno-loop-optimize (-1.868)
                     -fregmove (-1.461)
                 -ffloat-store (-2.224)
                   -fno-inline (-2.122)
                      -fnew-ra (-2.275)
                  -mfpmath=387 (-1.156)
                  -mfpmath=sse (-1.156)
              -mfpmath=sse,387 (-1.258)
     -momit-leaf-frame-pointer (-1.207)

evo
optimistic options:
          -fno-merge-constants (1.347)
                        -fgcse (3.053)
     -fexpensive-optimizations (1.182)
             -fschedule-insns2 (2.558)
                -falign-labels (1.237)
   -funsafe-math-optimizations (1.953)
           -fno-signaling-nans (1.292)

pessimistic options:
              -fschedule-insns (-1.955)
                     -fregmove (-1.239)
                      -fnew-ra (-1.239)
                -funroll-loops (-1.019)
                  -mfpmath=sse (-2.285)
              -mfpmath=sse,387 (-2.12)
          -fomit-frame-pointer (-1.955)
fft
optimistic options:
 -fno-guess-branch-probability (1.992)
        -frerun-cse-after-loop (1.219)
              -frerun-loop-opt (1.457)
             -fschedule-insns2 (1.219)
             -fstrict-aliasing (3.122)
        -fprefetch-loop-arrays (1.279)

pessimistic options:
            -fno-loop-optimize (-2.587)
                        -fgcse (-1.517)
            -frename-registers (-1.517)
                -funroll-loops (-2.111)
            -funroll-all-loops (-1.814)
                  -mfpmath=387 (-2.052)
                  -mfpmath=sse (-1.219)
              -mfpmath=sse,387 (-1.814)
          -fomit-frame-pointer (-1.636)
     -momit-leaf-frame-pointer (-1.517)

huff
optimistic options:
                -fno-defer-pop (1.129)
            -fno-loop-optimize (1.04)
        -frerun-cse-after-loop (1.129)
                   -fforce-mem (1.757)
              -freorder-blocks (1.174)
                 -falign-loops (2.026)
                      -fnew-ra (1.802)
          -fomit-frame-pointer (1.847)
           -fno-signaling-nans (1.309)

pessimistic options:
 -fno-guess-branch-probability (-1.516)
          -fno-cprop-registers (-1.022)
            -fno-if-conversion (-2.098)
      -foptimize-sibling-calls (-1.247)
              -fschedule-insns (-1.829)
                     -fregmove (-1.202)
            -fsched-interblock (-1.067)
                      -ftracer (-1.022)
                -funroll-loops (-1.471)
                  -mfpmath=387 (-1.605)
                  -mfpmath=sse (-1.471)
              -mfpmath=sse,387 (-1.247)
     -momit-leaf-frame-pointer (-1.919)
lin
optimistic options:
           -fno-if-conversion2 (1.23)
             -fno-crossjumping (1.23)
      -foptimize-sibling-calls (1.476)
             -fstrength-reduce (1.045)
             -fstrict-aliasing (2.092)
                  -fsched-spec (1.106)
        -fprefetch-loop-arrays (2.401)
             -freduce-all-givs (1.106)
                   -fno-inline (1.353)
                      -fnew-ra (1.599)

pessimistic options:
        -frerun-cse-after-loop (-1.112)
                -falign-labels (-1.112)
                -funroll-loops (-2.283)
                  -mfpmath=387 (-2.098)
                  -mfpmath=sse (-1.975)
              -mfpmath=sse,387 (-2.16)
     -momit-leaf-frame-pointer (-1.975)
               -fno-math-errno (-1.359)
mat1
optimistic options:
             -fno-crossjumping (1.131)
        -frerun-cse-after-loop (1.131)
              -frerun-loop-opt (1.195)
                -fcaller-saves (1.321)
                   -fforce-mem (1.828)
                   -fpeephole2 (2.018)
              -fschedule-insns (1.068)
  -fdelete-null-pointer-checks (1.195)
            -fsched-interblock (1.005)
                  -fsched-spec (1.005)
                      -fnew-ra (1.068)
    -maccumulate-outgoing-args (1.702)
           -fno-signaling-nans (1.005)

pessimistic options:
            -fno-loop-optimize (-2.797)
                 -ffloat-store (-2.734)
                  -mfpmath=387 (-2.164)
                  -mfpmath=sse (-1.403)
              -mfpmath=sse,387 (-1.783)
          -fomit-frame-pointer (-1.847)
     -momit-leaf-frame-pointer (-1.023)

tree
optimistic options:
                        -fgcse (1.411)
        -frerun-cse-after-loop (1.452)
             -fschedule-insns2 (1.164)
             -fstrict-aliasing (2.192)
              -freorder-blocks (1.329)
                      -ftracer (2.028)
                     -mieee-fp (1.082)
          -fomit-frame-pointer (2.028)

pessimistic options:
 -fno-guess-branch-probability (-1.713)
            -fno-loop-optimize (-1.878)
             -fstrength-reduce (-1.015)
                   -fforce-mem (-1.508)
                      -fnew-ra (-1.919)
                -funroll-loops (-1.919)
            -funroll-all-loops (-1.919)
                  -mfpmath=sse (-1.549)
              -mfpmath=sse,387 (-1.426)
     -momit-leaf-frame-pointer (-1.878)

Back to top
View user's profile Send private message
Hypnos
Advocate
Advocate


Joined: 18 Jul 2002
Posts: 2887
Location: Omnipresent

PostPosted: Fri Apr 30, 2004 5:31 am    Post subject: Reply with quote

Andrej wrote:
Why the perl script does not work for me?

Well, it has some weaknesses:

* It expects the output files to have very specific names, and in the same directory: <bench>bench.run

* You have to run for at least 20 generations for each benchmark -- the script looks for this generation in the .run file to extract the subsequent table of options.
_________________
Personal overlay | Simple backup scheme
Back to top
View user's profile Send private message
Daagar
Tux's lil' helper
Tux's lil' helper


Joined: 14 Mar 2003
Posts: 78

PostPosted: Fri Apr 30, 2004 1:22 pm    Post subject: Reply with quote

Hypnos wrote:
Andrej wrote:
Why the perl script does not work for me?

Well, it has some weaknesses:

* It expects the output files to have very specific names, and in the same directory: <bench>bench.run

* You have to run for at least 20 generations for each benchmark -- the script looks for this generation in the .run file to extract the subsequent table of options.


I know I shouldn't be correcting the author himself, but it is really looking for <bench>.run, not <bench>bench.run ;)

Oh, and thanks to wilburpan for his clarification as well. Unless I see an extremely large number of "no's" for -O2 options, I'll likely go the same route based on experiences here.
Back to top
View user's profile Send private message
Hypnos
Advocate
Advocate


Joined: 18 Jul 2002
Posts: 2887
Location: Omnipresent

PostPosted: Fri Apr 30, 2004 6:53 pm    Post subject: Reply with quote

Daagar wrote:
I know I shouldn't be correcting the author himself, but it is really looking for <bench>.run, not <bench>bench.run ;)

You're right :oops:
_________________
Personal overlay | Simple backup scheme
Back to top
View user's profile Send private message
scoobydu
Veteran
Veteran


Joined: 16 Feb 2003
Posts: 1076
Location: 'Mind the Gap'

PostPosted: Wed May 05, 2004 8:54 am    Post subject: Reply with quote

Am I missing something here?

I'm running gcc3.4 with no problems, but I don't get any Yes or Maybe's?

tia

This;

Quote:

#BENCHES="alma evo fft huff lin mat1 tree"
BENCHES="mat1"

for bench in $BENCHES; do
echo ""
echo "*** $bench ***"
time runacovea -config gcc34_opteron.acovea -bench ${bench}bench.c\
1> ${bench}.run 2> ${bench}.err
done


Gives me this .. (after about 3 hrs!)

Quote:

Score | So? | Switch (annotation)
------------------------------------------------------------------------------
14.7 | No | -mno-push-args
14.5 | No | -maccumulate-outgoing-args
13.9 | No | -fschedule-insns (-O2)
13.1 | No | -fno-if-conversion (! -O1)
12.7 | No | -fno-defer-pop (! -O1)
12.6 | No | -ftracer
12.4 | No | -frerun-cse-after-loop (-O2)
12.1 | No | -fdelete-null-pointer-checks (-O2)
11.5 | No | -fno-delayed-branch (! -O1)
11.5 | No | -falign-loops (-O2 GCC 3.3)
11.4 | No | -fno-cprop-registers (! -O1)
11.3 | No | -fweb
11.3 | No | -fstrength-reduce (-O2)
11.3 | No | -fcse-skip-blocks (-O2)
11.1 | No | -fno-if-conversion2 (! -O1)
10.9 | No | -fno-omit-frame-pointer (! -O1)
10.8 | No | -funswitch-loops
10.8 | No | -foptimize-sibling-calls (-O2)
10.8 | No | -fno-thread-jumps (! -O1)
10.4 | No | -fexpensive-optimizations (-O2)
10.3 | No | -freorder-functions (-O2 GCC 3.3)
10.2 | No | -falign-labels (-O2 GCC 3.3)
10.1 | No | -falign-jumps (-O2 GCC 3.3)
9.9 | No | -fgcse (-O2)
9.8 | No | -minline-all-stringops
9.7 | No | -fcse-follow-jumps (-O2)
9.6 | No | -frerun-loop-opt (-O2)
9.4 | No | -fmove-all-movables
9.4 | No | -fsched-spec (-O2 GCC 3.3)
9.1 | No | -frename-registers (-O3)
9.1 | No | -fstrict-aliasing (-O2)
8.8 | No | -fcaller-saves (-O2)
8.7 | No | -fsched-interblock (-O2 GCC 3.3)
8.0 | No | -freorder-blocks (-O2)
7.7 | No | -finline-functions (-O3)
7.6 | No | -fpeephole2 (-O2)
7.5 | No | -fschedule-insns2 (-O2)
7.3 | No | -fforce-mem (-O2)
6.9 | No | -funroll-loops
6.9 | No | -falign-functions
6.4 | No | -fregmove (-O2)
6.2 | No | -freduce-all-givs
6.0 | No | -mno-align-stringops
5.9 | No | -funroll-all-loops
5.7 | No | -mfpmath=387
5.3 | No | -mieee-fp
5.1 | No | -fnew-ra
3.4 | No | -funsafe-math-optimizations (fast math)
0.0 | No | -fno-merge-constants (! -O1)
0.0 | No | -fno-guess-branch-probability (! -O1)
0.0 | No | -fno-loop-optimize (! -O1)
0.0 | No | -fno-crossjumping (! -O1)
0.0 | No | -funit-at-a-time
0.0 | No | -ffloat-store
0.0 | No | -fprefetch-loop-arrays
0.0 | No | -fno-inline
0.0 | No | -fpeel-loops
0.0 | No | -fbranch-target-load-optimize
0.0 | No | -fbranch-target-load-optimize2
0.0 | No | -mfpmath=sse
0.0 | No | -mfpmath=sse,387
0.0 | No | -fno-math-errno (fast math)
0.0 | No | -fno-trapping-math (fast math)
0.0 | No | -ffinite-math-only (fast math)
0.0 | No | -fno-signaling-nans (fast math)
0.0 | No | -finline-limit

_________________
Tyan Tiger K8W, 2xOpteron 240,Powerbook5,6 15" 1.5g, Macbook Black 2g, Mac Mini 1g, Ipod P60g.
| Linux - From a windows user perspective|
Back to top
View user's profile Send private message
Hypnos
Advocate
Advocate


Joined: 18 Jul 2002
Posts: 2887
Location: Omnipresent

PostPosted: Wed May 05, 2004 10:18 am    Post subject: Reply with quote

scoobydu wrote:
Am I missing something here?

Yes -- you haven't accumulated sufficient counts, so the script says no on all of them (in Poisson statistics, the std. dev. goes as the square-root of the total). It takes most people about 48 hrs. to run through all the benchmarks.
_________________
Personal overlay | Simple backup scheme
Back to top
View user's profile Send private message
scoobydu
Veteran
Veteran


Joined: 16 Feb 2003
Posts: 1076
Location: 'Mind the Gap'

PostPosted: Wed May 05, 2004 10:47 am    Post subject: Reply with quote

Hypnos wrote:
scoobydu wrote:
Am I missing something here?

Yes -- you haven't accumulated sufficient counts, so the script says no on all of them (in Poisson statistics, the std. dev. goes as the square-root of the total). It takes most people about 48 hrs. to run through all the benchmarks.


I see, thanks.

So do I have to run all of the tests at the same time, one after the other?

As I only ran mat1 in this case (I said 3hrs, but that was before I went to sleep!, so it could have been alot more than that)
_________________
Tyan Tiger K8W, 2xOpteron 240,Powerbook5,6 15" 1.5g, Macbook Black 2g, Mac Mini 1g, Ipod P60g.
| Linux - From a windows user perspective|
Back to top
View user's profile Send private message
Hypnos
Advocate
Advocate


Joined: 18 Jul 2002
Posts: 2887
Location: Omnipresent

PostPosted: Wed May 05, 2004 10:52 am    Post subject: Reply with quote

scoobydu wrote:

So do I have to run all of the tests at the same time, one after the other?

As I only ran mat1 in this case (I said 3hrs, but that was before I went to sleep!, so it could have been alot more than that)

My recommenation is to hit Ctl-Z to suspend/resume the script and have it run mostly while you are sleeping :wink:
_________________
Personal overlay | Simple backup scheme
Back to top
View user's profile Send private message
scoobydu
Veteran
Veteran


Joined: 16 Feb 2003
Posts: 1076
Location: 'Mind the Gap'

PostPosted: Wed May 05, 2004 11:46 am    Post subject: Reply with quote

Hypnos wrote:
scoobydu wrote:

So do I have to run all of the tests at the same time, one after the other?

As I only ran mat1 in this case (I said 3hrs, but that was before I went to sleep!, so it could have been alot more than that)

My recommenation is to hit Ctl-Z to suspend/resume the script and have it run mostly while you are sleeping :wink:


Thanks Hypnos, I'll do that.

I'm building everything with gcc3.4 atm, so I'll wait for that to finish, then kick it off.
_________________
Tyan Tiger K8W, 2xOpteron 240,Powerbook5,6 15" 1.5g, Macbook Black 2g, Mac Mini 1g, Ipod P60g.
| Linux - From a windows user perspective|
Back to top
View user's profile Send private message
Hypnos
Advocate
Advocate


Joined: 18 Jul 2002
Posts: 2887
Location: Omnipresent

PostPosted: Wed May 05, 2004 11:52 am    Post subject: Reply with quote

scoobydu wrote:
I'm building everything with gcc3.4 atm, so I'll wait for that to finish, then kick it off.

Note that my script only does annotations for gcc3.3 options ("! -O1", "-O2", etc.), not gcc3.4.

Have fun!
_________________
Personal overlay | Simple backup scheme
Back to top
View user's profile Send private message
Display posts from previous:   
Reply to topic    Gentoo Forums Forum Index Gentoo Chat All times are GMT
Goto page Previous  1, 2, 3 ... 6, 7, 8 ... 14, 15, 16  Next
Page 7 of 16

 
Jump to:  
You cannot post new topics in this forum
You cannot reply to topics in this forum
You cannot edit your posts in this forum
You cannot delete your posts in this forum
You cannot vote in polls in this forum