View previous topic :: View next topic |
Author |
Message |
feffi Apprentice
Joined: 29 Mar 2003 Posts: 216 Location: Sol->Earth->Germany->Giessen
|
Posted: Sat Aug 21, 2004 11:32 pm Post subject: Re: what the....?! |
|
|
hovenko wrote: | why the **** did the Acer Travelmate 803LCi wiki-page got split up? now it's all just a mess... perhaps it's just the best if kev updates the forum-posts again, and we add things to this forumpage if we come up with new info. |
it is not split up!?!? It's all in one place, fine layouted and very actual. I decided to give everybody a chance to provide his knowledge on this howto. We (the main contributors to the howto and me) decided this step. I think it's a step forward as the howto slowly becomes too big for me alone to handle. Nevertheless all major changes and new hints and tricks are posted here before it is added to the gentoo-wiki, so you can still follow the media. For myself, it's a lot easier to handle updates and layouting all that stuff so you can read it in the gentoo-wiki. And even people not involved in this thread can add usefull information.
I was a coordinating instance at the beginning of this howto, now we are all the coordinating instance, watching over the howto and correcting and adding stuff. I really think this is a progress of the community. _________________ have fun
feffi
/(bb|[^b]{2})/ that is the Question!
Gentoo-Wiki: Acer Travelmate 803 LCi manual |
|
Back to top |
|
|
feffi Apprentice
Joined: 29 Mar 2003 Posts: 216 Location: Sol->Earth->Germany->Giessen
|
Posted: Tue Aug 24, 2004 12:14 pm Post subject: acovea testing |
|
|
Added 4.2.1 Acovea
Added 4.2.1.1 Pentium-M
Added 4.2.1.2 Pentium-4
Added 4.2.1.3 Pentium-3 _________________ have fun
feffi
/(bb|[^b]{2})/ that is the Question!
Gentoo-Wiki: Acer Travelmate 803 LCi manual |
|
Back to top |
|
|
pjv Guru
Joined: 02 Jul 2003 Posts: 353 Location: Belgium
|
Posted: Tue Aug 24, 2004 6:01 pm Post subject: |
|
|
Added Pentium 3
Did forget to save the flags themselves though! A little bit stupid. Does acovea have a log anywhere?
Also _kev_, where did you get the config file for pentium-M for use with acovea?
Last edited by pjv on Tue Aug 24, 2004 9:02 pm; edited 1 time in total |
|
Back to top |
|
|
fatrix n00b
Joined: 09 Oct 2003 Posts: 4 Location: Norway
|
Posted: Tue Aug 24, 2004 7:11 pm Post subject: |
|
|
Ahh..... I see... it was just talk about splitting up the howto....
There are some talk about this:
http://gentoo-wiki.com/Talk:Gentoo_Acer_Travelmate_803LCi_Manual
That's good _________________ Fatrix
*Trust is a Weakness*
Acer Travelmate 803LCi / Gentoo Linux
Take controll of your operating system, dont let the operating system take controll of you. |
|
Back to top |
|
|
feffi Apprentice
Joined: 29 Mar 2003 Posts: 216 Location: Sol->Earth->Germany->Giessen
|
Posted: Tue Aug 24, 2004 10:46 pm Post subject: |
|
|
pjv wrote: | Added Pentium 3
Did forget to save the flags themselves though! A little bit stupid. Does acovea have a log anywhere?
Also _kev_, where did you get the config file for pentium-M for use with acovea? |
Cool thanks, then I don't have to run it myself. acovea has no log I know of, but just copy together the optimistic ones. (edit: I just did it myself ) The Pentium-M config was from a friend of mine. He patched the Pentium-4 and Pentium-3 together and it worked so far. It's astounding how all the CFLAGS differ on different layouts. I also added a analysis section to discuss the CFLAGS one by one, so any help would be appreciated. Happy testing! _________________ have fun
feffi
/(bb|[^b]{2})/ that is the Question!
Gentoo-Wiki: Acer Travelmate 803 LCi manual |
|
Back to top |
|
|
feffi Apprentice
Joined: 29 Mar 2003 Posts: 216 Location: Sol->Earth->Germany->Giessen
|
Posted: Tue Aug 24, 2004 11:23 pm Post subject: |
|
|
Why is that good?
Yes, I stated my thought about this howto and I don't want to split it up! The main purpose was to have all this info in one place and not scattered around, so why do a split-up? If some people want to see our information in other tutorials, why don't they just copy it or crosslink? _________________ have fun
feffi
/(bb|[^b]{2})/ that is the Question!
Gentoo-Wiki: Acer Travelmate 803 LCi manual |
|
Back to top |
|
|
fatrix n00b
Joined: 09 Oct 2003 Posts: 4 Location: Norway
|
Posted: Wed Aug 25, 2004 7:26 am Post subject: |
|
|
i ment.... That's good... that it's not split up
It's great to have a howto for gentoo on this laptop, and that it's all placed in one howto.
Sorry for bringing this misunderstanding up at the first place. _________________ Fatrix
*Trust is a Weakness*
Acer Travelmate 803LCi / Gentoo Linux
Take controll of your operating system, dont let the operating system take controll of you. |
|
Back to top |
|
|
pjv Guru
Joined: 02 Jul 2003 Posts: 353 Location: Belgium
|
Posted: Wed Aug 25, 2004 10:29 pm Post subject: |
|
|
I'm posting my (long) findings about acovea here. This is my second (and first proper) benchmark. I think we shouldn't put discussions in the wiki, but here. And be careful, I'm not asking for flame wars. Au contrary, those should be in the CFLAGS central thread or something. Hope you don't mind this length!
I've made my own config, that I think is better fit than the standard ones for pentium-m. It's also for gcc 3.3. You can find it here: gcc33_mymachine.acovea. I used runacovea with the huffbench. I hope to do this all again when gcc 3.4 and 3.5 come out.
The config is based on my current CFLAGS, which are: Code: | CFLAGS="-march=pentium3 -mcpu=pentium4 -O2 -pipe -fomit-frame-pointer -fprefetch-loop-arrays -fmerge-constants"
| and which I've always thought to be very fast.
These are the last 5 generations of the output:
Code: |
runacovea -config gcc33_mymachine.acovea -bench huffbench.c -n 5 -p 40 -g 30
iteration 25: average fitness = 3.09802, best fitnesses = 2.7761 2.75346 2.75687 2.74798 2.76187
best options for population 0
gcc -lrt -lm -std=gnu99 -march=pentium3 -mcpu=pentium4 -O2 -pipe -fomit-frame-pointer -fprefetch-loop-arrays -fmerge-constants -o TEMP -fno-merge-constants -fno-thread-jumps -fno-if-conversion2 -fno-delayed-branch -fno-crossjumping -fno-optimize-sibling-calls -fno-cse-follow-jumps -fno-gcse -fno-expensive-optimizations -fno-rerun-cse-after-loop -fno-rerun-loop-opt -fno-caller-saves -fno-peephole2 -fno-schedule-insns2 -fno-strict-aliasing -fno-sched-interblock -fno-align-loops -fno-align-jumps -fno-align-labels -frename-registers -ffloat-store -fno-prefetch-loop-arrays -fmove-all-movables -fnew-ra -funroll-all-loops -mfpmath=sse -momit-leaf-frame-pointer -mmmx -msse2 -mno-ieee-fp -mpush-args -mno-accumulate-outgoing-args -mno-inline-all-stringops -mno-omit-leaf-frame-pointer -mno-mmx -mno-sse -mno-sse2 -fno-math-errno -funsafe-math-optimizations -fno-trapping-math -ffinite-math-only -fno-signaling-nans -finline-limit=600 /usr/share/acovea/benchmarks/huffbench.c
best options for population 1
gcc -lrt -lm -std=gnu99 -march=pentium3 -mcpu=pentium4 -O2 -pipe -fomit-frame-pointer -fprefetch-loop-arrays -fmerge-constants -o TEMP -fno-merge-constants -fno-optimize-sibling-calls -fno-cse-skip-blocks -fno-gcse -fno-expensive-optimizations -fno-strength-reduce -fno-rerun-cse-after-loop -fno-rerun-loop-opt -fno-caller-saves -fno-peephole2 -fno-schedule-insns -fno-regmove -fno-sched-interblock -fno-sched-spec -fno-align-loops -fno-align-jumps -fno-align-functions -finline-functions -frename-registers -ffloat-store -fno-prefetch-loop-arrays -fmove-all-movables -fnew-ra -funroll-all-loops -malign-double -mno-push-args -maccumulate-outgoing-args -minline-all-stringops -mfpmath=387 -mmmx -msse2 -mno-ieee-fp -mno-align-double -mno-accumulate-outgoing-args -malign-stringops -mno-inline-all-stringops -mno-sse -mno-sse2 -funsafe-math-optimizations -fno-trapping-math /usr/share/acovea/benchmarks/huffbench.c
best options for population 2
gcc -lrt -lm -std=gnu99 -march=pentium3 -mcpu=pentium4 -O2 -pipe -fomit-frame-pointer -fprefetch-loop-arrays -fmerge-constants -o TEMP -fno-defer-pop -fno-thread-jumps -fno-cprop-registers -fno-if-conversion2 -fno-delayed-branch -fno-cse-skip-blocks -fno-gcse -fno-gcse-lm -fno-gcse-sm -fno-rerun-cse-after-loop -fno-peephole2 -fno-schedule-insns -fno-schedule-insns2 -fno-strict-aliasing -fno-reorder-blocks -fno-sched-interblock -fno-sched-spec -finline-functions -frename-registers -fno-prefetch-loop-arrays -fmove-all-movables -fno-inline -ftracer -fnew-ra -funroll-all-loops -mieee-fp -malign-double -maccumulate-outgoing-args -minline-all-stringops -momit-leaf-frame-pointer -msse -mno-align-double -mno-omit-leaf-frame-pointer -mno-sse -fno-trapping-math -finline-limit=600 /usr/share/acovea/benchmarks/huffbench.c
best options for population 3
gcc -lrt -lm -std=gnu99 -march=pentium3 -mcpu=pentium4 -O2 -pipe -fomit-frame-pointer -fprefetch-loop-arrays -fmerge-constants -o TEMP -fno-merge-constants -fno-defer-pop -fno-thread-jumps -fno-cprop-registers -fno-if-conversion2 -fno-optimize-sibling-calls -fno-cse-skip-blocks -fno-gcse -fno-gcse-sm -fno-strength-reduce -fno-rerun-cse-after-loop -fno-rerun-loop-opt -fno-peephole2 -fno-schedule-insns -fno-regmove -fno-align-loops -frename-registers -ffloat-store -fno-prefetch-loop-arrays -fmove-all-movables -ftracer -fnew-ra -funroll-all-loops -malign-double -minline-all-stringops -mmmx -msse2 -mno-ieee-fp -mno-align-double -mno-accumulate-outgoing-args -mno-inline-all-stringops -mno-omit-leaf-frame-pointer -mno-mmx -mno-sse2 -fno-trapping-math -ffinite-math-only -finline-limit=600 /usr/share/acovea/benchmarks/huffbench.c
best options for population 4
gcc -lrt -lm -std=gnu99 -march=pentium3 -mcpu=pentium4 -O2 -pipe -fomit-frame-pointer -fprefetch-loop-arrays -fmerge-constants -o TEMP -fno-merge-constants -fno-defer-pop -fno-thread-jumps -fno-omit-frame-pointer -fno-if-conversion2 -fno-cse-follow-jumps -fno-cse-skip-blocks -fno-gcse -fno-rerun-cse-after-loop -fno-caller-saves -fno-peephole2 -fno-schedule-insns2 -fno-reorder-blocks -fno-sched-interblock -ffloat-store -fno-prefetch-loop-arrays -fmove-all-movables -fno-inline -ftracer -fnew-ra -funroll-all-loops -mieee-fp -malign-double -mno-push-args -maccumulate-outgoing-args -fomit-frame-pointer -msse -mno-align-double -mpush-args -mno-accumulate-outgoing-args -malign-stringops -mno-omit-leaf-frame-pointer -mno-mmx -mno-sse -mno-sse2 -fno-trapping-math -fno-signaling-nans /usr/share/acovea/benchmarks/huffbench.c
common options (found in all populations)
gcc -lrt -lm -std=gnu99 -march=pentium3 -mcpu=pentium4 -O2 -pipe -fomit-frame-pointer -fprefetch-loop-arrays -fmerge-constants -o TEMP -fno-gcse -fno-rerun-cse-after-loop -fno-peephole2 -fno-prefetch-loop-arrays -fmove-all-movables -fnew-ra -funroll-all-loops -fno-trapping-math /usr/share/acovea/benchmarks/huffbench.c
rejected options (rejected by all populations)
gcc -lrt -lm -std=gnu99 -march=pentium3 -mcpu=pentium4 -O2 -pipe -fomit-frame-pointer -fprefetch-loop-arrays -fmerge-constants -o TEMP -fno-guess-branch-probability -fno-if-conversion -fno-loop-optimize -fno-force-mem -fno-reorder-functions -freduce-all-givs -mno-align-stringops /usr/share/acovea/benchmarks/huffbench.c
Option counts:
-fno-merge-constants 25 9 19 19 9 81
-fno-defer-pop 18 2 10 14 25 69
-fno-thread-jumps 11 19 19 9 9 67
-fno-omit-frame-pointer 1 0 15 1 7 24
-fno-guess-branch-probability 0 0 2 3 0 5
-fno-cprop-registers 10 22 20 14 9 75
-fno-if-conversion 0 0 0 0 0 0
-fno-if-conversion2 8 19 20 21 15 83
-fno-delayed-branch 5 1 19 4 11 40
-fno-loop-optimize 3 20 3 1 4 31
-fno-crossjumping 7 19 12 6 14 58
-fno-optimize-sibling-calls 3 25 7 21 14 70
-fno-cse-follow-jumps 21 6 18 6 24 75
-fno-cse-skip-blocks 8 25 8 17 18 76
-fno-gcse 11 25 25 22 20 103
-fno-gcse-lm 0 0 21 6 6 33
-fno-gcse-sm 6 0 10 18 12 46
-fno-expensive-optimizations 19 24 16 13 9 81
-fno-strength-reduce 4 4 17 22 0 47
-fno-rerun-cse-after-loop 14 19 5 20 12 70
-fno-rerun-loop-opt 7 16 16 19 1 59
-fno-caller-saves 23 8 12 14 18 75
-fno-force-mem 4 6 0 2 0 12
-fno-peephole2 25 13 17 23 24 102
-fno-schedule-insns 9 20 24 21 11 85
-fno-schedule-insns2 10 0 8 0 9 27
-fno-regmove 14 18 3 21 8 64
-fno-strict-aliasing 7 0 15 5 12 39
-fno-reorder-blocks 7 10 12 0 19 48
-fno-sched-interblock 25 13 22 13 25 98
-fno-sched-spec 2 20 16 10 9 57
-fno-reorder-functions 1 9 5 2 12 29
-fno-align-loops 22 7 16 23 4 72
-fno-align-jumps 10 25 10 11 1 57
-fno-align-labels 23 18 1 3 6 51
-fno-align-functions 18 19 9 10 5 61
-finline-functions 0 23 20 2 9 54
-frename-registers 21 22 10 23 18 94
-ffloat-store 15 3 7 19 10 54
-fno-prefetch-loop-arrays 22 22 25 23 25 117
-fmove-all-movables 21 24 17 14 21 97
-freduce-all-givs 0 12 3 3 4 22
-fno-inline 3 1 7 8 14 33
-ftracer 18 18 14 13 23 86
-fnew-ra 15 25 22 24 20 106
-funroll-loops 0 4 3 0 0 7
-funroll-all-loops 25 8 18 21 21 93
-mieee-fp 17 0 15 3 24 59
-malign-double 5 4 22 24 17 72
-mno-push-args 10 4 8 4 17 43
-maccumulate-outgoing-args 4 4 19 14 13 54
-mno-align-stringops 4 0 2 7 1 14
-minline-all-stringops 1 20 19 24 8 72
-mfpmath=387 1 3 1 5 1 11
-mfpmath=sse 23 8 0 0 4 35
-mfpmath=sse,387 0 0 3 1 0 4
-fomit-frame-pointer 0 0 16 0 8 24
-momit-leaf-frame-pointer 12 0 6 2 16 36
-mmmx 6 12 11 24 3 56
-msse 4 0 14 2 13 33
-msse2 3 7 3 25 5 43
-mno-ieee-fp 8 7 20 22 18 75
-mno-align-double 9 20 25 23 20 97
-mpush-args 16 4 5 14 24 63
-mno-accumulate-outgoing-args 22 22 5 17 24 90
-malign-stringops 0 25 14 10 8 57
-mno-inline-all-stringops 16 20 2 22 12 72
-fno-omit-frame-pointer 1 1 0 0 0 2
-mno-omit-leaf-frame-pointer 16 8 7 16 23 70
-mno-mmx 20 18 16 14 14 82
-mno-sse 15 4 6 9 19 53
-mno-sse2 20 3 20 21 13 77
-fno-math-errno 12 6 5 7 15 45
-funsafe-math-optimizations 21 4 1 11 10 47
-fno-trapping-math 12 8 21 16 15 72
-ffinite-math-only 21 20 17 12 11 81
-fno-signaling-nans 24 15 6 2 12 59
-finline-limit 15 3 16 14 17 65
value options:
-finline-limit: 600 0 600 600 0 , average = 600 across 3 populations
cc1: let op: SSE instruction set disabled, using 387 arithmetics
cc1: let op: SSE instruction set disabled, using 387 arithmetics
cc1: let op: SSE instruction set disabled, using 387 arithmetics
cc1: let op: SSE instruction set disabled, using 387 arithmetics
cc1: let op: SSE instruction set disabled, using 387 arithmetics
cc1: let op: SSE instruction set disabled, using 387 arithmetics
cc1: let op: SSE instruction set disabled, using 387 arithmetics
cc1: let op: SSE instruction set disabled, using 387 arithmetics
cc1: let op: SSE instruction set disabled, using 387 arithmetics
cc1: let op: SSE instruction set disabled, using 387 arithmetics
cc1: let op: SSE instruction set disabled, using 387 arithmetics
cc1: let op: SSE instruction set disabled, using 387 arithmetics
cc1: let op: SSE instruction set disabled, using 387 arithmetics
cc1: let op: SSE instruction set disabled, using 387 arithmetics
cc1: let op: SSE instruction set disabled, using 387 arithmetics
cc1: let op: SSE instruction set disabled, using 387 arithmetics
cc1: let op: SSE instruction set disabled, using 387 arithmetics
cc1: let op: SSE instruction set disabled, using 387 arithmetics
cc1: let op: SSE instruction set disabled, using 387 arithmetics
cc1: let op: SSE instruction set disabled, using 387 arithmetics
cc1: let op: SSE instruction set disabled, using 387 arithmetics
cc1: let op: SSE instruction set disabled, using 387 arithmetics
cc1: let op: SSE instruction set disabled, using 387 arithmetics
cc1: let op: SSE instruction set disabled, using 387 arithmetics
cc1: let op: SSE instruction set disabled, using 387 arithmetics
cc1: let op: SSE instruction set disabled, using 387 arithmetics
cc1: let op: SSE instruction set disabled, using 387 arithmetics
------------------------------------------------------------
iteration 26: average fitness = 3.07784, best fitnesses = 2.77781 2.75852 2.75939 2.75241 2.76716
best options for population 0
gcc -lrt -lm -std=gnu99 -march=pentium3 -mcpu=pentium4 -O2 -pipe -fomit-frame-pointer -fprefetch-loop-arrays -fmerge-constants -o TEMP -fno-merge-constants -fno-thread-jumps -fno-delayed-branch -fno-cse-follow-jumps -fno-gcse -fno-strength-reduce -fno-caller-saves -fno-peephole2 -fno-schedule-insns2 -fno-regmove -fno-reorder-blocks -fno-sched-interblock -fno-align-labels -fno-align-functions -frename-registers -fno-prefetch-loop-arrays -fmove-all-movables -fno-inline -ftracer -fnew-ra -funroll-all-loops -mieee-fp -maccumulate-outgoing-args -mfpmath=sse -mmmx -mpush-args -mno-inline-all-stringops -mno-omit-leaf-frame-pointer -mno-sse -mno-sse2 -ffinite-math-only -fno-signaling-nans /usr/share/acovea/benchmarks/huffbench.c
best options for population 1
gcc -lrt -lm -std=gnu99 -march=pentium3 -mcpu=pentium4 -O2 -pipe -fomit-frame-pointer -fprefetch-loop-arrays -fmerge-constants -o TEMP -fno-merge-constants -fno-optimize-sibling-calls -fno-cse-skip-blocks -fno-gcse -fno-expensive-optimizations -fno-strength-reduce -fno-rerun-cse-after-loop -fno-rerun-loop-opt -fno-caller-saves -fno-peephole2 -fno-schedule-insns -fno-regmove -fno-sched-interblock -fno-sched-spec -fno-align-loops -fno-align-jumps -fno-align-functions -finline-functions -frename-registers -ffloat-store -fno-prefetch-loop-arrays -fmove-all-movables -fnew-ra -funroll-all-loops -malign-double -mno-push-args -maccumulate-outgoing-args -minline-all-stringops -mfpmath=387 -mmmx -msse2 -mno-ieee-fp -mno-align-double -mno-accumulate-outgoing-args -malign-stringops -mno-inline-all-stringops -mno-sse -mno-sse2 -funsafe-math-optimizations -fno-trapping-math /usr/share/acovea/benchmarks/huffbench.c
best options for population 2
gcc -lrt -lm -std=gnu99 -march=pentium3 -mcpu=pentium4 -O2 -pipe -fomit-frame-pointer -fprefetch-loop-arrays -fmerge-constants -o TEMP -fno-defer-pop -fno-thread-jumps -fno-cprop-registers -fno-if-conversion2 -fno-delayed-branch -fno-crossjumping -fno-cse-skip-blocks -fno-gcse -fno-gcse-lm -fno-gcse-sm -fno-rerun-loop-opt -fno-peephole2 -fno-schedule-insns2 -fno-strict-aliasing -fno-reorder-blocks -fno-sched-interblock -fno-sched-spec -finline-functions -frename-registers -fno-prefetch-loop-arrays -fmove-all-movables -fno-inline -ftracer -fnew-ra -funroll-all-loops -mieee-fp -malign-double -mno-push-args -maccumulate-outgoing-args -minline-all-stringops -momit-leaf-frame-pointer -msse -msse2 -mno-align-double -malign-stringops -mno-omit-leaf-frame-pointer -fno-trapping-math -ffinite-math-only -finline-limit=600 /usr/share/acovea/benchmarks/huffbench.c
best options for population 3
gcc -lrt -lm -std=gnu99 -march=pentium3 -mcpu=pentium4 -O2 -pipe -fomit-frame-pointer -fprefetch-loop-arrays -fmerge-constants -o TEMP -fno-merge-constants -fno-defer-pop -fno-thread-jumps -fno-cprop-registers -fno-if-conversion2 -fno-optimize-sibling-calls -fno-cse-skip-blocks -fno-gcse -fno-strength-reduce -fno-rerun-cse-after-loop -fno-rerun-loop-opt -fno-peephole2 -fno-schedule-insns -fno-regmove -fno-strict-aliasing -fno-sched-spec -fno-align-loops -fno-align-jumps -frename-registers -ffloat-store -fno-prefetch-loop-arrays -fnew-ra -funroll-all-loops -maccumulate-outgoing-args -mmmx -msse2 -mno-ieee-fp -mno-align-double -mno-inline-all-stringops -mno-omit-leaf-frame-pointer -mno-mmx /usr/share/acovea/benchmarks/huffbench.c
best options for population 4
gcc -lrt -lm -std=gnu99 -march=pentium3 -mcpu=pentium4 -O2 -pipe -fomit-frame-pointer -fprefetch-loop-arrays -fmerge-constants -o TEMP -fno-merge-constants -fno-defer-pop -fno-thread-jumps -fno-omit-frame-pointer -fno-if-conversion2 -fno-cse-follow-jumps -fno-cse-skip-blocks -fno-gcse -fno-rerun-cse-after-loop -fno-caller-saves -fno-peephole2 -fno-schedule-insns2 -fno-reorder-blocks -fno-sched-interblock -ffloat-store -fno-prefetch-loop-arrays -fmove-all-movables -fno-inline -ftracer -fnew-ra -funroll-all-loops -mieee-fp -malign-double -mno-push-args -maccumulate-outgoing-args -fomit-frame-pointer -msse -mno-align-double -mpush-args -mno-accumulate-outgoing-args -malign-stringops -mno-omit-leaf-frame-pointer -mno-mmx -mno-sse -mno-sse2 -fno-trapping-math -fno-signaling-nans /usr/share/acovea/benchmarks/huffbench.c
common options (found in all populations)
gcc -lrt -lm -std=gnu99 -march=pentium3 -mcpu=pentium4 -O2 -pipe -fomit-frame-pointer -fprefetch-loop-arrays -fmerge-constants -o TEMP -fno-gcse -fno-peephole2 -fno-prefetch-loop-arrays -fnew-ra -funroll-all-loops -maccumulate-outgoing-args /usr/share/acovea/benchmarks/huffbench.c
rejected options (rejected by all populations)
gcc -lrt -lm -std=gnu99 -march=pentium3 -mcpu=pentium4 -O2 -pipe -fomit-frame-pointer -fprefetch-loop-arrays -fmerge-constants -o TEMP -fno-guess-branch-probability -fno-if-conversion -fno-loop-optimize -fno-force-mem -fno-reorder-functions -freduce-all-givs -mno-align-stringops -fno-math-errno /usr/share/acovea/benchmarks/huffbench.c
Option counts:
-fno-merge-constants 26 10 19 20 10 85
-fno-defer-pop 18 2 11 15 26 72
-fno-thread-jumps 12 19 20 10 10 71
-fno-omit-frame-pointer 1 0 15 1 8 25
-fno-guess-branch-probability 0 0 2 3 0 5
-fno-cprop-registers 10 22 21 15 9 77
-fno-if-conversion 0 0 0 0 0 0
-fno-if-conversion2 8 19 21 22 16 86
-fno-delayed-branch 6 1 20 4 11 42
-fno-loop-optimize 3 20 3 1 4 31
-fno-crossjumping 7 19 13 6 14 59
-fno-optimize-sibling-calls 3 26 7 22 14 72
-fno-cse-follow-jumps 22 6 18 6 25 77
-fno-cse-skip-blocks 8 26 9 18 19 80
-fno-gcse 12 26 26 23 21 108
-fno-gcse-lm 0 0 22 6 6 34
-fno-gcse-sm 6 0 11 18 12 47
-fno-expensive-optimizations 19 25 16 13 9 82
-fno-strength-reduce 5 5 17 23 0 50
-fno-rerun-cse-after-loop 14 20 5 21 13 73
-fno-rerun-loop-opt 7 17 17 20 1 62
-fno-caller-saves 24 9 12 14 19 78
-fno-force-mem 4 6 0 2 0 12
-fno-peephole2 26 14 18 24 25 107
-fno-schedule-insns 9 21 24 22 11 87
-fno-schedule-insns2 11 0 9 0 10 30
-fno-regmove 15 19 3 22 8 67
-fno-strict-aliasing 7 0 16 6 12 41
-fno-reorder-blocks 8 10 13 0 20 51
-fno-sched-interblock 26 14 23 13 26 102
-fno-sched-spec 2 21 17 11 9 60
-fno-reorder-functions 1 9 5 2 12 29
-fno-align-loops 22 8 16 24 4 74
-fno-align-jumps 10 26 10 12 1 59
-fno-align-labels 24 18 1 3 6 52
-fno-align-functions 19 20 9 10 5 63
-finline-functions 0 24 21 2 9 56
-frename-registers 22 23 11 24 18 98
-ffloat-store 15 4 7 20 11 57
-fno-prefetch-loop-arrays 23 23 26 24 26 122
-fmove-all-movables 22 25 18 14 22 101
-freduce-all-givs 0 12 3 3 4 22
-fno-inline 4 1 8 8 15 36
-ftracer 19 18 15 13 24 89
-fnew-ra 16 26 23 25 21 111
-funroll-loops 0 4 3 0 0 7
-funroll-all-loops 26 9 19 22 22 98
-mieee-fp 18 0 16 3 25 62
-malign-double 5 5 23 24 18 75
-mno-push-args 10 5 9 4 18 46
-maccumulate-outgoing-args 5 5 20 15 14 59
-mno-align-stringops 4 0 2 7 1 14
-minline-all-stringops 1 21 20 24 8 74
-mfpmath=387 1 4 1 5 1 12
-mfpmath=sse 24 8 0 0 4 36
-mfpmath=sse,387 0 0 3 1 0 4
-fomit-frame-pointer 0 0 16 0 9 25
-momit-leaf-frame-pointer 12 0 7 2 16 37
-mmmx 7 13 11 25 3 59
-msse 4 0 15 2 14 35
-msse2 3 8 4 26 5 46
-mno-ieee-fp 8 8 20 23 18 77
-mno-align-double 9 21 26 24 21 101
-mpush-args 17 4 5 14 25 65
-mno-accumulate-outgoing-args 22 23 5 17 25 92
-malign-stringops 0 26 15 10 9 60
-mno-inline-all-stringops 17 21 2 23 12 75
-fno-omit-frame-pointer 1 1 0 0 0 2
-mno-omit-leaf-frame-pointer 17 8 8 17 24 74
-mno-mmx 20 18 16 15 15 84
-mno-sse 16 5 6 9 20 56
-mno-sse2 21 4 20 21 14 80
-fno-math-errno 12 6 5 7 15 45
-funsafe-math-optimizations 21 5 1 11 10 48
-fno-trapping-math 12 9 22 16 16 75
-ffinite-math-only 22 20 18 12 11 83
-fno-signaling-nans 25 15 6 2 13 61
-finline-limit 15 3 17 14 17 66
value options:
-finline-limit: 0 0 600 0 0 , average = 600 across 1 populations
cc1: let op: SSE instruction set disabled, using 387 arithmetics
cc1: let op: SSE instruction set disabled, using 387 arithmetics
cc1: let op: SSE instruction set disabled, using 387 arithmetics
cc1: let op: SSE instruction set disabled, using 387 arithmetics
cc1: let op: SSE instruction set disabled, using 387 arithmetics
cc1: let op: SSE instruction set disabled, using 387 arithmetics
cc1: let op: SSE instruction set disabled, using 387 arithmetics
cc1: let op: SSE instruction set disabled, using 387 arithmetics
cc1: let op: SSE instruction set disabled, using 387 arithmetics
cc1: let op: SSE instruction set disabled, using 387 arithmetics
cc1: let op: SSE instruction set disabled, using 387 arithmetics
cc1: let op: SSE instruction set disabled, using 387 arithmetics
cc1: let op: SSE instruction set disabled, using 387 arithmetics
cc1: let op: SSE instruction set disabled, using 387 arithmetics
cc1: let op: SSE instruction set disabled, using 387 arithmetics
cc1: let op: SSE instruction set disabled, using 387 arithmetics
cc1: let op: SSE instruction set disabled, using 387 arithmetics
cc1: let op: SSE instruction set disabled, using 387 arithmetics
cc1: let op: SSE instruction set disabled, using 387 arithmetics
cc1: let op: SSE instruction set disabled, using 387 arithmetics
cc1: let op: SSE instruction set disabled, using 387 arithmetics
cc1: let op: SSE instruction set disabled, using 387 arithmetics
cc1: let op: SSE instruction set disabled, using 387 arithmetics
cc1: let op: SSE instruction set disabled, using 387 arithmetics
cc1: let op: SSE instruction set disabled, using 387 arithmetics
cc1: let op: SSE instruction set disabled, using 387 arithmetics
cc1: let op: SSE instruction set disabled, using 387 arithmetics
cc1: let op: SSE instruction set disabled, using 387 arithmetics
cc1: let op: SSE instruction set disabled, using 387 arithmetics
------------------------------------------------------------
iteration 27: average fitness = 3.10247, best fitnesses = 2.77602 2.75773 2.7616 2.75396 2.76823
best options for population 0
gcc -lrt -lm -std=gnu99 -march=pentium3 -mcpu=pentium4 -O2 -pipe -fomit-frame-pointer -fprefetch-loop-arrays -fmerge-constants -o TEMP -fno-merge-constants -fno-thread-jumps -fno-if-conversion2 -fno-delayed-branch -fno-crossjumping -fno-optimize-sibling-calls -fno-cse-follow-jumps -fno-gcse -fno-expensive-optimizations -fno-rerun-cse-after-loop -fno-rerun-loop-opt -fno-caller-saves -fno-peephole2 -fno-schedule-insns2 -fno-strict-aliasing -fno-sched-interblock -fno-align-loops -fno-align-jumps -fno-align-labels -frename-registers -ffloat-store -fno-prefetch-loop-arrays -fmove-all-movables -fnew-ra -funroll-all-loops -mfpmath=sse -momit-leaf-frame-pointer -mmmx -msse2 -mno-ieee-fp -mpush-args -mno-accumulate-outgoing-args -mno-inline-all-stringops -mno-omit-leaf-frame-pointer -mno-mmx -mno-sse -mno-sse2 -fno-math-errno -funsafe-math-optimizations -fno-trapping-math -ffinite-math-only -fno-signaling-nans -finline-limit=600 /usr/share/acovea/benchmarks/huffbench.c
best options for population 1
gcc -lrt -lm -std=gnu99 -march=pentium3 -mcpu=pentium4 -O2 -pipe -fomit-frame-pointer -fprefetch-loop-arrays -fmerge-constants -o TEMP -fno-merge-constants -fno-optimize-sibling-calls -fno-cse-skip-blocks -fno-gcse -fno-expensive-optimizations -fno-strength-reduce -fno-rerun-cse-after-loop -fno-rerun-loop-opt -fno-caller-saves -fno-peephole2 -fno-schedule-insns -fno-regmove -fno-sched-interblock -fno-sched-spec -fno-align-loops -fno-align-jumps -fno-align-functions -finline-functions -frename-registers -ffloat-store -fno-prefetch-loop-arrays -fmove-all-movables -fnew-ra -funroll-all-loops -malign-double -mno-push-args -maccumulate-outgoing-args -minline-all-stringops -mfpmath=387 -mmmx -msse2 -mno-ieee-fp -mno-align-double -mno-accumulate-outgoing-args -malign-stringops -mno-inline-all-stringops -mno-sse -mno-sse2 -funsafe-math-optimizations -fno-trapping-math /usr/share/acovea/benchmarks/huffbench.c
best options for population 2
gcc -lrt -lm -std=gnu99 -march=pentium3 -mcpu=pentium4 -O2 -pipe -fomit-frame-pointer -fprefetch-loop-arrays -fmerge-constants -o TEMP -fno-defer-pop -fno-thread-jumps -fno-cprop-registers -fno-if-conversion2 -fno-delayed-branch -fno-cse-skip-blocks -fno-gcse -fno-gcse-lm -fno-gcse-sm -fno-rerun-cse-after-loop -fno-peephole2 -fno-schedule-insns -fno-schedule-insns2 -fno-strict-aliasing -fno-reorder-blocks -fno-sched-interblock -fno-sched-spec -finline-functions -frename-registers -fno-prefetch-loop-arrays -fmove-all-movables -fno-inline -ftracer -fnew-ra -funroll-all-loops -mieee-fp -malign-double -maccumulate-outgoing-args -minline-all-stringops -momit-leaf-frame-pointer -msse -mno-align-double -mno-omit-leaf-frame-pointer -mno-sse -fno-trapping-math -finline-limit=600 /usr/share/acovea/benchmarks/huffbench.c
best options for population 3
gcc -lrt -lm -std=gnu99 -march=pentium3 -mcpu=pentium4 -O2 -pipe -fomit-frame-pointer -fprefetch-loop-arrays -fmerge-constants -o TEMP -fno-merge-constants -fno-defer-pop -fno-cprop-registers -fno-if-conversion2 -fno-optimize-sibling-calls -fno-cse-follow-jumps -fno-cse-skip-blocks -fno-gcse -fno-strength-reduce -fno-rerun-cse-after-loop -fno-rerun-loop-opt -fno-caller-saves -fno-peephole2 -fno-schedule-insns -fno-regmove -fno-sched-spec -fno-align-loops -frename-registers -ffloat-store -fno-prefetch-loop-arrays -fnew-ra -funroll-all-loops -mno-push-args -maccumulate-outgoing-args -mmmx -msse2 -mno-ieee-fp -mno-align-double -mpush-args -malign-stringops -mno-inline-all-stringops -mno-omit-leaf-frame-pointer -mno-mmx -ffinite-math-only /usr/share/acovea/benchmarks/huffbench.c
best options for population 4
gcc -lrt -lm -std=gnu99 -march=pentium3 -mcpu=pentium4 -O2 -pipe -fomit-frame-pointer -fprefetch-loop-arrays -fmerge-constants -o TEMP -fno-defer-pop -fno-thread-jumps -fno-delayed-branch -fno-gcse -fno-gcse-lm -fno-rerun-loop-opt -fno-peephole2 -fno-schedule-insns2 -fno-regmove -fno-strict-aliasing -fno-sched-interblock -fno-sched-spec -fno-reorder-functions -ffloat-store -fno-prefetch-loop-arrays -fmove-all-movables -fnew-ra -funroll-all-loops -malign-double -mno-push-args -maccumulate-outgoing-args -mno-align-stringops -minline-all-stringops -fomit-frame-pointer -mmmx -msse2 -mno-ieee-fp -mno-align-double -mpush-args -malign-stringops -mno-inline-all-stringops -mno-sse2 -fno-trapping-math -ffinite-math-only -finline-limit=600 /usr/share/acovea/benchmarks/huffbench.c
common options (found in all populations)
gcc -lrt -lm -std=gnu99 -march=pentium3 -mcpu=pentium4 -O2 -pipe -fomit-frame-pointer -fprefetch-loop-arrays -fmerge-constants -o TEMP -fno-gcse -fno-peephole2 -fno-prefetch-loop-arrays -fnew-ra -funroll-all-loops /usr/share/acovea/benchmarks/huffbench.c
rejected options (rejected by all populations)
gcc -lrt -lm -std=gnu99 -march=pentium3 -mcpu=pentium4 -O2 -pipe -fomit-frame-pointer -fprefetch-loop-arrays -fmerge-constants -o TEMP -fno-omit-frame-pointer -fno-guess-branch-probability -fno-if-conversion -fno-loop-optimize -fno-force-mem -freduce-all-givs /usr/share/acovea/benchmarks/huffbench.c
Option counts:
-fno-merge-constants 27 11 19 21 10 88
-fno-defer-pop 18 2 12 16 27 75
-fno-thread-jumps 13 19 21 10 11 74
-fno-omit-frame-pointer 1 0 15 1 8 25
-fno-guess-branch-probability 0 0 2 3 0 5
-fno-cprop-registers 10 22 22 16 9 79
-fno-if-conversion 0 0 0 0 0 0
-fno-if-conversion2 9 19 22 23 16 89
-fno-delayed-branch 7 1 21 4 12 45
-fno-loop-optimize 3 20 3 1 4 31
-fno-crossjumping 8 19 13 6 14 60
-fno-optimize-sibling-calls 4 27 7 23 14 75
-fno-cse-follow-jumps 23 6 18 7 25 79
-fno-cse-skip-blocks 8 27 10 19 19 83
-fno-gcse 13 27 27 24 22 113
-fno-gcse-lm 0 0 23 6 7 36
-fno-gcse-sm 6 0 12 18 12 48
-fno-expensive-optimizations 20 26 16 13 9 84
-fno-strength-reduce 5 6 17 24 0 52
-fno-rerun-cse-after-loop 15 21 6 22 13 77
-fno-rerun-loop-opt 8 18 17 21 2 66
-fno-caller-saves 25 10 12 15 19 81
-fno-force-mem 4 6 0 2 0 12
-fno-peephole2 27 15 19 25 26 112
-fno-schedule-insns 9 22 25 23 11 90
-fno-schedule-insns2 12 0 10 0 11 33
-fno-regmove 15 20 3 23 9 70
-fno-strict-aliasing 8 0 17 6 13 44
-fno-reorder-blocks 8 10 14 0 20 52
-fno-sched-interblock 27 15 24 13 27 106
-fno-sched-spec 2 22 18 12 10 64
-fno-reorder-functions 1 9 5 2 13 30
-fno-align-loops 23 9 16 25 4 77
-fno-align-jumps 11 27 10 12 1 61
-fno-align-labels 25 18 1 3 6 53
-fno-align-functions 19 21 9 10 5 64
-finline-functions 0 25 22 2 9 58
-frename-registers 23 24 12 25 18 102
-ffloat-store 16 5 7 21 12 61
-fno-prefetch-loop-arrays 24 24 27 25 27 127
-fmove-all-movables 23 26 19 14 23 105
-freduce-all-givs 0 12 3 3 4 22
-fno-inline 4 1 9 8 15 37
-ftracer 19 18 16 13 24 90
-fnew-ra 17 27 24 26 22 116
-funroll-loops 0 4 3 0 0 7
-funroll-all-loops 27 10 20 23 23 103
-mieee-fp 18 0 17 3 25 63
-malign-double 5 6 24 24 19 78
-mno-push-args 10 6 9 5 19 49
-maccumulate-outgoing-args 5 6 21 16 15 63
-mno-align-stringops 4 0 2 7 2 15
-minline-all-stringops 1 22 21 24 9 77
-mfpmath=387 1 5 1 5 1 13
-mfpmath=sse 25 8 0 0 4 37
-mfpmath=sse,387 0 0 3 1 0 4
-fomit-frame-pointer 0 0 16 0 10 26
-momit-leaf-frame-pointer 13 0 8 2 16 39
-mmmx 8 14 11 26 4 63
-msse 4 0 16 2 14 36
-msse2 4 9 4 27 6 50
-mno-ieee-fp 9 9 20 24 19 81
-mno-align-double 9 22 27 25 22 105
-mpush-args 18 4 5 15 26 68
-mno-accumulate-outgoing-args 23 24 5 17 25 94
-malign-stringops 0 27 15 11 10 63
-mno-inline-all-stringops 18 22 2 24 13 79
-fno-omit-frame-pointer 1 1 0 0 0 2
-mno-omit-leaf-frame-pointer 18 8 9 18 24 77
-mno-mmx 21 18 16 16 15 86
-mno-sse 17 6 7 9 20 59
-mno-sse2 22 5 20 21 15 83
-fno-math-errno 13 6 5 7 15 46
-funsafe-math-optimizations 22 6 1 11 10 50
-fno-trapping-math 13 10 23 16 17 79
-ffinite-math-only 23 20 18 13 12 86
-fno-signaling-nans 26 15 6 2 13 62
-finline-limit 16 3 18 14 18 69
value options:
-finline-limit: 600 0 600 0 600 , average = 600 across 3 populations
cc1: let op: SSE instruction set disabled, using 387 arithmetics
cc1: let op: SSE instruction set disabled, using 387 arithmetics
cc1: let op: SSE instruction set disabled, using 387 arithmetics
cc1: let op: SSE instruction set disabled, using 387 arithmetics
cc1: let op: SSE instruction set disabled, using 387 arithmetics
cc1: let op: SSE instruction set disabled, using 387 arithmetics
cc1: let op: SSE instruction set disabled, using 387 arithmetics
cc1: let op: SSE instruction set disabled, using 387 arithmetics
cc1: let op: SSE instruction set disabled, using 387 arithmetics
cc1: let op: SSE instruction set disabled, using 387 arithmetics
cc1: let op: SSE instruction set disabled, using 387 arithmetics
cc1: let op: SSE instruction set disabled, using 387 arithmetics
cc1: let op: SSE instruction set disabled, using 387 arithmetics
cc1: let op: SSE instruction set disabled, using 387 arithmetics
cc1: let op: SSE instruction set disabled, using 387 arithmetics
cc1: let op: SSE instruction set disabled, using 387 arithmetics
cc1: let op: SSE instruction set disabled, using 387 arithmetics
cc1: let op: SSE instruction set disabled, using 387 arithmetics
cc1: let op: SSE instruction set disabled, using 387 arithmetics
cc1: let op: SSE instruction set disabled, using 387 arithmetics
cc1: let op: SSE instruction set disabled, using 387 arithmetics
cc1: let op: SSE instruction set disabled, using 387 arithmetics
cc1: let op: SSE instruction set disabled, using 387 arithmetics
cc1: let op: SSE instruction set disabled, using 387 arithmetics
cc1: let op: SSE instruction set disabled, using 387 arithmetics
cc1: let op: SSE instruction set disabled, using 387 arithmetics
------------------------------------------------------------
iteration 28: average fitness = 3.10248, best fitnesses = 2.7779 2.75557 2.76284 2.75121 2.76684
best options for population 0
gcc -lrt -lm -std=gnu99 -march=pentium3 -mcpu=pentium4 -O2 -pipe -fomit-frame-pointer -fprefetch-loop-arrays -fmerge-constants -o TEMP -fno-merge-constants -fno-thread-jumps -fno-if-conversion2 -fno-delayed-branch -fno-crossjumping -fno-optimize-sibling-calls -fno-cse-follow-jumps -fno-gcse -fno-expensive-optimizations -fno-rerun-cse-after-loop -fno-rerun-loop-opt -fno-caller-saves -fno-peephole2 -fno-schedule-insns2 -fno-strict-aliasing -fno-sched-interblock -fno-align-loops -fno-align-jumps -fno-align-labels -frename-registers -ffloat-store -fno-prefetch-loop-arrays -fmove-all-movables -fnew-ra -funroll-all-loops -mfpmath=sse -momit-leaf-frame-pointer -mmmx -msse2 -mno-ieee-fp -mpush-args -mno-accumulate-outgoing-args -mno-inline-all-stringops -mno-omit-leaf-frame-pointer -mno-mmx -mno-sse -mno-sse2 -fno-math-errno -funsafe-math-optimizations -fno-trapping-math -ffinite-math-only -fno-signaling-nans -finline-limit=600 /usr/share/acovea/benchmarks/huffbench.c
best options for population 1
gcc -lrt -lm -std=gnu99 -march=pentium3 -mcpu=pentium4 -O2 -pipe -fomit-frame-pointer -fprefetch-loop-arrays -fmerge-constants -o TEMP -fno-merge-constants -fno-optimize-sibling-calls -fno-cse-skip-blocks -fno-gcse -fno-expensive-optimizations -fno-strength-reduce -fno-rerun-cse-after-loop -fno-rerun-loop-opt -fno-caller-saves -fno-peephole2 -fno-schedule-insns -fno-regmove -fno-sched-interblock -fno-sched-spec -fno-align-loops -fno-align-jumps -fno-align-functions -finline-functions -frename-registers -ffloat-store -fno-prefetch-loop-arrays -fmove-all-movables -fnew-ra -funroll-all-loops -malign-double -mno-push-args -maccumulate-outgoing-args -minline-all-stringops -mfpmath=387 -mmmx -msse2 -mno-ieee-fp -mno-align-double -mno-accumulate-outgoing-args -malign-stringops -mno-inline-all-stringops -mno-sse -mno-sse2 -funsafe-math-optimizations -fno-trapping-math /usr/share/acovea/benchmarks/huffbench.c
best options for population 2
gcc -lrt -lm -std=gnu99 -march=pentium3 -mcpu=pentium4 -O2 -pipe -fomit-frame-pointer -fprefetch-loop-arrays -fmerge-constants -o TEMP -fno-defer-pop -fno-thread-jumps -fno-cprop-registers -fno-if-conversion2 -fno-delayed-branch -fno-crossjumping -fno-cse-skip-blocks -fno-gcse -fno-gcse-lm -fno-gcse-sm -fno-rerun-loop-opt -fno-peephole2 -fno-schedule-insns2 -fno-strict-aliasing -fno-reorder-blocks -fno-sched-interblock -fno-sched-spec -finline-functions -frename-registers -fno-prefetch-loop-arrays -fmove-all-movables -fno-inline -ftracer -fnew-ra -funroll-all-loops -mieee-fp -malign-double -mno-push-args -maccumulate-outgoing-args -minline-all-stringops -momit-leaf-frame-pointer -msse -msse2 -mno-align-double -malign-stringops -mno-omit-leaf-frame-pointer -fno-trapping-math -ffinite-math-only -finline-limit=600 /usr/share/acovea/benchmarks/huffbench.c
best options for population 3
gcc -lrt -lm -std=gnu99 -march=pentium3 -mcpu=pentium4 -O2 -pipe -fomit-frame-pointer -fprefetch-loop-arrays -fmerge-constants -o TEMP -fno-merge-constants -fno-defer-pop -fno-thread-jumps -fno-cprop-registers -fno-if-conversion2 -fno-optimize-sibling-calls -fno-cse-skip-blocks -fno-gcse -fno-strength-reduce -fno-rerun-cse-after-loop -fno-rerun-loop-opt -fno-caller-saves -fno-peephole2 -fno-schedule-insns -fno-regmove -fno-sched-interblock -fno-sched-spec -fno-align-loops -fno-align-jumps -fno-align-functions -frename-registers -ffloat-store -fno-prefetch-loop-arrays -fnew-ra -funroll-all-loops -maccumulate-outgoing-args -minline-all-stringops -mmmx -msse2 -mno-ieee-fp -mno-align-double -mno-inline-all-stringops -mno-omit-leaf-frame-pointer -mno-mmx -fno-trapping-math /usr/share/acovea/benchmarks/huffbench.c
best options for population 4
gcc -lrt -lm -std=gnu99 -march=pentium3 -mcpu=pentium4 -O2 -pipe -fomit-frame-pointer -fprefetch-loop-arrays -fmerge-constants -o TEMP -fno-merge-constants -fno-defer-pop -fno-thread-jumps -fno-omit-frame-pointer -fno-if-conversion2 -fno-cse-follow-jumps -fno-cse-skip-blocks -fno-gcse -fno-rerun-cse-after-loop -fno-caller-saves -fno-peephole2 -fno-schedule-insns2 -fno-reorder-blocks -fno-sched-interblock -ffloat-store -fno-prefetch-loop-arrays -fmove-all-movables -fno-inline -ftracer -fnew-ra -funroll-all-loops -mieee-fp -malign-double -mno-push-args -maccumulate-outgoing-args -fomit-frame-pointer -msse -mno-align-double -mpush-args -mno-accumulate-outgoing-args -malign-stringops -mno-omit-leaf-frame-pointer -mno-mmx -mno-sse -mno-sse2 -fno-trapping-math -fno-signaling-nans /usr/share/acovea/benchmarks/huffbench.c
common options (found in all populations)
gcc -lrt -lm -std=gnu99 -march=pentium3 -mcpu=pentium4 -O2 -pipe -fomit-frame-pointer -fprefetch-loop-arrays -fmerge-constants -o TEMP -fno-gcse -fno-peephole2 -fno-sched-interblock -fno-prefetch-loop-arrays -fnew-ra -funroll-all-loops -fno-trapping-math /usr/share/acovea/benchmarks/huffbench.c
rejected options (rejected by all populations)
gcc -lrt -lm -std=gnu99 -march=pentium3 -mcpu=pentium4 -O2 -pipe -fomit-frame-pointer -fprefetch-loop-arrays -fmerge-constants -o TEMP -fno-guess-branch-probability -fno-if-conversion -fno-loop-optimize -fno-force-mem -fno-reorder-functions -freduce-all-givs -mno-align-stringops /usr/share/acovea/benchmarks/huffbench.c
Option counts:
-fno-merge-constants 28 12 19 22 11 92
-fno-defer-pop 18 2 13 17 28 78
-fno-thread-jumps 14 19 22 11 12 78
-fno-omit-frame-pointer 1 0 15 1 9 26
-fno-guess-branch-probability 0 0 2 3 0 5
-fno-cprop-registers 10 22 23 17 9 81
-fno-if-conversion 0 0 0 0 0 0
-fno-if-conversion2 10 19 23 24 17 93
-fno-delayed-branch 8 1 22 4 12 47
-fno-loop-optimize 3 20 3 1 4 31
-fno-crossjumping 9 19 14 6 14 62
-fno-optimize-sibling-calls 5 28 7 24 14 78
-fno-cse-follow-jumps 24 6 18 7 26 81
-fno-cse-skip-blocks 8 28 11 20 20 87
-fno-gcse 14 28 28 25 23 118
-fno-gcse-lm 0 0 24 6 7 37
-fno-gcse-sm 6 0 13 18 12 49
-fno-expensive-optimizations 21 27 16 13 9 86
-fno-strength-reduce 5 7 17 25 0 54
-fno-rerun-cse-after-loop 16 22 6 23 14 81
-fno-rerun-loop-opt 9 19 18 22 2 70
-fno-caller-saves 26 11 12 16 20 85
-fno-force-mem 4 6 0 2 0 12
-fno-peephole2 28 16 20 26 27 117
-fno-schedule-insns 9 23 25 24 11 92
-fno-schedule-insns2 13 0 11 0 12 36
-fno-regmove 15 21 3 24 9 72
-fno-strict-aliasing 9 0 18 6 13 46
-fno-reorder-blocks 8 10 15 0 21 54
-fno-sched-interblock 28 16 25 14 28 111
-fno-sched-spec 2 23 19 13 10 67
-fno-reorder-functions 1 9 5 2 13 30
-fno-align-loops 24 10 16 26 4 80
-fno-align-jumps 12 28 10 13 1 64
-fno-align-labels 26 18 1 3 6 54
-fno-align-functions 19 22 9 11 5 66
-finline-functions 0 26 23 2 9 60
-frename-registers 24 25 13 26 18 106
-ffloat-store 17 6 7 22 13 65
-fno-prefetch-loop-arrays 25 25 28 26 28 132
-fmove-all-movables 24 27 20 14 24 109
-freduce-all-givs 0 12 3 3 4 22
-fno-inline 4 1 10 8 16 39
-ftracer 19 18 17 13 25 92
-fnew-ra 18 28 25 27 23 121
-funroll-loops 0 4 3 0 0 7
-funroll-all-loops 28 11 21 24 24 108
-mieee-fp 18 0 18 3 26 65
-malign-double 5 7 25 24 20 81
-mno-push-args 10 7 10 5 20 52
-maccumulate-outgoing-args 5 7 22 17 16 67
-mno-align-stringops 4 0 2 7 2 15
-minline-all-stringops 1 23 22 25 9 80
-mfpmath=387 1 6 1 5 1 14
-mfpmath=sse 26 8 0 0 4 38
-mfpmath=sse,387 0 0 3 1 0 4
-fomit-frame-pointer 0 0 16 0 11 27
-momit-leaf-frame-pointer 14 0 9 2 16 41
-mmmx 9 15 11 27 4 66
-msse 4 0 17 2 15 38
-msse2 5 10 5 28 6 54
-mno-ieee-fp 10 10 20 25 19 84
-mno-align-double 9 23 28 26 23 109
-mpush-args 19 4 5 15 27 70
-mno-accumulate-outgoing-args 24 25 5 17 26 97
-malign-stringops 0 28 16 11 11 66
-mno-inline-all-stringops 19 23 2 25 13 82
-fno-omit-frame-pointer 1 1 0 0 0 2
-mno-omit-leaf-frame-pointer 19 8 10 19 25 81
-mno-mmx 22 18 16 17 16 89
-mno-sse 18 7 7 9 21 62
-mno-sse2 23 6 20 21 16 86
-fno-math-errno 14 6 5 7 15 47
-funsafe-math-optimizations 23 7 1 11 10 52
-fno-trapping-math 14 11 24 17 18 84
-ffinite-math-only 24 20 19 13 12 88
-fno-signaling-nans 27 15 6 2 14 64
-finline-limit 17 3 19 14 18 71
value options:
-finline-limit: 600 0 600 0 0 , average = 600 across 2 populations
cc1: let op: SSE instruction set disabled, using 387 arithmetics
cc1: let op: SSE instruction set disabled, using 387 arithmetics
cc1: let op: SSE instruction set disabled, using 387 arithmetics
cc1: let op: SSE instruction set disabled, using 387 arithmetics
cc1: let op: SSE instruction set disabled, using 387 arithmetics
cc1: let op: SSE instruction set disabled, using 387 arithmetics
cc1: let op: SSE instruction set disabled, using 387 arithmetics
cc1: let op: SSE instruction set disabled, using 387 arithmetics
cc1: let op: SSE instruction set disabled, using 387 arithmetics
cc1: let op: SSE instruction set disabled, using 387 arithmetics
cc1: let op: SSE instruction set disabled, using 387 arithmetics
cc1: let op: SSE instruction set disabled, using 387 arithmetics
cc1: let op: SSE instruction set disabled, using 387 arithmetics
cc1: let op: SSE instruction set disabled, using 387 arithmetics
cc1: let op: SSE instruction set disabled, using 387 arithmetics
cc1: let op: SSE instruction set disabled, using 387 arithmetics
cc1: let op: SSE instruction set disabled, using 387 arithmetics
cc1: let op: SSE instruction set disabled, using 387 arithmetics
cc1: let op: SSE instruction set disabled, using 387 arithmetics
cc1: let op: SSE instruction set disabled, using 387 arithmetics
cc1: let op: SSE instruction set disabled, using 387 arithmetics
------------------------------------------------------------
iteration 29: average fitness = 3.11371, best fitnesses = 2.77334 2.76221 2.76497 2.75455 2.75575 |
Last edited by pjv on Wed Aug 25, 2004 10:37 pm; edited 1 time in total |
|
Back to top |
|
|
pjv Guru
Joined: 02 Jul 2003 Posts: 353 Location: Belgium
|
Posted: Wed Aug 25, 2004 10:35 pm Post subject: |
|
|
Code: |
best options for population 0
gcc -lrt -lm -std=gnu99 -march=pentium3 -mcpu=pentium4 -O2 -pipe -fomit-frame-pointer -fprefetch-loop-arrays -fmerge-constants -o TEMP -fno-merge-constants -fno-thread-jumps -fno-if-conversion2 -fno-delayed-branch -fno-crossjumping -fno-optimize-sibling-calls -fno-cse-follow-jumps -fno-gcse -fno-expensive-optimizations -fno-rerun-cse-after-loop -fno-rerun-loop-opt -fno-caller-saves -fno-peephole2 -fno-schedule-insns2 -fno-strict-aliasing -fno-sched-interblock -fno-align-loops -fno-align-jumps -fno-align-labels -frename-registers -ffloat-store -fno-prefetch-loop-arrays -fmove-all-movables -fnew-ra -funroll-all-loops -mfpmath=sse -momit-leaf-frame-pointer -mmmx -msse2 -mno-ieee-fp -mpush-args -mno-accumulate-outgoing-args -mno-inline-all-stringops -mno-omit-leaf-frame-pointer -mno-mmx -mno-sse -mno-sse2 -fno-math-errno -funsafe-math-optimizations -fno-trapping-math -ffinite-math-only -fno-signaling-nans -finline-limit=600 /usr/share/acovea/benchmarks/huffbench.c
best options for population 1
gcc -lrt -lm -std=gnu99 -march=pentium3 -mcpu=pentium4 -O2 -pipe -fomit-frame-pointer -fprefetch-loop-arrays -fmerge-constants -o TEMP -fno-merge-constants -fno-optimize-sibling-calls -fno-cse-skip-blocks -fno-gcse -fno-expensive-optimizations -fno-strength-reduce -fno-rerun-cse-after-loop -fno-rerun-loop-opt -fno-caller-saves -fno-peephole2 -fno-schedule-insns -fno-regmove -fno-sched-interblock -fno-sched-spec -fno-align-loops -fno-align-jumps -fno-align-functions -finline-functions -frename-registers -ffloat-store -fno-prefetch-loop-arrays -fmove-all-movables -fnew-ra -funroll-all-loops -malign-double -mno-push-args -maccumulate-outgoing-args -minline-all-stringops -mfpmath=387 -mmmx -msse2 -mno-ieee-fp -mno-align-double -mno-accumulate-outgoing-args -malign-stringops -mno-inline-all-stringops -mno-sse -mno-sse2 -funsafe-math-optimizations -fno-trapping-math /usr/share/acovea/benchmarks/huffbench.c
best options for population 2
gcc -lrt -lm -std=gnu99 -march=pentium3 -mcpu=pentium4 -O2 -pipe -fomit-frame-pointer -fprefetch-loop-arrays -fmerge-constants -o TEMP -fno-defer-pop -fno-thread-jumps -fno-cprop-registers -fno-if-conversion2 -fno-delayed-branch -fno-crossjumping -fno-cse-skip-blocks -fno-gcse -fno-gcse-lm -fno-gcse-sm -fno-rerun-loop-opt -fno-peephole2 -fno-schedule-insns2 -fno-strict-aliasing -fno-reorder-blocks -fno-sched-interblock -fno-sched-spec -finline-functions -frename-registers -fno-prefetch-loop-arrays -fmove-all-movables -fno-inline -ftracer -fnew-ra -funroll-all-loops -mieee-fp -malign-double -mno-push-args -maccumulate-outgoing-args -minline-all-stringops -momit-leaf-frame-pointer -msse -msse2 -mno-align-double -malign-stringops -mno-omit-leaf-frame-pointer -fno-trapping-math -ffinite-math-only -finline-limit=600 /usr/share/acovea/benchmarks/huffbench.c
best options for population 3
gcc -lrt -lm -std=gnu99 -march=pentium3 -mcpu=pentium4 -O2 -pipe -fomit-frame-pointer -fprefetch-loop-arrays -fmerge-constants -o TEMP -fno-merge-constants -fno-defer-pop -fno-if-conversion2 -fno-optimize-sibling-calls -fno-cse-skip-blocks -fno-gcse -fno-gcse-lm -fno-strength-reduce -fno-rerun-cse-after-loop -fno-rerun-loop-opt -fno-peephole2 -fno-schedule-insns -fno-regmove -fno-strict-aliasing -fno-sched-interblock -fno-align-loops -fno-align-jumps -ffloat-store -fno-prefetch-loop-arrays -fmove-all-movables -fno-inline -ftracer -fnew-ra -funroll-all-loops -mieee-fp -mno-push-args -maccumulate-outgoing-args -mno-align-stringops -minline-all-stringops -mfpmath=387 -msse2 -mno-ieee-fp -mno-align-double -mpush-args -malign-stringops -mno-inline-all-stringops -mno-mmx -ffinite-math-only -finline-limit=600 /usr/share/acovea/benchmarks/huffbench.c
best options for population 4
gcc -lrt -lm -std=gnu99 -march=pentium3 -mcpu=pentium4 -O2 -pipe -fomit-frame-pointer -fprefetch-loop-arrays -fmerge-constants -o TEMP -fno-merge-constants -fno-defer-pop -fno-thread-jumps -fno-cprop-registers -fno-if-conversion2 -fno-optimize-sibling-calls -fno-cse-skip-blocks -fno-gcse -fno-strength-reduce -fno-rerun-cse-after-loop -fno-rerun-loop-opt -fno-caller-saves -fno-peephole2 -fno-schedule-insns -fno-regmove -fno-sched-interblock -fno-sched-spec -fno-align-loops -fno-align-jumps -fno-align-functions -frename-registers -ffloat-store -fno-prefetch-loop-arrays -fnew-ra -funroll-all-loops -maccumulate-outgoing-args -minline-all-stringops -mmmx -msse2 -mno-ieee-fp -mno-align-double -mno-inline-all-stringops -mno-omit-leaf-frame-pointer -mno-mmx -fno-trapping-math /usr/share/acovea/benchmarks/huffbench.c
common options (found in all populations)
gcc -lrt -lm -std=gnu99 -march=pentium3 -mcpu=pentium4 -O2 -pipe -fomit-frame-pointer -fprefetch-loop-arrays -fmerge-constants -o TEMP -fno-gcse -fno-rerun-loop-opt -fno-peephole2 -fno-sched-interblock -fno-prefetch-loop-arrays -fnew-ra -funroll-all-loops -msse2 /usr/share/acovea/benchmarks/huffbench.c
rejected options (rejected by all populations)
gcc -lrt -lm -std=gnu99 -march=pentium3 -mcpu=pentium4 -O2 -pipe -fomit-frame-pointer -fprefetch-loop-arrays -fmerge-constants -o TEMP -fno-omit-frame-pointer -fno-guess-branch-probability -fno-if-conversion -fno-loop-optimize -fno-force-mem -fno-reorder-functions -freduce-all-givs /usr/share/acovea/benchmarks/huffbench.c
Option counts:
-fno-merge-constants 29 13 19 23 12 96
-fno-defer-pop 18 2 14 18 29 81
-fno-thread-jumps 15 19 23 11 13 81
-fno-omit-frame-pointer 1 0 15 1 9 26
-fno-guess-branch-probability 0 0 2 3 0 5
-fno-cprop-registers 10 22 24 17 10 83
-fno-if-conversion 0 0 0 0 0 0
-fno-if-conversion2 11 19 24 25 18 97
-fno-delayed-branch 9 1 23 4 12 49
-fno-loop-optimize 3 20 3 1 4 31
-fno-crossjumping 10 19 15 6 14 64
-fno-optimize-sibling-calls 6 29 7 25 15 82
-fno-cse-follow-jumps 25 6 18 7 26 82
-fno-cse-skip-blocks 8 29 12 21 21 91
-fno-gcse 15 29 29 26 24 123
-fno-gcse-lm 0 0 25 7 7 39
-fno-gcse-sm 6 0 14 18 12 50
-fno-expensive-optimizations 22 28 16 13 9 88
-fno-strength-reduce 5 8 17 26 1 57
-fno-rerun-cse-after-loop 17 23 6 24 15 85
-fno-rerun-loop-opt 10 20 19 23 3 75
-fno-caller-saves 27 12 12 16 21 88
-fno-force-mem 4 6 0 2 0 12
-fno-peephole2 29 17 21 27 28 122
-fno-schedule-insns 9 24 25 25 12 95
-fno-schedule-insns2 14 0 12 0 12 38
-fno-regmove 15 22 3 25 10 75
-fno-strict-aliasing 10 0 19 7 13 49
-fno-reorder-blocks 8 10 16 0 21 55
-fno-sched-interblock 29 17 26 15 29 116
-fno-sched-spec 2 24 20 13 11 70
-fno-reorder-functions 1 9 5 2 13 30
-fno-align-loops 25 11 16 27 5 84
-fno-align-jumps 13 29 10 14 2 68
-fno-align-labels 27 18 1 3 6 55
-fno-align-functions 19 23 9 11 6 68
-finline-functions 0 27 24 2 9 62
-frename-registers 25 26 14 26 19 110
-ffloat-store 18 7 7 23 14 69
-fno-prefetch-loop-arrays 26 26 29 27 29 137
-fmove-all-movables 25 28 21 15 24 113
-freduce-all-givs 0 12 3 3 4 22
-fno-inline 4 1 11 9 16 41
-ftracer 19 18 18 14 25 94
-fnew-ra 19 29 26 28 24 126
-funroll-loops 0 4 3 0 0 7
-funroll-all-loops 29 12 22 25 25 113
-mieee-fp 18 0 19 4 26 67
-malign-double 5 8 26 24 20 83
-mno-push-args 10 8 11 6 20 55
-maccumulate-outgoing-args 5 8 23 18 17 71
-mno-align-stringops 4 0 2 8 2 16
-minline-all-stringops 1 24 23 26 10 84
-mfpmath=387 1 7 1 6 1 16
-mfpmath=sse 27 8 0 0 4 39
-mfpmath=sse,387 0 0 3 1 0 4
-fomit-frame-pointer 0 0 16 0 11 27
-momit-leaf-frame-pointer 15 0 10 2 16 43
-mmmx 10 16 11 27 5 69
-msse 4 0 18 2 15 39
-msse2 6 11 6 29 7 59
-mno-ieee-fp 11 11 20 26 20 88
-mno-align-double 9 24 29 27 24 113
-mpush-args 20 4 5 16 27 72
-mno-accumulate-outgoing-args 25 26 5 17 26 99
-malign-stringops 0 29 17 12 11 69
-mno-inline-all-stringops 20 24 2 26 14 86
-fno-omit-frame-pointer 1 1 0 0 0 2
-mno-omit-leaf-frame-pointer 20 8 11 19 26 84
-mno-mmx 23 18 16 18 17 92
-mno-sse 19 8 7 9 21 64
-mno-sse2 24 7 20 21 16 88
-fno-math-errno 15 6 5 7 15 48
-funsafe-math-optimizations 24 8 1 11 10 54
-fno-trapping-math 15 12 25 17 19 88
-ffinite-math-only 25 20 20 14 12 91
-fno-signaling-nans 28 15 6 2 14 65
-finline-limit 18 3 20 15 18 74
value options:
-finline-limit: 600 0 600 600 0 , average = 600 across 3 populations
cc1: let op: SSE instruction set disabled, using 387 arithmetics
cc1: let op: SSE instruction set disabled, using 387 arithmetics
cc1: let op: SSE instruction set disabled, using 387 arithmetics
cc1: let op: SSE instruction set disabled, using 387 arithmetics
cc1: let op: SSE instruction set disabled, using 387 arithmetics
cc1: let op: SSE instruction set disabled, using 387 arithmetics
cc1: let op: SSE instruction set disabled, using 387 arithmetics
cc1: let op: SSE instruction set disabled, using 387 arithmetics
cc1: let op: SSE instruction set disabled, using 387 arithmetics
cc1: let op: SSE instruction set disabled, using 387 arithmetics
cc1: let op: SSE instruction set disabled, using 387 arithmetics
cc1: let op: SSE instruction set disabled, using 387 arithmetics
cc1: let op: SSE instruction set disabled, using 387 arithmetics
cc1: let op: SSE instruction set disabled, using 387 arithmetics
cc1: let op: SSE instruction set disabled, using 387 arithmetics
cc1: let op: SSE instruction set disabled, using 387 arithmetics
cc1: let op: SSE instruction set disabled, using 387 arithmetics
cc1: let op: SSE instruction set disabled, using 387 arithmetics
cc1: let op: SSE instruction set disabled, using 387 arithmetics
cc1: let op: SSE instruction set disabled, using 387 arithmetics
cc1: let op: SSE instruction set disabled, using 387 arithmetics
cc1: let op: SSE instruction set disabled, using 387 arithmetics
cc1: let op: SSE instruction set disabled, using 387 arithmetics
cc1: let op: SSE instruction set disabled, using 387 arithmetics
cc1: let op: SSE instruction set disabled, using 387 arithmetics
------------------------------------------------------------
iteration 30: average fitness = 3.1094, best fitnesses = 2.75282 2.75088 2.75468 2.75421 2.7497
best options for population 0
gcc -lrt -lm -std=gnu99 -march=pentium3 -mcpu=pentium4 -O2 -pipe -fomit-frame-pointer -fprefetch-loop-arrays -fmerge-constants -o TEMP -fno-merge-constants -fno-thread-jumps -fno-cprop-registers -fno-delayed-branch -fno-crossjumping -fno-cse-follow-jumps -fno-cse-skip-blocks -fno-gcse -fno-expensive-optimizations -fno-rerun-loop-opt -fno-caller-saves -fno-peephole2 -fno-schedule-insns -fno-schedule-insns2 -fno-regmove -fno-sched-spec -fno-align-loops -fno-align-functions -ffloat-store -fno-prefetch-loop-arrays -fmove-all-movables -fnew-ra -funroll-all-loops -malign-double -minline-all-stringops -mfpmath=sse -momit-leaf-frame-pointer -mno-ieee-fp -mno-accumulate-outgoing-args -mno-omit-leaf-frame-pointer -mno-sse -mno-sse2 -ffinite-math-only -fno-signaling-nans /usr/share/acovea/benchmarks/huffbench.c
best options for population 1
gcc -lrt -lm -std=gnu99 -march=pentium3 -mcpu=pentium4 -O2 -pipe -fomit-frame-pointer -fprefetch-loop-arrays -fmerge-constants -o TEMP -fno-merge-constants -fno-defer-pop -fno-cprop-registers -fno-if-conversion2 -fno-optimize-sibling-calls -fno-cse-follow-jumps -fno-cse-skip-blocks -fno-gcse -fno-strength-reduce -fno-rerun-cse-after-loop -fno-rerun-loop-opt -fno-caller-saves -fno-peephole2 -fno-schedule-insns -fno-regmove -fno-sched-spec -fno-align-loops -frename-registers -ffloat-store -fno-prefetch-loop-arrays -fnew-ra -funroll-all-loops -mno-push-args -maccumulate-outgoing-args -mmmx -msse2 -mno-ieee-fp -mno-align-double -mpush-args -malign-stringops -mno-inline-all-stringops -mno-omit-leaf-frame-pointer -mno-mmx -ffinite-math-only /usr/share/acovea/benchmarks/huffbench.c
best options for population 2
gcc -lrt -lm -std=gnu99 -march=pentium3 -mcpu=pentium4 -O2 -pipe -fomit-frame-pointer -fprefetch-loop-arrays -fmerge-constants -o TEMP -fno-merge-constants -fno-defer-pop -fno-thread-jumps -fno-cprop-registers -fno-if-conversion2 -fno-optimize-sibling-calls -fno-cse-skip-blocks -fno-gcse -fno-strength-reduce -fno-rerun-cse-after-loop -fno-rerun-loop-opt -fno-peephole2 -fno-schedule-insns -fno-regmove -fno-strict-aliasing -fno-sched-spec -fno-align-loops -fno-align-jumps -frename-registers -ffloat-store -fno-prefetch-loop-arrays -fnew-ra -funroll-all-loops -maccumulate-outgoing-args -mmmx -msse2 -mno-ieee-fp -mno-align-double -mno-inline-all-stringops -mno-omit-leaf-frame-pointer -mno-mmx /usr/share/acovea/benchmarks/huffbench.c
best options for population 3
gcc -lrt -lm -std=gnu99 -march=pentium3 -mcpu=pentium4 -O2 -pipe -fomit-frame-pointer -fprefetch-loop-arrays -fmerge-constants -o TEMP -fno-merge-constants -fno-defer-pop -fno-cprop-registers -fno-if-conversion2 -fno-optimize-sibling-calls -fno-cse-skip-blocks -fno-gcse -fno-strength-reduce -fno-rerun-cse-after-loop -fno-rerun-loop-opt -fno-peephole2 -fno-schedule-insns -fno-regmove -fno-strict-aliasing -fno-sched-spec -fno-align-loops -fno-align-jumps -frename-registers -ffloat-store -fno-prefetch-loop-arrays -fnew-ra -funroll-all-loops -mno-push-args -maccumulate-outgoing-args -mmmx -msse2 -mno-ieee-fp -mno-align-double -mno-inline-all-stringops -mno-omit-leaf-frame-pointer -mno-mmx /usr/share/acovea/benchmarks/huffbench.c
best options for population 4
gcc -lrt -lm -std=gnu99 -march=pentium3 -mcpu=pentium4 -O2 -pipe -fomit-frame-pointer -fprefetch-loop-arrays -fmerge-constants -o TEMP -fno-defer-pop -fno-thread-jumps -fno-delayed-branch -fno-cse-skip-blocks -fno-gcse -fno-rerun-cse-after-loop -fno-peephole2 -fno-schedule-insns2 -fno-regmove -fno-reorder-blocks -fno-sched-interblock -fno-sched-spec -fno-reorder-functions -ffloat-store -fno-prefetch-loop-arrays -fmove-all-movables -fnew-ra -funroll-all-loops -malign-double -mno-push-args -maccumulate-outgoing-args -mno-align-stringops -minline-all-stringops -fomit-frame-pointer -mmmx -mno-ieee-fp -mno-align-double -mpush-args -mno-accumulate-outgoing-args -malign-stringops -mno-inline-all-stringops -mno-mmx -mno-sse -fno-trapping-math -ffinite-math-only -fno-signaling-nans -finline-limit=600 /usr/share/acovea/benchmarks/huffbench.c
common options (found in all populations)
gcc -lrt -lm -std=gnu99 -march=pentium3 -mcpu=pentium4 -O2 -pipe -fomit-frame-pointer -fprefetch-loop-arrays -fmerge-constants -o TEMP -fno-cse-skip-blocks -fno-gcse -fno-peephole2 -fno-regmove -fno-sched-spec -ffloat-store -fno-prefetch-loop-arrays -fnew-ra -funroll-all-loops -mno-ieee-fp /usr/share/acovea/benchmarks/huffbench.c
rejected options (rejected by all populations)
gcc -lrt -lm -std=gnu99 -march=pentium3 -mcpu=pentium4 -O2 -pipe -fomit-frame-pointer -fprefetch-loop-arrays -fmerge-constants -o TEMP -fno-omit-frame-pointer -fno-guess-branch-probability -fno-if-conversion -fno-loop-optimize -fno-gcse-lm -fno-gcse-sm -fno-force-mem -fno-align-labels -finline-functions -freduce-all-givs -fno-inline -ftracer -mieee-fp -msse -fno-math-errno -funsafe-math-optimizations /usr/share/acovea/benchmarks/huffbench.c
Option counts:
-fno-merge-constants 30 14 20 24 12 100
-fno-defer-pop 18 3 15 19 30 85
-fno-thread-jumps 16 19 24 11 14 84
-fno-omit-frame-pointer 1 0 15 1 9 26
-fno-guess-branch-probability 0 0 2 3 0 5
-fno-cprop-registers 11 23 25 18 10 87
-fno-if-conversion 0 0 0 0 0 0
-fno-if-conversion2 11 20 25 26 18 100
-fno-delayed-branch 10 1 23 4 13 51
-fno-loop-optimize 3 20 3 1 4 31
-fno-crossjumping 11 19 15 6 14 65
-fno-optimize-sibling-calls 6 30 8 26 15 85
-fno-cse-follow-jumps 26 7 18 7 26 84
-fno-cse-skip-blocks 9 30 13 22 22 96
-fno-gcse 16 30 30 27 25 128
-fno-gcse-lm 0 0 25 7 7 39
-fno-gcse-sm 6 0 14 18 12 50
-fno-expensive-optimizations 23 28 16 13 9 89
-fno-strength-reduce 5 9 18 27 1 60
-fno-rerun-cse-after-loop 17 24 7 25 16 89
-fno-rerun-loop-opt 11 21 20 24 3 79
-fno-caller-saves 28 13 12 16 21 90
-fno-force-mem 4 6 0 2 0 12
-fno-peephole2 30 18 22 28 29 127
-fno-schedule-insns 10 25 26 26 12 99
-fno-schedule-insns2 15 0 12 0 13 40
-fno-regmove 16 23 4 26 11 80
-fno-strict-aliasing 10 0 20 8 13 51
-fno-reorder-blocks 8 10 16 0 22 56
-fno-sched-interblock 29 17 26 15 30 117
-fno-sched-spec 3 25 21 14 12 75
-fno-reorder-functions 1 9 5 2 14 31
-fno-align-loops 26 12 17 28 5 88
-fno-align-jumps 13 29 11 15 2 70
-fno-align-labels 27 18 1 3 6 55
-fno-align-functions 20 23 9 11 6 69
-finline-functions 0 27 24 2 9 62
-frename-registers 25 27 15 27 19 113
-ffloat-store 19 8 8 24 15 74
-fno-prefetch-loop-arrays 27 27 30 28 30 142
-fmove-all-movables 26 28 21 15 25 115
-freduce-all-givs 0 12 3 3 4 22
-fno-inline 4 1 11 9 16 41
-ftracer 19 18 18 14 25 94
-fnew-ra 20 30 27 29 25 131
-funroll-loops 0 4 3 0 0 7
-funroll-all-loops 30 13 23 26 26 118
-mieee-fp 18 0 19 4 26 67
-malign-double 6 8 26 24 21 85
-mno-push-args 10 9 11 7 21 58
-maccumulate-outgoing-args 5 9 24 19 18 75
-mno-align-stringops 4 0 2 8 3 17
-minline-all-stringops 2 24 23 26 11 86
-mfpmath=387 1 7 1 6 1 16
-mfpmath=sse 28 8 0 0 4 40
-mfpmath=sse,387 0 0 3 1 0 4
-fomit-frame-pointer 0 0 16 0 12 28
-momit-leaf-frame-pointer 16 0 10 2 16 44
-mmmx 10 17 12 28 6 73
-msse 4 0 18 2 15 39
-msse2 6 12 7 30 7 62
-mno-ieee-fp 12 12 21 27 21 93
-mno-align-double 9 25 30 28 25 117
-mpush-args 20 5 5 16 28 74
-mno-accumulate-outgoing-args 26 26 5 17 27 101
-malign-stringops 0 30 17 12 12 71
-mno-inline-all-stringops 20 25 3 27 15 90
-fno-omit-frame-pointer 1 1 0 0 0 2
-mno-omit-leaf-frame-pointer 21 9 12 20 26 88
-mno-mmx 23 19 17 19 18 96
-mno-sse 20 8 7 9 22 66
-mno-sse2 25 7 20 21 16 89
-fno-math-errno 15 6 5 7 15 48
-funsafe-math-optimizations 24 8 1 11 10 54
-fno-trapping-math 15 12 25 17 20 89
-ffinite-math-only 26 21 20 14 13 94
-fno-signaling-nans 29 15 6 2 15 67
-finline-limit 18 3 20 15 19 75
value options:
-finline-limit: 0 0 0 0 600 , average = 600 across 1 populations
run complete time: 2004 Aug 25 04:47:46
optimistic options:
-fno-gcse (1.765)
-fno-peephole2 (1.735)
-fno-sched-interblock (1.433)
-frename-registers (1.313)
-fno-prefetch-loop-arrays (2.188)
-fmove-all-movables (1.373)
-fnew-ra (1.856)
-funroll-all-loops (1.464)
-mno-align-double (1.433)
pessimistic options:
-fno-omit-frame-pointer (-1.312)
-fno-guess-branch-probability (-1.946)
-fno-if-conversion (-2.096)
-fno-loop-optimize (-1.161)
-fno-force-mem (-1.734)
-fno-reorder-functions (-1.161)
-freduce-all-givs (-1.433)
-funroll-loops (-1.885)
-mno-align-stringops (-1.584)
-mfpmath=387 (-1.614)
-mfpmath=sse,387 (-1.976)
-fomit-frame-pointer (-1.252)
-fno-omit-frame-pointer (-2.036)
|
This gives me the direct advised CFLAGS from acovea. However I went to the last 5 generations, and only selected the common ones, so not the ones that were added by mutation in the very last generation or so:
Code: |
common options (found in all populations)
gcc -lrt -lm -std=gnu99 -march=pentium3 -mcpu=pentium4 -O2 -pipe -fomit-frame-pointer -fprefetch-loop-arrays -fmerge-constants -o TEMP -fno-gcse -fno-peephole2 -fno-prefetch-loop-arrays -fnew-ra -funroll-all-loops /usr/share/acovea/benchmarks/huffbench.c
rejected options (rejected by all populations)
gcc -lrt -lm -std=gnu99 -march=pentium3 -mcpu=pentium4 -O2 -pipe -fomit-frame-pointer -fprefetch-loop-arrays -fmerge-constants -o TEMP -fno-guess-branch-probability -fno-if-conversion -fno-loop-optimize -fno-force-mem -freduce-all-givs /usr/share/acovea/benchmarks/huffbench.c
|
Next I've read the GCC manual's flag descriptions and made my own list based on all this knowledge. I also just found out that a higher -mcpu isn't doing anything. Sse2 is used in 3.4 in the pentium-m flag, so I'm adding this to my "pentium-3" too.
Code: |
-fno-gcse: This is also what I feel when I read the description. I guess there are a lot of computed gotos.
-fno-peephole2: Haven't a clue what this means. Are these the processor-specific flags? In that case it would be strange.
-fno-sched-interblock: Wasn't constant over the 5 latest generations. Also small distance. Might be though, but didn't expect it. Not adding this.
-frename-registers: Wasn't constant over the 5 latest generations. Also small distance. It's kind of a dangerous flag from -O3. Not adding this.
-fno-prefetch-loop-arrays: Strange, except that maybe it now avoids creating large code. Os is also effective, you know. Large distance so very decisively.
-fmove-all-movables: Wasn't constant over the 5 latest generations. Also small distance. Might be. Not adding this.
-fnew-ra: Logical. This would only make compile time longer. Another optimization algorithm ;-) Hey, graph coloring is nice.
-funroll-all-loops: Very very strange. I personally refuse to add this. I would add -funroll-loops though. I don't trust this benchmark on uncertain loops. Also small distance.
-mno-align-double: Wasn't constant over the 5 latest generations. I suppose this also is a space consideration. Also small distance. Wasn't added anyway.
===
Most of this confirms O2:
-fno-omit-frame-pointer: Wasn't constant over the 5 latest generations. Also small distance. You really seem to need omit-frame-pointer, as I expected. Certainly very good advice. However some weird things happen, because it appears more than once, both optimistically as pessimistically.
-fno-guess-branch-probability: You really seem to need guess-branch-probability, good advice.
-fno-if-conversion: Very large distance. Good advice.
-fno-loop-optimize: Small distance. Good advice.
-fno-force-mem: Good advice.
-fno-reorder-functions: Wasn't constant over the 5 latest generations. Also small distance. Good advice.
-freduce-all-givs: Also small distance. Very very strange I would say, because they don't seem to know any code that ran more slowly with reduce-all-givs turned on :-S. Not removing this.
-funroll-loops: Wasn't constant over the 5 latest generations. As I told above, I'm adding this instead of funroll-all-loops. Ofcourse the relationship between these two params is OR, so if one is used, the other isn't totally, which might explain the bad score.
-mno-align-stringops: Wasn't constant over the 5 latest generations. Seems good advice. Adding -malign-stringops.
-mfpmath=387 -mfpmath=sse,387: Wasn't constant over the 5 latest generations. Logical: 387 is slow and sse,387 is not well supported, so that leaves us with sse. Adding -mfpmath=sse.
Also adding -msse2.
|
Keep in mind that what I'm doing here actually is wrong. Acovea will find the global minimum of a function in lots of dimensions. But changing one flag may take you very far away from that minimum. The correlation between the flags also matters. Also, IF Acovea finds the absolute minimum.
This is why I tested some things manually:
Code: |
=====manual tests: huffbench.c
So, what I had:
CFLAGS="-march=pentium3 -mcpu=pentium4 -O2 -pipe -fomit-frame-pointer -fprefetch-loop-arrays -fmerge-constants"
gcc -lrt -lm -std=gnu99 -march=pentium3 -mcpu=pentium4 -O2 -pipe -fomit-frame-pointer -fprefetch-loop-arrays -fmerge-constants -o TEMP /usr/share/acovea/benchmarks/huffbench.c
huffbench (Std. C) run time: 3.865910
What it advised:
gcc -lrt -lm -std=gnu99 -march=pentium3 -mcpu=pentium4 -O2 -pipe -fomit-frame-pointer -fprefetch-loop-arrays -fmerge-constants -o TEMP -fno-cse-skip-blocks -fno-gcse -fno-peephole2 -fno-regmove -fno-sched-spec -ffloat-store -fno-prefetch-loop-arrays -fnew-ra -funroll-all-loops -mno-ieee-fp /usr/share/acovea/benchmarks/huffbench.c
huffbench (Std. C) run time: 3.163981
My purged advise:
CFLAGS="-march=pentium3 (-mcpu=pentium4) -O2 -pipe -fomit-frame-pointer -fmerge-constants -fno-gcse -fno-peephole2 -fnew-ra -funroll-loops -malign-stringops -mfpmath=sse -msse2"
gcc -lrt -lm -std=gnu99 -march=pentium3 -mcpu=pentium4 -O2 -pipe -fomit-frame-pointer -fmerge-constants -fno-gcse -fno-peephole2 -fnew-ra -funroll-loops -malign-stringops -mfpmath=sse -msse2 -o TEMP /usr/share/acovea/benchmarks/huffbench.c
huffbench (Std. C) run time: 3.430710
Other tryouts:
gcc -lrt -lm -std=gnu99 -march=pentium3 -mcpu=pentium4 -O2 -pipe -fomit-frame-pointer -fmerge-constants -fno-gcse -fno-peephole2 -fnew-ra -funroll-loops -malign-stringops -mfpmath=sse -o TEMP /usr/share/acovea/benchmarks/huffbench.c
huffbench (Std. C) run time: 3.343637
gcc -lrt -lm -std=gnu99 -march=pentium3 -mcpu=pentium4 -O2 -pipe -fomit-frame-pointer -fmerge-constants -fno-gcse -fno-peephole2 -fnew-ra -funroll-loops -malign-stringops -o TEMP /usr/share/acovea/benchmarks/huffbench.c
huffbench (Std. C) run time: 3.296118
gcc -lrt -lm -std=gnu99 -march=pentium3 -mcpu=pentium4 -O2 -pipe -fomit-frame-pointer -fmerge-constants -fno-gcse -fno-peephole2 -fnew-ra -funroll-loops -o TEMP /usr/share/acovea/benchmarks/huffbench.c
huffbench (Std. C) run time: 3.147447 !!!!!!!
gcc -lrt -lm -std=gnu99 -march=pentium3 -mcpu=pentium4 -O2 -pipe -fomit-frame-pointer -fmerge-constants -fno-gcse -fno-peephole2 -fnew-ra -funroll-all-loops -o TEMP /usr/share/acovea/benchmarks/huffbench.c
huffbench (Std. C) run time: 3.637689
gcc -lrt -lm -std=gnu99 -march=pentium3 -mcpu=pentium4 -O2 -pipe -fomit-frame-pointer -fmerge-constants -fno-gcse -fno-peephole2 -fnew-ra -o TEMP /usr/share/acovea/benchmarks/huffbench.c
huffbench (Std. C) run time: 3.550826
gcc -lrt -lm -std=gnu99 -march=pentium3 -mcpu=pentium4 -O2 -pipe -fomit-frame-pointer -fmerge-constants -fno-gcse -fno-peephole2 -fnew-ra -funroll-loops -fprefetch-loop-arrays -o TEMP /usr/share/acovea/benchmarks/huffbench.c
voyager benchmarks # ./TEMP huffbench (Std. C) run time: 3.625788
=====manual tests: evobench.c
gcc -lrt -lm -std=gnu99 -march=pentium3 -mcpu=pentium4 -O2 -pipe -fomit-frame-pointer -fprefetch-loop-arrays -fmerge-constants -o TEMP /usr/share/acovea/benchmarks/evobench.c
evobench (Std. C) run time: 7.816683
gcc -lrt -lm -std=gnu99 -march=pentium3 -mcpu=pentium4 -O2 -pipe -fomit-frame-pointer -fprefetch-loop-arrays -fmerge-constants -o TEMP -fno-cse-skip-blocks -fno-gcse -fno-peephole2 -fno-regmove -fno-sched-spec -ffloat-store -fno-prefetch-loop-arrays -fnew-ra -funroll-all-loops -mno-ieee-fp /usr/share/acovea/benchmarks/evobench.c
evobench (Std. C) run time: 8.040342
gcc -lrt -lm -std=gnu99 -march=pentium3 -mcpu=pentium4 -O2 -pipe -fomit-frame-pointer -fmerge-constants -fno-gcse -fno-peephole2 -fnew-ra -funroll-loops -malign-stringops -mfpmath=sse -msse2 -o TEMP /usr/share/acovea/benchmarks/evobench.c
evobench (Std. C) run time: 10.507957
gcc -lrt -lm -std=gnu99 -march=pentium3 -mcpu=pentium4 -O2 -pipe -fomit-frame-pointer -fmerge-constants -fno-gcse -fno-peephole2 -fnew-ra -funroll-loops -malign-stringops -mfpmath=sse -o TEMP /usr/share/acovea/benchmarks/evobench.c
evobench (Std. C) run time: 10.800310
gcc -lrt -lm -std=gnu99 -march=pentium3 -mcpu=pentium4 -O2 -pipe -fomit-frame-pointer -fmerge-constants -fno-gcse -fno-peephole2 -fnew-ra -funroll-loops -malign-stringops -o TEMP /usr/share/acovea/benchmarks/evobench.c
evobench (Std. C) run time: 10.664235
gcc -lrt -lm -std=gnu99 -march=pentium3 -mcpu=pentium4 -O2 -pipe -fomit-frame-pointer -fmerge-constants -fno-gcse -fno-peephole2 -fnew-ra -funroll-all-loops -o TEMP /usr/share/acovea/benchmarks/evobench.c
evobench (Std. C) run time: 8.471214
gcc -lrt -lm -std=gnu99 -march=pentium3 -mcpu=pentium4 -O2 -pipe -fomit-frame-pointer -fmerge-constants -fno-gcse -fno-peephole2 -fnew-ra -o TEMP /usr/share/acovea/benchmarks/evobench.c
evobench (Std. C) run time: 10.620707
gcc -lrt -lm -std=gnu99 -march=pentium3 -mcpu=pentium4 -O2 -pipe -fomit-frame-pointer -fmerge-constants -fno-gcse -fno-peephole2 -fnew-ra -funroll-loops -fprefetch-loop-arrays -o TEMP /usr/share/acovea/benchmarks/evobench.c
evobench (Std. C) run time: 9.908448
=====manual tests: mat1bench.c
gcc -lrt -lm -std=gnu99 -march=pentium3 -mcpu=pentium4 -O2 -pipe -fomit-frame-pointer -fprefetch-loop-arrays -fmerge-constants -o TEMP /usr/share/acovea/benchmarks/mat1bench.c
mat1bench (Std. C) run time: 4.189033
gcc -lrt -lm -std=gnu99 -march=pentium3 -mcpu=pentium4 -O2 -pipe -fomit-frame-pointer -fprefetch-loop-arrays -fmerge-constants -o TEMP -fno-cse-skip-blocks -fno-gcse -fno-peephole2 -fno-regmove -fno-sched-spec -ffloat-store -fno-prefetch-loop-arrays -fnew-ra -funroll-all-loops -mno-ieee-fp /usr/share/acovea/benchmarks/mat1bench.c
mat1bench (Std. C) run time: 4.140465
gcc -lrt -lm -std=gnu99 -march=pentium3 -mcpu=pentium4 -O2 -pipe -fomit-frame-pointer -fmerge-constants -fno-gcse -fno-peephole2 -fnew-ra -funroll-loops -malign-stringops -mfpmath=sse -msse2 -o TEMP /usr/share/acovea/benchmarks/mat1bench.c
mat1bench (Std. C) run time: 4.764191
gcc -lrt -lm -std=gnu99 -march=pentium3 -mcpu=pentium4 -O2 -pipe -fomit-frame-pointer -fmerge-constants -fno-gcse -fno-peephole2 -fnew-ra -funroll-loops -malign-stringops -mfpmath=sse -o TEMP /usr/share/acovea/benchmarks/mat1bench.c
mat1bench (Std. C) run time: 3.704792
gcc -lrt -lm -std=gnu99 -march=pentium3 -mcpu=pentium4 -O2 -pipe -fomit-frame-pointer -fmerge-constants -fno-gcse -fno-peephole2 -fnew-ra -funroll-loops -malign-stringops -o TEMP /usr/share/acovea/benchmarks/mat1bench.c
mat1bench (Std. C) run time: 4.183927
gcc -lrt -lm -std=gnu99 -march=pentium3 -mcpu=pentium4 -O2 -pipe -fomit-frame-pointer -fmerge-constants -fno-gcse -fno-peephole2 -fnew-ra -funroll-all-loops -o TEMP /usr/share/acovea/benchmarks/mat1bench.c
mat1bench (Std. C) run time: 4.092431
gcc -lrt -lm -std=gnu99 -march=pentium3 -mcpu=pentium4 -O2 -pipe -fomit-frame-pointer -fmerge-constants -fno-gcse -fno-peephole2 -fnew-ra -o TEMP /usr/share/acovea/benchmarks/mat1bench.c
mat1bench (Std. C) run time: 5.136133
gcc -lrt -lm -std=gnu99 -march=pentium3 -mcpu=pentium4 -O2 -pipe -fomit-frame-pointer -fmerge-constants -fno-gcse -fno-peephole2 -fnew-ra -funroll-loops -fprefetch-loop-arrays -o TEMP /usr/share/acovea/benchmarks/mat1bench.c
mat1bench (Std. C) run time: 3.677457
|
With the huffbench I would settle for the 3.430710. But the rest of the benchmarks aren't giving good results at all. I guess we would need to test these manual tests on some regular big apps, like glibc or something. Anybody know how we can do this.
Also, it would be nice if _kev_ or someone could do these exact same things all over again. So do the huff test with the same config. So we could see if there are many differences, or if it is all very stable.
Also, I had those strange high-tone chirping sounds again, each time in the three seconds I was running ./TEMP in the last evo and mat1 benchmarks. Sounded kinda like a cd drive starting up, but then with a high pitch. Strange... some flags that trigger some sounds?? You know what I am talking about: thé whistling sounds of this notebook.
To conclude: I'm not sure if these things are really reliable, but hey what the heck, I'm learning a lot about these flags.
Oh yeah, and no split up of the wiki!!
Ciao,
pjv
[edit]Thread was put in 2 posts because of the (almost flooding) length. Really hope, you don't mind.[/edit] |
|
Back to top |
|
|
feffi Apprentice
Joined: 29 Mar 2003 Posts: 216 Location: Sol->Earth->Germany->Giessen
|
Posted: Wed Aug 25, 2004 10:45 pm Post subject: |
|
|
@pjv: THIS is AWESOME !!! very good work. Shall i quote your post??
on the GCC there are some stress testing and benchmarking programms. Maybe I'll give them a try. I definetly will keep an eye on acovea, but I currently have no time to do all the research that's neccessary.
I currently sometime get this chirping sounds, when I scroll my console or ls a big dir, so I belive it either the HD or what I mostly believe in the ATI again. Strange, strange...
Quote: |
To conclude: I'm not sure if these things are really reliable, but hey what the heck, I'm learning a lot about these flags.
|
Thant's exactly what "Gentoo" means... _________________ have fun
feffi
/(bb|[^b]{2})/ that is the Question!
Gentoo-Wiki: Acer Travelmate 803 LCi manual |
|
Back to top |
|
|
matttions Guru
Joined: 19 Jul 2004 Posts: 338
|
Posted: Wed Sep 01, 2004 12:46 pm Post subject: |
|
|
hi to everybody ...
today I've got an external monitor working with the ati-drivers [3.11.1] under Xorg.
It was a LG 795FT Flatron Plus.
I try to explaine how I get it and, of course, if your computer or your monitor explose when you follow the informations display below , well, don't blame me!!!
Well .. so the first thing you have to do is to
and configure your xorg.conf as display in the graphics-section.
At this point just Code: | emerge ati-drivers-extra |
The qt must to be installed to have the graphical tool "fireglcontrol" working.
After that go to and, after yuo become root just launch it
If everything is ok you can see a graphical menu where you can choose four different output :
1) Disable Second Monitor
2) Clone Mode
3) Extended Desktop Horizontal
4) Extended Desktop Vertical
Be aware ... for me just works the 1) of course, this is quite simple and the 2) --> this is the most feature that I was interested on so that's quite all-right ..
You can also see a tab for the Tv-out ... this didn't work form me (my tv layout is PAL-B so, maybe someone of you are more lucky than me, with this option..)
For the tv-out just go on in investigation ... _________________ gentoo ... l'unico pinguino da corsa |
|
Back to top |
|
|
kwaakie n00b
Joined: 07 Jun 2003 Posts: 18 Location: Malden, Netherlands
|
Posted: Fri Sep 03, 2004 12:02 pm Post subject: irda |
|
|
I added a little more info in the irda part of the guide since I wasn't able to get irda to work without the module parameters.
edit: Also added a line about 1400x1050 framebuffer. _________________ Come out, come out, wherever you are ! |
|
Back to top |
|
|
thepi Guru
Joined: 06 Jan 2004 Posts: 352 Location: Tokyo, Japan
|
Posted: Sun Sep 05, 2004 10:34 am Post subject: |
|
|
okay, sorry guys, i've been away for a while, so i didn't come to adding the udev section. in fact, i'll have to start over since i already forgot how to do it
anyway i just updated to kernel 2.6.8-gentoo-r3, and it runs well. still, i'd recommend to any of you to _not_ use vesafb-tng yet, as it seems not to support 1400x1050 mode. for me, the display was messed up (could see only blinking cursor at the top of the screen, no text whatsoever - blind typing, yay.
moreover, supermount is taken out , and i can't dbus etc. to work (doesn't compile, presumably "masked" although listed in packacke.unmask).
if anyone could help on those subjects, you're welcome to do so
regarding the "tm80x live cd", i put up some precompiled packages. let's see where we get.
pi~ _________________ Violence is the last resort of the incompetent --Isaac Asimov
Gentoo on Acer TM800 | Help! | Vote! | Get FF! |
|
Back to top |
|
|
matttions Guru
Joined: 19 Jul 2004 Posts: 338
|
Posted: Sun Sep 05, 2004 11:21 am Post subject: |
|
|
thepi wrote: |
anyway i just updated to kernel 2.6.8-gentoo-r3, and it runs well. still, i'd recommend to any of you to _not_ use vesafb-tng yet, as it seems not to support 1400x1050 mode. for me, the display was messed up (could see only blinking cursor at the top of the screen, no text whatsoever - blind typing, yay.
pi~ |
Yes, also for me ...
just use the vesa 2.0 for now --> the framebuffer works with this driver...
There isn't the picture in the background because they removed the bootsplash from the 2.6.8-dev-source .. [I don't know why...]
So just have the small fonts .. but it works nice ..
... _________________ gentoo ... l'unico pinguino da corsa |
|
Back to top |
|
|
feffi Apprentice
Joined: 29 Mar 2003 Posts: 216 Location: Sol->Earth->Germany->Giessen
|
Posted: Mon Sep 06, 2004 9:53 am Post subject: |
|
|
thepi wrote: | okay, sorry guys, i've been away for a while |
You're always welcome!!
Quote: | so i didn't come to adding the udev section. in fact, i'll have to start over since i already forgot how to do it |
Yeah me too But if you could do it, it would be really nice...
Quote: | still, i'd recommend to any of you to _not_ use vesafb-tng yet, as it seems not to support 1400x1050 mode. |
naaa, don't think so, i've it up and running with 1400x1050 and fbsplash. Just follow the guide on fbsplash in gentoo-wiki
Quote: |
moreover, supermount is taken out |
Yeah, I'm working on that...
my prrrrrecious... But, we should also diskuss about architectures, do we want Pentium-4 or Pentium-3 or both? _________________ have fun
feffi
/(bb|[^b]{2})/ that is the Question!
Gentoo-Wiki: Acer Travelmate 803 LCi manual
Last edited by feffi on Sun Sep 12, 2004 4:03 pm; edited 1 time in total |
|
Back to top |
|
|
feffi Apprentice
Joined: 29 Mar 2003 Posts: 216 Location: Sol->Earth->Germany->Giessen
|
Posted: Mon Sep 06, 2004 4:19 pm Post subject: |
|
|
@matttions: thx, but I would like to ask you to put it into the Wiki-Howto
@kwaakie: thx too
@both: welcome to the community _________________ have fun
feffi
/(bb|[^b]{2})/ that is the Question!
Gentoo-Wiki: Acer Travelmate 803 LCi manual |
|
Back to top |
|
|
den_RDC Apprentice
Joined: 25 Aug 2002 Posts: 166 Location: beercountry, Belgium;)
|
Posted: Mon Sep 06, 2004 4:54 pm Post subject: |
|
|
I am planning on buying an acer 8003LMib. I suppose it is a lot like the 803, except it has a radeon 9700, a cpu with 2m cache, a dvdwriter, bluetooth and some other stuff. Anybody any experience with this one? I am betting it will be much like the 803, but i'd like to be a little sure before i blow a lot off hard earned cache on this one
I don't wanna end up running *winxp* (the horror). _________________ Fan of the "Survivor Warriors of the Evil Empire of Bloody Destruction and Bloody Darkness" |
|
Back to top |
|
|
matttions Guru
Joined: 19 Jul 2004 Posts: 338
|
Posted: Mon Sep 06, 2004 7:43 pm Post subject: |
|
|
_kev_ wrote: | @matttions: thx, but I would like to ask you to put it into the Wiki-Howto |
Yes ... I'm going to do it .... _kev_ wrote: |
@both: welcome to the community |
Thanks for the welcome ..
Just a thing ...
I have the bluetooth running fine on my gentoo-box, but there isn't any section in the wiki ...
If you make this section maybe I will fill it !
---EDIT---
Added Beamer & external device _________________ gentoo ... l'unico pinguino da corsa |
|
Back to top |
|
|
corey_s Apprentice
Joined: 29 Jun 2003 Posts: 264
|
Posted: Thu Sep 09, 2004 6:28 am Post subject: xorg-x11-6.8 |
|
|
xorg-x11-6.8...
So anyone got it running yet? I'm emerge -f'ing it right now, but thought I'd hold up on the actual build until I see if braver souls have any troubles with it...
Actualy, I'm in the midst of emerging KDE-3.3, so I'm gonna wait until that's through first. I'll report my results soon as I'm able. |
|
Back to top |
|
|
julienp n00b
Joined: 25 Nov 2003 Posts: 10
|
Posted: Thu Sep 09, 2004 8:01 am Post subject: |
|
|
den_RDC wrote: | I am planning on buying an acer 8003LMib. I suppose it is a lot like the 803, except it has a radeon 9700, a cpu with 2m cache, a dvdwriter, bluetooth and some other stuff. Anybody any experience with this one? I am betting it will be much like the 803, but i'd like to be a little sure before i blow a lot off hard earned cache on this one
I don't wanna end up running *winxp* (the horror). |
I bought an 8003LMib and it works quite well with linux. However I don't use the ATI binary driver since it doesn't seem to work with software-suspend, so I don't have hardware 3D acceleration.
Another difference is, that it has the ipw2200 wlan chip instead of the ipw2100 one, which isn't a problem, since there is an ipw2200 ebuild in portage.
One thing I haven't got to work is standby/suspend to ram (software-suspend, aka hibernate works well), though I don't think it works reliably on the 800 series.
Thinks I haven't tested: firewire, pcmcia, ir, smartcard, memory card reader. (most of these are probably the same hardware as in the 800 series)
DVD and bluetooth work.
[Edit]: I don't know if the acerhk module works, but all(?) of the special keys (including lauch keys) work (by setting the correct keycodes in ~/.Xmodmap) |
|
Back to top |
|
|
corey_s Apprentice
Joined: 29 Jun 2003 Posts: 264
|
Posted: Thu Sep 09, 2004 6:26 pm Post subject: xorg-6.8 |
|
|
xorg-6.8 didn't give me any troubles.
It's up and running with no hassle - however I'm yet to actualy try the Damage and drop-shadows and all that... I'll try those next and see what happens.
corey@scanner ~ $ X -version
X Window System Version 6.8.0
Release Date: 8 September 2004
X Protocol Version 11, Revision 0, Release 6.8
Build Operating System: Linux 2.6.9-rc1 i686 [ELF]
Current Operating System: Linux scanner 2.6.9-rc1 #1 Tue Sep 7 12:54:45 Local time zone must be set--see zic manua i686
Build Date: 09 September 2004
Before reporting problems, check http://wiki.X.Org
to make sure that you have the latest version.
Module Loader present |
|
Back to top |
|
|
feffi Apprentice
Joined: 29 Mar 2003 Posts: 216 Location: Sol->Earth->Germany->Giessen
|
Posted: Sun Sep 12, 2004 4:14 pm Post subject: |
|
|
matttions wrote: |
I have the bluetooth running fine on my gentoo-box, but there isn't any section in the wiki ...
If you make this section maybe I will fill it !
|
Done, go ahead, it's in section 12
matttions wrote: | Added Beamer & external device |
Thanks alot, I layouted it just a bit to fit the overall layout. BTW. It would be nice to have the xorg.conf part for the four different layouts in the chapter, just for people like me, who edit the file directly, without any GUI.
den_RDC wrote: |
I am planning on buying an acer 8003LMib. I suppose it is a lot like the 803, except it has a radeon 9700, a cpu with 2m cache, a dvdwriter, bluetooth and some other stuff. |
yeah, keep on, I'm getting jealous
den_RDC wrote: |
Anybody any experience with this one? I am betting it will be much like the 803, but i'd like to be a little sure before i blow a lot off hard earned cache on this one Smile. |
It's much like the 803 LCi, you're right, but as I'm viewing the service manual, there is no hardware in this little thingy, thats incompatible with the 803LCI. You will definetly have some troubles here and there, but the overall coverage should work for you.
den_RDC wrote: |
I don't wanna end up running *winxp* (the horror). |
That would really be a horror _________________ have fun
feffi
/(bb|[^b]{2})/ that is the Question!
Gentoo-Wiki: Acer Travelmate 803 LCi manual |
|
Back to top |
|
|
feffi Apprentice
Joined: 29 Mar 2003 Posts: 216 Location: Sol->Earth->Germany->Giessen
|
Posted: Sun Sep 12, 2004 4:21 pm Post subject: Re: xorg-x11-6.8 |
|
|
corey_s wrote: | xorg-x11-6.8...
So anyone got it running yet? |
up and running, no problems so far. even the CVS version with transparency works like charm... _________________ have fun
feffi
/(bb|[^b]{2})/ that is the Question!
Gentoo-Wiki: Acer Travelmate 803 LCi manual |
|
Back to top |
|
|
matttions Guru
Joined: 19 Jul 2004 Posts: 338
|
Posted: Mon Sep 13, 2004 1:34 pm Post subject: Re: xorg-x11-6.8 |
|
|
feffi wrote: | corey_s wrote: | xorg-x11-6.8...
So anyone got it running yet? |
up and running, no problems so far. even the CVS version with transparency works like charm... |
Are you using the open driver?
I try with the closed on by ATI, but emerge complains tha the xorg-x11-6.8.0 is too new for this driver, so it's no possible to use ati-driver and any kind of opengl application [glxgears crash all the system]
now I'm going to downgrade to xorg-x11-6.7.... again _________________ gentoo ... l'unico pinguino da corsa |
|
Back to top |
|
|
|T5| n00b
Joined: 13 Dec 2003 Posts: 33
|
Posted: Mon Sep 13, 2004 9:10 pm Post subject: |
|
|
hi everybody,
unfortunately i had to have my acer TM801LCiB repaired and the mainborad got changed. Now i got a different revision of the ATI Mobility 9000 graphics chip
Code: | VGA compatible controller: ATI Technologies Inc Radeon R250 Lf [Radeon Mobility 9000 M9] (rev 02) |
now the ati-binarydriver is not loadable any more (module not found). any ideas?
regards |
|
Back to top |
|
|
|
|
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
|
|