oh boy. Maybe this is a really bad gimmick site, or I've had one too many drinks tonight but I have to ask myself, WTF is this trying to be? Since when did coke-sniffing, gun concealing gansters think "hey, OpenSource is all that, y'know whad I mean? This shit be HOT". Well no.
> Maybe this is a really bad gimmick site
Ya think?
rofl. I think some nerds there need to get out more. ;-)
No, just looking at that site, I'd say they were clutching at straws when comming up with a name that wasn't already taken.
The site is suppose to be meant as a big joke, making fun of all of those "Hardcore" linux users. We encourage people to poke fun of others and start stupid flamewars. Its not meant to be serious, so go to the site and start some stupid flame war and have fun :)
Don't play >>4 backwards, it's really dangerous.
SOME HARDCORE PERSON HAS BEEN LOOKING AT THEIR REFERRER LOGS HAVEN'T THEY OH MY
>>6
Flame THERE! Not HERE!
Haha, that reminds me of good ole' http://funroll-loops.org/
> Watching shit scroll by for hours makes me a Linux expert overnight!
The joys of compiling.
lol. God I love ricer Gentoo users.
Especially being able to mark their bug as RESOLVED::INVALID and telling them to reformat and reinstall without the rediculous CFLAGS.
http://bugs.gentoo.org/show_bug.cgi?id=74072
http://bugs.gentoo.org/show_bug.cgi?id=104437
>>9
Great links, I can't stop laughing at
> CFLAGS="-g0 -DTT_CONFIG_OPTION_BYTECODE_INTERPRETER -pipe -O3 -march=pentium4 -fweb -funswitch-loops -funroll-all-loops -funit-at-a-time -fsched2-use-traces -fsched2-use-superblocks -fsched-stalled-insns=12 -frename-registers -fprefetch-loop-arrays -fpeel-loops -fomit-frame-pointer -fmerge-all-constants -finline-limit=32768 -finline-functions -ffunction-sections -ffast-math -fdata-sections -fbranch-target-load-optimize2"
and also,
> CXXFLAGS="-g0 -DTT_CONFIG_OPTION_BYTECODE_INTERPRETER -pipe -O3 -march=pentium4 -fweb -funswitch-loops -funroll-all-loops -funit-at-a-time -fsched2-use-traces -fsched2-use-superblocks -fsched-stalled-insns=12 -frename-registers -fprefetch-loop-arrays -fpeel-loops -fomit-frame-pointer -fmerge-all-constants -finline-limit=32768 -finline-functions -ffunction-sections -ffast-math -fdata-sections -fbranch-target-load-optimize2 -fvisibility-inlines-hidden -fabi-version=0 -fpermissive -fno-enforce-eh-specs"
Are these really legitimate bugzilla entries? Do some Gentoo users really have this many CFLAGS ?
> Do some Gentoo users really have this many CFLAGS ?
Yes. They are called "Gentoo Ricers" for a very good reason.
Yes we do... and no, the developers (and I speak as one of them) do NOT condone this sort of behaviour.
We point and laugh instead. ;-)
What's worse though, is when people think you're their personal slave.
e.g. http://bugs.gentoo.org/show_bug.cgi?id=100588
That user caused a real stir on the gentoo-core mailing list.
This thread needs to be about which gcc options sound the funniest.
I think -fpeephole2 is the best!
May 21 11:27:26 --- Cardoe has changed the topic to: Gentoo Development || bugs.g.o upgraded, have at it | Bugs slow at 1400UTC? Backups are running, so chill | Do NOT use epkgmove, you'll catch syphilis || Please check over dev.gentoo.org/distfile-mirroring/failure.xml for failed ebuild fetches, and contact ferringb about issues || nemo.amd64.dev is back online! || Trustees 2005 election results at http://article.gmane.org/gmane.linux.gentoo.nfp/252 || new gcc-4 CFLAG... -fomfg-fast-speed
^^ #gentoo-dev, irc.freenode.net. That little tidbit at the end stayed for some weeks.
Ricers indeed. -O2 or -O3 should be enough for everyone.
I have to write MPI code for a cluster (this thing specifically: http://www.vpac.org/content/systems_and_support/facility/linux_cluster.php), so obviously I'm looking for every little speedup possible. Compiler switches are a simple way to squeak out some further gains, so I've done a fair bit of testing on them.
At least for GCC 3.2.*, switches other than -O2 (and maybe -march) vanish into the noise. They're just not worth considering unless you're doing something unusual. Of course, if it's speed you need, you don't use GCC to begin with. You'd use Intel or IBM.
I could really use a -fomfg-fast-speed though.
But GCC is open sores! That's why I run Gentoo instead of Debian.
Yep, and on MIPS architectures, if you want speed, you use MIPSPro rather than gcc. On DEC/Compaq Alpha, ccc. On Intel 80x86, icc.
Not sure about others.
However, I use gcc because, well, MIPSPro doesn't run on Linux (it's for IRIX only AFAIK), and most Linux apps are designed with gcc in mind.
> and most Linux apps are designed with gcc in mind
I can understand using GCC instead of other compilers, since I favour it myself (outside of clusters). However, apps depending on a particular compiler bother me. The main reason for C's creation was portability.
C code that is meant for widespread use should be tested on at least two different compilers. While not having different architectures lying around is excusable, getting a copy of icc is easy. :{
>>16
i've gotten up to a 75% speed up on some programs from -mfpmath=sse,387 (using gcc 3)
>>13
i'd have to go with -fdelete-null-pointer-checks
-fsched-spec-load-dangerous
-fstrength-reduce
>>21
SSE is only useful for programs in specific domains.
I'd be interested in knowing which programs you got a 75% speedup.
-fGA!
>>22,25
hahaha, win
I've traveled in seedy enough circles to be surprised a few times a geek and thug mixing together. Dealers who have DnD books on their "workspace" left over from the other night's game, things like that. Years ago, in high school, there was a guy who sat on the other side of class who was know to be involved with all sorts of shit, killing time reading Animerica. This wasn't too long after they had their first interview with Takeuchi Naoko, so that must've been... 1995 or 96, I think.
>>15
GAPING_SECURITY_HOLE is the best use flag
http://gentoo-portage.com/net-analyzer/netcat/USE
I've never seen much of an appreciable (more than say 20%) increase from any compiler flags. Gprof and long hours however work wonders. Also compiling with a branch prodiction profile has also worked wonders for me (Like I'll ever be able to get my if's to around 50% on each branch when most of them are edge cases and error conditions!)
On x86, you'll get piss all performance increase. Let's face it, today's latest x86 CPU's are still little more than overclocked 386's with additional instruction sets thrown in. (such as SSE{,2}, MMX, 3DNow, etc)
It's still very much the same register-starved architecture it was back in 1990. This isn't the case with other architectures.
On MIPS, take, say, an Indy, chuck on Debian/MIPS, and have a fiddle. When you're done, throw on Gentoo/MIPS, and compare.
Debian/MIPS is compiled for MIPS1-class CPUs, with a little MIPS2-class code thrown in.
Gentoo/MIPS is compiled for either MIPS3-class or MIPS4-class CPUs -- depending on the stage tarball you download.
I've found my MIPS boxes much more responsive under Gentoo than under Debian, largely for that very reason.
What's the difference between MIPS1 and MIPS3/4? Well, more instructions implemented in hardware (as per the RISC philosophy), and also MIPS3, MIPS4, MIPS5 and MIPS64 are all 64-bit, whereas MIPS1, MIPS2 and MIPS32 are all 32-bit.
Yes, the Pentium4 has more instructions than the 386 -- many of them implemented in microcode, which breaks down to RISC-like instructions in the end. So you're just replacing a bunch of compiler-generated instructions in a program, with just one instruction in the program, that ends up translating to the same microcode program in the end.
> more instructions implemented in hardware (as per the RISC philosophy)
That's a funny interpretation of "reduced instruction set".
> CPU's are still little more than overclocked 386's
Only to the programmer. Underneath it's a whole different story: register renaming, out-of-order execution, superscalar execution, pipelining, branch prediction, floating-point, and who knows what else.
As >>30 points out, the RISC philosophy is fewer instructions, more registers. Except that there are very few CPUs anymore that fit the definition of RISC. We have too many transistors to know what to do with them.
Well, when I say more instructions... you'd count the number on one hand. (and a badly mutilated one at that in some cases)
>>29
You can compile stuff yourself on Debian too, or use an apt repos with non-generic binaries... hell comparing gentoo and debian for speed is borderline ignorant
System designed for stablity and put through lots of testing --versus--System with minimal testing and 20039 different ways for you to screw it up
>>10
Even with the ricer CFLAGs one of them is a legitimate aliasing bug
>>33 I never said it was the bees knees... or that only speed counts.
We were discussing CFLAGS -- and I was just making a point on how much difference certain CFLAGS make on specific CPU architectures.
I was not necessarily trying to compare Debian to Gentoo specifically.