[RC5] P4 Speed Question

Peter Cordes peter at llama.nslug.ns.ca
Mon Jun 18 04:14:48 EDT 2001


On Fri, Jun 15, 2001 at 09:44:31AM -0500, Ryan Malayter wrote:
> >> Somewhat ironically, it's the Linux folks - who can recompile most of
> their
> >> software to their liking - that might find they actually like performance
> >> (maybe not price) of the P4 right now.
> 
> > I pity the foo' who can't recompile her OS (apologies to Mr. T :).  I
> don't
> >find it ironic at all, since one of the reasons for using a Free OS is that
> >you can do anything you want with it, and aren't dependent on any company
> to
> >get around to doing stuff for you.  Could you explain what you meant there?
> 
> Well... haven been an admin/DBA/programmer for an exclusively NT environment
> (well, one HP/UX box) for the last 5 years, I guess well I'm out of the **IX
> scene. I suppose I was thinking - fondly, mind you - of what we outsiders
> think of as the "slashdot culture," that core of underdog-loving Linux
> diehards
> who loathe Microsoft first and Intel second. To me, it's somewhat ironic
> that
> they who promote love for AMD and hate of Intel's monopolistic tendencies
> can benefit most from Intel's new beast.

 Ah, I see what you mean.  It's only ironic when you think about the
goofs who think AMD would act differently if they were as big and
had as much market share as Intel.  The only reason to hate Intel is
their marketing department.  They get stuff pushed out before its
ready, and they cripple the budget stuff so it doesn't compete with
the high-end stuff.  (compare duron vs. celeron for the difference
between the budget product of AMD and Intel...)

 This is pointed out often enough on /. that I've thought about it,
and concluded there's nothing inherently evil about Intel.  They have
a dumb marketing dept. and currently a single-CPU system will run
faster for the same money if you use an AMD CPU.  That's all, even if
it is pretty important when helping people buy a new computer.  Intel
has higher prices, so don't buy them.  They aren't evil for having
higher prices.  They just don't have stuff worth buying unless you
want a cheap dual-cpu machine (or you want stuff other than IA32 CPUs,
like eepro NICs.  err, not a great example, since there are other good,
cheap, NICs that don't charge you for the brand name.  Both of my
Network Everywhere NC-100 NICs work wonderfully for me.)

 Anyway, yeah, now I get it, but I'm mature enough to realize
that rooting for AMD is just plain underdog loving, not some kind of
fight against an evil empire (unlike the case with Microsoft and
software... http://bbspot.com/News/2000/4/MS_Buys_Evil.html :-)

 I'm sure you know this already, but there are a lot of people who
believe strongly in Free software but aren't such mental midgets that
they think like the most stereotypical /. zealots.

> 
> >> Assuming, of course, that these Lunix
> >> folks have access to good P4-optimized compilers and script
> interpreters...
> 
> > Oh, you mean this?
> >http://developer.intel.com/software/products/compilers/linuxbeta.htm
> 
> It's a start, I guess. Not being a Linux guy, I didn't know about this
> offering
> from Intel. But from my Solaris days in college, I remember how compiler
> choice
> sparked minor holy wars amongst our C programmers, a sun vs. gcc battle
> waged in
> the Makefiles. Nobody wanted to handle all those compiler flag differences;
> everyone
> swore that their compiler choice produced the tightest, fastest code. I
> would think
> Intel has to make their compiler fully gcc compatible to make any P4
> Linux/FreeBSD
> inroads. But maybe things have changed in this regard since I last used two
> forward
> slashes to comment my code...

 Hehe :) My school has Solaris machines with gcc and Sun's workshop
compilers both installed.  Last time I checked, I think I found that
both give similar performance on integer code, but that the workshop C
compiler generates faster floating point code, and has some fancier
machine-specific and optimization options.  It also has an option to
handle C++ style // comments.  It can do stuff like automatically
parallelize some loops and compile the program to start multiple
threads.  For a Mandelbrot fractal generator, this gives a very real
speed boost when running on all but one of your CPUs (to leave room
for other processes without interrupting one of your threads).  Of
course, I had to wait until the middle of the night for the load
average to be low enough for this to work well, since the only MP
machine random students like me could use is THE server that everyone
uses to do their homework.  For my homework, I always had to do stuff
in my Makefile like if $(uname ... ) (sorry don't remember Makefile
conditional syntax right now...)   Actually, that was because of
Solaris vs. Linux in other ways, since I just compiled my homework
with GCC.  Being able to use GNU C (as opposed to ISO C) features was
nice, but not all of them were available since the GNU C library is
needed for some of them.

 Sorry to be so far off topic.  I hope you didn't read this far if you
didn't like this message :)

-- 
#define X(x,y) x##y
Peter Cordes ;  e-mail: X(peter at llama.nslug. , ns.ca)

"The gods confound the man who first found out how to distinguish the hours!
 Confound him, too, who in this place set up a sundial, to cut and hack
 my day so wretchedly into small pieces!" -- Plautus, 200 BCE
--
To unsubscribe, send 'unsubscribe rc5' to majordomo at lists.distributed.net
rc5-digest subscribers replace rc5 with rc5-digest



More information about the rc5 mailing list