Jump to content

DIY Street Linux Thread.


Guest culinist_merged

Recommended Posts

Wow... so many responses.. :P

Hey AG, I respect your decision of not wanting to use Linux anymore as it is fairly dificult, and yes.. some stuff require the "command line". Still... imho the command line is more powerful than graphical mode, and there are lots of things you do faster on the command line than in the graphical mode.

 

Thing is .. I think each OS has it's own good and bad points and for me they are:

Windows's Good Points: Easy to use, Gaming, Support for most modern hardware.

Windows's Bad Points: Security, Stability, Commercial.

 

Linux's Good Points: Stability, Security, Open-Source, Free.

Linux's Bad Points: Not-so-easy to use, almost no gaming, and the support for the latest stuff is not so great.

 

I can live without Gaming on the PC, and I sure like to know and learn about things (specially when related to computers). I've learned more in these 2 weeks I've been running Linux solo than in the last 5 years of using Windows.

 

ExRoadie's saying that there seem to be 2 distinct versions of linux (the mainstream and the nerd's one) and I think so aswell. There are wonderful ready-to-use linux distros like Ubuntu, Kurumin (brazilian), etc... and then comes Debian, Gentoo and Slackware.

Each one better for its niche of people. I also think that Linux will only emerge when it's really easy to use, but strong efforts are being made in that way and Ubuntu is a clear show of that.

 

Regarding to king of nothing post.. nope.. I have never been able to show my fan speeds under lm-sensors altough they displayed the speed fine in the BIOS an inside Windows. And speaking of windows .. go check your pm's or the Toastie Siggies thread. :P

Share this post


Link to post
Share on other sites

  • Replies 518
  • Created
  • Last Reply

Top Posters In This Topic

Guest culinist

I just run Linux because it's fun and free. I also like the idea of overclocking your software using gcc compiler optimizations:)

 

Linux will never take over Windows for the top being the top dog in the yard. It takes money to do that and for the most part the folks that do the coding for *nix are unpaid geeks like us.:)

 

As for it being to difficult to use......I'm not the sharpest tack in the box by a long shot....but I enjoy it, so that makes it easier. If I can get a Debian true 64bit install workin...anyone can. It's alot like learnin to overclock, you do your "homework", google your butt off, and get ta' bizzness.

 

Linux is my OS of choice, but I will probably always keep a small Windows partition on which ever comp is my top dog at the time. Gots ta play my games, ya know.

Share this post


Link to post
Share on other sites

linux = star trek geeks who go to conventions and then argue the merits of bifurcated plasma shielding and pulsar radiation penetration ratings

 

windows = like that molester uncle every family has but no one talks about and is ashamed of...you got to have him cuz he's family, you'd like to get rid of him, but can't, but damn if I just belonged to another family I'd not be neurotic and blurt out loud curses on a regular basis thinking of him.

 

mac = people still use macs???

 

heh

Share this post


Link to post
Share on other sites

Macs .... yes people still use Macs...some of these people actually run Linux on their Macs...

 

Oddly enough, I've never gone to a Star Trek convention...

 

by the way, it now only takes me 2 minutes or so to recompile the kernel (@2300 MHz). Man is that ever nice and quite a jump from the 45 to 60 minutes on my Pentium 200. I like the fact that drivers I do not use never get loaded (and sometimes not even built!).

Share this post


Link to post
Share on other sites

Guest culinist
by the way, it now only takes me 2 minutes or so to recompile the kernel (@2300 MHz). Man is that ever nice and quite a jump from the 45 to 60 minutes on my Pentium 200. I like the fact that drivers I do not use never get loaded (and sometimes not even built!).

 

Ahh kernel compilation....., the ultimate benchmark.:)

Share this post


Link to post
Share on other sites

Guest culinist

Here a cool little benchmark that Spankin Partier over at pcper posted.....

 

 

1. Make a new directory in your home directory to preform the test. Then change into that directory.

# mkdir ~/pi

# cd ~/pi

2. Copy and past the following code into your favourite text editor and save the file as ~/pi/pi.c

 

/*

** PI.C - Computes Pi to an arbitrary number of digits

**

** Uses far arrays so may be compiled in any memory model

*/



#include <stdio.h>

#include <stdlib.h>

#include <sys/time.h>

#include <unistd.h>



long secstart, usecstart;

long kf, ks;

long *mf, *ms;

long cnt, n, temp, nd;

long i;

long col, col1;

long loc, stor[40];



void shift(long *l1, long *l2, long lp, long lmod)

{

long k;



k = ((*l2) > 0 ? (*l2) / lmod: -(-(*l2) / lmod) - 1);

*l2 -= k * lmod;

*l1 += k * lp;

}



void yprint(long m)

{

if (cnt<n)

{

	if (++col == 11)

	{

		col = 1;

		if (++col1 == 6)

			{

				col1 = 0;

				printf("n");

				printf("%4ld",m%10);

			}

		else printf("%3ld",m%10);

	}

	else printf("%ld",m);

	cnt++;

}

}



void xprint(long m)

{

long ii, wk, wk1;



if (m < 8)

{

	for (ii = 1; ii <= loc; )

	yprint(stor[(int)(ii++)]);

	loc = 0;

}

else

{

	if (m > 9)

	{

		wk = m / 10;

		m %= 10;

		for (wk1 = loc; wk1 >= 1; wk1--)

		{

			wk += stor[(int)wk1];

			stor[(int)wk1] = wk % 10;

			wk /= 10;

		}

	}

}

stor[(int)(++loc)] = m;

}



void memerr(int errno)

{

printf("anOut of memory error #%dn", errno);

if (2 == errno)

free(mf);

_exit(2);

}



int main(int argc, char *argv[])

{

int i=0;

char *endp;



stor[i++] = 0;

if (argc < 2)

{

	puts("aUsage: PI <number_of_digits>");

	return(1);

}

n = strtol(argv[1], &endp, 10);

mf = malloc((size_t)(n + 3L)*(size_t)sizeof(long));

if (!mf)

memerr(1);

ms = malloc((size_t)(n + 3L)*(size_t)sizeof(long));

if (!ms)

memerr(2);

printf("nApproximation of PI to %ld digitsn", (long)n);



struct timeval tv;

struct timezone tz;

	gettimeofday(&tv, &tz);

	secstart=tv.tv_sec;

	usecstart=tv.tv_usec;

cnt = 0;

kf = 25;

ks = 57121L;

mf[1] = 1L;

for (i = 2; i <= (int)n; i += 2)

{

	mf[i] = -16L;

	mf[i+1] = 16L;

}

for (i = 1; i <= (int)n; i += 2)

{

	ms[i] = -4L;

	ms[i+1] = 4L;

}

printf("n 3.");

while (cnt < n)

{

	for (i = 0; ++i <= (int)n - (int)cnt; )

	{

		mf[i] *= 10L;

		ms[i] *= 10L;

	}

	for (i =(int)(n - cnt + 1); --i >= 2; )

	{

		temp = 2 * i - 1;

		shift(&mf[i - 1], &mf[i], temp - 2, temp * kf);

		shift(&ms[i - 1], &ms[i], temp - 2, temp * ks);

	}

	nd = 0;

	shift((long *)&nd, &mf[1], 1L, 5L);

	shift((long *)&nd, &ms[1], 1L, 239L);

	xprint(nd);

}

printf("nnCalculations Completed!n");

gettimeofday(&tv, &tz);

printf("Time: %ld secondsn",tv.tv_sec-secstart);

free(ms);

free(mf);

return(0);

}

 

3. Now compile the code

# gcc -o pi pi.c -lm

4. Shut down any CPU intensive applications (Seti, FaD, etc)

5. Run the benchmark for 100,000 decimal places.

# ./pi 100000

6. When it's done, post your time and your system specs (CPU, motherboard, over clock speed if applicable).

 

 

585 sec on my LP w/A64 3000+ stock.

Share this post


Link to post
Share on other sites

okay, I'm game and will try it out on my A64 system as soon as its stable at 2500.

but in the meantime I tried this little benchmark out on my AthlonXP 2000+ (stock).

 

without gcc optimizations pi.c does 717 s.

 

with "-O3 -mtune=athlon-xp -march=athlon-xp -mmmx -msse -m3dnow -pipe -mfpmath=sse" added to "gcc -o pi pi.c -lm" pi.c does 502 s.

 

I didn't think there would be *that* much difference. Can't wait to try it out on the A64 once I've set up a 64 bit Linux on it!

Share this post


Link to post
Share on other sites

Guest culinist

Yeah on my 1700+ I did 1174 secs without gcc opts. and with "gcc -march=athlon-xp -O3 -fomit-frame-pointer -falign-functions=4 -fprefetch-loop-arrays -funroll-loops -fforce-addr -ffast-math -ftracer -o pi pi.c -lm" it was 613 secs. I was real surprised at the difference. Nothin like overclockin code:)

 

Just put Debian true64 back on with windors and pclinuxos so I'll start foolin around with some gcc opts on her and see what she'll say.

Share this post


Link to post
Share on other sites

Hey all-

Mine just calculated it in 473 seconds without the gcc optimizers. I'm running my LPb xp-m 2600+, hardware OCed to 2.25ghz with only 512 ram.

 

I'm running RedHat Enterprise Linux, Version 3...the ES build.

 

I've been working with linux remotely with alot of the webhosting work I do, and thought I've give it a run...my boss at work gave it to me. And those of you with vast knowledge of linux...is this OS I'm runnnin any good, or should I use something else?

Share this post


Link to post
Share on other sites

is this OS I'm runnnin any good, or should I use something else?

Red Hat is a good solid product (and has customer support). I have tried a number of different distros over the past few years and have found the one I like - LinuxFromScratch.

 

My advice: get to know the distro you have now really, really well. And then if you find something you don't like - see what else is out there or if it can be configured/tweaked to your liking.

 

Most of the time, the different distros are based on a similar set of software, but it is packaged and set it up differently.

Share this post


Link to post
Share on other sites

The main things I'd love to have but from my googling arnt readily available, are drivers for Sata Radi (Sil 3114), and my video card...ati radeon 9800pro/xt.

 

Ran the pi calculator as "gcc -O3 -march=athlon-xp -mmmx -msse -m3dnow -pipe -mfpmath=sse -o pi pi.c -lm", and it finished in 324 seconds. :nod: :D

Share this post


Link to post
Share on other sites

Please sign in to comment

You will be able to leave a comment after signing in



Sign In Now

×
×
  • Create New...