The collected works of jan-jaap - Page 1

edefault wrote: dear iconsmith - where to put these to make them functional?


(as root):
copy *.fti /usr/lib/filetype/local/iconlib/
copy *.ftr /usr/lib/filetype/local/

cd /usr/lib/filetype/ && make

firefox has tag 0x100013, you have to tag your /usr/local/firefox/firefox (or whereever it lives):
cd /usr/local/firefox && tag 0x100013 firefox

Ditto for tbird, except the tag is 0x100031

Voila !

Very nice icons, btw.
skylark wrote: I'm writing a Really Nice (tm) MIPS disassembler Very Soon Now (I plan to start working on it at July 6)


What's wrong with "objdump --disassemble" or even "objdump --disassemble-all".
There are some "disassemblers" for Linux that are really just perl scripts that wrap around objdump. You may be able to adapt one for IRIX.

Also, there's "dis" on IRIX. Depending on the status of IRIX binary compatibility in Linux/MIPS, you may be able to run it.

NetBSD has IRIX compatibility. The article series is an excellent read: http://www.onlamp.com/pub/a/bsd/2002/08/08/irix.html
squeen wrote:
Appropriately, I can not see the link (flash?) from an IRIX based browser. :razz:


No, but Acroread/IRIX can still be used to read the specsheet :o
Intel-OUTSIDE wrote: the xbox is finished. :(


That's just too funny for somebody that goes by the handle Intel-OUTSIDE :D :D
I guess you meant XBOW. And, sorry for your loss.
87Porsche wrote:


That's the x86 version of Maya.
As we all know, the x86 code execution speed of the Itanic is best described as "pathetic" . :twisted:

Futhermore, this little blurb doesn't sound good:
Code:
Maya 6.0 is currently Unsupported on the following 64-bit systems:
Itanium:    Red Hat Enterprise Linux WS (Intel Itanium)
RageX wrote:
GCC as always is behind compared to the official compilers.


You may be right for non-x86 platforms. The fact that itanic leaves a lot of work to the compiler, and the companies involved then patented that work doesn't help either. But on x86, gcc-3.4 will perform roughly on par with icc.

In fact, numerical kernels like atlas perform very close to cpu peak performance when compiled with either of them. If you want to bash a compiler, try MSVC.

Don't take my word for it, but do some tests yourself. I did :D
warerat wrote: I've got an official SGI Price Book from February 1998. An Octane with a R10K-250, 128MB RAM, 4GB disk, MXE graphics, and a 20" monitor had a list price of $47,995! Double the memory and add dual-CPUs and the price went up to $61,495!


I've got an R10K MaxImpact Indigo2 with TRAM option and 512M RAM. I wonder what that cost when it was new, I'm sure it was not exactly for free.

And I'm pretty sure the G160 board I put in it cost more than my last diving holidays :shock:
clavileno wrote: I just missed a 4xR10K 195 Onyx on eBay UK last week for 500.00 (ukp)


I saw that one. I would have bought it if I had the space for it. But all my stuff is in the attic, and the largest I can fit through the staircase is a deskside... For my next house I want something with a basement under the whole house, and a loading dock to unload a truck :D :D
See also:
http://www.schrotthal.de/sgi/indigo2/index.html
nekonoko wrote: I rarely drink except socially around holidays, but even then I stick with beer - Guinness being my favorite.


I just returned from two weeks of diving in the Red Sea, and I was real happy to find my favorite Image in the fridge at home :)
Egyptian beer is not worth mentioning, Egyptian whiskey is probably only usable as anti-freeze. Ah, well, all-inclusive... :wink:
Satoru wrote:
Unix workstations aren't dead at HP.

Who knows, with Carly out of the way , they might make a comeback ....
Well, you could always go for GNU Pascal (gpc), http://www.gnu-pascal.de/gpc/h-index.html
Several (5 or 6) years ago, I worked on that compiler and even in those days it ran on IRIX.

I know Debian has a package of it for MIPS/Linux so it's probably still possible to build it on IRIX. I wouldn't know; I haven't used Pascal in years...
I heard rumours of people porting XFS to FreeBSD.
I've never seen anything released, though. (Unfortunately, since I might seriously consider retiring my Linux server if there were a serious journaling FS for a *BSD).

Since the OS-X kernel is based on the FreeBSD kernel, at least theoretically, it should then be possible to port XFS to OS-X.

But if your aim is to read the disk of an Indy, it's probably still less work to stick the disk in a peecee and boot it from a Knoppix CD. Or stick it back in the Indy, and do something with it!

PS: If any of this has anything to do with removing the (root) password from the Indy's OS, that's a FAQ.
MooglyGuy wrote: All MIPS processors have their boot ROMs located at 0xBFC00000, so yes, it will work.


MIPS, but not MIPS64. I know at least some SGI's have their PROM at an address too large for a 32bit pointer, which makes it impossible to run a 32bit kernel on them (assuming you would like to, that is).
Now this is a deep dark secret, so everybody keep it quiet :)
It turns out that when reset, the WD33C93 defaults to a SCSI ID of 0, and it was simpler to leave it that way... -- Dave Olson, in comp.sys.sgi

Currently in commercial service: Image :Onyx2: (2x) :O3x02L:
In the museum : almost every MIPS/IRIX system.
Wanted : GM1 board for Professional Series GT graphics (030-0076-003, 030-0076-004)
dexter1 wrote: The best C++ book out there is: "The C++ Programming Language, Third Edition" by Bjarne Stroustrup.


It's the definitive resource, but to read it is like reading the telephone book.
As a starter in C++, "Accelerated C++" isn't bad.

Other must-reads are the "Exceptional C++" series and the "Effective C++" series. One or two books on the STL won't hurt either.

If you're serious about programming, you want to read something like "Design patterns". I'm currently halfway through "Generic programming and the STL" which is the best C++ book I've read for a while.
Now this is a deep dark secret, so everybody keep it quiet :)
It turns out that when reset, the WD33C93 defaults to a SCSI ID of 0, and it was simpler to leave it that way... -- Dave Olson, in comp.sys.sgi

Currently in commercial service: Image :Onyx2: (2x) :O3x02L:
In the museum : almost every MIPS/IRIX system.
Wanted : GM1 board for Professional Series GT graphics (030-0076-003, 030-0076-004)
Brombear wrote: AFAIK you cannot use an irix version higher than 6.5.19 (or was it 20) to develop with 7.3.1.3m. The later version had special (fucked up) header files targeted to the 7.4 compiler which caused huge compile problems over several os revisions. Personally I stay away from the 7.4 series. Simply not worth the hassle IMHO.

You could always install MIPSpro 7.3.x as a 'module' in /opt, with it's own IDF and IDL. I did that on my Octane, and never noticed any brokeness. I even had 7.2.1 on it (as a module), pretty sure it was there untill about the time I upgraded to 6.5.23
Now this is a deep dark secret, so everybody keep it quiet :)
It turns out that when reset, the WD33C93 defaults to a SCSI ID of 0, and it was simpler to leave it that way... -- Dave Olson, in comp.sys.sgi

Currently in commercial service: Image :Onyx2: (2x) :O3x02L:
In the museum : almost every MIPS/IRIX system.
Wanted : GM1 board for Professional Series GT graphics (030-0076-003, 030-0076-004)
I can get you a dump file, you could burn your own....

Look in these forums for threads about PROM dumping, "the dumping project" etc. I'm sure this one has been posted here sometime in the last 6 months or so.
To accentuate the special identity of the IRIS 4D/70, Silicon Graphics' designers selected a new color palette. The machine's coating blends dark grey, raspberry and beige colors into a pleasing harmony. ( IRIS 4D/70 Superworkstation Technical Report )
There are two gotchas when using a Linux box as an install server:
1) local port range. You obviously got that covered, or the thing wouldn't boot at all.

2) During the install process, the IRIX miniroot rexec's several shell scripts on the install server. On Linux, /bin/sh is GNU bash which isn't quite the same. Read the faq again about using csh or ksh or recompiling bash with a special flag. Be advised that most Linux distributions don't boot when /bin/sh != bash, which is pretty bad in itself.

I'm currently rebuilding my Linux server and plan to setup a virtual install server using vserver (similar to BSD jails and Solaris zones). This way I can hopefully find a combination that works for IRIX without causing crazy mods to the Linux side and doesn't require wide open BSD r* services on the actual server. If I figure out a recipe I'll turn it into a FAQ.
Now this is a deep dark secret, so everybody keep it quiet :)
It turns out that when reset, the WD33C93 defaults to a SCSI ID of 0, and it was simpler to leave it that way... -- Dave Olson, in comp.sys.sgi

Currently in commercial service: Image :Onyx2: (2x) :O3x02L:
In the museum : almost every MIPS/IRIX system.
Wanted : GM1 board for Professional Series GT graphics (030-0076-003, 030-0076-004)
Me, I'm a senior software engineer at a company building FireWire (IEEE1394) analyzers. I spend my days writing code in C++ and making people follow the ANSI standards instead of the Microsoft interpretation of it...

I'm old enough to have seen the sixties, or at least the last three weeks of 'em. :wink:
Now this is a deep dark secret, so everybody keep it quiet :)
It turns out that when reset, the WD33C93 defaults to a SCSI ID of 0, and it was simpler to leave it that way... -- Dave Olson, in comp.sys.sgi

Currently in commercial service: Image :Onyx2: (2x) :O3x02L:
In the museum : almost every MIPS/IRIX system.
Wanted : GM1 board for Professional Series GT graphics (030-0076-003, 030-0076-004)
Gerhard.Lenerz wrote: Oops... it's that easy? Hmmm... is there any reason they specifically mention Big Endian? ACE/ARC?


Rumour has it that there was a modified version of the Indigo2 that ran Windows NT, and that the modification was mostly a different PROM.
NT on MIPS hardware was little endian, and so any PROM used to boot into NT must have been LE as well.

There is also a rumour that NT ran on a modified Indigo(1), and late Indigo PROM revisions have the BE tag as well. Coincidence? Let the conspiracy theories begin ...
Now this is a deep dark secret, so everybody keep it quiet :)
It turns out that when reset, the WD33C93 defaults to a SCSI ID of 0, and it was simpler to leave it that way... -- Dave Olson, in comp.sys.sgi

Currently in commercial service: Image :Onyx2: (2x) :O3x02L:
In the museum : almost every MIPS/IRIX system.
Wanted : GM1 board for Professional Series GT graphics (030-0076-003, 030-0076-004)
gandalf wrote: getting glut to work on IrisGL? A Dream.

One of the key differences between IrisGL and OpenGL was that OpenGL was platform independant and thus didn't contain the support for window management and fonts that was in the original GL. GLUT was an attempt to recreate (some of?) the missing functionality using X11 and OpenGL. So, it should be possible to create a GLUT over GL; in fact it should not be more than some kind of a wrapper. Then again, it's been awhile that I did serious work with IrisGL.

gandalf wrote: skywriter has that blender classic project. Imagine blender working on your 4D series, indigo1 or Crimson? wouldn't that rock?
running blender on a skywriter?

Stop teasing me 8)

lewis wrote: FYI I think it's impossible to build Blender with gcc. The problem is that it needs to link in the GLU library, which brings the SGI C++ library libC with it, which conflicts horribly with gcc's own C++ library, libstdc++. Someone else had similar problems with other things, and as far as I know there's no solution - you can't compile C++ programs that use GLU with gcc on Irix. If anyone knows otherwise, pipe up!

What about compiling your own from source, with GCC. Make it a static library and you don't have a clash with what's on IRIX already.
To accentuate the special identity of the IRIS 4D/70, Silicon Graphics' designers selected a new color palette. The machine's coating blends dark grey, raspberry and beige colors into a pleasing harmony. ( IRIS 4D/70 Superworkstation Technical Report )
GeneratriX wrote: ...the right cable to connect a DB-9 based SGI Workstation as serial console is slightly more complex than the one required to connect a Characters Terminal:

[...]

You'll not getting working, unless you can buy/built one of theses.


My experience is different.

A few years ago, I hacked together a serial cable for an Indy or Indigo2, and it didn't work properly.
I then disconnected all hardware flow control lines (only TX/RX and GND left, so the most basic serial null modem cable), and it worked.

I use a generic windows PC with HyperTerm to connect. Of course you have to disable hardware flow control before you try to connect.

These days, I have a bunch of cables:
1) DB9 => DB9 nullmodem (for O2 and Octane)
2) DB9 => mini-DIN8 nullmodem (for 4D/35, Indigo, Indy, Indigo2)
3) DB9 => DB9 (male->female, special wiring for 4D series, looks suspiciously like a serial extension cable).
4) A real serial extension (male->female) to reach boxen that are too far away otherwise.
I'm pretty sure none of #1 ... #3 has more than 3 wires inside.

What I should have done was something like this .
But when I need just one more cable, I get just that cable, rather than redoing the entire collection. But still, this is pretty ingenious. One day ....

Hope this helps.
Now this is a deep dark secret, so everybody keep it quiet :)
It turns out that when reset, the WD33C93 defaults to a SCSI ID of 0, and it was simpler to leave it that way... -- Dave Olson, in comp.sys.sgi

Currently in commercial service: Image :Onyx2: (2x) :O3x02L:
In the museum : almost every MIPS/IRIX system.
Wanted : GM1 board for Professional Series GT graphics (030-0076-003, 030-0076-004)
I gave it a try, and
* I got a re-issue for MIPSpro (C/C++/F77/F90/Workshop) for my Power Indigo2 (duh -- that one did have license already). I didn't get licenses for APO and roboinst (which it didn't have before either).

* I didn't get one for the Crimson (it asks for entitlement IDs), even if I register it as an Indy :lol:

So far it looks like what they're saying is correct: if the hostid is in their database, they generate a new license. Lots of my systems are from universities and were previously covered by Varsity. I'll try to get licenses for my Indies, see what happens :D

_________________
Now this is a deep dark secret, so everybody keep it quiet :)
It turns out that when reset, the WD33C93 defaults to a SCSI ID of 0, and it was simpler to leave it that way... -- Dave Olson, in comp.sys.sgi
GIJoe wrote:
i tried for one of my boxes as well. it didn't spit out a single license, bastards!
but the box came pre-loaded with lics for the whole dev-suite and XVM when i got it (and some weirdo software for modelling and simulating car interior safety bits ;) ), so i guess these licenses where deleted when the original contract ended?
ah well, i don't do C/C++ but somehow it hurts when you're left out in he cold like that. :D

Let me guess, the "issuer" filed of the license says "HaxForce 2000" or something like that :lol:

I tried two more Indigos and an Indigo2 and got nothing as well. It looks like xmas is not in april this year :P

_________________
Now this is a deep dark secret, so everybody keep it quiet :)
It turns out that when reset, the WD33C93 defaults to a SCSI ID of 0, and it was simpler to leave it that way... -- Dave Olson, in comp.sys.sgi
I'm sure there must have been a nice spike in the network traffic @ sgi.com yesterday, I refreshed my mirror as well :wink:

This is probably not the best way, but here's how I did it. Various Linuxisms (debian 3.1) may be hidden in here.

Code:
#!/bin/bash

#set -x

# Freeware (fw) doesn't have books in it's collection
COLLECTIONS="0530 0620 0630 0640 0650 hdwr linux nt"

WGETOPT="-m -nv -T60 -t0 -nH --cut-dirs=2"

for coll in $COLLECTIONS; do
mkdir manuals_$coll
echo "#!/bin/sh" > wget_$coll.sh
chmod 755 wget_$coll.sh
echo "cd  manuals_$coll" >> wget_$coll.sh
lynx -dump -width=999 "http://techpubs.sgi.com/library/tpl/cgi-bin/browse.cgi?db=bks&coll=$coll&pth=ALL" > dump_$coll.txt
# Get part numbers
grep "download.cgi" < dump_$coll.txt | cut -d '=' -f4 | sort > manuals_$coll.txt
MANUALS=`cat manuals_$coll.txt`
for book in $MANUALS; do
major=`echo $book | cut -c5`
echo "wget "$WGETOPT" http://techpubs.sgi.com/library/manuals/"$major"000/"$book"/pdf/"$book".pdf" >> wget_$coll.sh
echo "wget "$WGETOPT" http://techpubs.sgi.com/library/manuals/"$major"000/"$book"/dl/"$book".html.tgz" >> wget_$coll.sh
done
done;


This creates dirs 'manuals_0530' etc. and scripts 'wget_0530.sh' etc.

Scripts look like this:
Code:
#!/bin/sh
cd  manuals_0530
wget -m -nv -T60 -t0 -nH --cut-dirs=2 http://techpubs.sgi.com/library/manuals/0000/007-0603-100/pdf/007-0603-100.pdf
wget -m -nv -T60 -t0 -nH --cut-dirs=2 http://techpubs.sgi.com/library/manuals/0000/007-0603-100/dl/007-0603-100.html.tgz
...

After inspecting the "wget_*.sh you run them.

Expect these download volumes (kB)
184928 manuals_0530
247840 manuals_0620
252076 manuals_0630
284396 manuals_0640
562524 manuals_0650
805100 manuals_hdwr
210080 manuals_linux
68356 manuals_nt

This is both the online (html.tgz) and pdf versions.

_________________
Now this is a deep dark secret, so everybody keep it quiet :)
It turns out that when reset, the WD33C93 defaults to a SCSI ID of 0, and it was simpler to leave it that way... -- Dave Olson, in comp.sys.sgi

Currently in commercial service: Image :Onyx2: (2x) :O3x02L:
In the museum : almost every MIPS/IRIX system.
Wanted : GM1 board for Professional Series GT graphics (030-0076-003, 030-0076-004)
squeen wrote: Running modified code (below) on 1 1Ghz R16K with MIPSpro C 7.4.4m, I get

Elapsed time to multiply two matrices of order 128: 0.242758



And a humble R10k Indigo2 running IRIX 6.2, MIPSpro 7.2.1.3m:
Elapsed time to multiply two matrices of order 128: 2.199671


Code: Select all

void matrix_mult(long double A[][MAX_DIM], long double B[][MAX_DIM], long double C[][MAX_DIM], int n){
int i, j, k;
for(i = 0; i < n; i++)
for(j = 0; j < n; j++)
for(k = 0; k < n; k++)
C[i][j] +=  A[i][k]*B[k][j];
}


That is the most naive matrix mult routine I've seen in awhile :wink:
Expect 9 ~ 11x speedup if you replace this with optimized routines, such as those found in SCSL or ATLAS. Or, to put it differently, I would be faster with my lowly Indigo2 than the 1GHz Tezro. Algorithm optimization rules -- you can't beat that with compiler switches!
Now this is a deep dark secret, so everybody keep it quiet :)
It turns out that when reset, the WD33C93 defaults to a SCSI ID of 0, and it was simpler to leave it that way... -- Dave Olson, in comp.sys.sgi

Currently in commercial service: Image :Onyx2: (2x) :O3x02L:
In the museum : almost every MIPS/IRIX system.
Wanted : GM1 board for Professional Series GT graphics (030-0076-003, 030-0076-004)
squeen wrote:
jan-jaap wrote: Algorithm optimization rules -- you can't beat that with compiler switches!


Amen brother!

BTW what clock speed on the Indigo2?


R10k @ 195MHz.

squeen wrote: Hmm...looks like blas3 (man INTRO_BLAS3) doesn't support quad precision, just single (SGEMM) and double (DGEMM). Still, a double precision test might be fun.


Yes, and given that BLAS, LAPACK and co. are the defacto standard libraries for this sort of business should tell you something about the real world value of quad precision floating point. In my experience with numerical solvers, everything beyond the 13th decimal was noise, and you can do that easily with double precision.
Now this is a deep dark secret, so everybody keep it quiet :)
It turns out that when reset, the WD33C93 defaults to a SCSI ID of 0, and it was simpler to leave it that way... -- Dave Olson, in comp.sys.sgi

Currently in commercial service: Image :Onyx2: (2x) :O3x02L:
In the museum : almost every MIPS/IRIX system.
Wanted : GM1 board for Professional Series GT graphics (030-0076-003, 030-0076-004)

Code: Select all

#include <math.h>
#include <stdio.h>
#include <sys/time.h>
#include <stdlib.h>
#include <cblas.h>

void matrix_mult(double* a, double* b, double* c, int n);
void matrix_rand(double* d, int nn);

int main(int argc, char ** argv)
{
int n;
double t1=0.0, t2=0.0, elapsed;
timespec_t tp[2];
double *a, *b, *c;

n = 128; /* default order of matrix */
a = (double*)calloc(n*n, sizeof(double));
b = (double*)calloc(n*n, sizeof(double));
c = (double*)calloc(n*n, sizeof(double));
printf("Using default value, n = %d\n", n);

matrix_rand(a, n*n);
matrix_rand(b, n*n);

clock_gettime(CLOCK_SGI_CYCLE, &tp[0]);
matrix_mult(a, b, c, n);
clock_gettime(CLOCK_SGI_CYCLE, &tp[1]);

t1 += (double)tp[0].tv_sec+(1.e-9)*tp[0].tv_nsec;
t2 += (double)tp[1].tv_sec+(1.e-9)*tp[1].tv_nsec;
/*    matrix_print(C, nout, n); */
elapsed = t2 -t1;
printf("\nElapsed time to multiply two matrices of order %d: %f\n", n, elapsed);

return(0);
}
/*****************************************************************************/
void matrix_mult(double* a, double* b, double* c, int n)
{
double alpha=1.0, beta=0.0;
dgemm(NoTranspose, NoTranspose, n, n, n, alpha, &a[0], n, &b[0], n, beta, &c[0], n);
}
/*****************************************************************************/
void matrix_rand(double* d, int nn)
{
int i;
for(i = 0; i < nn; i++)
d[i] = (double)rand()/12.3;
}


Then:

Code: Select all

cc -fullwarn -O3 -Ofast=IP28 -IPA -o test_double test_double.c -lblas -lftn
ld32: WARNING 85: definition of main in test_double.o preempts that definition in /usr/lib32/mips4/libftn.so.
ld32: WARNING 85: definition of main in test_double.ipaa0064e/test_double.o preempts that definition in /usr/lib32/mips4/libftn.so.

And voila:

Code: Select all

./test_double
Using default value, n = 128

Elapsed time to multiply two matrices of order 128: 0.014334

Eat dust, you ugly Dell :shock: :shock:

And finally, I'd like to see you do a problem of slightly less trivial size :twisted:

Code: Select all

Elapsed time to multiply two matrices of order 1024: 8.305315
Now this is a deep dark secret, so everybody keep it quiet :)
It turns out that when reset, the WD33C93 defaults to a SCSI ID of 0, and it was simpler to leave it that way... -- Dave Olson, in comp.sys.sgi

Currently in commercial service: Image :Onyx2: (2x) :O3x02L:
In the museum : almost every MIPS/IRIX system.
Wanted : GM1 board for Professional Series GT graphics (030-0076-003, 030-0076-004)
This being an "old game for old 4D hardware", could you eventually share the source so I can build it for IRIX 4 or something like that? My 4D/4xx don't run IRIX 6.5 :lol:
Now this is a deep dark secret, so everybody keep it quiet :)
It turns out that when reset, the WD33C93 defaults to a SCSI ID of 0, and it was simpler to leave it that way... -- Dave Olson, in comp.sys.sgi

Currently in commercial service: Image :Onyx2: (2x) :O3x02L:
In the museum : almost every MIPS/IRIX system.
Wanted : GM1 board for Professional Series GT graphics (030-0076-003, 030-0076-004)
foetz wrote:
, Firewire


dm10.


Yeah, right.

I've got a DM10 (actually, a red adaptec) in my O2. Recently, I got myself a LaCie d2 Extreme 500GB disk with firewire, and for kicks I plugged it into the O2 -- BLAM instant kernel crash. Oops, I guess it wasn't hot-pluggable. Restart with the firewire disk connected, and it seemed to be fine -- until halfway through a 'diskperf' run -- BLAM another kernel crash. OK, I get it, that's enough punishment for me -- disk goes back to server backup duty (what I bought it for in the first place).
Now this is a deep dark secret, so everybody keep it quiet :)
It turns out that when reset, the WD33C93 defaults to a SCSI ID of 0, and it was simpler to leave it that way... -- Dave Olson, in comp.sys.sgi

Currently in commercial service: Image :Onyx2: (2x) :O3x02L:
In the museum : almost every MIPS/IRIX system.
Wanted : GM1 board for Professional Series GT graphics (030-0076-003, 030-0076-004)
In 2001 SGI released 'open64', which is basically MIPSpro, with the frontends (licensed from EDG) replaced with GNU, and the MIPS code generator replaced with one targeted at the IA64.

'open64' was never a very popular OSS project, probably because of it's complexity, lack of support by SGI, and lack of people with IA64 hardware.

Then some of the key people of the original SGI compiler team created a new company, retargeted open64 from IA64 to x86-64 and voila: the pathscale EKOpath compiler was born. They added all sorts of goodies (openMP etc.), but the source code of the compiler is still covered by the SGI OSS license. If you google for 'open64-alchemy' you can find it. And it builds, not only on an opteron (which I don't have), but even on an ordinary x86! With IPA, LNO and all the other goodies I loved so much in MIPSpro :D

Right now it's busy rebuilding itself (I bootstrapped it with GCC), if that works I'll throw some numerical codes at it. I'm just wondering why it seems to be relatively unknown. And I wish they hadn't removed the MIPS backend. Just imagine :(

So, does anybody have any experience with this compiler (on x86 or x86-64)?
Now this is a deep dark secret, so everybody keep it quiet :)
It turns out that when reset, the WD33C93 defaults to a SCSI ID of 0, and it was simpler to leave it that way... -- Dave Olson, in comp.sys.sgi

Currently in commercial service: Image :Onyx2: (2x) :O3x02L:
In the museum : almost every MIPS/IRIX system.
Wanted : GM1 board for Professional Series GT graphics (030-0076-003, 030-0076-004)
kshuff wrote: Wait a minute, gotta wipe the slobber from my lower lip...x86? Does that mean I can get rid of Borland?

:)

Only if you use Borland on Linux :wink:

Code: Select all

janjaap@twister:~$ uname -a
Linux twister 2.6.12-1-k7 #1 Wed Nov 9 00:30:20 CET 2005 i686 GNU/Linux
janjaap@twister:~$ pathcc --version
PathScale EKOPath(TM) Compiler Suite: Version 2.1
Built on: 2006-09-05 15:19:11 +0200
Thread model: posix
GNU gcc version 3.3.1 (PathScale 2.1 driver)

Copyright 2000, 2001 Silicon Graphics, Inc.  All Rights Reserved.
Copyright 2002, 2003, 2004, 2005 PathScale, Inc.  All Rights Reserved.
See complete copyright, patent and legal notices in the
/vol/home/janjaap/open64/share/doc/pathscale-compilers-2.1/LEGAL.pdf file.
janjaap@twister:~$


My attempt to rebuild it with itself was less than successful, but that was probably due to my own incompetence/inexperience with the thing.

I tried to compare compilation speed between this and GNU g77 by timing both when building the NAG Fortran library (2626 files, ~800K lines), but failed because g77 doesn't compile all of it :roll:

Next attempt is to compare the speed of the code generated by building ATLAS with it, and benching it against a GCC built one. An issue came up where the compiler is too smart for a calibration program. 8) But I'd seen that one with MIPSpro on IRIX already. Right now it's building. That's going to take a couple of hours, we'll see what comes out.

If I were doing HPC type of things with an AMD64 or EM64T box or cluster I'd definately check them out. http://www.pathscale.com/ekopath.html
Now this is a deep dark secret, so everybody keep it quiet :)
It turns out that when reset, the WD33C93 defaults to a SCSI ID of 0, and it was simpler to leave it that way... -- Dave Olson, in comp.sys.sgi

Currently in commercial service: Image :Onyx2: (2x) :O3x02L:
In the museum : almost every MIPS/IRIX system.
Wanted : GM1 board for Professional Series GT graphics (030-0076-003, 030-0076-004)
Here's a testrun of ATLAS, benching GCC (3.3.x) against pathscale (a.k.a. MIPSpro/x86), solving complex double matrixes on an Athlon XP 1600:

Code: Select all

./xzl3blastst


----------------------------------- GEMM ---------------------------------------
TST# A B    M    N    K     ALPHA  LDA  LDB      BETA  LDC TIME MFLOP SpUp  TEST
==== = = ==== ==== ==== ==== ==== ==== ==== ==== ==== ==== ==== ===== ==== =====
0 N N  100  100  100  1.0  0.0 1000 1000  1.0  0.0 1000  0.0 1600.3 1.00 -----
0 N N  100  100  100  1.0  0.0 1000 1000  1.0  0.0 1000  0.0 2000.5 1.25 PASS
1 N N  200  200  200  1.0  0.0 1000 1000  1.0  0.0 1000  0.0 1828.8 1.00 -----
1 N N  200  200  200  1.0  0.0 1000 1000  1.0  0.0 1000  0.0 1882.6 1.03 PASS
2 N N  300  300  300  1.0  0.0 1000 1000  1.0  0.0 1000  0.1 1928.9 1.00 -----
2 N N  300  300  300  1.0  0.0 1000 1000  1.0  0.0 1000  0.1 1963.9 1.02 PASS
3 N N  400  400  400  1.0  0.0 1000 1000  1.0  0.0 1000  0.3 1917.9 1.00 -----
3 N N  400  400  400  1.0  0.0 1000 1000  1.0  0.0 1000  0.3 1917.9 1.00 PASS
4 N N  500  500  500  1.0  0.0 1000 1000  1.0  0.0 1000  0.5 1887.1 1.00 -----
4 N N  500  500  500  1.0  0.0 1000 1000  1.0  0.0 1000  0.5 1887.1 1.00 PASS
5 N N  600  600  600  1.0  0.0 1000 1000  1.0  0.0 1000  0.9 1963.9 1.00 -----
5 N N  600  600  600  1.0  0.0 1000 1000  1.0  0.0 1000  0.9 1959.5 1.00 PASS
6 N N  700  700  700  1.0  0.0 1000 1000  1.0  0.0 1000  1.4 1981.5 1.00 -----
6 N N  700  700  700  1.0  0.0 1000 1000  1.0  0.0 1000  1.4 1977.2 1.00 PASS
7 N N  800  800  800  1.0  0.0 1000 1000  1.0  0.0 1000  2.1 1986.7 1.00 -----
7 N N  800  800  800  1.0  0.0 1000 1000  1.0  0.0 1000  2.1 1991.6 1.00 PASS
8 N N  900  900  900  1.0  0.0 1000 1000  1.0  0.0 1000  2.9 2002.4 1.00 -----
8 N N  900  900  900  1.0  0.0 1000 1000  1.0  0.0 1000  2.9 2003.7 1.00 PASS
9 N N 1000 1000 1000  1.0  0.0 1000 1000  1.0  0.0 1000  4.1 1964.0 1.00 -----
9 N N 1000 1000 1000  1.0  0.0 1000 1000  1.0  0.0 1000  4.1 1963.0 1.00 PASS

10 tests run, 10 passed


Always GCC first, then pathscale. Most of the math kernel is handcoded, so for large cases this is where the computation time goes and both score equal. In small cases, pathscale wins. In fact, it keeps the cpu very close to peak performance for the whole range.

PS: Yeah, this is hardly a numbercruncher box. Shoot me :lol:
Now this is a deep dark secret, so everybody keep it quiet :)
It turns out that when reset, the WD33C93 defaults to a SCSI ID of 0, and it was simpler to leave it that way... -- Dave Olson, in comp.sys.sgi

Currently in commercial service: Image :Onyx2: (2x) :O3x02L:
In the museum : almost every MIPS/IRIX system.
Wanted : GM1 board for Professional Series GT graphics (030-0076-003, 030-0076-004)
I'd say it's time to update my list a bit. Here's my collection:

Big iron:
* 4D/210 GTX (pinstripe of death :cry: )
* 4D/420 VGX
* 4D/380 VGX, FDDI, IPI, VideoLab (currently being resurrected)
* Crimson Elan
* SGI 2100, 8x R10000, 8GB RAM
* Onyx2, 1x R10k, IR

Workstations:
* Personal Iris 4D/35 TG
* Indigo R3k Entry
* Indigo R4k XS24-Z
* Indigo R4400 Elan
* Indigo2 R4400 Extreme
* Indigo2 R8000 Extreme, FDDI
* Indigo2 R10000 MaxImpact+TRAM, G160
* Indy R4400 XZ
* Indy R5000 NG1-24bit, FDDI
* 2x Indy R4400 NG1-24bit, to toy with Linux and NetBSD
* O2 R10000, AV, DM10 (red adaptec)
* Octane R12000 SE

There is one old PC acting as a server, and a newer one acting as, ahem, a PC.

There are several more or less complete Indy/Indigo/Indigo2's around still that I have to get rid off of strip for parts. Several boxes full of 4D series boards. I'll probably get rid of the 4D/210 because it takes up too much space and I have better toys now.

Plans for the future:
* Get the Predator back in shape.
* Upgrade as much old (pre 100MBit) hardware to FDDI as possible
* Get an RE boardset for the Crimson
* Upgrade the 4D/420 to a 4D/440
* Get 400MHz cpus for the Origin and move the old ones to the Onyx2
* Continue installing bigger and more quiet harddisks
* Sort out the huge pile of spares/books/cables/cds/tapes etc. I have.

* Build a website for all of this.
* Either convert the garage into an office, or (better) build an extension to the house so I have enough space :lol:

Looks like I won't be bored for a while :wink:
Now this is a deep dark secret, so everybody keep it quiet :)
It turns out that when reset, the WD33C93 defaults to a SCSI ID of 0, and it was simpler to leave it that way... -- Dave Olson, in comp.sys.sgi

Currently in commercial service: Image :Onyx2: (2x) :O3x02L:
In the museum : almost every MIPS/IRIX system.
Wanted : GM1 board for Professional Series GT graphics (030-0076-003, 030-0076-004)
Our cat came from Switzerland originally, and was moved to the Netherlands ~ 5 years ago.

We moved this year and for two months lived with my parents who don't like cats so she went back to Switzerland for that time. Recently we took her back to the Netherlands. So she's an experienced traveller by now :wink:

A few observations:
1) Driving at night means the cat freaks out everytime you overtake a truck due to the lights flashing by.
2) When we moved her to Switzerland she spent a few nights in an asylum before we took her to Switzerland, so she was already somewhat stressed. Then we gave her the pill and before it could properly calm her down we took of. She was very stressed because of that, and by the time we arrived the cat had almost no voice left and we were exhausted as well.

I would definately recommend giving the cat a tranquilizer, but I'm not sure giving them two days in a row is a good idea. especially if it's an old cat.

Btw: Netherlands - Switzerland takes about 8 hours.
To accentuate the special identity of the IRIS 4D/70, Silicon Graphics' designers selected a new color palette. The machine's coating blends dark grey, raspberry and beige colors into a pleasing harmony. ( IRIS 4D/70 Superworkstation Technical Report )
SAQ wrote:
major hacking (which would probably look something like: install on an Octane or other 64-bit architecture, move disk to IP27, move over at least /unix, /usr/sysgen and /usr/gfx, attempt boot whilst praying, and most likely move over some more files.

An Octane (IP30) kernel wouldn't contain hardware support for IP28 hardware such as the GIO64 and EISA busses, SCSI, seeq ethernet etc. etc. so wouldn't boot.

A better approach would be to add the arch specific files from .22 to 6.5.30 and rebuild the inst images. With a little luck the kernel API didn't change. Of course the advantage of 6.5.30 over .22 for this old hardware doesn't make the effort worthwhile or someone would have done it already.

_________________
Now this is a deep dark secret, so everybody keep it quiet :)
It turns out that when reset, the WD33C93 defaults to a SCSI ID of 0, and it was simpler to leave it that way... -- Dave Olson, in comp.sys.sgi

Currently in commercial service: Image :Onyx2: (2x) :O3x02L:
In the museum : almost every MIPS/IRIX system.
Wanted : GM1 board for Professional Series GT graphics (030-0076-003, 030-0076-004)
AX wrote: Anybody know why patch 7110 is not available to everyone who has access to download .30?

I recently had my developer online application approved and there isn't a patch I can't download, including this one.
Now this is a deep dark secret, so everybody keep it quiet :)
It turns out that when reset, the WD33C93 defaults to a SCSI ID of 0, and it was simpler to leave it that way... -- Dave Olson, in comp.sys.sgi

Currently in commercial service: Image :Onyx2: (2x) :O3x02L:
In the museum : almost every MIPS/IRIX system.
Wanted : GM1 board for Professional Series GT graphics (030-0076-003, 030-0076-004)
hinv -v

Code: Select all

4 40 MHZ IP7 Processors
FPU: MIPS R2010A/R3010 VLSI Floating Point Chip Revision: 4.0
CPU: MIPS R2000A/R3000 Processor Chip Revision: 3.0
On-board serial ports: 2 per CPU board
Data cache size: 64 Kbytes
Instruction cache size: 64 Kbytes
Secondary data cache size: 1 Mbyte
Main memory size: 64 Mbytes
I/O board, slot F: IO3
Integral Ethernet: et0, IO3
VGX Graphics option installed
Tape drive: unit 7 on SCSI controller 1: QIC 150
Disk drive: unit 3 on SCSI controller 1
Integral SCSI controller 1: Version WD33C93A, revision 9
CDROM: unit 6 on SCSI controller 0
Disk drive: unit 1 on SCSI controller 0
Integral SCSI controller 0: Version WD33C93A, revision 9


gfxinfo -v

Code: Select all

Graphics board 0 is "VGX" graphics.
Managed (":0.0") 1280x1024
10 spans,       IMP3
EV mask = 0x0
1 screen(s) on this pipe


This is the system that had a broken power supply a couple of weeks ago. As you can see, this problem has been fixed :)
When I got this one it was a 4D/420 and it had only a 5 span VGX.

System is in perfect running condition, with complete and perfect looking skins.

The following upgrades are in the pipeline (I have the parts, but haven't had the time to install and test them):
* VME FDDI network board
* VME dual differential scsi board
* Upgrade MC2 to 256MB RAM.

At that time I may have to upgrade it to IRIX 5.3 also. Not that I want to, but I don't have the drivers for the FDDI and SCSI board for IRIX 4.x. For IRIX 5.3 they're all on the CD, but for IRIX4 they were on seperate discs and of course these are long gone. Even SGI can't seem to find them anymore :(


In this configuration, with two 5.25" full height harddisks it consumes about 875 Watts electricty, or about 2.35 EUR/day
To accentuate the special identity of the IRIS 4D/70, Silicon Graphics' designers selected a new color palette. The machine's coating blends dark grey, raspberry and beige colors into a pleasing harmony. ( IRIS 4D/70 Superworkstation Technical Report )
R-ten-K wrote: I had to debug the VGXT boardset through the serial port of the command/control processor (VGX* had their own dedicated 68020 I believe), man it was fun!

I tried once, but didn't get very far:
viewtopic.php?t=6129&postdays=0&postorder=asc&start=28

Any memories? I'm assuming the host side is pushing commands into the memory space of the GM so if you know where to set a breakpoint (on the host and/or GM) you might actually see this happening?
To accentuate the special identity of the IRIS 4D/70, Silicon Graphics' designers selected a new color palette. The machine's coating blends dark grey, raspberry and beige colors into a pleasing harmony. ( IRIS 4D/70 Superworkstation Technical Report )
Updated. Now has the maximum memory (256MB), 100MBit network (FDDI) and an additional dual channel differential SCSI board (Interphase 'Jaguar').

hinv -v

Code: Select all

4 40 MHZ IP7 Processors
FPU: MIPS R2010A/R3010 VLSI Floating Point Chip Revision: 4.0
CPU: MIPS R2000A/R3000 Processor Chip Revision: 3.0
On-board serial ports: 2 per CPU board
Data cache size: 64 Kbytes
Instruction cache size: 64 Kbytes
Secondary data cache size: 1 Mbyte
Main memory size: 256 Mbytes
I/O board, slot F: IO3
VGX Graphics option installed
Interphase 4210 VME-SCSI controller 0: Firmware revision 01D
Integral Ethernet: et0, IO3
FDDIXPress controller: ipg0, version 1
Integral SCSI controller 1: Version WD33C93A, revision 9
Tape drive: unit 7 on SCSI controller 1: QIC 150
Disk drive: unit 1 on SCSI controller 1
Integral SCSI controller 0: Version WD33C93A, revision 9
CDROM: unit 6 on SCSI controller 0
Disk drive: unit 1 on SCSI controller 0
VME bus: adapter 0
To accentuate the special identity of the IRIS 4D/70, Silicon Graphics' designers selected a new color palette. The machine's coating blends dark grey, raspberry and beige colors into a pleasing harmony. ( IRIS 4D/70 Superworkstation Technical Report )