[help!] problem with sage-vmware 3.4.1 on windows xp

[Alessandro Torre]

Hello to everybody.
I can't make Sage works on windows and I'm going crazy.

When typing "sage" or "notebook" on the welcome screen of the VMware
Player nothing happens.
If I type "manage" and then "sage" the following message appears:

------------------------------------------
| Sage Version 3.4.1, (etc..) |
------------------------------------------
usr/local/sage/local/bin/sage-sage: line 198: 3558 Illegal
instruction sage-ipython "$@" -i


What have I missed?

VMware 2.5.2 on Windows XP sp3 .
AMD Athlon 64 Processor 3000+ 2.00 GHz, 1.50 GB Ram
0 MB Paging File

Is it possible that my processor is obsolete for this version of Sage?
I've set my paging file to 0 MB. Could it be a problem?

Thanks in advance.
Alessandro

[William Stein]

On Sat, May 2, 2009 at 8:53 AM, Alessandro Torre
<adessobas...@gmail.com> wrote:
>
> Hello to everybody.
> I can't make Sage works on windows and I'm going crazy.
>
> When typing "sage" or "notebook" on the welcome screen of the VMware
> Player nothing happens.
> If I type "manage" and then "sage" the following message appears:
>
> ------------------------------------------
> | Sage Version 3.4.1, (etc..) |
> ------------------------------------------
> usr/local/sage/local/bin/sage-sage: line 198:  3558 Illegal
> instruction     sage-ipython "$@" -i
>
>
> What have I missed?
>
> VMware 2.5.2 on Windows XP sp3 .
> AMD Athlon 64 Processor 3000+ 2.00 GHz, 1.50 GB Ram
> 0 MB Paging File
>
> Is it possible that my processor is obsolete for this version of Sage?

Yes.

What is the output of

cat /proc/cpuinfo

after you login as "manage"?

Anyway, you might have to build Sage from source. This is actually
easier than you might think. Do all of the following after login as
"manage" and typing "sudo su".

1. Delete /usr/local/sage, then
cd /usr/local; rm -rf sage
2. download sage-3.4.1.tar by typing
wget http://sagemath.org/src/sage-3.4.1.tar
3. extract it,
tar xf sage-3.4.1.tar
4. move it to be /usr/local/sage, and
mv sage-3.4.1 sage
5. build Sage -- this will take a few hours (no manual intervention):
cd sage; make

> I've set my paging file to 0 MB. Could it be a problem?

You could try setting it larger and see what happens.

-- William

[mabshoff]

On May 2, 2:04 pm, William Stein <wst...@gmail.com> wrote:
> On Sat, May 2, 2009 at 8:53 AM, Alessandro Torre

<SNIP>

> > Is it possible that my processor is obsolete for this version of Sage?
>
> Yes.
>
> What is the output of
>
>    cat /proc/cpuinfo

I doubt this will be helpful. Having thought about this for a while my
guess is that some of the instructions used in MPIR might cause this.
When I wrote a script to detect "advanced" instructions I only
concentrated on SSE instructions since back then it was the major (and
only known) problem. If gdb is installed in the VMWare image we can
start Sage under gdb, get a back trace (to figure out which library it
is) and then disassemble around the current pc so that I can fix
this :)

So far no one has provided information to what is wrong, i.e. which
instruction[s] in what library causes this, so I cannot even start to
fix this.

Cheers,

Michael

[Alessandro Torre]

>     wgethttp://sagemath.org/src/sage-3.4.1.tar

> 3. extract it,
>      tar xf sage-3.4.1.tar
> 4. move it to be /usr/local/sage, and
>      mv sage-3.4.1 sage
> 5. build Sage -- this will take a few hours (no manual intervention):
>       cd sage; make
>
> > I've set my paging file to 0 MB. Could it be a problem?
>
> You could try setting it larger and see what happens.
>
>  -- William

Thanks a lot for this step to step procedure.
I found your reply on a similar post too, but I didn't figure out I
could make it in winxp thanks to vmware.

The output you asked for:

processor : 0
vendor_id : AuthenticAMD
cpu family : 15
model : 12
model name : AMD Athlon(tm) 64 Processor 3000+
stepping : 0
cpu MHz : 1999.515
cache size : 512 KB
fdiv_bug : no
hlt_bug : no
f00f_bug : no
coma_bug : no
fpu : yes
fpu_exception: yes
cpuid level : 1
wp : yes
flags : fpu vme de pse tsc msr pae mce cx8 apic sep mtrr
pge mca cmov pat pse36 clflush mmx fxsr sse sse2 syscall nx mmxext
3dnowext 3dnow up ts fid vid ttp
bogomips : 4016.23


P.S: Setting a proper paging file didn't help.

Alessandro

[Alessandro Torre]

On 2 Mag, 23:29, mabshoff <Michael.Absh...@mathematik.uni-dortmund.de>
wrote:

Can I help somehow?

[mabshoff]

On May 3, 12:54 am, Alessandro Torre <adessobastadavv...@gmail.com>
wrote:

> On 2 Mag, 23:29, mabshoff <Michael.Absh...@mathematik.uni-dortmund.de>

<SNIP>

> "So far no one has provided information to what is wrong, i.e. which
> instruction[s] in what library causes this, so I cannot even start to
> fix this."
>
> Can I help somehow?

Absolutely :). I posted this in some other thread, but cannot find it
at the moment.

I do not believe gdb is installed as is in the VMWare image and since
that version of Ubuntu is no longer supported it really ought to be
updated and gdb installed per default. It won't make much of a
difference size wise and for debugging problems like this is is rather
helpful since otherwise we are otherwise poking around in the dark.
Assuming you have gdb run

./sage -gdb

It should fail due to illegal instruction and drop you to the gdb
prompt:

(gdb)

Then punch in "bt" (no quotes) and you will see a backtrace. This
might exceed the length of the screen, so there might be another
prompt to continue. Copy and paste the backtrace (I assume only the
first 20 lines will be of interest, but to be safe take the whole
first page). Then hit enter until you get a prompt again.

enter "disassemble $pc+32,$pc-32" at the gdb prompt and copy and paste
the output.

Mail all of the above to the list and we will fix the problem :)

Cheers,

Michael

[Alessandro Torre]

@William
I've build sage as you said and it works. Great!

@Michael
I've installed gdb and followed your instruction, but I dunno how to
copy the output.
(I'm sorry for such a trivial question, but this is my first
experience with a unix shell)
And which address should I mail to what you need?

Thanks to both of you!

[William Stein]

On Sun, May 3, 2009 at 8:32 AM, Alessandro Torre
<adessobas...@gmail.com> wrote:
>
> @William
> I've build sage as you said and it works. Great!
>
> @Michael
> I've installed gdb and followed your instruction, but I dunno how to
> copy the output.
> (I'm sorry for such a trivial question, but this is my first
> experience with a unix shell)

VMware doesn't by default make this easy at all. One thing you can do
is use a standard windows ssh program (e.g., putty is a good free
one), and ssh to the vmware machine.
Use the login "login" and password "sage". The address that you ssh
to is the same one
that you use to connect to the sage notebook. Copy and paste, etc.,
should work very well with putty.

> And which address should I mail to what you need?
>
> Thanks to both of you!
> >
>

--
William Stein
Associate Professor of Mathematics
University of Washington
http://wstein.org

[Alessandro Torre]

putty cannot connect to the address you say.
I think it is normal if sage fails to start.

I tried to use standard gdb logging to file, but I can't find gdb.txt
I tried to set another file, but it seems it doesn't work (I used "set
logging file mylog.txt")
Any hints?
I could ftp the log file to a server of mine and then download it from
within winxp.

On 3 Mag, 20:20, William Stein <wst...@gmail.com> wrote:
>
>
> VMware doesn't by default make this easy at all.  One thing you can do
> is use a standard windows ssh program (e.g., putty is a good free
> one), and ssh to the vmware machine.
> Use the login "login" and password "sage".  The address that you ssh
> to is the same one
> that you use to connect to the sage notebook.    Copy and paste, etc.,
> should work very well with putty.
>

[William Stein]

On Mon, May 4, 2009 at 10:51 AM, Alessandro Torre
<adessobas...@gmail.com> wrote:
>
> putty cannot connect to the address you say.
> I think it is normal if sage fails to start.

putty should connect fine if you get a login prompt for the sage
virtual machine. Connecting via ssh has nothing to do with sage.

[Alessandro Torre]

Hi.

Here is the requested output

(http://groups.google.com/group/sage-support/browse_thread/thread/df502c9e16565886)

Good Work.

Alessandro

[Alessandro Torre]

I forgot the files...

[mabshoff]

On May 4, 12:49 pm, Alessandro Torre <adessobastadavv...@gmail.com>
wrote:
> Hi.Here is the requested output
> (http://groups.google.com/group/sage-support/browse_thread/thread/df50...
> )
>
> Good Work.
> Alessandro

Hi Alessandro,

in the subsequent email you disassembled the whole file. What I wanted
was the output from

disassemble $pc+32,$pc-32

That will make it a lot less cumbersome to determine what the problem
is. You should also send that output inline and not attach small text
files to emails IMHO :)

Cheers,

Michael

[Alessandro Torre]

Hi.

disassemble.txt does contain the output from "disassemble $pc
+32,$pc-32"

Instead "disassemble $pc+32 $pc-32" produces no output:

Dump of assembler code from 0xb76ce853 to 0xb76ce813:
End of assembler dump.

And the output of "disassemble $pc-32 $pc+32":

Dump of assembler code from 0xb76ce813 to 0xb76ce853:
0xb76ce813 <__gmpz_set_str+451>: rorl $0x31,-0x378ad42a(%ecx)
0xb76ce81a <__gmpz_set_str+458>: rclb %cl,0x56(%edx)
0xb76ce81d <__gmpz_set_str+461>: fildll (%esp)
0xb76ce820 <__gmpz_set_str+464>: add $0x8,%esp
0xb76ce823 <__gmpz_set_str+467>: mov 0x10(%ebp),%edi
0xb76ce826 <__gmpz_set_str+470>: lea (%edi,%edi,4),%edx
0xb76ce829 <__gmpz_set_str+473>: mov -0x2c(%ebx),%eax
0xb76ce82f <__gmpz_set_str+479>: fdivl 0x4(%eax,%edx,4)
0xb76ce833 <__gmpz_set_str+483>: fisttpl -0x20(%ebp)
0xb76ce836 <__gmpz_set_str+486>: mov -0x20(%ebp),%eax
0xb76ce839 <__gmpz_set_str+489>: lea 0x1f(%eax),%edx
0xb76ce83c <__gmpz_set_str+492>: cmp $0xffffffff,%eax
0xb76ce83f <__gmpz_set_str+495>: cmovle %edx,%eax
0xb76ce842 <__gmpz_set_str+498>: sar $0x5,%eax
0xb76ce845 <__gmpz_set_str+501>: add $0x2,%eax
0xb76ce848 <__gmpz_set_str+504>: mov -0x3c(%ebp),%edx
0xb76ce84b <__gmpz_set_str+507>: cmp (%edx),%eax
0xb76ce84d <__gmpz_set_str+509>: jg 0xb76ce917 <__gmpz_set_str
+711>
End of assembler dump.

I hope that's all.

Cheers

[Alessandro Torre]

Any news?
Do you think you will need me for future testing or can I delete the
bugged vmware image?
And how are things going?
Have you solved the problem?

Cheers
A.

On 5 Mag, 12:13, Alessandro Torre <adessobastadavv...@gmail.com>
wrote: