It is not whether the CPU has any bug. The sage binaries and the
libraries it depends on are (I believe) compiled with some minimal level
of optimization. Maybe some optimization got applied in the atlas
library which is not present in your cpu. Typically, the atlas library
will run fine as long as you don't call that one particular command that
is using that optimized path. And then, you will get a segv or illegal
instruction when it does hit the optimized path.
Why dima asked you for the output of your /proc/cpuinfo is to determine
what instructions your processor supports. It will help tone down or
omit that specific optimization that the atlas library is being compiled
with.
On 10/14/2012 03:38 PM, Georgi Guninski wrote:
I am pretty sure all modern CPUs have a lot of bugs.
Don't have time to debug this bug.
Checked it on ubuntu 12.04 with binary sage and didn't crash,
so it is probably a "feature". 12.04 made me install libgfortran3
and it is a different version from 10.04.
On Sat, Oct 13, 2012 at 06:25:16AM -0700, Dima Pasechnik wrote:
On Saturday, 13 October 2012 20:58:46 UTC+8, Georgi Guninski wrote:
On Sat, Oct 13, 2012 at 05:45:29AM -0700, Dima Pasechnik wrote:
On Saturday, 13 October 2012 19:26:17 UTC+8, Georgi Guninski wrote:
On Fri, Oct 12, 2012 at 10:17:29AM -0700, kcrisman wrote:
What is the name of the binary (i.e., is it 64 or 32 bit, and so
forth)?
It easy is to check if the problem is in my boxen - install ubuntu
10.04
in
a virtual machine, download the 64 bit binary from sagemath and
run the testcases.
Not all x86_64 boxes are equal. Different models of x86_64 processors
have
different sets of commands, and VMs are even worse in this case, as we
saw
situations where not all the processor capabilities are allowed by the
VM,
but for the software it looks as if these capabilities are allowed...
So you are implying my two boxen (intel and amd) are all buggy so i
should compile from source on both?
buggy? I never said that. It just so happened that the binary you run on
them was built for
a slightly different architecture.
I should have be warned before downloading the binary, please make
this statement explicit on the binary download page :)))
Well, Sage keeps shooting itself in the foot here. Few releases ago there
was a similar problem with OSX binaries :-(.
It would help if you post specifications of your processors.
That is, the output of
cat /proc/cpuinfo
--
You received this message because you are subscribed to the Google Groups
"sage-support" group.
To post to this group, send email to [email protected].
To unsubscribe from this group, send email to
[email protected].
Visit this group at http://groups.google.com/group/sage-support?hl=en.
--
You received this message because you are subscribed to the Google Groups
"sage-support" group.
To post to this group, send email to [email protected].
To unsubscribe from this group, send email to
[email protected].
Visit this group at http://groups.google.com/group/sage-support?hl=en.
