On 11/20/10 3:45 PM, William Stein wrote:
On Sat, Nov 20, 2010 at 12:07 PM, Jeroen Demeyer<jdeme...@cage.ugent.be> wrote:
I think somebody should seriously look into www.sagenb.org. It is so
slow that it is essentially unusable. Unfortunately, it is also very
bad advertising for Sage, because for many people their first contact
with Sage might be www.sagenb.org. I've actually been directing my
students to alpha.sagenb.org, but of course that is not a good solution.
Somebody who knows the notebook might be able to understand why it is so
slow. The problem is not the machine, because the load is not high and
other notebooks on the same server are much faster. So it is truly an
implementation problem.
Hi,
To solve this slowness problem will require probably at least a month
of focused work to completely reimplement basically from scratch the
storage and server architecture of the notebook in a much more
scalable way. Nobody is working on this, though if nobody else does
it, then I'll do it sometime in the next year. So the options are:
(1) I simply totally disable www.sagenb.org,
(2) I leave things as they are.
(3) Somebody reimplements the notebook in a way that scales.
Writing scalable web applications is not easy. It's possible, but it
requires much different techniques than are used for the current
notebook, e.g., one must have multiple server instances, and all data
has to be stored in a database that can accessed from any server, etc.
But IIRC, we aren't sure exactly where the bottlenecks even are right
now. You've mentioned before that the first step is someone developing
some test suites that will hammer a server and profile the server to see
where the bottlenecks are. Has anyone done this?
Thanks,
Jason
--
To post to this group, send an email to sage-devel@googlegroups.com
To unsubscribe from this group, send an email to
sage-devel+unsubscr...@googlegroups.com
For more options, visit this group at http://groups.google.com/group/sage-devel
URL: http://www.sagemath.org
