Many thanks, Bart, (esp. for answering on the Labor Day weekend!).
Pointing to examples where people have done similar tracing would be
sufficient (and also time-efficient for you). I'll let Saeed follow-thru.
Best,
Ganesh
On 9/7/15 9:52 AM, Barton Miller wrote:
We'll do our best. Ganesh or Saeed, please send a summary of what's going on.
--bart
On Sep 7, 2015, at 10:48 AM, Dorian Arnold <darnold@xxxxxxxxxx> wrote:
Dyninst folks, please help Ganesh and his student Saeed as they try to
use dyninst in some of their research investigations. They can give
you more details as necessary, but the digest is that they have been
trying to use Dyninst to collect traces from multiple MPI processes,
each spawning threads. This effort only partially succeeds as they
can't seem to collect traces from all threads -- they can get traces
from all processes or one thread at a time.
Thanks and I'm sure they look forward to hearing from you,
- dorian
|
![[Computer Systems Lab]](http://www.cs.wisc.edu/pics/csl_logo.gif){kind=logo}