Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom

Seems OK... #2

Closed
laprej opened this issue Jul 16, 2015 · 10 comments
Closed

Seems OK... #2

laprej opened this issue Jul 16, 2015 · 10 comments

Comments

@laprej
Copy link
Collaborator

laprej commented Jul 16, 2015

So when I run it like this:

mpirun -np 2 ./tnt_benchmark --synch=3 --buddy_size=24

Things seem to work OK? It was looking for an undefined function (neuronReverseState()) so I just commented those calls out.
@markplagge
Copy link
Owner

That is very interesting. Very interesting indeed. First off, the buddy size thing shouldn't need to be enabled -- it's only on if you use --delta as an option. Also, neuronReverseState() should be something important.

So you're saying that running in synch=3 gives the same results as synch=1?

Mind==blown

  • Mark Plagge
    Sent from my iPhone

On Jul 16, 2015, at 3:09 PM, Justin M. LaPre [email protected] wrote:

So when I run it like this:

mpirun -np 2 ./tnt_benchmark --synch=3 --buddy_size=24
Things seem to work OK? It was looking for an undefined function (neuronReverseState()) so I just commented those calls out.

—
Reply to this email directly or view it on GitHub.

@markplagge
Copy link
Owner

Okay, i'm going to take another look at my coat, make sure there's no major errors or omissions and doublecheck everything. If you're available tomorrow to do a quick glance at my code I would really appreciate a second set of eyes. Thank you!

  • Mark Plagge
    Sent from my iPhone

On Jul 16, 2015, at 4:07 PM, Mark Plagge [email protected] wrote:

That is very interesting. Very interesting indeed. First off, the buddy size thing shouldn't need to be enabled -- it's only on if you use --delta as an option. Also, neuronReverseState() should be something important.

So you're saying that running in synch=3 gives the same results as synch=1?

Mind==blown

  • Mark Plagge
    Sent from my iPhone

On Jul 16, 2015, at 3:09 PM, Justin M. LaPre [email protected] wrote:

So when I run it like this:

mpirun -np 2 ./tnt_benchmark --synch=3 --buddy_size=24
Things seem to work OK? It was looking for an undefined function (neuronReverseState()) so I just commented those calls out.

—
Reply to this email directly or view it on GitHub.

@laprej
Copy link
Collaborator Author

laprej commented Jul 16, 2015

Leaving off --buddy_size still runs... I just saw some of the delta calls in there so I figured maybe you were using that by default.

@carothersc-zz
Copy link
Collaborator

Mark -- it's winter .. you don't need a "coat" :-)).

Happy "coating" ..

Enjoy,
Chris

On Thu, Jul 16, 2015 at 4:16 PM, Mark Plagge [email protected]
wrote:

Okay, i'm going to take another look at my coat, make sure there's no
major errors or omissions and doublecheck everything. If you're available
tomorrow to do a quick glance at my code I would really appreciate a second
set of eyes. Thank you!

  • Mark Plagge
    Sent from my iPhone

On Jul 16, 2015, at 4:07 PM, Mark Plagge [email protected] wrote:

That is very interesting. Very interesting indeed. First off, the buddy
size thing shouldn't need to be enabled -- it's only on if you use --delta
as an option. Also, neuronReverseState() should be something important.

So you're saying that running in synch=3 gives the same results as
synch=1?

Mind==blown

  • Mark Plagge
    Sent from my iPhone

On Jul 16, 2015, at 3:09 PM, Justin M. LaPre [email protected]
wrote:

So when I run it like this:

mpirun -np 2 ./tnt_benchmark --synch=3 --buddy_size=24
Things seem to work OK? It was looking for an undefined function
(neuronReverseState()) so I just commented those calls out.

—
Reply to this email directly or view it on GitHub.

—
Reply to this email directly or view it on GitHub
#2 (comment)
.

Christopher D. Carothers

Director, Center for Computational Innovations
Professor, Department of Computer Science
Rensselaer Polytechnic Institute
110 8th Street
Troy, New York 12180-3590

e-mail: [email protected]
web page: www.cs.rpi.edu/~chrisc http://www.cs.rpi.edu/%7Echrisc
phone: (518) 276-2930

fax: (518) 276-4033

@markplagge
Copy link
Owner

Hahah, well I did leave some pointers in my coat - I’m always leaving them around

— Mark P. Plagge —
Graduate Research Assistant
Department of Computer Science
Rensselaer Polytechnic Institute
[email protected]

On Jul 16, 2015, at 4:33 PM, Chris Carothers [email protected] wrote:

Mark -- it's winter .. you don't need a "coat" :-)).

Happy "coating" ..

Enjoy,
Chris

On Thu, Jul 16, 2015 at 4:16 PM, Mark Plagge [email protected]
wrote:

Okay, i'm going to take another look at my coat, make sure there's no
major errors or omissions and doublecheck everything. If you're available
tomorrow to do a quick glance at my code I would really appreciate a second
set of eyes. Thank you!

  • Mark Plagge
    Sent from my iPhone

On Jul 16, 2015, at 4:07 PM, Mark Plagge [email protected] wrote:

That is very interesting. Very interesting indeed. First off, the buddy
size thing shouldn't need to be enabled -- it's only on if you use --delta
as an option. Also, neuronReverseState() should be something important.

So you're saying that running in synch=3 gives the same results as
synch=1?

Mind==blown

  • Mark Plagge
    Sent from my iPhone

On Jul 16, 2015, at 3:09 PM, Justin M. LaPre [email protected]
wrote:

So when I run it like this:

mpirun -np 2 ./tnt_benchmark --synch=3 --buddy_size=24
Things seem to work OK? It was looking for an undefined function
(neuronReverseState()) so I just commented those calls out.

—
Reply to this email directly or view it on GitHub.

—
Reply to this email directly or view it on GitHub
#2 (comment)
.

Christopher D. Carothers

Director, Center for Computational Innovations
Professor, Department of Computer Science
Rensselaer Polytechnic Institute
110 8th Street
Troy, New York 12180-3590

e-mail: [email protected]
web page: www.cs.rpi.edu/~chrisc http://www.cs.rpi.edu/%7Echrisc
phone: (518) 276-2930

fax: (518) 276-4033

—
Reply to this email directly or view it on GitHub #2 (comment).

@markplagge
Copy link
Owner

Yeah - the delta is a toggle option. And that error with neuronReverseState should not happen - the function is there. I’ll see if there’s something I’m missing.
— Mark P. Plagge —
Graduate Research Assistant
Department of Computer Science
Rensselaer Polytechnic Institute
[email protected]

On Jul 16, 2015, at 4:19 PM, Justin M. LaPre [email protected] wrote:

Leaving off --buddy_size still runs... I just saw some of the delta calls in there so I figured maybe you were using that by default.

—
Reply to this email directly or view it on GitHub #2 (comment).

@laprej
Copy link
Collaborator Author

laprej commented Jul 17, 2015 

me->free_q.size <= 1024

in the optimistic debug scheduler never seems to trip??? Manually setting that variable to a way lower value caused it to fire and subsequently run the reverse event handler.

@markplagge
Copy link
Owner

Okay, what exactly is that variable and where is it?

@gonsie
Copy link
Collaborator

gonsie commented Jul 17, 2015

That's the trip condition to start doing rollbacks in optimistic debug mode. You have too much event memory allocated if that never trips.

@laprej
Copy link
Collaborator Author

laprej commented Jul 17, 2015

http://carothersc.github.io/ROSS/ROSS-docs/docs/html/tw-sched_8c.html#ad49988d8dcde406fac7f4a2e11eac17f

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Assignees
No one assigned
Labels
None yet
Projects
None yet
Milestone
No milestone
Development

No branches or pull requests
4 participants
@gonsie @laprej @carothersc-zz @markplagge