- Asg - mailman3.cels.anl.gov

Milestones on Wiki
by Richard Brooks 03 Feb '12

03 Feb '12

Just got around to checking the wiki list of milestones. Due dates on the wiki do not match the times on the SoW that I have seen to date. Am curious as to the when and why of the change. Is this an error or is there a reason for this? I believe the wiki is in conflict with our contract. Am hesitant to just change the wiki page back to the contractual milestone without any knowledge of the background for these changes. It would also be better to tell the people involved, instead of just changing the wiki documents. With the amount of e-mail traffic on this server, it would not be productive for us to follow all comments in depth. -- =================== R. R. Brooks Associate Professor Holcombe Department of Electrical and Computer Engineering Clemson University 313-C Riggs Hall PO Box 340915 Clemson, SC 29634-0915 USA Tel. 864-656-0920 Fax. 864-656-5910 email: rrb(a)acm.org web: http://www.clemson.edu/~rrb

2 1

SSM Request: Unexpected Messages
by Kevin Harms 03 Feb '12

03 Feb '12

UAB, we'd like to make another request for SSM. (And i think this may be re-treading old ground, but I wanted to document it) For our server to handle unexpected requests, we'd like to be able to post a match_bits that contains a list/array/pool of buffers. In conjunction with the SSM_ME_NEVER_UNLINK flag, we'd want SSM to select a free buffer and then process the completion callback using that buffer. The NEVER_UNLINK flag would ensure that a client would not get an error when trying to match while another client request was arriving at the same time. There should be some method to return the buffer to the list/array/pool once the callback was done. We would also want an explicit error code for the condition where the match entry is live but the list of buffers has been exhausted so our clients/servers can differentiate this condition. That was a bit of a implementation specific request, but in general we just want a solution for easy handling of unexpected requests. thanks, kevin

5 13

overhead of blocking aesop call
by Dries Kimpe 02 Feb '12

02 Feb '12

I wrote a small test program where a function with two arguments (one int, one void *) is called from a timed for loop. Results below. This is with -O2. lts@today aesop % tests/blocking-overhead (master‣2…) Making 40000000 calls. Starting regular test... regular calls: 9.53674e-07 (2.38419e-14 / call) Starting blocking test... blocking calls: 2.23765 (5.59412e-08 / call) First number of total time in seconds, second is / call. I haven't looked at the assembler output to make sure that the compiler isn't inlining the regular call or just dropping it all together. (The only thing done within the call is an assignment to a variable). I haven't tried moving the function in another object file (which would make inlining much much for a compiler (and impossible for gcc)). I'm going to check the assembler output first. I did put a malloc()/free() in the regular call, just as a test to see how much the memory allocations contribute to the time. For the results below, 'regular calls' refers to a non-blocking call doing a 'free (malloc (100))', while the aesop blocking call was not modified. Making 40000000 calls. Starting regular test... regular calls: 0.946443 (2.36611e-08 / call) Starting blocking test... blocking calls: 2.24125 (5.60312e-08 / call) Since we cannot really get rid of the malloc/free except in certain special cases, this basically shows the best we can possibly hope to do. Using tcmalloc (Google's alternative malloc library) this reduces to: Making 40000000 calls. Starting regular test... regular calls: 0.664825 (1.66206e-08 / call) Starting blocking test... blocking calls: 1.47122 (3.67805e-08 / call) (or about ~30% improvement for both the simple malloc/free and the aesop blocking call). Remember: this is a synthetic benchmark! Results not guaranteed. No refunds if you don't see 30% improvement. :-) Adding 'pthread_mutex_lock' and 'pthread_mutex_unlock' to the REGULAR call (so that it contains both the malloc and pthread operations) changes the results to: Making 40000000 calls. Starting regular test... regular calls: 1.15233 (2.88083e-08 / call) Starting blocking test... blocking calls: 1.50091 (3.75227e-08 / call) Instead of using a pthread mutex, using OPA compare-and-swap on an integer makes it: Making 40000000 calls. Starting regular test... regular calls: 0.782022 (1.95505e-08 / call) Starting blocking test... blocking calls: 1.48202 (3.70506e-08 / call) Note that aesop internally already uses OPA atomics in certain places, but not everywhere. So, summary (numbers are seconds / call): regular call: 2.384e-14 regular with malloc/free: 2.366e-08 regular, malloc/free, tcmalloc: 1.662e-08 regular, malloc/free,tcmalloc,mutex: 2.881e-08 regular, malloc/free,tcmalloc,OPA cas: 1.995e-08 aesop blocking: 5.594e-08 aesop blocking, tcmalloc: 3.678e-08 In all cases, the aesop generated code still contains ae_debug_blocking () calls which translate to 'if (aesop_dbg_blocking) { ...}', but I don't think this affects the performance too much. Also: gcc 4.5.3, glibc 2.13-r4. Test code is in the repository. Dries

1 0

Downselection Process
by Kevin Harms 02 Feb '12

02 Feb '12

I added the 'downselect' stuff into its own page. I thought it warranted it. Are we going to do some documentation/report on selected some technology or are we just going to decide and document the decision? https://wiki.mcs.anl.gov/asg-private/index.php/Downselect kevin

3 2

Wiki update: Downselect
by Wiki update: Asg-private: Harms 01 Feb '12

01 Feb '12

1 0

Wiki update: Main Page
by Wiki update: Asg-private: Harms 01 Feb '12

01 Feb '12

1 0

Wiki update: Downselect
by Wiki update: Asg-private: Harms 01 Feb '12

01 Feb '12

1 0

Wiki update: Main Page
by Wiki update: Asg-private: Harms 01 Feb '12

01 Feb '12

1 0

aesop trac site fixed
by Phil Carns 31 Jan '12

31 Jan '12

We had some minor technical difficulties with the aesop trac site over the last few days, but it is back up and running now. i just wanted to mention it in case anyone tried to submit tickets and wasn't able to. thanks, -Phil

1 0

apparently there is listserv archive here
by Jill Gemmill 31 Jan '12

31 Jan '12

For some reason Clemson prefers to keep people's mailboxes forever, but does not archive listservs. Do we still want to try to archive prior email list conversation from P2Pfs list? (which would involve searching/filtering/exporting old mail) ? - Jill

2 1