| Anonymous | Login | Signup for a new account | 2010-07-31 05:06 CEST |
| Main | My View | View Issues | Change Log | Roadmap | Docs |
| Viewing Issue Simple Details [ Jump to Notes ] | [ View Advanced ] [ Issue History ] [ Print ] | ||||||||
| ID | Category | Severity | Reproducibility | Date Submitted | Last Update | ||||
| 0000022 | [OMNeT++] simulation kernel | tweak | always | 2008-12-05 16:11 | 2009-07-25 18:38 | ||||
| Reporter | baumgart | View Status | public | ||||||
| Assigned To | andras | ||||||||
| Priority | normal | Resolution | fixed | ||||||
| Status | resolved | Product Version | 4.0b8 | ||||||
| Summary | 0000022: Perfomance issue: cMessage::getArrivalTime() returns simtime_t by value | ||||||||
| Description |
cMessage::getArrivalTime() returns simtime_t by value. This is a performance problem especially for the FES (shiftup and insert). After changing the method to return const simtime_t& we had performance improvements of about 10-20 percent (compiled with -O3). Probably this should be changed for all simtime_t setters and getters in cmessage.h. |
||||||||
| Additional Information | |||||||||
| Tags | No tags attached. | ||||||||
| Attached Files |
 simtime_by_reference.diff [^] (941 bytes) 2009-03-10 12:03 |
||||||||
|
|
|||||||||
 Relationships |
|
|
Notes |
|
|
(0000160) andras (administrator) 2009-07-25 14:22 edited on: 2009-07-25 14:24 |
I did some measurements as well, with interesting results. cqn example (100 tandems, 50 queues/tandem), cmdenv express mode, ubuntu 32-bit, g++ 4.3.2. In debug mode (make MODE=debug), I saw the performance hit you described (~ 15-20% spent in getArrivalTime()). In release builds (-O2), this difference disappeared: original: real 0m15.372s user 0m10.257s sys 0m0.144s patched: real 0m15.377s user 0m10.193s sys 0m0.124s kcachegrind runs showed similar results (difference was only ~0.2%). Still, going on with the change looks like a good idea, because debug builds benefit a lot. It is an interesting question how this would affect performance on 64-bit architectures, where a SimTime object fits into a single machine register. There, returning a reference (which is actually a pointer) instead of the value might be a performance penalty. I did not make measurements in this direction yet. However, today most users are still on 32-bit architectures, so the change would benefit them. |
|
(0000161) andras (administrator) 2009-07-25 18:38 |
Introduced simtime_t_retval which is a typedef for const simtime_t&, and changed cMessage method return types to that. Via simtime_t_retval, it will be easy to turn off this optimization when it adversely affects performance (i.e. 64-bit optimized builds) |
|
Issue History |
|||
| Date Modified | Username | Field | Change |
| 2008-12-05 16:11 | baumgart | New Issue | |
| 2009-03-10 12:03 | baumgart | File Added: simtime_by_reference.diff | |
| 2009-07-25 14:22 | andras | Note Added: 0000160 | |
| 2009-07-25 14:23 | andras | Assigned To | => andras |
| 2009-07-25 14:23 | andras | Status | new => in work |
| 2009-07-25 14:24 | andras | Note Edited: 0000160 | |
| 2009-07-25 18:38 | andras | Note Added: 0000161 | |
| 2009-07-25 18:38 | andras | Status | in work => resolved |
| 2009-07-25 18:38 | andras | Fixed in Version | => 4.1 |
| 2009-07-25 18:38 | andras | Resolution | open => fixed |
| Mantis 1.1.5[^] Copyright © 2000 - 2008 Mantis Group |  |
