| 01:10.50 | *** join/#brlcad guu (i=guu@myth.gibbscam.com) | |
| 01:17.32 | *** join/#brlcad CIA-16 (i=cia@cia.navi.cx) | |
| 01:39.00 | *** join/#brlcad CIA-5 (i=cia@cia.navi.cx) | |
| 01:46.40 | *** join/#brlcad CIA-5 (i=cia@cia.navi.cx) | |
| 02:03.59 | *** join/#brlcad CIA-4 (i=cia@cia.navi.cx) | |
| 02:30.55 | *** join/#brlcad tegtmeye (n=tegtmeye@pool-71-248-83-186.bltmmd.east.verizon.net) | |
| 02:41.02 | Twingy | http://js.cx/~justin/images/livingroom/ |
| 02:41.05 | *** join/#brlcad CIA-5 (i=cia@cia.navi.cx) | |
| 03:16.22 | *** join/#brlcad CIA-5 (i=cia@cia.navi.cx) | |
| 03:40.34 | *** join/#brlcad CIA-5 (i=cia@cia.navi.cx) | |
| 04:11.35 | *** join/#brlcad CIA-5 (i=cia@cia.navi.cx) | |
| 04:44.21 | *** join/#brlcad CIA-5 (i=cia@cia.navi.cx) | |
| 05:02.24 | *** join/#brlcad CIA-5 (i=cia@cia.navi.cx) | |
| 05:27.18 | *** join/#brlcad CIA-5 (i=cia@cia.navi.cx) | |
| 05:52.47 | *** join/#brlcad CIA-4 (i=cia@cia.navi.cx) | |
| 05:58.23 | *** join/#brlcad CIA-4 (i=cia@cia.navi.cx) | |
| 06:23.34 | *** join/#brlcad CIA-5 (i=cia@cia.navi.cx) | |
| 06:49.27 | *** join/#brlcad CIA-5 (i=cia@cia.navi.cx) | |
| 07:02.51 | brlcad | CIA is ill this evening |
| 07:03.03 | brlcad | navi is down |
| 07:23.17 | *** join/#brlcad CIA-5 (i=cia@cia.navi.cx) | |
| 07:40.09 | *** join/#brlcad CIA-5 (i=cia@cia.navi.cx) | |
| 08:01.07 | *** join/#brlcad CIA-5 (i=cia@cia.navi.cx) | |
| 08:26.30 | *** join/#brlcad CIA-5 (i=cia@cia.navi.cx) | |
| 08:46.33 | *** join/#brlcad CIA-4 (i=cia@cia.navi.cx) | |
| 08:57.13 | *** join/#brlcad CIA-5 (i=cia@cia.navi.cx) | |
| 10:20.49 | archivist | I expected to see the milling machine in the living room |
| 14:16.17 | *** join/#brlcad tegtmeye (n=tegtmeye@bz.bzflag.bz) | |
| 16:45.12 | *** join/#brlcad pra5ad-werk (n=803f2049@bz.bzflag.bz) | |
| 16:45.34 | pra5ad-werk | brlcad: at home? |
| 16:45.42 | brlcad | yes |
| 16:45.47 | pra5ad-werk | dammit |
| 16:45.55 | pra5ad-werk | where's jason |
| 16:46.23 | brlcad | in mexico all weekend |
| 16:46.37 | brlcad | hopefully still hung over |
| 16:46.44 | pra5ad-werk | oh |
| 16:46.50 | pra5ad-werk | u in tomorrow? |
| 16:48.33 | brlcad | possibly |
| 16:48.52 | brlcad | i'm always in somewhere |
| 16:49.46 | pra5ad-werk | ew |
| 16:49.52 | pra5ad-werk | too much info |
| 16:49.54 | pra5ad-werk | =) |
| 17:30.29 | *** join/#brlcad pra5ad-werk (n=803f2049@bz.bzflag.bz) | |
| 17:56.55 | ``Erik | heh |
| 17:57.33 | ``Erik | you shoulda came in today, sean, I was looking for someone to pawn racking a machine off on |
| 18:00.31 | tegtmeye | sean, you know gcc can do profiling opts like the intel compiler can right? |
| 18:13.27 | learner | yeah, would love to see the results |
| 18:16.08 | learner | i did mention that was without going into any great length on either compiler, add -O3 to each and what are the results, add no options, what are the results, etc |
| 18:17.12 | learner | the multi-pass gcc would be nice to see, though I would be still be incredibly surprised if it closes the gap fully |
| 18:19.53 | brlcad | ``Erik: aww, darn.. and I missed that? shucks |
| 19:06.37 | Maloeran | Just -O3 is a very poor benchmark with GCC, especially on AMD chips... |
| 19:33.32 | learner | give a list of ia-64 flags that should be optimal, and I'll give it a whirl |
| 19:33.57 | learner | fast-math, mtune, expensive-opts, not much different |
| 19:34.51 | learner | i.e. this is an itanium2, an altix |
| 19:39.27 | learner | i've gone through more than a dozen compiles now, and in each case the intel compiler is out-performing with the exception of specifying no optimization options gcc takes it |
| 19:40.38 | learner | which really just isn't so surprising. intel compiler only supports a couple platforms, you'd hope that they'd do a decent job |
| 19:43.42 | Maloeran | Ah yes, I didn't know the arch you were benchmarking, GCC's optimisation for IA-64 isn't anywhere near Intel's |
| 20:01.32 | *** join/#brlcad kierra (n=Les@fl-71-0-162-17.dyn.sprint-hsd.net) | |
| 20:01.54 | *** part/#brlcad kierra (n=Les@fl-71-0-162-17.dyn.sprint-hsd.net) | |
| 20:16.21 | ``Erik | <PROTECTED> |
| 20:16.23 | ``Erik | gheh |
| 20:18.01 | tegtmeye | we where talking today, you know that -fast-math turns off ieee math right? |
| 20:32.13 | brlcad | yep |
| 20:32.48 | brlcad | brl-cad doesn't rely on ieee floating point |
| 20:36.04 | brlcad | the benchmark test the actual results of the computations to ensure that they are within an accepted level of deviation (zero off-by-many pixel errors) |
| 20:36.20 | brlcad | if they don't match up, the benchmark will fail |
| 21:03.49 | ``Erik | make |
| 21:03.50 | ``Erik | ghah |
| 21:08.26 | brlcad | make: *** No targets specified and no makefile found. Stop. |
| 21:10.37 | ``Erik | gmake -j99999999 |
| 21:11.28 | brlcad | gmake: command not found |