02:09.32 |
Twingy |
hrm |
02:09.42 |
Twingy |
my precioussss |
02:11.44 |
Twingy |
hrm, 52 days till I move |
02:13.21 |
Twingy |
645,277 rays/sec on a 2.4ghz xeon, and
1,048,576 rays/sec on a 2ghz opteron |
02:13.25 |
Twingy |
hrm, that number looks funny |
02:13.34 |
Twingy |
hah |
02:13.37 |
Twingy |
it's 1024^2 |
02:13.45 |
Twingy |
that's highly amusing |
04:55.41 |
*** join/#brlcad PKMOBILE
(n=Apathy@pcp0011645240pcs.aberdn01.md.comcast.net) |
13:04.47 |
Maloeran |
Hehe Twingy, I wasn't kidding about the
floating point crunching power of the Opterons, and how they flow
through unpredictable branch with no penalty |
13:05.45 |
Maloeran |
I hope this benchmark was ran on 64 bits too,
with a -mtune=opteron so that branches are silently changed into
conditional move instructions ( unlike on all Intel chips,
conditional moves are as fast as regular moves ) |
17:21.26 |
Twingy |
hrm, a 1.5% speedup |
19:13.28 |
*** join/#brlcad CIA-5
(n=CIA@flapjack.navi.cx) |
19:17.02 |
CIA-5 |
BRL-CAD: 03twingy * 10brlcad/src/adrt/ (6
files in 3 dirs): stuff to fix 32/64 communication |
19:22.41 |
Twingy |
aha |
19:25.38 |
CIA-5 |
BRL-CAD: 03twingy *
10brlcad/src/adrt/isst/master/master.c: 32/64 comms fixed, time to
fix observer-fu |
19:32.19 |
CIA-5 |
BRL-CAD: 03twingy * 10brlcad/src/adrt/isst/
(master/master.c observer/observer.c): 32/64 observer comms should
work |
19:43.00 |
Twingy |
hrm |
19:43.10 |
Twingy |
that's highly amusing |
19:47.56 |
Twingy |
apparently I've only been running at 50% speed
the whole time I've been doing distributed stuff |
19:48.15 |
Twingy |
something with freebsd's threading I
think |
19:52.36 |
*** join/#brlcad iSteve
(n=isteve@isteve.bofh.cz) |
19:55.17 |
Twingy |
hrm |
20:18.58 |
Twingy |
mmm 12.25 mil rays/sec |
20:20.11 |
Twingy |
hrm, might be able to get that up to 14
mil/sec by tonight |
21:18.01 |
*** join/#brlcad cad478
(n=43b0389f@bz.bzflag.bz) |
21:18.23 |
Twingy |
back |
21:52.11 |
Twingy |
sean, you alive? |
22:10.57 |
Twingy |
hrm |