On Saturday, 23 May 2009 at 21:51, Christian Ohm wrote:
> After the first hour, the profile looked like:
> 16% openal sound mixing
> 6% path finding
> ...
> After the second hour, the pathfinding topped the list by far.
Last one for today: To see if this was the result of the camera position, I di
I also did some more testing with the OpenALsoft versions from 1.0 to 1.7. A
driver lockup ate the exact results, but it looked somewhat like this:
1.0 ~22%
1.1 ~22%
1.2 ~22%
1.3 ~11%
1.4 ~13%
1.5 ~15%
1.6 ~44%
1.7 ~51%
So 1.3 was the best, 1.4 and 1.5 got a bit worse, and 1.6 and 1.7 got a lot
w
Am Samstag, 23. Mai 2009 19:00:27 schrieb bugs buggy:
> This is with release build, same map, autogame, and at position 4
> again. ~15 mins before I quit the game:
> Pathfinding is still the colossal giant.
Could it be that this depends on the compiler and its ability to optimise?
--DevU
signat
On 5/23/09, Per Inge Mathisen wrote:
> On Sat, May 23, 2009 at 4:04 PM, Christian Ohm wrote:
> > This is from an optimized build (CFLAGS="-g -O2"), so fpathCompare and
> > fpathOpenGet are inlined. I didn't get to 30 minutes though, it crashed
> after
> > ~10 minutes (crashdump attached, auto
On 5/23/09, Christian Ohm wrote:
> After reading the IRC logs I did a test run with various OpenAL-soft versions
> (compiled from source with the default settings via "cmake .&& make"), and
> while 1.5 didn't seem quite as bad as the others, the results are:
>
> samples %linenr info
On Sat, May 23, 2009 at 4:04 PM, Christian Ohm wrote:
> This is from an optimized build (CFLAGS="-g -O2"), so fpathCompare and
> fpathOpenGet are inlined. I didn't get to 30 minutes though, it crashed after
> ~10 minutes (crashdump attached, autogame on ziggurat, with demomode to
> simulate playin
