(06-21-2011, 04:00 PM)Decimus10 Wrote: [ -> ]May I try out the previous L103 build? I didn't have the opportunity to download the x86 version (due to issues with Megaupload) and I have an Intel processor.
If there's anything that I can do to help, please give me some instructions on what 'tests' to run.
Sorry if I'm late to the scene.
(06-21-2011, 04:47 PM)KnilDark Wrote: [ -> ]But the L103 build has only extra optimizations for AMD processors, as I have understood, will not help you nothing compared with the L102 build.
KnilDark is right. Since you have an Intel processor, when you ran the program it
would act no different than it's L102 counterpart, as you would be running the same code
path. The L103 builds were useless, but the idea behind them wasn't. I have to find out
what code is incompatible with /arch:SSE3 and if/when I do, I can fix it and it will run better
on AMD processors. Even if that happens though, it will still run the same as an L102
build on Intel processors.
r7612 is up and ready for Download!
-had this one done day before yesterday, but then computer started acting up
and I had to format it and re-install everything in order to update the files on
my site. yay microsoft officelive buggy version of webdav
meh, oh well. I've pretty much become a pro at quickly wiping hard drive and
re-installing everything and copying all my files and stuff over. Synctoy comes
in mighty handy
I've an i7, so i've tried your release but it's as fast as the normal release, maybe even slower a bit :/
There's an other advantage ? or it changes nothing on some games and speed up others ??
Sorry for my english :/
Ps: Windows 7 64bit, i7 2630qm, Nvidia GT 540M
Edit: i've tried your 7608version in SSE4 (x64)
Edit2: i dont criticize your work, it's nice, i ask just some questions ^^"
Thx for the hard work Lectrode!
(06-23-2011, 10:20 PM)Armas Wrote: [ -> ]I've an i7, so i've tried your release but it's as fast as the normal release, maybe even slower a bit :/
There's an other advantage ? or it changes nothing on some games and speed up others ??
Couple things:
-OpenGL will show little (if any) speed improvements. This is because it is still compiled with Visual Studio
(So to see any speed difference you'll have to use DX9 & DX11)
- ... had other points but I completely forgot what I was going to say lol
(may edit this post once I remember them tomorrow)
ICC is the best compiler out there as far as optimizations go. If you're not witnessing speed
improvements, your computer may be so fast on its own that any speed improvement given
by ICC optimizations are negligible.
Also, make sure to adequately test the two against each other (no other programs running,
let run for ~15 seconds before you start keeping track of FPS). Fraps is awesome program
for testing.
Lastly, make sure you have both framelimit and audiothrottle OFF. They will skew the speed
results if left on.
Thx for you comment
(06-23-2011, 11:26 PM)Spit-Nick Wrote: [ -> ]Thx for the hard work Lectrode!
your welcome
I use Direct 3D9, i've disable the audio throttle and i've set frame limit to "off" but, on timesplitters 2 (i've test juste that game), the 2 builds up to 120 fps +/- 2 =x
I've tested with the same config: (3x native resolution, enable openCL,openMP, disable fog and fast mipmaps. Should i do something other to see any different ? Or maybe should i test on a game that's slow on the normal release ?
Thanks in advance for your reply ! =)
(06-24-2011, 08:39 PM)Armas Wrote: [ -> ]I've tested with the same config: (3x native resolution, enable openCL,openMP, disable fog and fast mipmaps.
There is NO point in doing that.
Both of those options usually slow down texture decoding. And texture decoding speed almost never affects performance.
(06-24-2011, 11:10 PM)NaturalViolence Wrote: [ -> ]Both of those options usually slow down texture decoding. And texture decoding speed almost never affects performance.
So it's a useless option :o
OpenCL sure is. The only reason it was kept is that some people argued that "some day someone might" come around and fix it to use the GPU in a more effective way...