Search found 58 matches

by mpeterson
Mon Apr 02, 2012 9:58 am
Forum: General Development
Topic: Nvidia limiting gtx680 compute speed?
Replies: 20
Views: 17477

Re: Nvidia limiting gtx680 compute speed?

madd wrote:What's the best performance that can be achieved for the fastest Intel CPUs for AO rays in conference room?
i can just tell you results for fairy. on dual sandy (3.06ghz) around 290mrays/s with ray length of 0.25 * scene-diag.

mp
by mpeterson
Fri Feb 03, 2012 12:30 pm
Forum: Visuals, Tools, Demos & Sources
Topic: complex rendering
Replies: 10
Views: 9506

Re: complex rendering

Wow, amazing quality at that speed! Is this Whitted-style ray tracing? more or less yes. we trace refl/refr at the same time (no mc here). so its exploding after a time (recursion depth 15). we block/discard rays when their energy drops below a threshold. hdr is pre-processed and mc sampled with 16...
by mpeterson
Fri Feb 03, 2012 11:16 am
Forum: Visuals, Tools, Demos & Sources
Topic: complex rendering
Replies: 10
Views: 9506

complex rendering

enclosed a first testshot of our new rt-engine. its a head lamp with around 2.5mio ploygons. here we used a combination of our multi-frustum traversal + ray sorting/streaming. on a dual sandy-bridge system 6fps are possible at 1920x1080. there is no progessive rendering like elsewhere. the final qua...
by mpeterson
Fri Jan 27, 2012 2:28 pm
Forum: General Development
Topic: Dissapointing building times from lbvh gpu based builder
Replies: 35
Views: 27693

Re: Dissapointing building times from lbvh gpu based builder

dr_eck wrote:
Optix now has a GPU based HLBVH builder, as of version 2.5. Has anyone tried to benchmark it yet?
Does anyone know if it is work queue based or original?
i think it is the work queue thingy because the original one has strong cpu dependencies.
by mpeterson
Fri Jan 13, 2012 12:25 pm
Forum: General Development
Topic: Dissapointing building times from lbvh gpu based builder
Replies: 35
Views: 27693

Re: Dissapointing building times from lbvh gpu based builder

first at all it is an academic paper... what we have seen is that it was never ever able
to produce high quality trees for 1mio polygons in 10ms. it was more in the range
of 100ms. and that is not that bad.
by mpeterson
Mon Jan 09, 2012 6:22 pm
Forum: General Development
Topic: AVX perf on dual sandy bridge
Replies: 5
Views: 7062

Re: AVX perf on dual sandy bridge

Interesting :-) So, I have a few questions too : - How many CPU do you have ? Is it on only 1 i7 CPU ? - Which frustrum based BVH technique have you used ? Is it something equivalent to the Kirill Garanzha paper ? - Have you test several scenes for coherent and incoherent rays ? Thx & congrats 2 cp...
by mpeterson
Mon Jan 09, 2012 3:18 pm
Forum: General Development
Topic: AVX perf on dual sandy bridge
Replies: 5
Views: 7062

Re: AVX perf on dual sandy bridge

Do you have any SSE (4-way) vs AVX (8-way) number ? It was discussed on old ompf.org and it was looking like AVX wasn't particularly better than standard SSE code. yes i have. avx is sucking a little bit because of missing integer ops, limited shuffles/permute etc. so i used different native implem...
by mpeterson
Fri Jan 06, 2012 3:28 pm
Forum: General Development
Topic: AVX perf on dual sandy bridge
Replies: 5
Views: 7062

AVX perf on dual sandy bridge

hallo, i wanted to share some early results from our multi-frustum bvh traversal on intels new dual socket sandy bridge machine (native avx implementation). on fairy we achieve up to 560mrays/s for coherent ray sets (1024x1024 img resolution). there are also some better instruction to convert from s...