Page 12 of 62
Re: OCTANE BENCHMARK: Submit performance scores for your rig
Posted: Sun Dec 30, 2012 4:46 pm
by igen
I try to make a list benchmark from all post here .... maybe not to right but I try to read again after I make the list and I little bit confusing after see it, because the benchmark use various version from octane and some member use single gpu and some member use multi gpu.
I find some vga with same spec get different score, do you think different manufacture cause that ??
Re: OCTANE BENCHMARK: Submit performance scores for your rig
Posted: Sun Dec 30, 2012 5:29 pm
by Refracty
Wow, well done research work.
Many manufacturers and users overclock their cards so some variations will happen.
Re: OCTANE BENCHMARK: Submit performance scores for your rig
Posted: Sun Dec 30, 2012 6:45 pm
by igen
because the benchmark use various octane version make little bit confused, why not make more better list (octane v.1) and put in page 1 or at home page.
maybe the benchmark suppose tested with 1 gpu (multi gpu for optional) and another detail such as cuda core, vga manufacture & motherboard.
cheers

Re: OCTANE BENCHMARK: Submit performance scores for your rig
Posted: Sun Dec 30, 2012 8:43 pm
by Refracty
An automated benchmark system like Maxwell's would be nice to record information easier.
Re: OCTANE BENCHMARK: Submit performance scores for your rig
Posted: Sun Dec 30, 2012 9:49 pm
by face_off
GTX 550Ti, 1GB. Win 7, i7, 8GB RAM.
1.11 Ms/sec.
Re: OCTANE BENCHMARK: Submit performance scores for your rig
Posted: Mon Dec 31, 2012 12:37 am
by aoktar
also write me with new rig

2 x gtx 590
pathtracing 15.40 mb/s
Re: OCTANE BENCHMARK: Submit performance scores for your rig
Posted: Wed Jan 02, 2013 10:05 pm
by ASyme1
Hi guys-
Here's an official benchmark for my rig with the four 680's humming away.
PT with alpha shadows turned off
14.18
Nice. Seems about right as Octane's 4X690's is about twice as fast.
Alec
Re: OCTANE BENCHMARK: Submit performance scores for your rig
Posted: Wed Jan 02, 2013 11:40 pm
by FrankPooleFloating
igen -- would it not make more sense to list them in order of Pathtracing, rather than cores?... and perhaps divide 600/Kepler series cores by 3 (and call attention to this at top - that the Keplers are tweaked to reflect Fermi cores)?
Re: OCTANE BENCHMARK: Submit performance scores for your rig
Posted: Thu Jan 03, 2013 12:22 am
by FooZe
cfrank78 wrote:Hi Buzby,
Here are the specs for that machine:
CPU: Intel i7 3770 with Intel liquid cooling DRX-B
MOBO: Gigabyte G1.Sniper 3 Socket 1155
http://www.gigabyte.co.nz/products/prod ... id=4169#ov
RAM: 16GB 2400MHz G.SKILL
PSU: EVGA SuperNOVA NEX1500 Classified 1500W
OS: Win7 64bit
Case: Cooler Master "Storm Trooper."
NVIDIA Drivers: 306.97
(And of course the 4x 690's, or whatever combination of cards we need to test.)
Thanks
Chris.
Hi Chris!
What about the heat in the Thunderstorm? 4 690 so near together? Can they breathe? What kind of 690 do you use with which kind of cooler? Windforce? Twinforce?
Thanks - Kind regards Chris!
It does get hot, we are using stock EVGA 690's.
EDIT: These are not EVGA's - they are Gigabytes (sorry most of our cards we get from EVGA so i just assumed they were!).
I noticed that they are designed a little bit thinner than a full 2 slots, especially in the area of the fan in the center.
They blow air from the center, out the back, but also out the other end into the case, this was a problem before some modifying of the position and direction of the fans in the case.
We aren't a graphic artist/VFX house so we don't punish this machine with days of non stop rendering, but from some 30min or so tests (long enough for temperatures to settle), without tweaking fans, i was getting about 92 deg C out of one of the cores (temps ranged down to about 80 deg, depending on the core). That was with an ambient of about 25 deg C here. For stock cooling etc i think this is pretty good, and within the tolerance limit. After tweaking the case fans to vent the hot air coming out of the case side of the GPU's straight out the front, the temps dropped to the high 80's.
Thanks
Chris.
Re: OCTANE BENCHMARK: Submit performance scores for your rig
Posted: Thu Jan 03, 2013 12:33 am
by FooZe
buzby wrote:FooZe wrote:Hi Buzby,
Here are the specs for that machine:
CPU: Intel i7 3770 with Intel liquid cooling DRX-B
MOBO: Gigabyte G1.Sniper 3 Socket 1155
http://www.gigabyte.co.nz/products/prod ... id=4169#ov
RAM: 16GB 2400MHz G.SKILL
PSU: EVGA SuperNOVA NEX1500 Classified 1500W
OS: Win7 64bit
Case: Cooler Master "Storm Trooper."
NVIDIA Drivers: 306.97
(And of course the 4x 690's, or whatever combination of cards we need to test.)
Thanks
Chris.
Hi Chris, and great thanks. I think you guys really are the benchmark that we should be going against in this render test as well as purchasing equipment as you are at ground zero. One question, is the liquid cooling DRX-B for the CPU overclocking just for the sake of having a speedier processor? I have read the CPU does not have any relevance for Octane Render.
I read about the G1.Sniper but opted for the ASUS P8B WS LGA 1155 Intel C206 ATX Intel Xeon E3 Server/Workstation Motherboard as it got much better stability reviews, glad to see yours is working well. Hope to be posting with my 3 GTX 580 and 1 GTX 670 soon.
Cheers!
Our needs are not necessarily the same as others - so our machine doesn't fit everyone, so it's best not to just run of this one. The liquid cooling is not a necessity. The main reason was heat, get as much out of the case as possible. Having a standard air cooler inside a hot case probably isn't the best environment for the CPU! The CPU is not overclocked, you are right, this has no bearing on render speed. For animation production a fast CPU for speeding up voxelization and more than 16GB of RAM may be more ideal, but this machine suits our purposes well; testing various combinations (and quantity) of GPU's.
Thanks
Chris.