OTOY Forums

Hey guys, I owe : Nvidia GTX 570 , Nvidia GTX 690 and NVidia GTX 670,

Cam Octorender Support al 3 of them and actually use them at all for best performance ?

Also, I do have this Available Mother Boards : Asus PX9X79 , 2X Gygabyte Z77x-UP4 TH , basically 3 computers. I have no idea how to arrange them based on the fact that the extension PCI express 3 on the Gybabyte is 16x 8x and 4x ( ellectronically ? - i dont what that means ).. Any sugestion Please ? I am using C4d and 64 gb ram.

Yes, you can use all 3 with Octane render and it will utilize all 3 at 100% scaling. (just combine the speed of all 3 individually and you will get your final speed)

I didn't take the time to look at the motherboards, but from what I understand the speed of the bus only affects loading everything into the GPU, but does not affect the render speed. I might be wrong on that though.

Can anyone confirm me that even if the Bus speed is 8x or 4x, the render speed will be the same as having all 3 of them on 16x ?

Hi Euandrei,

Yes, the render speed (Ms/sec) shouldn't be adversely affected by a 8x or 4x buss.
It should only affect scene load speed and even then it might not be noticeable.

Thanks
Chris.

Here are some notes from a system I am building at the moment testing with 3 GTX 580's. PCIe slot#2 (primary graphics) and #5 are x16 and #7 is x4 speed. Aircooled at the moment. Ran the benchmark with pt alpha shadows off @ 1000 iterations. There is a 2 slot gap between 2 and 5, while #5 is sandwiched between slot #7.

Test with 2 active GPU's
7.94ms
#2 (active) 83degrees
#5 (active) 89 degrees (temp rose quickly than the top slot)
#7 **turned-off** idle temp mid 30's-40's

7.77ms and 7.75ms avg
Using just two active GPU's either #2 + #7 and #5 + #7 (combination between x16 + x4 pcie)
Interestingly slot#7 x4 pcie temps stayed below 70 degrees, averaged mostly in the mid 60's throughout the test. Something to do with it being x4 speed? Or just because no hot card below it lol.

11.73ms all 3 active. Slot#5 temps skyrocketed quickly, scary. Wasn't paying attention on an earlier test and stopped it as it soared mid 90's. While the card above was mid 80's and lower card in the 70's, so definitely something to watch out for.

Power usage averaged 700w for 2 cards on load and apx 850~900 for all 3 cards during this quick test.
The higher the GPU temperature rose the more power usage increased incremently as I watched the watt gauge and gpu temp meter.

edit update: I swapped out the two hot cards after some testing. The 3rd GTX card I bought from a diff vender, and this card stayed leveled at 70degrees 5 min on the benchmark (tested on primary and second slot) whereas the pair from a another vendor temps just shot up at 80+degrees in 5min. I plan to take these apart to see if the cooling is off somewhere, but in any case a gpu waterblock is in order now.

Win 7 x64 | Gigabyte SOC GTX 580's | AMD 8350 8-core 4Ghz | 16GB ddr3