GPU benchmarks

[auke]

Some high-end server motherboards have 8 PCI/e x 16 slots spaced for double height GPU cards to be placed. The most I have seen on consumer and workstation boards is 7 though, and always spaced for single slot cards, which means you could fit 4 cards next to each other at most. If you want to be creative you could make a custom case to mount 7 double height cards in and connect them to the motherboard via PCI-risers. Seen mixed results with that method though so I would't go down that road with little experience in pc building.

[kosso_olli]

Sorry for the noob questions, but how do you manage to get up to 8 Titans working in a single machine?

[eyepiz]

Benchmark Version 2.0

Titan X: 10m 29.4s
Titan X + 3 Titans: 3m 27.9s

Nvidia Driver 355.82

[cray-zard]

my freakin friend fooled me, he wanted me to feel bad for purchasing the ti's, jokes on me.. feeling too stupid for posting this..

[super gnu]

its hard to tell from your post, if you realise that was a joke article..

[cray-zard]

cant wait for someone to bench this 990ti

http://wccftech.com/nvidia-gtx-990-t...rked-reviewed/

[3LP]

The driver version is written right next to the name of the card in the benchmark so it's easy for people to have that info as it can be quite important.
Anyways, I'm using the 350.12 on my 4x Titan X and the 353.30 on my 3x Titan X rig.
Both work as expected with time, but I use EVGA to force fans at 100% as I don't want and speed reduction managed from the card heat.
Actually I even OC my cards at +215 this means my 7 card OC rig is actually equivalent to a 8 default card rig

Stan

[super gnu]

im using 353.06 for my titan x and im getting results in line with expectations.

[cray-zard]

here it is,

[auke]

Nice times! Which Nvidia driver are you using?

[cray-zard]

i re-rendered both benchmark scene just to be sure with my results





1:32 version 1.0

9:56 version 2.0

and still waiting for my 2nd zotac 980ti ref.

[auke]

Could anybody who is using the Titan-X for rendering and getting times comparable to the times listed in the benchmarks tell me what driver they are using? I'm still having trouble getting it up to speed and heard there can be large differences between the used drivers. It's a bit disappointing to see the Titan-X only reach the benchmark scores of the GTX-970 :P.

[auke]

Thanks for the quick reply! Definatly no throttling going on though, while rendering it's at a steady 1130Mhz all the way through, temperatures stabilize around 77 to 82 degrees which is the normal operating temperature for the Titan X (the fan will only start serious work when it goes over 83 degrees, a function meant to keep the card silent while gaming.)

Will try to see if there are any differences while rendering it from another workstation using DR, otherwise I'll just wait for a new node with just 2 TitanX's to try again and see if it's the different GPU families causing the trouble.

[super gnu]

7000 mhz is the DDR memory rate. the actual clock to the chips is shown in nvinspector, which is 3500 mhz ( well actually i believe the actual clock is much lower than that, there are several multipliers involved. ) p2 slows the clocks to 3000 mhz, or 6000 mhz effective. so an oc of the memory to 4000 mhz gives you a final memory clock of 8 ghz

from my testing, the memory clock gives about the same, if not slightly more, benefit than a core clock increase, which is why a p0 state for RT would be quite desireable.

have you checked if your chip is throttling when you run RT? what are your temps/ clock speeds for the chip shown as in nvinspector or gpu-z when rendering?

pcie bandwith has been discussed here several times before. since rt loads the whole scene into ram, the only major penalty for a slower pcie link is longer load times, and less interactivity in activeshade, although i believe once you get down to pcie 1x levels things do start to suffer. no issue with a pcie3.0 8x link

i suspect your issue is either, as you mentioned, some driver issue related tot he quadro and the titan in the same setup, or clock throttling due to over power/overheat of the titan (although in my testing power draw when using RT is much lower than a heavy graphics benchmark such as furmark) could also be a particular issue with the driver revision you are using.


btw thanks for asking me to check my pcie link status.. for some reason my old gtx670 is running 3.0 @2.0 x16, but my titan x is running 3.0@1.1 x8 ! thats terrible. i need to check the slot spec i have it in. anyway its a good illustration as to how pcie bandwith isnt a major limit.. i get 7 minutes something on benchmark 2 on my oc titan x.

[auke]

How do I find out what P-state?

Not sure if there are any official utilities that check this but I'm using NVinspector.