Titan Black SLI. Some benches.

[Face2Face]

Hey Keys, try running the same test with one card enabled and compare from there.

 

[Keysplayr]

I can try that later tonight.

 

[toyota]

So... the CPU isn't making that big of a difference, really? What is Ivy's IPC advantage over Sandy, 5-10%?

as I already mentioned it's also HT that's helping out by about 10 percent with that level of GPU power it's having to push in this game.

 

[SlowSpyder]

as I already mentioned it's also HT that's helping out by about 10 percent with that level of GPU power it's having to push in this game.

Right, more threads (moar cores). I get that, but it would seem even with Keys older modestly clocked i5 the CPU really isn't that big of a deal on even these higher end builds.

Don't get me wrong, if I had two cards of that level in SLI, I would probably grab a faster CPU too.

Maybe I'm reading the benches wrong?

 

[Face2Face]

as I already mentioned it's also HT that's helping out by about 10 percent with that level of GPU power it's having to push in this game.

Toyota, do you own Metro LL? If so, do a run so we can compare. I would be interested to see what your GTX 780 does with a 4770K.

Here is Metro LL run again at 1920x1080, Tesselation off, and PhysX off. Everything else is same. So I'm the same as Desperado's first run.

Just did a run with a single 780 OC'd with the same settings.

 

[toyota]

Toyota, do you own Metro LL? If so, do a run so we can compare. I would be interested to see what your GTX 780 does with a 4770K.




Just did a run with a single 780 OC'd with the same settings.

http://i1118.photobucket.com/albums/k619/mwilding/Capture_zps5b9f2844.png

your in game boost is 20% higher than what I am running. your cpu is also oced enough to surpass what little IPC advantage I have. my only advantage is HT which is not going to make much difference on single card.

still I got 57 fps in the bench on those settings which is great considering what I just said above.

 

[tviceman]

wow so still 77 fps. so even taking into account the slightly lower res you are smoking his setup considering your cards are actually slower.

Overclocking gtx780 has great scaling though. The ROP performance is the same as Titan Black or 780 TI, so it's only in shader-bound situations that fully enabled GK110's can separate themselves from regular OC'd 780's.

 

[Face2Face]

your in game boost is 20% higher than what I am running. your cpu is also oced enough to surpass what little IPC advantage I have. my only advantage is HT which is not going to make much difference on single card.

still I got 57 fps in the bench on those settings which is great considering what I just said above.

Very nice, so what were you final clocks on the GTX 780 when running the bench? Also, what was your memory clock speed and do you have Hynix?

 

[toyota]

Very nice, so what were you final clocks on the GTX 780 when running the bench?

just 1124 boost and 6210 memory. I could oc more but anything above 1137 requires a bump in voltage. max the memory can do is 6500.

 

[Face2Face]

just 1124 boost and 6210 memory. I could oc more but anything above 1137 requires a bump in voltage. max the memory can do is 6500.

Thanks for the info. I just dropped the core clock down to 1200Mhz and it finished at 59 fps. Seems this game doesn't scale to well with overclocking.

Edit - Running it again at your same clock speeds.

Seems like you are faster than me at the same GTX 780 clock speeds - I am at AVG 55 fps.

 

[Keysplayr]

Allright, here is a run with a single TitanB on i5 2500K @ 4.5GHz
Same setting as you did above.

 

[Face2Face]

Allright, here is a run with a single TitanB on i5 2500K @ 4.5GHz
Same setting as you did above.

Nice fps, considering that's stock boost. I wonder how much better it would do in games like Tomb Raider - Bioshock Infinite? Both have built-in benchmarks. Do you happen to own those games?

 

[Keysplayr]

Heaven and Valley added.