Hey Guys,
Recently got my hands on this EVGA GTX 1070 SC. I'm running it at the stock (out of the box) speeds. I absolutely LOVE this card, and of course I felt like doing some benchmarks using the unigine tools.
 
Note: I am aware that my stock clock i7 2600K and the fact that it's running through PCI-E 2.0 means there will be some bottleneck, research indicates around 2-5%.
 
I ran Heaven 4.0 @ 1080p (Ultra, Extreme Tess, and 8xAA) and achieved:
 
92.5 FPS
2329 Score
 
I then ran Valley 1.0 @ 1080p (Ultra, 8xAA) and achieved:
 
83.3 FPS
3487 Score
 
As you can see the Heaven 4.0 scores are within normal range, while factoring in the possible bottleneck reasonably, however the valley scores are a decent amount lower than what I've been seeing around the forums, which is high 90s to 100s easily. Any thoughts on this matter?
GPU Usage in heaven 4.0 stays mostly pegged at 98/99 with occasional dips to 80s at the lowest that don't last more than a couple seconds.
GPU Usage in valley is maxed a much lower amount of the time, though valley has some more long-distance scenes like the one high up in the sky where GPU usage drops to as low as 30% momentarily, and overall it spends a lot more of it's time in the 60-80% during this benchmark, and fluctuates a lot, especially during scene changes it drops.
 
Note: GPUZ Render test does not peg usage at max, it fluctuates.
 
Note: In Witcher 3 Ultra with absolutely everything maxed except for Nvidia Hairworks which is OFF I get between 70-85 FPS and GPU usage of between 70-90%. This result correlates better with the expectations due to the bottleneck , and with Vsync on it's incredibly smooth and never dips below 60FPS so I'm not complaining but the valley result has me somewhat intrigued, since it seems a bit out of proportion.
 
I have set Prefer Maximum Performance in Nvidia control panel for all the above programs, including GPUZ.
 
Thanks in advance for your thoughts on whether these results are nominal.

Well this is very disturbing ... my Gtx 1080 FTW getting 4154 score in Valley and 2820 in Heaven... looks low to me if your only short of 600 with gtx 1070. i guess something is very wrong with my setup also :S. 

arestavo
What you see is the best that you can expect for a stock clocked 2600K CPU.

It is holding back that 1070 a bit, and if you want to minimise that you'll need to overclock. The 2600K over clocks easily, why aren't you? Running a stock Intel cooler?

Yeah this is what I was thinking, it correlates almost perfectly with the expected bottleneck, except for Valley where there's a greater differential (which is why I posted mainly), although still great.
I have an aftermarket CPU cooler, keeps it at a nice 45C~ under full load and I USED to run it overclocked. However I realised that on a gaming standpoint I didn't need the extra boost from the overclock, even with this slight bottleneck I see in the benchmarks I'm still running say Witcher 3, Ultra w/ Vsync and not a single dip in FPS at any time.

Rails646

lordaeron1
Well this is very disturbing ... my Gtx 1080 FTW getting 4154 score in Valley and 2820 in Heaven... looks low to me if your only short of 600 with gtx 1070. i guess something is very wrong with my setup also :S. 

This is actually reassuring, our CPUs should have similar potential and both on a Z68 chipset, I think it's simply the rest of our hardware isn't able to use the full potential of these monstrous cards, even then the performance is good so there shouldn't really be any rush to go upgrading our CPU/Mobos.

wait just a sec ... if we are using the same mobo chipset why did you say that you are running the card through PCIx 2.0? ... mine is 3.0

Rails646

lordaeron1

Rails646

lordaeron1
Well this is very disturbing ... my Gtx 1080 FTW getting 4154 score in Valley and 2820 in Heaven... looks low to me if your only short of 600 with gtx 1070. i guess something is very wrong with my setup also :S. 

This is actually reassuring, our CPUs should have similar potential and both on a Z68 chipset, I think it's simply the rest of our hardware isn't able to use the full potential of these monstrous cards, even then the performance is good so there shouldn't really be any rush to go upgrading our CPU/Mobos.

wait just a sec ... if we are using the same mobo chipset why did you say that you are running the card through PCIx 2.0? ... mine is 3.0

Yours is more recent, mine should be from a year or two prior I believe, it's a Z68 but with PCI E x 2.0 support. The board itself though should be fairly similar overall.

ow i see. my mobo is a revision of an older model so they added a (b3) revision at the end... an upgrade from a none b3 version. guess they added the 3.0 with the revision and some other stuff.
 
but now i wonder... if i switch to 6700K + latest mobo + ddr4 will i see a huge improvement or in a rage of 100 score points +\-. it's like a 800-1000$(+ new PSU) upgrade i wonder if it's worth it :(

arestavo
It's only woth it if you aren't getting the FPS that you want.

For me, that means maintaining >60 FPS on a 60Hz monitor.

good point there ... all this chase after FPS driving me nuts lol :)
i am playing at 1440p and i am not even touching 70FPS minumum on all games so... well i am not counting Crysis 3 in because there i dip into 45FPS in some huge grass areas :S (this game is insane)... even witcher 3 fps around the 70fps. maybe crysis 3 is not well optimized or something. (my GPU load is not 100% there)
 
 
no point in upgrade just yet :)

lordaeron1

arestavo
It's only woth it if you aren't getting the FPS that you want.

For me, that means maintaining >60 FPS on a 60Hz monitor.

good point there ... all this chase after FPS driving me nuts lol :)
i am playing at 1440p and i am not even touching 70FPS minumum on all games so... well i am not counting Crysis 3 in because there i dip into 45FPS in some huge grass areas :S (this game is insane)... even witcher 3 fps around the 70fps. maybe crysis 3 is not well optimized or something. (my GPU load is not 100% there)
 
 
no point in upgrade just yet :)

That's the very definition of a CPU bottleneck there - the CPU can't feed the GPU instructions fast enough (GPU utilization is low). If that's the only game that you play where it happens, then you are sitting pretty good with what you have.

Try this: post # 15    LINK
 

EVGATech_RayH
This happens a lot with GTA V specifically if you are running with a single threaded Quad-Core CPU. We made this guide based off of my own experience with the issue with the game running an OC'd i5-4670K and a GTX Titan where my FPS was stuttering and awful at 1080p and was smoother at 4k due to the bottleneck being placed on the GPU rather than the CPU at the higher resolution. Likely it is just the buggy game not utilizing the CPU like it should and this guide should help:
http://www.evga.com/support/faq/afmmain.aspx?faqid=59641
Making that batch file worked wonders for my performance in GTA V