Decent model of GTX 970 these days?

[Vellinious]

Stock clocks from a strong CPU still does give you alot of points. FS for example, even the graphic score itself is affected by how fast your CPU is. There is no true way of comparing gpus if you're looking at benchmarks like them. Having a strong CPU already gives one the edge and over the other. Even more so for 390 since dx11 overhead is quite alot. I'd imagine that when the first dx12 benchmarks comes, the playing field will be more even.

Games are the best benchmarks. Since at the end of the day, that's what you'll be playing. Not fs or valley.

No, it's really not.  I can run the exact same graphics score with my CPU at stock clocks, and with the benchmark overclock.  Try it, you'll see.  I've done quite a bit of testing in it.  = )  The core count of the Haswell E processors is the only advantage it has, and that only comes into play in the physics test and the combined test.  The IPC on the Haswell E is actually lower than the standard Haswell, even the high end Ivy quad cores, because the base clock speed is slower.  They actually have the advantage in Valley and Heaven.  The graphics test in firestrike was made to be as much GPU as humanly possible.  There's very little there that the CPU does.  

Gaming benchmarks are great, but..nobody ever has the same game with which to compare.  There's nothing standardized there.  

[Kinda Bottlenecked]

No, it's really not.  I can run the exact same graphics score with my CPU at stock clocks, and with the benchmark overclock.  Try it, you'll see.  I've done quite a bit of testing in it.  = )  The core count of the Haswell E processors is the only advantage it has, and that only comes into play in the physics test and the combined test.  The IPC on the Haswell E is actually lower than the standard Haswell, even the high end Ivy quad cores, because the base clock speed is slower.  They actually have the advantage in Valley and Heaven.  The graphics test in firestrike was made to be as much GPU as humanly possible.  There's very little there that the CPU does.  

Gaming benchmarks are great, but..nobody ever has the same game with which to compare.  There's nothing standardized there.  

 

Alright, though I do not make baseless assumptions base off nothing. 

 

Here's my Highest scoring FS run with my 4670k OC to 4.2ghz. Both scores have my 390x at 1180/1600mhz

 

 

 

And here's my FS score without CPU OC

 

 

That's too great of an effect for the GPU test. 

[Vellinious]

Alright, though I do not make baseless assumptions base off nothing. 

 

Here's my Highest scoring FS run with my 4670k OC to 4.2ghz. Both scores have my 390x at 1180/1600mhz

 

 

 

And here's my FS score without CPU OC

 

 

That's too great of an effect for the GPU test. 

That doesn't look right at all

[Kinda Bottlenecked]

That doesn't look right at all

 

Yea, you tell me. Though I did notice the the same trend when I went through my 7970 scores. But the difference was only 50~30 points. This however probably shows the huge driver overhead. 

[Vellinious]

The graphics score seems low for those kinds of clocks with a 390X.  Hmmmm

[Kinda Bottlenecked]

That doesn't look right at all

 

Oh hold on hold on

 

the 10202 score isn't the right one, that's with my 390x at 1180/1595 mhz 

 

The bottom one is 1180/1600 without CPU OC. 

[Vellinious]

That's even lower with a higher physics score. I know scores fluctuate, but......

[Kinda Bottlenecked]

That's even lower with a higher physics score. I know scores fluctuate, but......

You can discount ~100 points off those scores. But that near 800 point difference is what I want explanied.

[Vellinious]

Yeah, still tryin to figure out how that's even possible.  Some of the highest graphics scores I've run were during testing with stock clocks on the CPU.  

[Kinda Bottlenecked]

Very likely driver overhead at this point. And your card is a 970. Totally different architecture.

[Purplehazeffc]

My Gigabyte G1 Gaming done 1615mhz on air.  And does 1650mhz on water.

Though the Samsung memory only gets to 2053mhz. Didn't matter on air or water.

And while there were some 3d Mark Firestrike scores shown.  Here is mine..
I haven't checked for a month or so. The highest for a 4690k & 970 combo.

Higher scores have only been with a more powerful CPU to help the Physics score

 

[Vellinious]

That's a really good score.  I'm pulling high 14.7k and low 14.8k graphics scores with the 80% ASIC GPU with a custom bios and water cooling.  I'm pushing to break 15k graphics score.....I'll get it there.  Eventually lol