Sli gtx 970 barely plows

iMelnikoff92
makes no sense at all if we are talking about i7 4790k because it is more expensive
to try to overclock the processor and mother all the way to the death march.
and there you look you can buy what i7 7770k and new filling 2.5 times faster.

The processor in overclocking 4.5 is already higher than the core under 90 and the frenzied roar of the fan, it is possible to exchange your own for 4790k with a surcharge of 10 chunks, but you don't need it, Sergei Viktorovich claims that he has almost stable 60 frames at 2560 * 1080 on his 4670, my goal is equal 60 with full d, who will say what to this, to take a risk with an exchange? And why is there such a difference in performance, because 4690k is the same 4670 only with a modified thermal interface?

I start to sin for a percentage, with whom I did not communicate, everyone says that 4690 is still enough for 60 frames with full HD, while mine starts to drain in processor-dependent games, although there are enough cards with a margin

if it is possible to change the processor with a small surcharge, then this can be used,
however, I do not know how fast this i7 patam is that I personally did not feel them, but I know the current test results. so, in theory, it should be faster from 20 to 40 percent at identical frequencies.
however, the processor requirements grow relatively slowly and the GPU requirements grow rather quickly. therefore, most likely in a year or a year and a half, if the 970 should reach its full potential in the performance rank in the 40fps region,
and 60fps is a rather expensive performance standard. maintaining it is somewhat costly. therefore, in my opinion, all these pitch adventures are not particularly worth it.
also according to my pitch research, the MPGU mode still affects the upper load of the processor. Those MPGU modes during tests are more dependent on sotorona not GPU than, for example, solo with one extremely cool video card. the difference should not be significant there

You have an overclocking mother and a middle-level percentage, of course it won't be enough for a slide of two new-generation video cards. In general, the load of one of the video cards by 99% and the other by half means that the application is not quite demanding. Start Minecraft for fun.
p, s, and your lid temperature is 73 or cores?

fOXY9
By cores at a frequency of 4.5, the edge of 73 degrees in stress tests. With the load, I solved the problem, now the cards are puffing synchronously by 90% in zhta well ultras at msaa8, and even 60 fps while the character is standing or just running, as soon as the movement begins, like chases and fierce shootings, drawdowns begin up to 20, so what's the point of the cards a little, one hell of a setting had to be reduced. I noticed that the quality of particles, high-resolution shadows and an increased drawing range, are especially heavy on the percentage of the quality of particles, another proof that ah 5 does not cope

MunchkiN 616
I'm not going to chase permanent 60 shots, but in current projects I want stable ones, so that for the next 3 years I somehow play at least at medium high

MunchkiN 616
I agree with you, when the second card was installed, the load per percent increased by 5-10 percent

iMelnikoff92 in
order to identify the load of the MPGU, it is necessary to create such testing conditions when no card is a bottleneck under any circumstances and the fps is limited only by the performance of the cpu and platform,
and it is difficult to create such conditions due to the absence of super low resolutions or a pair of titans or tees.
I can mark the current wachdogs where the second GPU cut me off at once up to 10fps with wires and got tired of the friezes