Thread Rating:
  • 0 Vote(s) - 0 Average
  • 1
  • 2
  • 3
  • 4
  • 5
OCaholic Challenges Trollpoppin's CPU Bottleneck Claim
#1
http://www.ocaholic.co.uk/modules/smarts...26&page=20
The challenge has been made with Ivy Bridge-E and Haswell-E, plus a GTX 980.
Quote:Stock Clocks

First, we’re going to talk about the average performance differences of the two CPUs we've tested here. To keep things as simple as possible, we've averaged all the results that we gathered at different resolutions. This allows us to create an overall performance index. The Core i7-4960X is defining the 100.00% baseline, to which we're going to compare all other results. On average the Core i7-5820K is 2.92% quicker than its the former highest-end CPU.

A closer look at the different resolutions reveals, that at FullHD the Core i7-5820K performs 2.62 percent faster. In the case of 1440p the 5820K offers on average 3.47 % higher frame rates. Last but not least we’ve tested the games at UHD and the differences are very similar. The system with Core i7-5820K runs games 2.69 % quicker than equipped with the i7-4960X.

Another interesting value to investigate is power consumption. In idle our test system with Core i7-4960X burns 32.35 % more power than the setup with Core i7-5820K. At this point we have to mention, that the two CPUs have to be tested with different motherboards. With the system under full load – the graphics card is also loaded – the percentage differences are smaller. The system with Core i7-4960X needs 14.12 percent more power than the setup with i7-5820K, despite the fact new generation are more power saver than the old generation.

Overclocked
Following the same approach – but shorter - like with the CPUs at stock clocks we’re now going to talk about the results gathered with the CPUs overclocked to 4.5 GHz. Again, the Core i7-4960X is setting the 100.00% baseline. On average the new Core i7-5820K opens a 2.14 percent gap. In other words, overclocking the CPUs to 4.5 GHz has no effect on performance scaling and it doesn’t open any bottleneck. The reason for this is simple: the CPU isn't the limiting factor, it's the graphics card and this is the case even at FullHD.
Valve hater, Nintendo hater, Microsoft defender, AMD hater, Google Fiber hater, 4K lover, net neutrality lover.
Reply
#2
Yeah, duh

I am not surprised the least bit by the results. But if i remember correctly, apoppin was trashing the 5820k and saying that it sucks compared to the 4770/4790k when it comes to gaming.

I dont think he was able to admit that the reason the quads do better in some games was due to the higher clocks right out of the box. See, the 3.3ghz haswell E will not be faster that the 4ghz devils canyon in most games because it is very rare that 8 threads would be limiting.
But since 5820k chips can overclock to 4fgz and beyond (in this review 4.5ghz), the haswell E is absolutely great for gaming. Clock per clock, there is no haswell quad advantage
Reply


Forum Jump:


Users browsing this thread: 1 Guest(s)