Building An Intel Xeon E3 v5 "Skylake" Linux System

[Michael]

Michael

Until you get an i7-6700K please use this system for gaming benchmarks - if you rely on stock speeds (no gamer does). But I don't get why you don't oc your i7-5960X, this cpu is 100% made for oc - look at

http://www.tomshardware.com/reviews/overclocking-retail-intel-core-i7-5960x-cpu,4237.html

Could you please use the i7-5960 as a "normal" gamer would do with this cpu? I mean running at 4.5 GHz+

This Xeon system is being dedicated for a PTS customer's daily benchmarking once I'm done running my PTS tests on it.

[pal666]

MichaelBut I don't get why you don't oc

because nobody uses overclocking

[Linuxxx]

because nobody uses overclocking

Not sure if serious, but if you are, I bought a Pentium Anniversary Edition for ~60 EUR because I am running it @4.4 GHz on my ~50 EUR mainboard.
And I'm pretty sure that when it comes to gaming, my dual-core is beating Michael's ~1000 USD octa-core with hyper-threading (because, you know, single core performance is still king [and by the looks of it, will probably stay that way even with Vulkan...])

[Kano]

@pal666

How many i7-5960X users do you know who did not get this cpu for oc? It is the same with the K series, those cpus are made for oc, nothing else.

[pal666]

@pal666

How many i7-5960X users do you know who did not get this cpu for oc? It is the same with the K series, those cpus are made for oc, nothing else.

so you have zero statistics, but somehow claim what you like

[Zan Lynx]

@pal666

How many i7-5960X users do you know who did not get this cpu for oc? It is the same with the K series, those cpus are made for oc, nothing else.

I have a 5960X and yes I do overclock it. Only to 4.1 GHz which is more than enough. So there's a statistic of at least 1 for pal666.

[hoohoo]

REGARDING OVERCLOCKING AND BENCHMARKS

Sorry for shouting...

You don't want to benchmark overclocked CPUs and compare result against other processors in an article like this because every CPU overclocks differently.

If you stick to stock clocks then you can compare CPU model to CPU model.

If you overclock your CPU and then benchmark it all you can do is compare it to every other CPU regardless of model, not to other CPU models - because each individual CPU behaves differently. This is not useful when making a buying decision.

Example: back in the day I built a machine using an Abit BP6 board. One of my CPUs would happily clock to 550 MHz, but the other one was not happy above about 480 MHz; in a benchmark comparison should I have said the CPU model I bought two instances of was capable of 550 MHz or of 480 MHz or of it's stock 360 MHz? The only reasonable choice is the baseline stock clock.

[hoohoo]

@pal666

How many i7-5960X users do you know who did not get this cpu for oc? It is the same with the K series, those cpus are made for oc, nothing else.

Yes, Kano but how do you make a meaningful comparison of say, i7-5960X, to say, FX-9590 when you overclock one or both CPUs? Your 5960X might o/c 50% and mine might o/c only 33%, but my FX might o/c 50% and your FX only 33%. So which apple do you want to compare to which orange? And for that matter your o/c'd 5960X and mine are not the same fruit any more because they run at different clocks, same for the FX's.

By all means o/c your CPU and upload your PTS scores if you want.

But don't start claiming that those results are useful for comparing CPU models to other CPU models because not all instances of a given model o/c to the same degree.

O/c's benchmarks are only useful for comparing different instances of the same CPU model to each other, or your o/c'd CPU to all other individual CPUs.

[Kano]

If you don't oc this cpu the gfx card comparisons are non-sense, did you look at the results?

[pal666]

I have a 5960X and yes I do overclock it. Only to 4.1 GHz which is more than enough. So there's a statistic of at least 1 for pal666.

it is called anecdotal evidence, statistics is something different