OcUK Staff
Exactly as I did with Haswell-E, X99 Launch I wanted to compile a short review showing the Overclocking of the new Skylake i7 6700K on very basic air cooling and mid range MB. Then potential buyers from OCUK can see what kind of overclocking potential this new chip has with very minimum effort. This is the highest priced Skylake offering running at 4ghz stock with hyper threading so you have a total of 8 threads can be handled simultaneously in applications supporting multi threading.
With this in mind I chose the well priced with decent features ASUS Z170 Pro Gamer MB, Kingston 3000mhz Predator DDR4 mems and a Zotac 980 Ti amp edition. I chose the Ti to remove the GPU bottleneck while benching at 1080p and shift the bottleneck back to the CPU.
All Skylake 6700K chips I have tested can run 4.5ghz+Core Clock, 4.1-4.5ghz Uncore on basic air without any bump from the 1.3v stock voltage. On this board I simply set 1.3v for core. 45 multi and XMP and there I was with stability in Windows running all I could throw at it. Good samples can of course go higher but all the samples I tried did have consistency in how they clocked and like I say 4.5-4.6 was a minimum which anyone could achieve by making a couple of settings.
Here are the benchmark results for this chip running at 4.5g with 3000mhz mems and 4.1 uncore on air.
3D mark 11
Firestrike
Terragen 3
Cinebench R15
To give the reader some idea of how Skylake i7 6700K at 4.5ghz Compares to Haswell E 5820K at 4.2ghz on air with the same cooling, 2400mems and 3.5 uncore I have provided screen shots of the Haswell E's performance below.
3D mark 11
Firestrike
Terragen 3
Cinebench R15
What was surprising for me about these results is that as expected on all multi threaded stuff the X99 platform smashes the Z170 Skylake. Terregan and Cinebench both heavily multi threaded benefiting from both cores and threads as do games based physics or combined tests.
What I did not expect and what is certain is in 3D game based benches the FPS was very very close indeed with Skylake faster yes but only by around 2-3% in 3D mark 11 and Firestrike. This makes X99 look amazing value considering its other benefits.
Wanting to explore this further I switched out the 980 Ti for a 970 which is the current mid priced gaming card of choice for the masses. In both Heaven and Valley which are only game tests run maxed out at 1080p the X99 system with a 970 clocked equally on both gave less than 1 FPS difference when benched.
For me at this specific price point X99 is king and does not need high speeed DDR4 to achieve this position.
i5's are not really my realm of interest but I can confirm the i5 Skylake samples I tried using these same basic settings all hit at least 4.5ghz. This consistency is much welcomed.
X99 Anyone
With this in mind I chose the well priced with decent features ASUS Z170 Pro Gamer MB, Kingston 3000mhz Predator DDR4 mems and a Zotac 980 Ti amp edition. I chose the Ti to remove the GPU bottleneck while benching at 1080p and shift the bottleneck back to the CPU.
All Skylake 6700K chips I have tested can run 4.5ghz+Core Clock, 4.1-4.5ghz Uncore on basic air without any bump from the 1.3v stock voltage. On this board I simply set 1.3v for core. 45 multi and XMP and there I was with stability in Windows running all I could throw at it. Good samples can of course go higher but all the samples I tried did have consistency in how they clocked and like I say 4.5-4.6 was a minimum which anyone could achieve by making a couple of settings.
Here are the benchmark results for this chip running at 4.5g with 3000mhz mems and 4.1 uncore on air.
3D mark 11
Firestrike
Terragen 3
Cinebench R15
To give the reader some idea of how Skylake i7 6700K at 4.5ghz Compares to Haswell E 5820K at 4.2ghz on air with the same cooling, 2400mems and 3.5 uncore I have provided screen shots of the Haswell E's performance below.
3D mark 11
Firestrike
Terragen 3
Cinebench R15
What was surprising for me about these results is that as expected on all multi threaded stuff the X99 platform smashes the Z170 Skylake. Terregan and Cinebench both heavily multi threaded benefiting from both cores and threads as do games based physics or combined tests.
What I did not expect and what is certain is in 3D game based benches the FPS was very very close indeed with Skylake faster yes but only by around 2-3% in 3D mark 11 and Firestrike. This makes X99 look amazing value considering its other benefits.
Wanting to explore this further I switched out the 980 Ti for a 970 which is the current mid priced gaming card of choice for the masses. In both Heaven and Valley which are only game tests run maxed out at 1080p the X99 system with a 970 clocked equally on both gave less than 1 FPS difference when benched.
For me at this specific price point X99 is king and does not need high speeed DDR4 to achieve this position.
i5's are not really my realm of interest but I can confirm the i5 Skylake samples I tried using these same basic settings all hit at least 4.5ghz. This consistency is much welcomed.
X99 Anyone
