News for hackers. Stuff that matters.
Re:Intel Kaby Lake Potentially an Overclocker's Dream, Seen ... by jankotek on Wednesday December 28, 2016 @10:29AM (#13271072)
I was not very precise. I actually meant heat-spreader instead of heat-sink.

Kaby Lake has bad thermal conductivity between silicone and heat-spreader (metal top of CPU). If you crack-open the CPU and replace thermal paste, it will reduce temperature under load significantly.

It is common problem for Intel CPUs since Ivy Bridge. i7 are affected. Extreme version is soldered and does not have this issue.

Re:Intel Kaby Lake Potentially an Overclocker's Dream, Seen ... by floatboth on Wednesday December 28, 2016 @10:39AM (#13271144)
I know about delidding.

I thought this:

> You literally have to cut this CPU in half, to make it usable for OC.

referred to turning off the cores, not delidding :D

But you don't need to do either to "make it usable for OC". Just get a good cooler.

Re:Intel Kaby Lake Potentially an Overclocker's Dream, Seen ... by jankotek on Wednesday December 28, 2016 @10:52AM (#13271235)
Good cooler wont help, the bottleneck will be in CPU.
Re:Intel Kaby Lake Potentially an Overclocker's Dream, Seen ... by floatboth on Wednesday December 28, 2016 @10:56AM (#13271275)
I'm running 4.5GHz @ 1.4V on an i5-6400, doesn't get hotter than 60 degrees under full load with a watercooler.
Re:Intel Kaby Lake Potentially an Overclocker's Dream, Seen ... by jankotek on Wednesday December 28, 2016 @11:34AM (#13271600)
60 degrees is not much, not really reaching full potential of that chip.

I had i7 2600K at 5.0 GHzx. Water cooled, 90 degrees under full load. It worked for 5 years 24/7, until motherboard died last year, CPU still works.

It is simple to set higher frequency, but that does not mean CPU is actually faster. Power management is very tricky. Sometimes it will put idle instructions, or turbo boost would not work.

You have to run benchmark for several hours to verify CPU is actually faster.