Tuesday, July 8, 2014

My 4790K is here and it can boot to desktop at 4.6GHz with 1.25V!

When the doorbell rang in the morning of Saturday, the 5th of July, I knew what I had hoped for had come true. I ordered the Core i7 4790K CPU on the previous evening from Amazon and the CPU arrived before noon. I had already dug up the toolset, the Arctic Cooling MX-5 thermal paste and the box of the old CPU by this time, and was ready to replace the CPU and test it as soon as possible.

But first things first. Photo-shoot!

The batch number of this particular CPU is L418C209.
  • L means it is made in Malaysia.

  • 4 means it is made in the year 2014. (You would expect all the Decil's Canyon chips to be made in 2014, but that's not the case. Some of them had been made as early as mid 2013. Seems like Intel had been binning the chips for a year to get to this point.)

  • 18 means it is made in the 14th week.

  • And the rest of it means the lot number.

Then it was time to install the CPU. The top cover and the two side panels of the case were removed quickly. While removing the panel behind the motherboard, a thumbscrew fell into the case and it could not be found ever again!!! It was like finding a needle in the haystack, with all the cables behind the motherboard tray. If I was a little bit more patient, I probably could have found it but I had a brand new CPU right next to me begging to be overclocked!! What did you expect? But I did a quick glance to see if there were any kind of short circuiting.

As the double sided tape that was holding the back-plate of the CPU cooler is no longer present in my Antec Kuhler 620 cooler, I knew that installation of it would be an irritable experience. Thus I tried a new trick this time. Instead of removing the whole mounting cage, I decided to just loosen it and sneak the CPUs in and out. Plus, I decided to leave the radiator mounted on the case while doing all this. The water block could be easily removed but the retention lever of the CPU could not be pulled to take off the CPU as the frame of the mounting cage was blocking its motion. If seemed as if it was impossible to remove the CPU without removing the mounting cage.

But I didn't give up. Instead of removing the whole mounting cage, I removed three screws of it and kept one in. By doing that, it was possible to move the cage a bit so that it could give me enough freedom to move the lever. I did all this while the case was standing upright, hence the CPU fell off in to the case when the lever was pulled out. I hope it didn't do any harm to the CPU, and it shouldn't have. I wiped the thermal paste off it using a tissue paper, packed it in its box and kept aside.

Then I quickly installed the new CPU and lightly tightened the other three screws of the mounting cage so that it didn’t fall off. Then it was time to apply some thermal paste. As mentioned earlier, it was going to be Arctic Cooling MX-4 which I bought off eBay many years ago. I'm sure there are better pastes available now – like the Gelid GC Extreme, but this is still right up there with the leaders of the pack.

First, the residue on the water block had to be removed. I used surgical spirit to clean it up; didn't really check the composition of it. Once it was clean and tidy, I applied a pebble sized amount on the middle of the CPU, and let the pressure do the rest. The water block definitely doesn't sound as sturdy as my old Silver Arrow, so this alone could affect the cooling up to a certain degree. But this is all that could be done with this cooler.

Finally, it was time to fire up the PC. I actually forgot to reset the BIOS settings before installing this new CPU. But the board sensed the new CPU and loaded up the default settings, and asked me to go into BIOS and do whatever I had in mind. I straight away set the Vcore to 1.25V and clock multiplier to 46x and booted. And it booted fine!!! This was impossible with my old 4770K. I was feeling great. Could this be a 5GHz do-able chip? Is that a far cry?

When the Windows desktop loaded, I decided to run OCCT and see how the temps were. It actually hit similar temps as my 4770K @4.3GHz/1.25V did in OCCT. That definitely was disappointing. At least a temp drop of 6C was expected, but it could not be observed. However, it is possible that the 300MHz clock speed deficit made the older CPU run cooler. It is a known thing that Haswell runs a lot hotter not just with higher voltages, but with higher clock speeds at the same voltage. I sense that the 3D Transistors are at it again.

So what did I do next? Overclock, of course. But that for another post.

No comments:

Post a Comment

Related Posts Plugin for WordPress, Blogger...