Figured I should bring this old thread back:

1) One trick I learned which searching the interent, was a guy who decided he would try to see if he could eliminate the "dead spot" on his fan. In order to do this, he bought multiple cheap 120mm fans, and using a dremel, cut them out and smoothed out the insides. He then attached these hollow "extensions" to the fan in order to see if they would increase airflow. What he discovered was that you could add up to 3 of these, with each one increasing airflow by roughly 6%, to a max. of 18%, after 3, the pressure began to drop, and it started to decrease air flow.

So I tested this out on my H60 which was at the time using (and still is) using 2 Ultra Kaze 3000 (rpm at 2600 each). By adding 2 extensions to the push, I was able to drop my temps by several degrees.

2) While doing some further reading, I also came across someone tested how much airflow is blocked by the case filters "metal holes in the case", and it blocked air flow A LOT. So, I decided to cut those holes out using a pair of metal cutting sheers. IT MADE A HUGE DIFFERENCE. The best analogy (metaphor?) I can think to describe the difference is this, "Think of the difference between blowing air out through your mouth, and then try blowing air out through a straw!"

3) Your push should be slightly higher than your pull. This goes against popular wisdom that they should be running at the same speed, but if since the push fan is going through the radiator, this creates a disturbance. So showed another website with pretty graphs.

So I decided to put all this together (of course after doing one at a time) when I learned these new things, thereby taking about a week to put it all together and figure all this out. So now my current setup is as follows Fan@2500-->housing-->housing-->case (with hole cut out)-->radiator-->housing-->Fan@2300. All of this hot air is then sucked out through a fan at the top.

This setup allows my 2500K to run at 4.2, LLC=20%, Vcore=1.24, at idle to be at 1-2 degrees above ambient (21C), and at 100%=40-42C, with "hot core" at 44C. This is also with a GPU in case as well.

I applied the concept of cutting the holes out for air flow to other cases as well, attached a fan, and my average temps on my AMD chips when running never exceed 40C, (these use Hyper212, without extensions on those blocks for the fans.) OC one 3.5 and other way I have at stock since it's on a POS motherboard.

As a side note, I guy over at PrimeGrid was quite amazed at my low temps, since his setup was with a 2600K, never stated what OC, ambient@18C with a quad radiator, and he was at about 65C, after some tinkering he got his down to 51C. Although, I will state he said he had recently opened his window, and the temps were 8C outside, but he still seemed satisfied with the results.

Hope this helps someone over here as well.

It really isn't, but since these chips will be VERY GOOD for quite some time, temps are all you can control for longevity. (and voltage of course)

I do not think that you told us the temp limit on the P4.

I am running five new or newish machines, i5-520M,i7-920, i7-840QM, i7-2720QM, i7-3930k. I had heat problems on all of them. I have Coretemp software temp gauge on all of my machines. I played with the % of the processors, and the % CPU time. I got them all down to about 20% below the stated max for the processor.

The most helpful thing I have done to help cooling is what 5pot said...


For me too "IT MADE A HUGE DIFFERENCE"

Interesting memory fans. May have some research to do.