• Welcome to Overclockers Forums! Join us to reply in threads, receive reduced ads, and to customize your site experience!

P4 3.2GHz OC to 3.7Ghz Write-up

Overclockers is supported by our readers. When you click a link to make a purchase, we may earn a commission. Learn More.

AlcatrazT2

Member
Joined
Nov 26, 2004
To those individuals with the Intel Pentium 4 Northwood 3.2GHz/512K/800F/478P-SL6WG processor seeking some additional information regarding the overclocking possibilities of such exceptionally engineered piece of hardware.

Knowing that the processor stock comes with the 16 x 200FSB = 3,200Mhz settings, I decided to purchase some good RAM and unleash all of the power. Special thanks to the Corsair XMS TWINX1024-3200C2PT 1GB (2 x 512MB) DDR 400 (PC 3200) Dual Channel Kit System Memory, I was able to bump my processor to a total of 16 x 231FSB = 3,696Mhz at a 1:1 ratio with timings set to 2-3-3-5

Note, that the memory was set to defaults at 1:1 200Mhz with timings set to 3-3-3-8. I went ahead and changed them to the defaults specified at Corsair and went on from there, 2-3-3-6.

After much research, I was able to figure out that the computer will boot maximum to a 16 x 237FSB = 3792Mhz on both 5:4 and 3:2 ratios. However, the computer would not boot to 16 x 238FSB = 3808Mhz on either 5:4 or 3:2, thus giving us the idea that most likely the maximum speed of the processor is almost a clock speed of 3,800Mhz.

Knowing that the RAM has some potential, I decided to start testing 1:1 right from the 238FSB. As many of you may imagine, 238, 237, 236 would not boot at all but once I reached the 235FSB level, I was able to boot into Windows XP Professional SP2 and do some testing.

I decided to test the RAM out through MemTest-86 v3.2 and through much testing, 235-232 testing gave errors but 231FSB looked like the sweet spot for the RAM on 1:1. Not only was this a great overclock of 496Mhz for the processor, but I had also achieved a 62Mhz overclock in terms of my RAM and I could not ask for anything more. I figured it wasn’t worth testing the stability on the 5:4 16 x 237 = 3,792Mhz because in reality, I would be only gaining a total of 96Mhz with my RAM running at 380Mhz (190Mhz per stick) instead of 400Mhz (200Mhz per stick).

At the end, it a far better choice to be running 496Mhz stronger with 64Mhz more RAM rather than benefiting 96Mhz in terms of the processor and loosing the 20Mhz of my RAM.

All of the testing was done via the specifications listed below in my signature. Hope this is of any help to anyone and good luck on the overclocking journey.
 
They had that GeIL DDR500 PC4000 on sale at Tiger Direct a bit ago for $138 a gig.
I got it and have yet to use the rating due to laziness and my P4 is set like yours. I will try 240 soon. I have a main board that says 1200mhz+. So that prospect is tempting. GeIL sets at SPD at 2.5 up to 250 frq, so they say.

I have been using mine for a while now after maxinf out at 3.57GHz with DDR400 .
memory. I have an XFX 6800 stock plain vanilla and 3DMark '01 ran a 19600.

Run SANDRA Memory Bandwidth for me. I wondere how they compare. No entusiast site runs those benchmarks.
So things sound similar to my experience.
 
I've got a 3.2c 30-capper that I've been able to squeeze up to 3.93 using some Kingston HyperX with 2-2-2-6 timings. In order to get there I did have to raise my vcore to 1.7v, with temps climbing to just over 50C. I so wanted to hit 4GHz on air but it does not look like its going to happen.

I don't see any mention (may have missed it) of how much voltage you've had to give the 3.2c to get to your current overclock. Have you increased it, or are you still running stock voltage? If you are still running stock voltage then I suspect that you've got more left in that proc. If you have raised it...what levels are you currently at?
 
Thanks for the comments guys.
Well, my ram is at 2-3-3-6 now.
Turned out that I was getting errors on both
2-3-2-5
2-2-3-5
2-3-3-5
So I decided to keep timings at stock
2-3-3-6 vs. 3-3-3-8 (detected settings).
I ran the memtest86 for about 16 hours on default test and no fails...decided to keep it at that.

My CPU is at 1.6volts...decided to bump it up to be safe. Don't play on hittin any higher since people mentioned the problem with SNDS.
 
Back