Originally Posted by mirage
Hi Everyone,
I used a lot of information in this long thread for my system, which is not a common combination here. Pentium-M 705 (Banias 1.5G@400M FSB) + CT-479 + P4P800-VM. So, although it may not be interesting to the Dothan users here, I thought it is worthwhile to add my experience.
You might ask why Banias when the latest and greatest is Dothan (or Core Duo actually). Because I found a $50 deal at mwave several weeks ago. Then, I needed a MB to build a Media Center PC, and there came P4P800-VM(newegg refurb) and CT-479 (zipzoomfly). I also bought a MATX Athenatech case with 180W power supply. I started with setting the FSB at 533 MHz on CT-479. Used the default heatsink of CT-479 with Arctic Silver 5. When the system first posted, it reported FSB as 400 MHz on the screen and in the BIOS. There was no indication that the FSB was 533 MHz. I tried the u-wire trick to set the FID to 533 but the post screen and BIOS did not indicate any success. So I removed the u-wire and left the 533 MHZ setting on the CT-479 adapter, and continued with installing the OS. After Windows XP installation was completed, to my surprise, I was greeted with a 2GHz Pentium-M running at 533 MHz FSB. Apparently, BIOS was only reporting the processor specs not the actual applied frequency. That was good news, I was happy but not for long. The stability tests using Prime95 were failing after several minutes and it was not a CPU heating problem or related to memory. I had to increase the vcore but BIOS VID change had no effect and the software I used, namely EIST and RMClock was also not effective in increasing the voltage (though they can decrease the voltage). So, for the second time, I used u-wire, this time for VID mod. The default voltage of this Banias is 1.484v, using P1, p2, p3 pins high. The best option seemed as setting p3 low to achieve 1.612v. After booting with the new VID (1.612v) I saw in CPUZ that the actual voltage was 1.58v and everything was stable in this case. Also, the CPU temperature under load was not exceeding 56C.
I tested to go higher than 2 GHz (using Clockgen) but it was not stable beyond 2.1GHz. I have settled at 2GHz with a safety limit and in the final configuration I used 166MHz DDR frequency (single channel) which was slightly faster than 133MHz (1:1). And finally, the SuperPI 1M is 49.5 seconds. By using RMClock I was also able to activate Speedstep, it worked perfectly by scaling the system between 800MHz(6x133)@1v and 2GHz(15x133)@1.58v depending on the workload.
Please let me know if I can supply any further information or I am not aware of something that can increase my overclock. I will eventually use dual channel when I can buy the other stick.
Cheers and thanks to everyone for all the information I found here.:toast:
Summary:
Pentium-M 705 (Banias 1.5G@400M FSB)
overclocked to 2.0G@533M FSB using jumpers on CT-479
u-wire VID mod P3 high->low, 1.484v->1.612v (motherboard supply 1.58v)
P4P800-VM 1017 BIOS
CT-479 adapter set at 533MHz
Activate Speedstep using RMClock
single channel 512MB @DDR333
SuperPI-1M : 49.5 seconds