Comparo on 2600 mobiles

magratton

Senior member
Mar 16, 2004
523
0
0
Okay, so I don't run the risk of looking like as big of a fool as the heatsink guy (for those of you who know what I mean), I would like to ask you all for a little help with a mini-comparo I am looking to do this weekend.

I am going to be testing two 2600 mobiles. I will have a 0346 production chip (already have that one from FSFT forum). Tonight the other chip will be arriving from Newegg. Barring the instance that this chip from will be from the same production week as the other, I am assuming it is going to be an 04 chip. There has been a lot of conjecture that the more recently manufactured chips are not oc'ing as well and since I am building some systems for a cousin, a friend and me too! (YAY!) I figured I would look into this, even if the manufacturing difference in age is not all that great.

The only nuisance to all this is that the mobo's (among other parts) are not the same between systems so to truely test this out I will have to work on only one system.
So my test setup will be:

Abit NF7-S V2.0
WD 80GB SATA HD (Windows XP Pro)
Mushkin Level II 2x256 PC3500 (2-3-3-11@2.6v) in Dual Channel mode (though I may change this to single channel)
Antec True430
ThermalRight SLK-947u w/Vantec Tornado 92mm

To begin with, and just for mention sakes, I already have an 2500 mobile (210x11.5@1.75v) prime stable for 8 hours (large fft's).

So, the question I have is, what are people looking for in terms of an overclockability comparo? Just raw MHz? Highest FSB? And also, so that I can get there a little faster than my random trial and error approach with my 2500 mobile, what is the "right" way to engage in overclocking? I have heard some folk say "set multiplier somewhat lower (e.g., 11x), then raise FSB 'til stable, then raise multiplier". Any tips in this area would be helpful as well.

Thanks!
 

myocardia

Diamond Member
Jun 21, 2003
9,291
30
91
Okay, here's how you'll want to do your testing. Start by raising your vcore to somewhere between 1.75v & 1.80v, then set your bios to around 12x175. Then, with this: ClockGen, once you've booted to Windows, crank up the fsb. If you go too far, no resetting the bios, just reboot the computer, and it will be right back to where you started. And most of us are interested in overall clockspeed, not in highest fsb, since that involves too many variables (memory, mobo, more). Now, for you personally, with your Winbond BH-5 chips, you should be running your mobile at ~220 fsb or higher, because the higher the fsb, the more performance.
 

Avalon

Diamond Member
Jul 16, 2001
7,571
178
106
I really don't know how helpful this is, but I've always done it trial and error by the following...

Start off trying 166x11 (stock barton) 1.45v
Then try hitting the voltage up to 1.5v and try 200x11
If all goes well, start increasing the bus little by little to around 210, adding an increment of voltage if needed (you should need about 1.55v here)
drop it back down to 200fsb and try a multiplier of 12 now
add more voltage if needed
then again, raise the fsb little by little until you can up the multiplier again
after you find the roof, use whatever fsb and multiplier combination you wish to run at a good speed you could document

whoops...this is assuming your memory can do 210fsb that is :X
sorry!
 

magratton

Senior member
Mar 16, 2004
523
0
0
Welp, CPU arrived from Newegg and wouldnt ya know it: 0348. Doesnt seem like much of a comparo to do. Oh well. I am sure others will post their experiences with the more recent 04 chips.

Now then playing with what I have :)