The Extreme Pin Modding Guide - Page 5

what should i do to cool it besides for cleaning it.

ok it found an error after 11 hours temps were at 65

I was able to run 3dmark05 all the way this time scored 6778

Well, it looks like your system is almost rock-stable. If you want completely stability, you're going to need to voltmod. However, your system should be stable for almost everything you do.

If you flash your GPU to 445/1245 you should get more like 7500 3dmark5s with that CPU speed

Yea i was thinking about flashing it once i get my cpu under control. That test was with a volt mod

Ok did it, hope it works Thanks a lot Mr. K6

Update: I started prime when i left for school, got back and i found that the screen had shut off and wouldn't turn back on, mabey it forze idk, So i re-started it again at 5 (central time) so far so good at its 11:40 now! I hope its still going in the morning.

Well it froze again, so im guessing, its not going to work.

Update: went back to the origional plan and it works! I also set my 7800gtx to 450/1200 and scored 4138 in 06'
I the compare seach with pentium M as the cpu and go 7800gtx and the highest one so far is 4167. Im in second place as of right now

Well it freezes up druing games still, so I don't know what to do, im running it a 1.8 now

Yeah I did just video 2 and 1 and it was freezing up on me for some reason, mabey I had the jumper in the worng spot? Its F 1&2 and F 3&4 right? I just upped the 7800gtx to 450/1200 because I thought everything was all stable, on stock cooling. And when i re-applied AS5 i lowered my temps about 5 degrees

Okay, bought a 1.8 from that Paylessss dude on ebay for $130 or something and slapped it in with a little speaker wire in the mod-holes. Booted up fine and ran through 3DMark05 including CPU tests without a hitch. Default volts are 1.340, and I'm currently Priming as per your instructions with large FFTs...big thanks to Mr. K6, prona, and everyone else who helped out here.

My question has to do with the CPUs themselves...Dell has some sort of adhesive tape on the bottom of the heatsink, which I assume is the 'pad' everyone has been talking about...also, there are some markings on this tape to indicate that a small amount of thermal grease was applied on the original CPU (a 2.13, btw) at the factory. Both the 1.8 CPU and the 2.13 Pentium M's look as if the core has been 'dipped' in some sort of goo as well, but now that I think of it, EVERY Pentium M core I've ever seen has had that beeswax-looking stuff dripped on it...anyhow, to make a long story even longer, I didn't have any isoprophyl alcohol handy, so I just took out the 2.13, slapped in the 1.8 with some speakerwire jammed into those holes and turned it on...so far, it has been priming through 20 iterations and the CPU temp has not gone above 67C...maybe Dell had the right idea? I know from experience with desktop processors that Artic Silver does make a difference when properly applied, but this appears to be a fundamentally different sort of CPU architecture...?

Finally, is there any way to find out FOR SURE whether this thing will eventually crash without leaving it Prime for 12 hours straight? I'd like to use it again today to actually do work, but I'm not screwing all that stuff back in until I know whether I need to jam another speakerwire into another hole or two.

Thanks guys!

Can you pin mod a 1.6 celeron m dothan to 2.13?
Would you use the same pin guild?

Okay, guys, problem...it crapped out at 1 hour 20 minutes in Prime95...this is going from 1.8->2.4Ghz and running at default 1.340v...no problem, I think...I'll just volt-mod the sucker...unfortunately, I cannot get the voltage to change AT ALL no matter what I do! I've grounded Vid1 and Vid2 as suggested by some, so I should be running at 1.436 (010001), right? Even if ONLY ONE of these groundings was successful, the default voltage shown by CHC should change to SOMETHING OTHER THAN 1.340, correct? In other words, pretend ONLY VID1 gets grounded because my VID2 mod is FUBAR for some random reason...this means that I should be running 1.372 in CHC (010101)...also, if ONLY VID2 gets grounded because I screw up VID1, there should still be SOME change in voltage because grounding EITHER of these pins changes to some voltage other than 1.340v...still, this is what CHC is reading everytime.

The odds that I did the initial mod (to go from 100 to 133 FSB) totally correct on the very first try and screwed up BOTH of the other volt mods TWICE EACH just seems wrong, somehow, although it is certainly not completely impossible...maybe I'm reading CHC incorrectly:

THe first page where it says STATUS, the CPU Clock reads 2399Mhz (24x), while the Voltage reads 1.340V. Also, I cannot change the voltage to anything higher than 1.34V in any of the drop-down boxes, and I can't change the multiplier to anything higher than 18x either (which seems weird since it's already at 24x)...RMClock gives me the same sorts of readings...

Maybe I just got really lucky with the initial mod and screwed up the voltage wires each time...any advice on getting them in? I usually get one end of the same small wire as far as I can into each hole (in a 'U' shape) and then 'mash' the whole 'extra' mass down with the tweezers so that it's not in the way of getting the CPU in or touching the other pins. Arrrghhh!

PS. Just so we're all on the same page, here, shorting VID1 is F2-G2 and shorting VID2 is F3-E3. Thanks!

No program is going to show the difference in voltage. It will run at the new voltage if you pinmod the right sockets though. You will only be able to tell by the increased stability in Prime or it not crashing after hours of use. CHC or RMClock will automatically read the stock voltage at 1.340, and if you voltmod it, it will run at the highest voltmodded voltage if you have the 1.340 volts selected in the program. If you lower the voltage through a program, every step will be down from your voltmodded voltage, and not the actual votage that it has listed in the box.

I can't recommend which to voltmod, but I wouldn't go right to 1.463 if it was stable for over an hour to begin with. I'd check the first post of the thread and look at the chart and just step up incrimentally. 1.463 is a lot of volts, and you could probably be lower and have a perfectly stable card. I am confident that you in fact did the voltmodding correctly, and just assumed that it did not work since CHC didn't report it.

Hopefully K6 is around here and can recommend a voltage for you, but hopefully that helps a little.

hammer, judging from your successful prime95 time on stock voltage, I don't think you'll need a big voltage increase. BTW, like mZimm said, the programs are not going to show you the actual voltage. If you grounded vid1 and vid2, then when the CPU attempts to set to 1.340v (since vid1 and 2 are now permanently 0), you get a real voltage of 1.436V. Programs like NHC and RMClock just tell you what the CPU is trying to set the voltage to, not the actual voltage. For that, you're going to need a voltmeter to physically measure it.

Anyhow, I would try grounding VID0 and VID2 (E2-E3 and F3-F4). That'll get you 1.420V which will hopefully stabilize your system. Note that since you grounded VID0 and VID2, the lowest actual voltage you can get when you set to 0.700V is 0.780V...which is still very good compared to default (0.988 i believe). So you can still underclock when you are running on battery.

Oh, holy mutherf*cking jeebus! I just spent the last four hours pissing off my wife by mucking about with this sh!t strewn all over the kitchen table and cursing like a goddang sailor! thanks, guys. will report back. sigh.