New Posts  All Forums:Forum Nav:

The Extreme Pin Modding Guide

post #1 of 565
Thread Starter 
Hells yah

So, you've got your new 9300/XPS2 and your CPU just isnt fast enough for you. You want more and sure as hell aren't going to pay $500 for a 780, or if you've already got one, are lookin for other options. Well look no further my friends, because there is a solution: [echo effect] EXTREME PIN MODDING [/echo effect]

Ok no more theatrics

UPDATED
The easy way to get the voltage you need -

Just as a quick guide, here's my general assessment of the probabilty of chips needing voltage mods (remember there are exceptions to ever generalization ):

715 (1.5GHz) -> 2.0GHz - Essentially guaranteed to pin-mod and even undervolt
725 (1.6GHz) -> 2.13GHz - Same as above
735 (1.7GHz) -> 2.26GHz - ~75% do not need a voltage mod, but for those of you who do you usually only need a little boost (<1.4V)
745 (1.8GHz) -> 2.4GHz - Only about 25% don't need a voltage mod, usually a voltage boost in the area of 1.388-1.436V gets the job done.
755 (2.0GHz) -> 2.66GHz - You're asking for trouble . I think ~5% can do it without a voltage mod, and maybe another ~50% can do it with one (there hasn't been much research here). Generally you can go up to 1.6V and still be OK.
765 (2.1GHz) -> 2.8GHz - Yeah right . The problem here is the limitation of the Pentium M architecture within the extreme environment of the laptop. The dinky cooling solution does nothing to help us out either .

Programs that will be helpful:
Prime95 - http://www.mersenne.org/freesoft.htm
RMClock - http://cpu.rightmark.org/download.shtml (1/2 down the page)
I9KFanGUI - http://www.notebookforums.com/showth...ight=I9KFanGUI (keep track of temps)
NHC - http://www.pbus-167.com/ (another program to keep track of temps in case I9KFanGUI doesnt work on your laptop)
TrueVolt - http://discord.anarchs.com/TrueVolt.exe Forum member Hazridi wrote this program to help those figure out what pin mods give what voltage

STEP BY STEP INSTRUCTION

STEP 1: Plug in your chip without any pin mods to boot up at its stock speed to make sure it A) it isn't DOA (dead on arrival) and B) find out your your stock voltage.

STEP 2: These are two figures taken from the official Intel technical documents on the Penitum M line:


The pin lay out of the P-M line, and


The Intel specification voltages and their associated switches/pins

The actual pin mod that triggers the Pentium M 400FSB to run at 533FSB is by grounding the C16 PIN by connecting it via a small piece of wire to either the B16 or the C15 VSS pins. I recommend doing this first to see if you can boot at stock voltage at the new speed. If you can get into windows, run prime95 and see how stable you are. Generally I go by the 12 hour rule: if you get no errors in 12 hours, you're stable. The quick and dirty way around this is to just lower your voltage via RMClock while prime95 is running. If you can lower it past say ~0.050V below your stock without instant erroring in prime, you're stable at stock voltage. If you are stable, CONGRATS, you're done. If not, the fun begins .

STEP 3: If you aren't stable, you're now going to want to check out the second figure in Step 2. Voltage is requested based on the six VID pins (found in the upper left hand corner of the processor pin map, right below the key). These six pins are either set to high (1) or low (0) and, kind of like a key, select the correct voltage based upon their sequence. Stock voltage for Pentium M's is generally in the 1.308-1.356V range. Once you know your stock voltage, you can figure out what pins you need to turn to 0 by looking at the chart to get a higher voltage. VID pins are turned low (0) by grounding them (connecting with another little piece of wire) to a nearby VSS pin. But how do you know what voltage you need? Here are a couple of quick ways to find it:

- Ask me (). I can give you a good estimation based on how prime stable you were at stock voltage

- If your stock voltage is 1.324V or you can boot at stock 1.340V and then select 1.324V using RMClock, ground VID3. When 1.324V and VID3 is grounded, this sends the voltage zooming up to 1.452V. You can then select an voltage between 1.340V and 1.452V by selecting the corresponding voltage between 1.212V and 1.324V. For instance, selecting 1.324V will give you 1.452V actual, 1.308V will give you 1.436V, etc. Note that after you find you stable voltage, I highly recommend that you hard wire it in via pin mods and DO NOT rely on software voltage modification because it can be a bit flaky and not send full voltage at all times (due to the lag in its overriding the hardware voltage request).

NOTE: When you pin mod, there is NO software that can read the actual voltage that you're getting. Software only reads requested voltage, not actual. Therefore, when you check your voltage and see that it doesn't change from stock, this is completely normal.

For more specific cases and other things I haven't addressed, post away .

Further reading/resources I used:
http://www2.ijib.com:1337/phpbb/viewtopic.php?t=3
ftp://download.intel.com/design/mobi...s/30218907.pdf
http://www.xtremesystems.org/forums/...4&page=1&pp=25
http://www.xtremesystems.org/forums/...7&page=2&pp=25
http://www.hardforum.com/showthread....oto=nextnewest

ADDITION: Fellow forum member ozzlo has written an extensive guide on voltage modding over at ocforums, give it a read!
http://www.ocforums.com/showthread.php?t=452126

Cheers all

Mr. K6
post #2 of 565
So can i overclock my 760? I guess no isnt it?
post #3 of 565
This sounds promising, if abit complicated but i like the sound of 2.5ghz of a 1.7ghz.
post #4 of 565
Thread Starter 
There is a chance that you can overclock your 760 once motherboards that support 667FSB come out (yonah maybe?). The possiblity seems to be there, the chipset and RAM aren't though.

The problem is you dont have total freedom with FSB and multipliers, but at least you have more options now. 1.7->2.26 is a definite yes (works well in mine, even with an "unlucky" chip). Also, many more 1.8->2.4's should also work now. The only problem would be if you got a 1.8 that WAS a bad overclocker, then you might have to throw some major voltage at it, which I doubt the Dell cooling solution can handle well (temps 60, 65, maybe even 70).

Anyways, I'm gaming and testing this beast out to see how it and the subsystems run together. I'm also testing it on battery. Battery life doesnt seem to be hurt that much, maybe 15 min. max from my undervolted 760. However, that all is completely determined by the applications used.
post #5 of 565
Great guide man! Now, will it work for 740 533FSB as well?

I put my 740 into my new beloved lapdesk, but SuperPI barks on anything above 1920Mhz. And my P4P800-VM does not have vcore adjustments, such a shame.

I am talking desktop here guys, don't get confused.
post #6 of 565
Thread Starter 
UPDATE: Yah, she needs WAAAAYYY too much vcore to pin-mod. Shame really. Oh well, back to 2.0 GHz =/

and Maine Coon, pin-modding is definitely the way you want to go then. Follow those charts and find what you want and then crank the voltage
post #7 of 565
Quote:
Originally Posted by Mr. K6
UPDATE: Yah, she needs WAAAAYYY too much vcore to pin-mod. Shame really. Oh well, back to 2.0 GHz =/

and Maine Coon, pin-modding is definitely the way you want to go then. Follow those charts and find what you want and then crank the voltage
Naah..I'll leave her the way she is. Dothan 1920mhz translates into 3456mhz P4. Good enough for media control HTPC. And she already controls accessDTV and Sony 400 DVD changer and DirecTV receiver. And can play DVD's and burn them at 8x.

Mind you, she is made mostly from the laptop spare parts: CPU, HDD, DVD burner

Sorry, didn't mean to hijack the thread, I am just sooo excited..
post #8 of 565
If the 667Mhz Yonah board came out... could I be able to pinmod the 750?
post #9 of 565
So if the 667 Mhz comes out, we have to purchase it and replace the recent board by ourself, dont we (Is that a hard job?)? Is there anyone know when it comes out and how much it will cost?

Cheers.
post #10 of 565
Thread Starter 
All are possbilities. It depends on the constraints of the new boards and if I am able to find the pin settings to change this. Right now I'm just speculating, so dont get too excited yet . Unfortunately I have to go back to college and this weekend at home was the only time I had to work on this physically, but I'll take a look around.
post #11 of 565
I feel that my chips are not in the limits which mean I could have a possibility to pinmod in the future.
post #12 of 565
Sorry fella's, but I truly don't get what the fuss is all about.
I just pinmodded my 745 1.8GHz to do 2.4GHz and even undervolted it to 1.276v
Runs as smoothly as a baby's bottom.

No pinmodding for intermediate voltages and definately not for any higher voltages. Next on my list is to make a 2.0 do 2.66GHz... with less than stock voltage of course.

Edit: I think I understand it's for the voltage-requirementally-challenged people amongst us... sorry about that.
post #13 of 565
Thread Starter 
Quote:
Originally Posted by tijgert
Edit: I think I understand it's for the voltage-requirementally-challenged people amongst us... sorry about that.
Ding ding ding, we have a winner

Unfortunately, not all CPU's are created equal, but this might add some functionality to less-than-spiffy CPU's.

Also good luck on that 2.0-2.66GHz, I guarantee you it's not going to happen on stock volts with the Dell cooling solution. Maybe if you put it on a cascade, but that kind of defeats the purpose of a laptop
post #14 of 565
Quote:
Originally Posted by Mr. K6
Ding ding ding, we have a winner

Unfortunately, not all CPU's are created equal, but this might add some functionality to less-than-spiffy CPU's.

Also good luck on that 2.0-2.66GHz, I guarantee you it's not going to happen on stock volts with the Dell cooling solution. Maybe if you put it on a cascade, but that kind of defeats the purpose of a laptop

Just ignore it Mr.K6.

That smell of air in Amsterdam..I will never forget it.
post #15 of 565
Thread Starter 
lol, come on now, group hug group hug

Hmm..., I think I'm definitely going to go for the 1.8 on this, 2.4 is just so shiny and cool
post #16 of 565
Thread Starter 
Just to post an update, I'm now officially at 2.4GHz. Unfortunately the chip wasn't totally stable at stock volts (1.34v), errored in prime after 15min and wouldnt do PI calcs, so I ripped here back open and grounded VID3. I'm now at 1.452v and she's perfectly stable so far, gonna test for lower volts. Still, she's loading at ~60C, which is damn fine in my book, and I'll lower the volts when I can. Here's a performance screeny for you guys:

Updates will follwo and I'll post benches if you guys want them. Once I get my Go 7800GTX, I hope I can get into the 8k club
post #17 of 565
So is this still for 7x5's only? Or are you saying that 533FSB chips can be pin modded too?
post #18 of 565
Thread Starter 
Yah, still for the 7x5's unfortunately . However, this does prove that you can have a less than stellar chip and still make out alright. Prime testing right now and she's great. Loading at 57C @ 1.372v and not a hiccup. Considering that 1.340v errored after ~15min, this is spot on. Gonna re-assemble my laptops and do some hardcore gaming and benching to stress test. Anyway, I highly recommend trying a 1.8
post #19 of 565
Ok, so what you're saying is that with normal pinmodding; like the kind listed in the big sticky at the top of the page the only reason a CPU wouldn't pinmod high would be because your mobo can't get enough power to it. Using this pinmodding technique, you can pinmod it and change the amount of power that is sent to the CPU so essentially you can pinmod any of the 7x5 series into a current Dothan mobo and you no longer have to risk that it isn't able to run stably?

If so, I need to go order my 745...now I just gotta come up with the $300...
post #20 of 565
I'm actually planning to sell my 745 sometime soon and upgrade to the next step up. I might get lucky and end up with a chip that can do 2.66 without breaking 70C on load (haha!).
New Posts  All Forums:Forum Nav:
  Return Home