Tualeron 1.2GHz at 1572MHz ,standard cooler!

[whitey]

Ok guys,I finally got around to doing the wire trick on my cpu.It seems rock stable at the moment,but i'm a bit worried about the 43 MHz PCI.I didn't try the next step up,because it would have taken the pci to 44 !!.I tried for 1.6 to bring my pci and ram back to spec,but it wouldn't load windows.Here's the info:
1572MHz
131 fsb
43 PCI
45 degrees load
41 degrees idle
Ran 3D Mark 2001 for about an hour,played Giants for 3 hours straight on maximum graphics(aren't holidays great?)with no problems at all.I'm going to order a good cooler for it this week,even though the standard cooler with arctic silver seems to be doing ok,i'd like lower temps.I'll try a cold boot in the morning to see if it is a heat issue with it not booting at 1.6gig.Anyone want benchies ? let me know. Heh ,what's this smoke coming out of my box? lol.

[Czechmate]

Excellent result, whitey. Always nice to see another dedicated Intel overclocker.

Why is your PCI speed so high?? Don't you have a 1/4 PCI divider?? If I recall only the Gigabyte 6OXET had issues with being unable to select a 1/4 PCI divider with 100MHz processors, and this was addressed with a new BIOS which features full AGP and PCI speed control at any FSB. Which motherboard are you running it in??

It would be interesting to see some Sandra benchmarks, even if it is just the raw numbers. I'd like to see the CPU Arithmetic and Multimedia performance, as I want to compare this to my Tualatin PIII-S at 1589MHz (for now, until I get my peltiers ).

It's good to see that the wire trick works on the Tualatins as well. How did you end up performing it?? Did you wire the VID pins together or did you stick a wire into the VID "holes" in the socket itself??

Finally, your temperatures are indicative of typical Tualatin processors: there is very little difference between the idle and load temperatures. I don't know why this occurs, but I see it as a positive, as the CPU isn't subjected to large thermal stresses with excessive temperature differences.

[flu!d]

Good job Whitey!! What voltage are you running?

I believe the TUSL2 mobos don't utilize the 1/4 pci divider until right on 133 mhz FSB....

[whitey]

I'm running it on an Asus TUSL2 ,I don't think i have a quarter divider under 133,as Flu!d mentioned.It's running at 1.72 volts with the wire trick,this board delivers slightly more voltage than default. I wrapped the pins ,two loops each end,so it sat down flush.Definately not a job you would do after a night on the piss!.Sandra 2001 results are:
Integer SSE = 8528
Floating point SSE = 10686
Dhrystone ALU = 4421 MIPS
Whetstone = 2108 MFLOPS

3856 3D Marks 2001
8552 3D Marks 2000
Asus Geforce 2 32mb Deluxe @ 220 core, 395 memory.
I'm going to try some case fan mods first,before i get the new
hsf.Do you reckon 45 degrees is too hot?

[blimpyboy]

Originally posted by whitey
It's running at 1.72 volts with the wire trick,this board delivers slightly more voltage than default. I wrapped the pins ,two loops each end,so it sat down flush.

Excellent result!

Do you mean you did the 1.675 V pin trick, but your mobo is delivering 1.72 instead?

Also do you know how thin (in micrometres) the wire you used is? Because it sounds as though you are not suffering from a slightly tiled cpu/hs that some do with your method of the pin trick (other put the wire in a U-shape and insert it into the actual holes on the mptherboard as czechmate mentioned)

[whitey]

Yes,this mobo supplies a bit more voltage than what it's set on.For instance,the standard voltage for the tualatin is 1.52volts on this board,instead of 1.475.I think this is the case with all of these Tualatin boards made by Asus.I'm not complaining though,every little bit helps as i still have no adjustment in the bios.At the standard voltage i could only reach 1440MHz,so the 0.2 volts gets me an extra 132MHz .As far as the wire size goes,I don't know what size it is,it was just one strand of copper from a car speaker wire.For anyone who hasn't done this before,this is how i did it.Grab a sewing needle slightly smaller than the pin,hold the wire against the needle and wrap it around twice.Cut the end off so you have a small coil,now gently slide the coil over the inner most pin.Now wrap the wire twice around the pin on the outside edge of the cpu,and cut the end off flush with the pin (I found nail scissors work best for this) .Use the sewing needle CAREFULLY to bend the wire if needed to clear the surrounding pins and to compress the coils on the pins so the wire is against the ceramic .This makes sure the cpu sits flat in the socket when fitted .I tried the other way ("U" shape wire in the socket)but i thought it was a bit dodgy.If you want the wire size measured,i could do it but my micrometer is at work and i'm on holidays but if you really need to know i can get it.Cheers.

[blimpyboy]

No need to cut your holidays short.

Thanks for the technique, very interesting.

Did you need to bend the wire to avoid any other pins like you mentioned?

[whitey]

Only very slightly,so it was exactly in between the two surrounding pins.To do the "U" wire method,you have to cut a small groove in the socket to keep the wire in place.I didn't want to do this in case i need warranty on the mobo.

[whitey]

I fitted some case fans today,a blowhole fan on top, and another fan on the side of the case with a tube I made up to blow directly onto the standard hsf.Temps have dropped by 3 degrees ,42 degrees loaded,38 degrees idle.I don't think i'll bother with an aftermarket cooler,the standard one is doing fine.Fluid,how are your temps at 1.5 ?? Czechmate is spot on with the temperature variation from idle to loaded,regardless of what frequency the cpu is at ,the variation is only ever 4 - 5 degrees.Now how can i get a little bit more voltage so i can try to hit 1.6MHz ?

[blimpyboy]

You can't get more than 1.675 volts without ripping out pins, which is not something I'd recommend. Its not a reversible procedure as far as I know.

As shown in this picture: you have to connect two wires (just like the 1.675V mod), and remove another pin.

Also this gives you 1.825 volts, the maximum tualatin processors can ask for. This is 3.5 Volts above default voltage, which will most definately require some serious cooling if you want to keep it around 45 degrees.

EDIT
-----
sorry about the huge image, its a small size to download but still a little big.

[whitey]

You mean .35 volts don't you ? I've heard of insulating the pin with several coats of nail polish instead of removing it.Might work,but i'm not willing to try it out.I hope Asus comes up with a new bios that allows voltage adjustment.The bastards did a bios for the Tusl2-c but not the Tusl2

[SyK]

If it helps, with the nail polish thing, I've done this (Not with a cpu mind you, it was in the guts of an old laptop), and found it easier, and much less messy, by poking the pin on the connector into one of those ear-cleaning cotton-wool-bud-on-a-stick thingos... (Sorry, bit technical there. ) then, pouring the nail polish into the hole, then sticking the pin into it and leaving it set for a while.

But, I'm still not sure how much of a good idea it was, it worked, but I dunno how it'd go with the heat on a cpu...

Anyway, worth a thought, if getting nail polish everywhere is what had you worried, it did me.

[flu!d]

Whitey....My temps vary with ambients of course but most of the time I sit at 30 degrees C at idle and climb to 34 degrees C under load....

I have one intake fan, 2 exhaust fans and the PSU fan pumping at through a very large EYO server case....

[whitey]

Luckily my ambient temp doesn't change much,I set it to 21 degrees on my air conditioner.What program do you use to check the temp?On Sandra 2002 it reports mine at 19 degrees!WRONG .I use asus probe software that comes with the motherboard and it seems to work ok.

[flu!d]

I use MBM 5.1. The Gigabyte mobo monitoring software dosen't work with Tualatin processors....