Overclocking a Dell XPS Gen4 650 3.4 to 4.5ghz EE!!

C

calibos

n00b
Joined
Apr 1, 2005
Messages
63
I'd like to head off the 'should have built your own' comments by saying that I got an amazing deal on an Dell XPS Gen 4 and 2405FPW monitor and shelved my self build plans.

Now to the point. Some very brave soul called socrilles from extreme forums tested out a solder mod overclock adapted from the original notebook version on his Dimension 8400 with 925xe modo and 6 series P4. Now we know a 6 series is identical to a 3.78EE except it runs with a 800mhz FSB instead of 1066mhz. We also know that a 925xe supports 1066mhz EE's. This guy put the two together and soldered pin bsel0 to bsel1 on his CPU tricking the mobo into 1066FSB mode which in combination with the multiplier of 15 on his 3ghz 630 gave him an overclock and equivelent of a 4ghz EE!! :D

Because everything else is locked on Dells this is the only way to overclock one and applies only to 925xe and 6 series owners and means each speed of CPU can only yield one specific overclock. Thing is, I went and spent the extra (seeing as I was getting the PC cheap anyway) on a 650 3.4. This has a multiplier of x17 which yields a 4.5EE!

I don't know whether this is a good or a bad thing? :D Good if I could cool one without spending a fortune on a peltier or vapo but it might actually be cheaper to pick up a 630 which would OC to 4ghz and at most just require water cooling.

So what would I need to cool a 4ghz EE or a 4.5ghz EE?? !! :D
 
Seems like an awfully big risk seeing as that is a VERY hefty jump for a P4 (the 4.5Ghz...4.0Ghz isn't quite as big) no matter what kind of cooling you're using. I'd imagine it would be hard to convince Dell that you were just holding a soldering iron over your open XPS and happened to drop it if you screwed it up :p
 
Oh, I know my warranty would be gone if I ever attemped this! :D

As it turns out Socrilles posted on this subject here too. The thread is a few pages back. Someone mentioned that a 6 series with its speed step tech actually has a downward unlocked multiplier. Does this mean the multiplier could be lowered with software or can it only be done by the chip itself under stress or by a bios setting? Obviously its a no go with a dell bios but if it could be done in software I could lower the multiplier to 15 to get the 4ghz instead of 4.5ghz?? Then when the liquid metal nano coolers come out for CPU's I could up it back to 17 and get my 4.5ghz :D :D
 
Well, the same software you could use to lower the multi can probably OC the FSB. Seems like a safer way to deal with the OC :).

Usually, multipier settings are a BIOS thing.
 
Do not solder it, either wire mod it or insulate it, I've done something similar, and with wiring or insulating it's reversable.
 
I don't know if something like ClockGen would work (I've only used it for NF4 based boards), but you can adjust the voltage from windows (nice feature).
 
Jerunk said:
Do not solder it, either wire mod it or insulate it, I've done something similar, and with wiring or insulating it's reversable.
Click to expand...

Thats the thing though. On the Dell version of the mobo AFAIK (and maybe it applies to all 775 sockets) you can't do a wire mod on the socket for some reason but must solder the processor pins to get this to work. I know its a warranty killer and very possibly a processor killer too! :D

As for the voltage, I'm hoping it'll do it on stock voltage as no doubt clockgen etc won't work on a Dell mobo. (Hey, I didn't go into this blind, I knew what the downsides to a Dell were :D ...accepted them because of the great deal...and was pleasantly suprised to hear about a new overclocking method!)

Does anyone have any info on whether its possible to lower the multiplier on the 6 series to get me a more managable 4ghz rather than 4.5ghz on my 650 and also what kind of cooling would be required for either.

Its not like I'll be doing this mod now anyway because tbh I don't really need the extra performance atm and by the time I might need it my warranty will have expired anyway, its just that I've just found about this Dell overclocking method and it sounds interesting so I might as well do some research on it now even if I don't intend to do it till next year anyway.
 
I guess the only way we can find out if it works is by soldering it ...

I would try o/c the cpu on another motherboard and see if it will reach that speed first of all. If it doesn't then forget about it.

Also ... is your cpu prime95 stable right now ? This has a great effect on o/c potential

well my man .. good luck on your adventure ...
 
Hehe, when I first saw this thread I thought you had already done it :), anyway if you really want to be brave you can adjust the voltage as well via soldering, I haven't looked at the pins that determine it, but I'm sure there is some you can ground to raise vcore so if your willing to put WCing in there is a decent chance you could get 4.5 ghz. Also the solder mod is reversible. The first time soldering I tried to boot up and nothing happened, so the solder was undone I tried to boot up again and still nothing. Which led to the eventual realization the mobo had to be seated perfectly or Dell's wont boot. After figuring this out I did boot at default first (successful unsolder) and then soldered to 4 ghz. When I run under stress it actually gets up to about 4.1 ghz on default and it is prime stable under two instances now. Also after soldering the second time it took a decent amount of force to snap the processor in, which I attribute to some of the solder getting in the hole - my brother in law (the EE) was doing the soldering and didnt seem worried at all about doing a perfect job (although was doing a very good job.
 
So you got it prime stable with two instances in the end Socrilles?? What made the difference. Did you improve cooling to get it stable. What cooling are you using now anyway. Standard 8400 air cooling, improved air cooling or Water cooling??
 
just the standard air cooling that came with it (I didn't even put any new thermal compound on) I think the old thermal goo just had time to find the crevices, either that or maybe it was a cooler day, but I got 3 hours two instances stable (I stopped after 3). The one other thing that may be an issue for you is the PSU, I don't know how beefy the XPS one is, but the 8400 is only a 350 watt.
 
run prime95 overnight and check how stable it is then .. 3 hrs in my opinion is never enough .. it has to last the WHOLE night ..

And use blend test
 
mohammedtaha said:
run prime95 overnight and check how stable it is then .. 3 hrs in my opinion is never enough .. it has to last the WHOLE night ..

And use blend test
Click to expand...
Do you mean one instance overnight or 2 instances like socrilles used for his 3 hour test??

Well impressed that you are stable on standard cooling Socrilles!! Just goes to show that Dells shroud method has its merits. They've had shrouds for years and its only now that you are seeing movement in that direction by others with their 'Wind Tunnels'. I always thought it made more sense myself, ie pull air from mobo area across heatsink through shroud and straight out the back rather than a fan sitting on a heatsink blowing hot air off a heatsink in all directions over the mobo and hoping the rear fans will suck some of that warm air out the back before it heats up the mobo or other components. The only differance with the wind tunnel now being adopted by others is that they are pulling air from outside straight through across the CPU heatsink and out the back.

This really gives me great hope for my Gen 4 XPS. Its got a 460w PSU and even better cooling than the 8400. 2 80mm Rear exhaust Fans for the shroud/CPU, PSU in a seperate compartment at the bottom with its own 2 fans and an intake fan mounted over the Graphics card/PCI slot area that pulls air from a vent at the front through an intake tunnel and blows this cool air over the GC/PCI slots. The Gen 4 has been replaced on the Dell Ireland site by the Gen 5 so they are the only XPS specs I could find, but the amount of fans and their configuration seems unchanged from the Gen 4. The heat dissipation specs for the Gen5 and I assume the Gen 4 are 931.2 BTU/hour compared to about 450BTU/hour for the Dimension 8400.

I take it this means I would be good to go with my XPS air cooling at 4ghz. Does anyone think with double the heat dissipation power of the 8400 that I might be able to do 4.5ghz on air?? How about mounting a fan on the side of the heatsink giving the rear fans a helping push of air through the heatsink into the shroud and the rears pulling it out the back (as they are doing already)

If 4.5ghz is still out of the question on air, can anyone clarify if I can manually lower the multiplier to get the more manageable 4ghz

[Edit] Socrilles, are Dell still using thermal pads??
 
I think your going to be stuck at 4.5 ghz at least for boot up, the dell mobo isn't going to have the option for changing the multiplier even if there is a program you can run in windows. Oh I also just got a PM from another guy who has done the same mod successfully I'll try to get him to post his results either a new thread, or in mine or yours. Oh and for my stability I did blend test on one instance, and max CPU usage on the other 3 hours is plenty for me. I've also seen 4.5 ghz on the 6xx series chips with relatively low volts about 1.4 to 1.5, but its always been with WC, I just think it might be too much for air
 
Just read a few threads on extreme overclockers where someone who seemed to talk with some authority reckoned a 6 series prescott would do 4.5 on air easy. Did the 4.5 guys you know of already have WC anyway or were they playing it safe with WC and never gave air a try?

As for the multiplier. So even software changing the 6 series multiplier(ie taking advantage of the speed step technology in the chip) has to do it through the bios?? Ah, I see. Dell locked bios strikes again!!

WRT v-core. You did your OC at standard V-core?? and you reckon I would need more to get to 4.5 stable on air or water.Can I raise v-core on my XPS?

So basically then its 4.5ghz or bust? Either get it stable by air or WC at 4.5 or sell my 650 for a 630 and mod that to its 4ghz overclock
 
yah I think that sums it up, I just don't really see changing the multiplier before windows if at all. I think I read somewhere XPS's do have some OCing options somewhere but I could be completely wrong or that might be a future model. As far as changing the Vcore goes I'm not sure where those pins/holes are on the chip, the nice thing about the bsel0/bsel1 is that its right on the outside of the chip so you don't have to worry as much about hitting other contacts.
 
Its the new XPS Gen 5 with the 955 mobo and dual core 840. What I gathered was that the 840 could be overclocked from 3.2 up to 3.6 in the Dell Bios. Imagine that, an officially overclockable Dell!! Question is whether the Bios would allow you to overclock any chip you put into that mobo. Would the bios lock out the overclocking if it detected any other chip other than the dualcore 840.

Still, although the Dell documentation issues warnings about the dangers of overclocking, ie Users may decrease the life expectancy of the processor by overclocking blah blah, I reckon thats just to cover themselves. I think this shows Dells confidence in the standard cooling system in their XPS Gen4/5.

Does anyone know how an overclocked Dual-Core 840(3.2) overclocked to 3.6 would compare temp wise with a single core 6 series overclocked to 4.5. If they are similar that gives me even more hope of cooling a 4.5 on air!

As for the Volt solder mod Socrilles, one would merely have to be a little more careful with the appropriate pins being closer to the centre with more pins surrounding them, but that otherwise it would be exactly the same process?? Is that what you mean?

Would one first just do the FSB solder mod and see if the system was stable and only if it wasn't stable one would try upping the voltage with the volt solder mod??
 
B/c Intel is going to be leaving the multipliers unlocked on their EE's (to compete with the unlocked multi's on FX's presumably).

Dell just lets you adjust the multiplier which won't work on just "any" chip.
It would have to be an Engineering Sample (unlocked) or possibly IT MIGHT ALLOW lowering the multiplier.... but with no FSB adjustments, thats useless!!

P.S. Do the Gen4 & Gen5's have a proprietary PSU connector to the mobo, or is it ATX standard pin?

I assume its ATX standard as most of the recent mobos.
 
Just had a search for some pics of the LGA775 sockets and cpu's.

Contact pads on the underside of the CPU and spring loaded pins in the socket. How on earth did your mate manage to solder these?? Wouldn't the solder just break once the pins moved when you seated the cpu and wouldn't that be dangerous, bits of broken solder joining other pins together? Why is it not possible to use a conductive pen I've heard mentioned to ground the appropriate contact pads on the cpu?

Say it was possible to use the pen (I'd be much more confident doing that than using a soldering iron) Would it be dangerous for the CPU to boot at 4.5ghz. ie. if my particular CPU just wasn't capable of that OC, or my cooling wasn't good enough or I didn't have enough default voltage for that speed would it just mean my PC wouldn't boot, I'd know it wasn't working and would just remove the conductive 'ink' to return the CPU/FSB to default with no harm done...or....would I think everything was working but my CPU would be acting strangely and actually be melting away! :D

BTW heres a pic of cpu-z



Would that voltage be enough for 4.5 or is cpuz reporting wrong. Also, the cpu speed never quite reads 3.4 but varies between 3389mhz and 3394mhz with the FSB never quite reading 800mhz either.
 
Ah, thought so. Saw some guys on extreme talking about upping their voltage on their six series to about 1.5 for their 4.5-5ghz overclocks on consumer mobo's. Thought, hey my Dell mobo/cpu is already running at that for some reason, woohoo!...but did figure it was too good to be true! :D

Getting cold feet now with all the delicate soldering involved and maybe extra cooling required. I suppose its not that I need the extra speed right now more that I would have loved the bragging rights of a cheap Dell XPS4 overclocked to 4.5! Definately a case of craving a bigger e-pen1s!! :D

I saved so much on this Dell that, its not going to be much of a financial hardship to pick up a new mobo and dual-core from Dell in a year or so. :D

Thanks for all your help Socrilles mate :)
 
Read this about Dell's new Gen 5 XPS:
http://www1.us.dell.com/content/lea...s&~lt=popup&~model=xps5&~series=dimen&ref=CFG


Intel® Pentium® D 840 Extreme Edition

3.20GHz with Hyper-Threading Technology. User-overclockable in BIOS to 3.40GHz and 3.60GHz
Click to expand...

"WHoa! Dude, you're getting a Dell..... that you can OVERCLOCK!!"

That's it I'm going to start underclocking.... too many people are jumping on the OC'ing bandwagon..... And I can't go Apple, b/c they just joined the Intel bandwagon... So I'm going to be uber-geek cool, by getting an FX-57 and underclocking it! ;)
 
You must log in or register to reply here.
Back
Top