Core 2 duo overclock suggestions - voltage

X

Xenozx

2[H]4U
Joined
Jun 16, 2004
Messages
2,127
Ok, I have a E6420 Core 2 Duo w/ 4 MB of cache. I currently have it overclocked to 3.5 ghz @ 1.425v. I would like to go higher, but only if I can be sure it can be used on a PC that is up and running usually 24/7. I dont want to go super high in the voltage department, and was currious what would be a safe voltage for every day use? Obviously Ive already made it from 2.13ghz to 3.5 ghz, so thats nothing to be ashamed about, but the 3.7ghz+ territory looks really good :) But I only want to proceed to a clock speed like that if I can run the PC 2/47 that way, and at a safe voltage.

I guess my questions are as follows:

1.) whats a safe voltage for a core 2 duo as far as overclocking with aftermarket cooling to be used on a mostly 24/7 basis?

2.) is it OK to overclock the processor to the limit of the FSB, but leave the memory locked @ 900 mhz?

3.) HT multiplier that starts at 5x, should this be reduced as you overclock, and if so at what FSB should it be reduced to 4x, 3x, 2x?

clocksspeeds.JPG


4.) I have corsair DDR-2 800mhz XMS2 6400 C4 4-4-4-12 memory and currently have it at the clocks above @ 2.1 volts. I was currious if anyone had any results overclocking this memory in the past, and if so about what the limits would be on 2.1v and the 4-5-5-12 clocks?

sorry if this is full of run on sentances and makes little sense. I would be happy to clarify anything if needed, hopefully its understandable. IM at work so typing this up kinda fast :)
 
Nice overclock so far.

The first question is are you Orthos stable at 3500 MHz with 1.425 volts? If you can run Orthos for 4+ hours you'll be the first 6420 in the [H] OC database.

Next question is what type of apps will you be running 24/7? If you're running video encoding apps 24/7 or running @home 24/7 then zero errors and 100% stability is your main goal.

There's generally not much performance to be gained reliably beyond your present speed. 3700 MHz is only a 6% gain over 3500 MHz but it usually takes a lot more core voltage and a lot more stress to run 3700 MHz reliably.
 
Yes, I ran orthos for a few hours to make sure my temps where OK. I Idle arround 35, but orthos had my processor under load @ about 61c. this was after about 2 hours. the each core itself was about 50-51 degrees.

It sits idle most of the time. I play lots of games, but while IM at work it just sits idle, and I may have Bit torrent up or something. I also VNC into it from work sometimes but other then that its pretty idle. Its not under load 24/7.

Yea, if its not worth it going any higher, I will probably just stay where I am. The speed increase is amazing compared to stock, so I am happy, but If I could have easily made it to 3.7, I would have went there as well. Is it possible I might make it there under my current voltage, or would I need to raise it?
 
The higher you go, the more volts it will need, which creates more heat to deal with. ;p

Id just leave it. 3.5ghz is good. If youre worried about the difference between 3.5ghz and 3.7ghz, or even 4ghz, dont... Its minimal. I wouldnt worry about it. 3.7ghz might give me 5-10 fps more in games....big whoop.
 
Xenozx said:
Yes, I ran orthos for a few hours to make sure my temps where OK. I Idle arround 35, but orthos had my processor under load @ about 61c. this was after about 2 hours. the each core itself was about 50-51 degrees.

It sits idle most of the time. I play lots of games, but while IM at work it just sits idle, and I may have Bit torrent up or something. I also VNC into it from work sometimes but other then that its pretty idle. Its not under load 24/7.

Yea, if its not worth it going any higher, I will probably just stay where I am. The speed increase is amazing compared to stock, so I am happy, but If I could have easily made it to 3.7, I would have went there as well. Is it possible I might make it there under my current voltage, or would I need to raise it?
Click to expand...

Only you can answer that question. Try 3.7ghz and see what happens....but as others have said, the voltage increase + temperatures going up is not worth the extra 200mhz.

I have an E6600 running @ 3.8ghz with 1.43V (Windows) and if I want to do 4ghz, it needs 1.54V and the temperatures go through the roof. Just not worth it to me!
 
unless your going for benchmarks its not worth it, keep it stable and living at 3.5. if you want to see how high you can get you can go furthur, those temps are not dangerous but for 24/7 its not worth the extra 200mhz
 
Here's an excellent article on how an E4300 scales with cpu voltage:
http://www.nordichardware.com/Reviews/?skrivelse=500&page=5

Those numbers are for SuperPi 8M on both cores benchmark stable but not Orthos stable.

In this document:
ftp://download.intel.com/design/processor/datashts/31327803.pdf

Intel® Core™2 Extreme Processor X6800Δ and Intel® Core™2 Duo Desktop Processor E6000 and E4000Δ Sequences

otherwise known as the Intel Core™2 Duo Datasheet it shows in Section 2.6.1 Table 4 that the maximum recommended core voltage for a C2D processor is 1.55 volts.

Going beyond that is foolish. If you don't care about your processor's life span and are simply going after a new world record then it's fine but for 24/7 use there's no reason to be above 1.55 volts.

To be honest, even going beyond 1.50 volts doesn't really get you a hell of a lot of extra MHz so why bother. Somewhere in the 1.45 volts to 1.50 volts range is generally sufficient.

The important numbers to watch are your core temperatures. The best program to use for that is CoreTemp 0.95.

Your processor has a maximum rated core temperature of 85C but will generally start to lose long term Orthos stability when fully overclocked at full load when the core temperature hits 75C. I've found that when my E6400 is fully overclocked if I keep the maximum temperature while running Orthos at about 70C to 72C, I never have a problem with stability.

If your memory is willing, I think 3600 MHz would be a good number to shoot for. You've got a great cpu cooler and your processor isn't screaming out for more voltage yet so see how much voltage you need to be stable at this level and decide if it is really worth it or not.

For overall performance you might be better staying at 3500 MHz and seeing if your memory can run with tighter timings like CL 4-4-4-12.
 
Thanks for the advice. I think I will take some time and try to go higher while staying under 1.5v. Once I get to the max stable I will probably stay there assuming temps are in line. I do agree that a 6% additional overclock is not worth additional heat, especially on a PC that will be running 24/7.

I do have one last question. I have a 2gb kit of the memory below:

CORSAIR XMS2 2GB (2 x 1GB) 240-Pin DDR2 SDRAM DDR2 800 (PC2 6400) Dual Channel Kit Desktop Memory - Retail

Its rated @ 2.1v. Is it OK to go up to 2.2 or 2.3 for stability purposes? Whats the general rule of thumb when upping the voltage of memory modules that would be considered safe?

I am having stability problems with my memory I can barely do 925mhz @ 5-5-5-13 where stock it handled 800 mhz @ 4-4-4-12. I seen many reviews get into the 1000mhz by laxing the timings and increasing voltage, and I would like to try this as well. I think the voltage side is what I am not doing, as Ive stuck w/ stock voltage.
 
If you are overclocking the RAM, then yes, you need to up the voltage to @ least 2.2V.

Also, Corsair RAM is not the best overclocking RAM, so don't expect to get what other's have. Crucial, Buffalo and G.SKill are known for good overclocks!
 
Here's my favorite review of how memory scales with additional voltage:
http://eclipseoc.com/index.php?id=1,72,0,0,1,0

Bumping your memory up from 2.1 volts to 2.2 or 2.25 volts shouldn't hurt anything and in the testing I've done with different memory modules, it has always performed better with a little more juice.

Corsair and OCZ change the actual memory chips they use on their modules from one version to the next so the only way to be 100% certain what chips your memory are using is to pop off a heat spreader and have a look.

The Micron D9GMH and D9GKX chips have a great reputation for going to DDR2-1000 and beyond at 2.2 volts but some of the competition like Elpida struggle at this level.
 
Well I think I hit my PC's limits. I cant get stable over 3550 mhz. I can boot into windows all the way up to 3.75 ghz but after running Orthos for more then 10 minutes I will get a BSOD.

I worked my way back from 3.75 to 3.7 to 3.65 to 3.6 then to 3.55 and the only place I was stable was 3.55. I up'ed the voltage up to 1.485 and even there I couldnt get any futher. I just decided to stay at 3.55 @ 1.45v. My memory appears to have issues over 900 mhz :( I cant get the memory stable above 900 mhz @ 2.2v :( I tried timings over 555-13 and over 900mhz was still a no go. The symptoms of this was my PC would just lock up, and I would have to hard reset :( I am @ 887mhz w/ 5-5-4-12 timings, so I will just leave it at that.

I dont know if my 480w PSU is a issue or not. When I run orthos I can see the 12v and the CPU voltage drop quite drastically when under full load. I have a silverstone 750w PSU I hope to get back very soon, and I may attempt to overclock further once I have that PSU.


So 4 hour Orthos stable 3550 mhz w/ 887 memory @ 5-5-4-12 timings :)

I guess thats not bad, and should have no complaints :)
 
Sorry to get off topic, but what version of the evga card do you have? N811 or N815 ? Superclocked or not ?
 
orion23 said:
Only you can answer that question. Try 3.7ghz and see what happens....but as others have said, the voltage increase + temperatures going up is not worth the extra 200mhz.

I have an E6600 running @ 3.8ghz with 1.43V (Windows) and if I want to do 4ghz, it needs 1.54V and the temperatures go through the roof. Just not worth it to me!
Click to expand...

out of curiousity is this on air? I imagine it would have to be water :\
 
currently at 8 x 425 = 3.4 ghz @1.45

i get errors at 1.425 at even 3.2.ghz, but they all go away at 1.45.

maybe i'll try to up the fsb tonight to see how our cpu's compare.
 
Just started work on my e6420 over the weekend. Currently at 8 x463 = 3.7 ghz@1.225v
I'm very excited to see where I can take this processor. New ram and a better radiator are on the way. Will post cpuz verification tonight

My only concern was that even on a Swiftech Apogee GT and single 120mm rad I was hitting as high as 56c.

With such a low voltage, how would I be generating so much heat?
 
Terc said:
Just started work on my e6420 over the weekend. Currently at 8 x463 = 3.7 ghz@1.225v
I'm very excited to see where I can take this processor. New ram and a better radiator are on the way. Will post cpuz verification tonight

My only concern was that even on a Swiftech Apogee GT and single 120mm rad I was hitting as high as 56c.

With such a low voltage, how would I be generating so much heat?
Click to expand...

did i read that right... 3.7ghz at 1.225v?

you undervolted the cpu to get a speed that only some people can achieve with 1.45v or higher?
 
Tzzird said:
did i read that right... 3.7ghz at 1.225v?

you undervolted the cpu to get a speed that only some people can achieve with 1.45v or higher?
Click to expand...
Sorry, I'm an idiot. We all have to start somewhere. I'm actually just over 1.4v. I read the voltage from cpuz(Me=noob)... I had left my voltage on auto and my p5b had given it a healthy bump.
I'm getting back to work on this rig now. Will post results when I have something better to offer.
 
Terc said:
Just started work on my e6420 over the weekend. Currently at 8 x463 = 3.7 ghz@1.225v
I'm very excited to see where I can take this processor. New ram and a better radiator are on the way. Will post cpuz verification tonight

My only concern was that even on a Swiftech Apogee GT and single 120mm rad I was hitting as high as 56c.

With such a low voltage, how would I be generating so much heat?
Click to expand...

aint no way man... Your processor is fudged, probably gonna die in less then a week if that statement was accurate.
 
your CPU voltage is 1.54v.

CPU-Z doesnt detect our processors right. Mine shows a lower voltage too, and it shows my 3500mhz overclock as 2600mhz.

Speedfan shows the corect voltage in the screenshot you posted. 1.54v.

I would have hoped you would have known thats what you selected in the BIOS, as a 1.54v vcore would in most cases of had to be manually set.
 
Xenozx said:
your CPU voltage is 1.54v.

CPU-Z doesnt detect our processors right. Mine shows a lower voltage too, and it shows my 3500mhz overclock as 2600mhz.

Speedfan shows the corect voltage in the screenshot you posted. 1.54v.

I would have hoped you would have known thats what you selected in the BIOS, as a 1.54v vcore would in most cases of had to be manually set.
Click to expand...

For whatever reason, my p5b deluxe AUTO volted me to 1.54vcore... the thing is, loosening timings on my ram and raising its voltage allowed me to reach the same speeds. Not sure what the board was doing, but I've since gone through and set all my own voltages (should have done this to begin with)

As far as cpuz showing 3500mhz overclock as 2600mhz, that's stepping, watch it when you run a processor intensive task, it'll go back to 3.5ghz, just look at the multi. I bet it's at 6x when you see your 2600mhz reading and 8x for 3500mhz
 
yup, sure enough 2662mhz x 6 x when in chill mode. I guess thats the answer to that question :D

Thanks for the heads up.

That auto set to 1.54v is kinda scary, I mean, imagine if it did that on stock cooling solution....
 
You must log in or register to reply here.
Back
Top