E6400 on a P5B-E overclocking help

keldegar

Limp Gawd
Joined
Dec 11, 2006
Messages
339
Hi, I wanted to know what you guys think I should do.

I currently have my rig at 3ghz (375x8) stable for Orthos overnight (8 hours, manually stopped it). VCore is set to 1.3875 in BIOS but CPU-Z reports 1.304v on idle and 1.264-1.288V on load. Temps are 36-37C idle 48C load.

my RAM is Corsair XMS2 PC5400 DDR2-675 and is rated at 4-4-4-12 @ 1.8v. It has the heat spreaders on it.

I have it currently running with 1.9v and set to 4-4-4-12 @ DDR2-750. Orthos (8 hours) and Super pi 32M stable.

I tried last night to set FSB to 400mhz (DDR2-800) and I left the RAM @ 4-4-4-12 @ 1.9v. It passed super Pi 1M but failed Super Pi 32M. It also fails Orthos Blend in <10Min. Same result with FSB to 401mhz, and I also tried giving the RAM 2.0V and VCore 1.4V in BIOS (1.32 in CPU-z idle), still same problems.

Do you guys think it's my RAM (ocing too much) or is it something else? I don't want to have to up the voltages too much, but I see a lot of other people running their E6400 @ 3.2Ghz.
 
You should try running your ram at CAS 5... It will help you get to higher FSB... Loosen your timings or voltage and try 400fsb...

Try 5-5-5-15 and 2.0 volts... You will go 800fsb and beyond easily...

I have ddr2 pc5400 1.8volts 5-5-5-15 667mhz ram... But I'm running them 850mhz right now at 2.1 volts...
 
2.1 or 2.2 volts is the sweet spot for a lot of DDR2 ram these days.

Go with a little extra voltage before dropping your timings any.

The biggest difference between high performance and regular ram is the voltage that it is rated at. The actual chips used are often times exactly the same so don't be afraid to run your ram up to 2.2 volts.

My Micron DDR2-667 1.8 volt memory runs fine at DDR2-940 and CL4 timings with 2.1 volts.
 
I will try 2.1V and 5-5-5-15 when I get home and see if that lets me run @ 3.2Ghz.

If that passes Orthos overnight i'll try to tighten the timings.

Will it harm the RAM at all to run @ that voltage 24/7?
 
^No It will not, I have my DDR2-800 with a E6400 @ 3.4ghz, at 2.1volts constantly.
 
I had a similar problem with the e6400 and Asus P5B-E and Deluxe. The P5B-E wasn't stable at 2.8ghz with any voltage increase, the P5B-E requires an extreme amount of volts to hit the same 3.2ghz my Gigabyte could hit at stock volts. Albeit, when I was trouble shooting my boot issue Intel suggested RMA'ing my CPU (which turned out not to be the problem), and I got a replacement E6400.
 
TWIN2X2048-6400C4, i'm using it at 4-4-4-12. It's running at 1100mhz or something like that. The higher the speed or FSB you use, the ram will also increase in speed causing you need to give it some volts to make it stable. Did you try the new bios update? I been messing with the bios update for the P5B deluxe and some are good overclockers while others are not.
 
aznpotpie said:
TWIN2X2048-6400C4, i'm using it at 4-4-4-12. It's running at 1100mhz or something like that. The higher the speed or FSB you use, the ram will also increase in speed causing you need to give it some volts to make it stable. Did you try the new bios update? I been messing with the bios update for the P5B deluxe and some are good overclockers while others are not.

I have 0706 i think. I just upped the VDIMM to 2.1V 4-4-4-12 DDR2-800 and it passed Super Pi 32M! I have 400FSB 8x multiplier 1.3875V in the BIOS (1.302V in CPU-Z)

I am going to do orthos overnight, but so far it looks like you guys were right, not enough voltage to the RAM.

Just curious what are you guys' super pi 1M mod 1.4 times? mine is 18.2xx seconds.
 
^ At 800 or higher you need 2.1V and higher. As the speed goes up so does the need for power but too much and it burns out.
 
aznpotpie said:
^ At 800 or higher you need 2.1V and higher. As the speed goes up so does the need for power but too much and it burns out.

I see. Yeah some of the super overclocker's rams needs 2.4V...

Man I am stoked! I thought I hit a wall @ 3ghz.. I think I will have all i need @ 3.2Ghz and DDR2-800. Wow this Corsair RAM really performs. CAS4 @ DDR2-800, when its only rated for DDR2-667! :)

My 3dMark06 went up 300 points too lol.
 
aznpotpie said:
^ At 800 or higher you need 2.1V and higher. As the speed goes up so does the need for power but too much and it burns out.
Does anyone out there actually have a memory module that they have applied too much voltage to and have burnt it out? How about a CPU?

As long as you keep an eye on temps it's surprising how much abuse these components can take before you physically damage them. You're likely to have your computer crash long before you do any sort of damage. I've been overclocking everything and anything that I can get my hands on for the last 10 years and I've yet to fry one electronic component.

I think some crazy people have been feeding their DDR2 as much as 2.6 volts without any problems. For the rest of us 2.2 volts is a good starting point and if you need to go to 2.3 v or 2.4 v for stability and better timings, you shouldn't have a problem.

FYI: Micron D9GMH chips are good for about DDR2-940 speeds at 2.1 volts and CL4-4-4 timings. Not bad for chips rated at DDR2-667. :D
 
unclewebb said:
Does anyone out there actually have a memory module that they have applied too much voltage to and have burnt it out? How about a CPU?
I have about 2 modules of Ram that i burn out by applying too much voltage. Not for cpu though because I always keey temperature lookout for that.
 
unclewebb said:
FYI: Micron D9GMH chips are good for about DDR2-940 speeds at 2.1 volts and CL4-4-4 timings. Not bad for chips rated at DDR2-667. :D

Do you know if my chip is the Micron D9GMH? Corsair XMS2 DDR2-675 from newegg. Here is the link: http://www.newegg.com/Product/Product.asp?Item=N82E16820145167

I paid $199 for it.

Currently Orthos stable 2 hours and climbing...

E6400 @ 400Mhz fsb, 3.2Ghz, Idle temp: 38-39C Orthos load temp: 54-56C
Ram is @ 1:1, 800Mhz 4-4-4-12-6-35 with 2.1v

Am I running too hot?
 
I bet it isn't even a problem with your RAM. Try 410 FSB and you'll see that board switch to a different strap. This very well may be your problem.

Keep your RAM voltage where it is.

I can't overclock my board at 400 at all but I can run 460x7 orthos stable.
 
Well Orthos passed overnight, 8 hours. Gonna let it run a few more. CPU hasn't budged past 56C and I didn't even leave my AC on overnight :)

So it was the RAM. Thanks for everyone's help.
 
I couldn't find out what chips your ram is using but it sounds like it is working good so I guess it's not that important!

Update: It's likely that they are Micron D9GCT which were originally designed for DDR2-533 at 1.8 volts. At 1.9v they are rated at DDR2-667 and at 2.1v you're getting DDR2-800 speeds out of them. At CL5-5-5 I've seen some users getting DDR2-900 out of these same chips. You can never be 100% sure what chips were used on your module without pulling off the heat spreader and having a look. I saw one Crucial Ballistix stick that actually had 7 D9GMH chips and 1 D9GCT on the SAME module so anything is possible.

Next stop, 425 X 8 = 3400 MHz.

At 2.1v your memory should be OK to get to this level. Just make sure that under full Orthos load that you are getting at least the Intel default of 1.325 volts for your CPU. You might need 1.35 volts under full load. The Asus board has some voltage droop so you will likely have to set it higher than this in the bios to end up with the required voltage.

Other users are getting 450 MHz out of similar Corsair DDR2-675 memory.

450 x 8 would really put a smile on your face. :)
 
Should I be worried, if I am stable?

Cuz right now at 3.2Ghz, under orthos load CPU-Z actually drops to 1.264V. BIOS is set to 1.3875V. That's a helluva drop!!

idle CPU-Z is reporting 1.304V.

Here's what I woke up to this morning:



Note: I ran Orthos for 4 hours before this and then played FEAR for an hour then re-ran Orthos and went to bed :)
 
Don't worry about it, if the tempeatures are good and it is stable its good because those readings are never accurate. It's call Vdroop. Theres not much to worry about.
 
his point was that he should bump up the vcore to a point where under load his voltage is still 1.325 (stock voltage), if he wants to go for max overclock.... maybe 3.4Ghz+ ?

B/c if it drops too much underload the CPU could be unstable.

But if he's happy at 8x400 = 3.2Ghz he shouldn't need to do anything. 8 hours Orthos is good. And 3.2Ghz is a smokin fast system.

BTW, you are running the Large TTF Orthos tests? right?

Just making sure, b/c that is the most stressful and will show if your system is truely stable.
 
chrisf6969 said:
his point was that he should bump up the vcore to a point where under load his voltage is still 1.325 (stock voltage), if he wants to go for max overclock.... maybe 3.4Ghz+ ?

B/c if it drops too much underload the CPU could be unstable.

But if he's happy at 8x400 = 3.2Ghz he shouldn't need to do anything. 8 hours Orthos is good. And 3.2Ghz is a smokin fast system.

BTW, you are running the Large TTF Orthos tests? right?

Just making sure, b/c that is the most stressful and will show if your system is truely stable.

I actually have been doing blend since it was my RAM that was in question. I can pass dual superPI 32M no problem.

I will try running it on large TTF or just do dual prime 95's tonight.
 
keldegar said:
I actually have been doing blend since it was my RAM that was in question. I can pass dual superPI 32M no problem.

I will try running it on large TTF or just do dual prime 95's tonight.

Prime is the same thing. (it will also ask you which one of the 3 test to run)
 
Some of the biggest CPU voltage droop seems to be on the Asus boards. If you have an Asus board then don't go by or trust the voltage numbers in the bios. I think CPUz accurately reports CPU voltage on that board but it might be a good idea to ask around just to make sure.

Nothing wrong with 3.2 GHz but if you are Orthos stable at 1.264 volts then I can pretty much guarantee you that there is still more left in that chip. I think 3.4 GHz is likely and 3.6 GHz is quite possible. Your memory might hold you back. If you have to drop the timings to CL5 then your faster processor might not equal a faster computer. Run some benchmarks as you adjust the MHz and play with the settings.

The important number to keep an eye on is temperature as reported by CoreTemp. Try to keep that under about 65C at full load and you should be OK.

If you want to run faster, you will likely have to boost the bios voltage a little higher. Keep it in the 1.325 v -1.350 v range and the temps under 65C and you won't have any problems.

Orthos is a nice looking GUI wrapped around the Prime95 code and is designed for dual core processors so you might as well run Orthos like everyone else is running these days.
 
Thanks for all the advice from everyone in this thread. I think I will stick with 3.2Ghz, 56C load sounds good and i'm actually running lower than stock voltage. :)

All my games play incredibly as it is.
 
unclewebb said:
Some of the biggest CPU voltage droop seems to be on the Asus boards. If you have an Asus board then don't go by or trust the voltage numbers in the bios. I think CPUz accurately reports CPU voltage on that board but it might be a good idea to ask around just to make sure

There's also that voltage bug, when you get to about 1.45-1.5v+ all the sudden CPU-Z starts reporting 1.2x v

Do you know what I'm talking about?!
 
Just sharing some info:

http://www.hardforum.com/showthread.php?t=1101150&page=2&pp=20

Intel Core 2 E6400
Asus P5B (non dlx)
(Geil) PC6400 800MHz 4.4.4.12
Zalman 9500 AT
(Asus 7900GT)

Bios update naar 0509
Bios setting CPU: 1,5250V
Bios setting RAM: 2.00V

Bus clock op 401MHz (400MHz = hang!)
Ram clock automaticly goes along to 802Mhz
PCI-E clock 105MHz

RESULT! &gt;:)
http://img238.imageshack.us/img238/9051/shotxar2.jpg

Windows Xp CPU at 1,213V (CPU-z)
De cpu(s) Temp 43 celsius
=Fully Stable!

Choosing the Right Memory for Core 2 Duo Platform
Conclusion:
First of all we have to acknowledge the high efficiency of data pre-fetch algorithms implemented in Core 2 Duo processors. It is thanks to them that platforms with such processors have data access latency comparable to that of Socket AM2 Athlon 64 systems, which feature an integrated memory controller. However, notwithstanding the impressive achievement of Intel's engineers, the memory subsystem of Core 2 Duo systems with an external memory controller, located in the chipset's North Bridge, cannot match the memory subsystem of Socket AM2 systems in overall efficiency. To be exact, the platforms with new Intel processors cannot provide as high memory bandwidth as the competing platforms do.

The memory bandwidth on Core 2 Duo systems is limited not by the characteristics of DDR2 SDRAM modules but by the bandwidth of the bus that connects the CPU with the chipset's North Bridge. This is why your changing the memory frequency or timings is going to have a small effect on performance of a non-overclocked Core 2 Duo system (but the frequency influences performance somewhat more than the timings do).

More interesting are the results of the overclocked platform. In this case, there is more sense in using fast memory and the optimal memory frequency divisor is 1:1 (FSB:DRAM) as has been shown in our tests. In other words, you can achieve maximum performance by using memory with lowest possible timings in synchronous mode. It means that if you overclock the FSB to 400MHz, DDR2-800 SDRAM with low timings is the optimal choice. If the FSB is overclocked more, DDR2-1000 or DDR2-1067 SDRAM is the best option. An additional argument in favor of using memory and the FSB in synchronous mode at overclocking is that the 1:1 divisor is the most stable one on a majority of mainboards.
http://www.xbitlabs.com/articles/memory/display/core2duo-memory-guide.html


Intel Core 2 Duo - Test
http://www.behardware.com/articles/623-1/dossier-intel-core-2-duo.html
 
unclewebb said:
I couldn't find out what chips your ram is using but it sounds like it is working good so I guess it's not that important!

Update: It's likely that they are Micron D9GCT which were originally designed for DDR2-533 at 1.8 volts. At 1.9v they are rated at DDR2-667 and at 2.1v you're getting DDR2-800 speeds out of them. At CL5-5-5 I've seen some users getting DDR2-900 out of these same chips. You can never be 100% sure what chips were used on your module without pulling off the heat spreader and having a look. I saw one Crucial Ballistix stick that actually had 7 D9GMH chips and 1 D9GCT on the SAME module so anything is possible.

Next stop, 425 X 8 = 3400 MHz.

At 2.1v your memory should be OK to get to this level. Just make sure that under full Orthos load that you are getting at least the Intel default of 1.325 volts for your CPU. You might need 1.35 volts under full load. The Asus board has some voltage droop so you will likely have to set it higher than this in the bios to end up with the required voltage.

Other users are getting 450 MHz out of similar Corsair DDR2-675 memory.

450 x 8 would really put a smile on your face. :)

welp to find out if my RAM was the bottleneck (currently max stable is 417fsb x 8) i tried 425x6 and guess what, I fail Orthos rather quickly.. I tried 5-5-5-15 latency and my RAM didn't like it!

Weird, huh? I can do 834mhz@ Cas4, but can't go any higher regardless of latency..

Or it could be I am reaching the max stable on my mobo without more northbridge juice.. which my mobo doesn't have a setting for :(
I also tried CAS6 and it posted but froze on the BIOS screen (had to reset CMOS).
 
My board is totally unstable above 400mhz too, though i'm trying mine with an e6600.
does your 8800GTS cooler cover the bios reset jumper like my 8800gtx does? I just about cried when i realised i was going to have to yank my 8800gtx every time i need to re-set my bios. It's made me timid about trying my higher o/cs and i've kinda settled on 8x400 even though i've got a 9x multi cpu :( My temps on my zalman suck, though, so it could just be a cpu issue...3.2ghz @ 1.35v and i'm hitting 63-64c under full load...and it's fucking winter time! Figure +5c for summer and i've already hit the heat barrier. This damn zalman barely puts any pressure on the IHS at all compared to my A64 heatsinks, i'm starting to wonder if something's fucked with my mounting bracket. I idle @ 48-50c! Which i find absurd considering my A64 @ 1.59v idled @ 38c. is this zalman 7000CU way shittier than the XP120 i had for my athlon or what :(
 
headless said:
My board is totally unstable above 400mhz too, though i'm trying mine with an e6600.
does your 8800GTS cooler cover the bios reset jumper like my 8800gtx does? I just about cried when i realised i was going to have to yank my 8800gtx every time i need to re-set my bios. It's made me timid about trying my higher o/cs and i've kinda settled on 8x400 even though i've got a 9x multi cpu :( My temps on my zalman suck, though, so it could just be a cpu issue...3.2ghz @ 1.35v and i'm hitting 63-64c under full load...and it's fucking winter time! Figure +5c for summer and i've already hit the heat barrier. This damn zalman barely puts any pressure on the IHS at all compared to my A64 heatsinks, i'm starting to wonder if something's fucked with my mounting bracket. I idle @ 48-50c! Which i find absurd considering my A64 @ 1.59v idled @ 38c. is this zalman 7000CU way shittier than the XP120 i had for my athlon or what :(

To reset my BIOS I just unplugged the system and yanked the CMOS battery for ~2mins.

Yeah your cooler sounds crappy. I'm at 417x8 = 3336mhz and I idle at 36~37C and load 50-51C @ 21C Ambient.
 
to share my experience with the e6400 and P5B-E, the P5B-E has a bad spot between about 360 and 400 FSB, I could not even run at all in that area. Then I got in the 400/401 range and started seeing more stable numbers.

In testing, I could usually get to around 8hrs no problem, but up around 10-12hrs it would fail. For me the solution (instead of trying to run at even higher voltage) was to drop the multi to x7 and run at 2.8ghz. Plenty fast for me and still smokes anything i throw at it.

If you are stable where you are though with 56c load, that is awesome :)
 
Just a note to anyone reading this thread wondering about the RAM I bought..

nintek said:

it's either

PCS 90nm ETT
PROMOS ETT Rev E
or Micron D9GCT

But then RedBeard says it is Micron since I have version 1.2

So I guess it must be Micron D9GCT.
 
Back
Top