E6750 & Gigabyte P35C-DS3R Overclocking Thread

[Conjugal Visit]

Thank you for this, biltong. Very interesting. I'm going to google around to find some comparable tests. BTW, how do you like that AC Pro 7? Are you running it in conjunction with other A C fans using their PWM system? And how's the clearance at the P35C's northbridge heatsink? You think there is enough clearance there to stick on a 40mm fan on the northbridge heatsink?

Regards and thanks for the testing,

 

[mrfatboy]

Interesting results. Thanks for doing all that testing. I wonder why everybody keeps saying 1:1 memory ratio with tighter timings is the best? If you search around for the same type of posts, we are not the only ones that have had this discussion. I wonder if there is something else that we are not considering but I can't imagine what that might be.

 

[biltong]

@Conjugal Visit - I'm pretty happy with the AC Pro 7, although I think when I set Fan control to Auto, its actually using voltage control rather than PWM and the speed varies between about 600rpm at idle to about 2000rpm at load (using Orthos small FTT).

If I switch to PWM, the idle speed seems to be about 1200rpm and load is about 2200rpm, so I've left it on Auto since it gives lower fan speeds and not much difference in temperature at load (about 1C). Load temperatures using Orthos small FTT are about 55 to 56 on both cores, but it is nice and cool here at the moment

I'm just using 2 sharkoon silent eagle 2000 fans with a Zalman fanmate 2 controlling each. The front input fan is set to 1300 rpm and the rear exhaust is set to 1700rpm (for some reason, I think because the front fan is blowing through a hard disk cage etc, increasing the front fan speed too much causes a noticeable whooshing noise. No such problem with the rear fan).

I have maybe a couple of mm between the bottom of the heatsink fins and the northbridge HS and maybe 1mm between the bottom corner of the fan mount and the northbridge HS (that's height wise clearance). The AC Pro 7 overhangs (sideways) the Northbridge HS by about 2mm. Here's a couple of not very good photos I've just taken, probably good enough to see what I mean about the overhang. Not sure how easy it would be to mount a fan over the northbridge HS......

AC Pro 7 Picture 1
AC Pro 7 Picture 2
AC Pro 7 Picture 3


With my AC Pro 7, I must confess that I think there's actually a very slight gap between the copper base of the HS and the CPU heatspreader. I know this because I removed the supplied thermal paste and put a thin layer of AS 5 on the CPU heatspreader (as I normally would) and my temperatures shot into the 80s under load (before I managed to shut it down). When I took the HS off, I noticed that the bottom of the copper base was still mostly spotless. I then applied AS5 to both the CPU and the copper base of the HS and the load temperatures dropped dramatically to about 57 or 58 under load. This was the same as when I was still using the MX-1 heat paste supplied with the HS.

I tried remounting the HS several times with different quantities of AS5, but the best load temperatures I could get were 57 to 58. Finally, I replaced the AS5 with AC MX-2 and that dropped the load temps a couple of degrees to 55 or 56 on both cores. If I run the folding app, the temperatures are generally 50 on one core and 52 on the other.

@mrfatboy - I've seen the same recommendation of using a 1:1 ratio, but no idea why. Looking at the memtest86+ results, I thought it interesting that there wasn't a HUGE difference between running at 1066 MHz 5-5-5-12 and 888 MHz 4-4-4-12. A much bigger difference between running at 888 MHz with 4-4-4-12 timings and at 888MHz with 5-5-5-15 timings. The other thing I don't fully understand is Northbridge "straps". Never really found a very good explanation.....

 

[Conjugal Visit]

@biltong - just excellent, excellent information - thank you for the pictures! That is exactly what I needed to see. I'd like to cool the northbridge at some point in time, either by using the current heatsink ad strapping a small fan to it, or using a Noctua or something. What I really want to do is to quiet an already pretty quiet rig. With your A C fan around 100-1200 rpm, can you hear it - side panel on? I am already planning to change out the fans that came stock with my Lian Li case, particularly the 80mm's at the top and in the back, but I have need of a 120mm to replace the side panel 120mm fan too. Since I'm doing all this, I've been looking at quieter heatsinks - one's that are a little more compact than the Tuniq, Extreme 120, Scythe, etc. Thanks again, biltong!

Regards,

 

[TallBill]

Ok, totally new to overclocking. I just built a rig with a DS3, Q6600, with 4 sticks of GSKILL 800 and a Tuniq Tower 120. With no load my temps are about 40 degrees, but I run a distributed computing program which puts all 4 cores at 100% 24 hours a day. My temps are around 60 degrees with this running.

Anways, I'm at 8x333 and CPU-Z is showing my core voltage at 1.264 although I swear I set it higher in the Bios. I guess I want to slowly work my way faster, but would be happy even at only 3.0 ghz. Thats like adding a 5th core for distributed computing.

 

[biltong]

No problem. Its all useful info in this forum !

I'd say the AC Pro fan would be practically inaudible at 1000-1200rpm. I just did a quick test with it running under load at 1600rpm. I disconnected the 2 case fans and briefly stopped the graphics card fan (so I had to do the test with the side panel off) and it was pretty quiet. That left the AC Pro 7 fan and the PSU fan running (Tagan modular 530) and very quiet indeed. The loudest fan is probably the graphics card fan (hence why I had to stop it) and even that isn't too bad, although I'll probably swap that out at some point for a Zalman VF900, again controlled through a fan mate for tweaking.

I guess if you go for a AC Pro 7, your temps will probably rise a little, but its cheap and easy to attach/remove. You might want to replace the stock thermal paste with AC MX-2 or similar which seems better than the supplied MX-1 or AS 5.

I'm not sure what temps my Northbridge is reaching, but the NB heatsink only feels slightly warm. Is there any tool that measures the Northbridge temperature ? Is it one of the temperatures in Speedfan ?

 

[Gaucho01]

Tonight I pushed the envelope a litle further.

I manage to 12.941 at 3dmark2006, using a 695x1055 and 3,6 and 900 mhz 4.4.4.12.

3DMark Score 12941 3DMarks

SM 2.0 Score 5771 Marks

SM 3.0 Score 5795 Marks

CPU Score 3246 Marks

I think I could go even further but didnt have enough time.
By de way ambient temp was about 18C, and it was real nice.
Lower temps make all the diference.


3dmark06 - 12941

Some screens of my custom case.

baby

baby 2

 

[mrfatboy]

Nice score! You still beat me even when my vid card was overclocked more. You should try to break the 13k mark.

Did you buy your 3dmark2006 pro? I'm just using the freeware version. I wonder if there is a difference. There shouldn't be.

When you are running the benchmark what do you turn off or disable? I noticed before you turned of explorer.exe. Anything else?

I see from your sig that it takes you quite a bit more juice to get to 3.6. Did you test for stablilty for 24/7 3.6?

 

[landonthedinosaur]

mrfatboy-
I'm currently putting together a system with the exact specs as yours. Have you made any changes to your bios since the beginning of this thread?

Is there anything that you would do differently on a similar budget?

 

[mrfatboy]

Originally posted by: landonthedinosaur
mrfatboy-
I'm currently putting together a system with the exact specs as yours. Have you made any changes to your bios since the beginning of this thread?

Is there anything that you would do differently on a similar budget?

I'm very happy with my system. I run it 24/7 at 3.6 and it's rock solid stable. I don't like to change my system constantly. It's a waste of money. I'm more of a long term hold type of guy. This system does it. Especially with DDR3 and Penryn upgradablilty. Also, buying quality componets like the Tuniq and hx520 will give me a legup on my next system in next serveral years. I'm extremely happy with all my components and I would buy them all again.

There seemed to be a bad batch of the P35C boards out there a couple of months ago but that has quieted down alot. I have not heard or seen any post about this in a while.

The board does overvolt it's ram but if you go into knowing this it's not a problem. Just following the instructions in the OP.

You might want to look at the new 8800GT 512 if you are trying to save $100 but I guess the 8800GTS still beats it in bandwidth tests. I'm playing all the new games on high right now. Your call.

I'm running everything in my sig and OP. My bios is still f4g

You should be able to get my OC quick and easy if you follow the first 3 post's in this thread.

Its been a really fun system to work with and experiment with OCing. I need my system to run 24/7 and this system does it beautifully. Also, it's been fun talking to a great bunch of guys in this thread sharing the experience.

Good Luck! There is a wealth of knowledge in this thread so you will do fine.

 

[mrfatboy]

@gaucho

I got a little more with my vid card OC'd to (707/1100). But you are still the winner


3DMark Score 12806 3DMarks

SM 2.0 Score 5689 Marks

SM 3.0 Score 5730 Marks

CPU Score 3236 Marks

 

[biltong]

Originally posted by: TallBill
Ok, totally new to overclocking. I just built a rig with a DS3, Q6600, with 4 sticks of GSKILL 800 and a Tuniq Tower 120. With no load my temps are about 40 degrees, but I run a distributed computing program which puts all 4 cores at 100% 24 hours a day. My temps are around 60 degrees with this running.

Anways, I'm at 8x333 and CPU-Z is showing my core voltage at 1.264 although I swear I set it higher in the Bios. I guess I want to slowly work my way faster, but would be happy even at only 3.0 ghz. Thats like adding a 5th core for distributed computing.

@TallBill - Is there a reason you're running 8x333 ? I though the standard CPU multiplier for the Q6600 was x9 ? CPU-Z reports the actual Vcore rather than what you specify in the BIOS, so it'll always report a lower voltage due to Vdrop and Vdroop. Basically, at idle, the Vcore you see in CPU-Z will be lower than the BIOS setting and at load, lower still.

Also, is your CPU the B3 stepping or the G0 stepping ?

There's several approaches to overclocking, but your temperatures already seem a little on the high side, so you'd need to watch those carefully.

I'll try and give an outline of how I did it when I get a bit of time, but other people may have better methods....

 

[Gaucho01]

I'm using 3dmark2006 pro. I updated to F6 Bios, and 169.04 Nvidia drivers.
I don't think there is any difference, just to make shure I tryed with 3dmark freeware and the only difference is you get your marks on internet.
I turned off antivirus, firewall and explorer.
My system runs 3.6 stable at 1,504 V, and nothing less.
I was worried about that, so Im not using 3.6 24/7.
I don't think I could pass 13k, but I'll try.
I didn't push the memories to the limit, I gues I could go close to 1100.
I'll try push GPU to 700 first, and then push memories.
I takes a lot of time, and I'm looking for another cold day to try it.

 

[mrfatboy]

I didn't realize that you had an aftermarket vid card cooler until I saw a picture of your rig. That thing looks like a monster However, youre XFX card is still clock for clock faster than the EVGA. Have you used ATITool for your vid card overclocking? It's nice. Just press "find max core" and then "find max mem". That should save you alot of time. Once I found my max's I just use the nvidia software to set it nowadays.

I remember reading about adjusting a secondary clock on these cards. I'll see if I can find it today and post the info.

 

[mrfatboy]

I had the big scare this morning--- the dreaded "cycle reboot" problem I was running a 3dmark06 test and came back to my computer ---- black screen, fans running. I did a hardware reset and it would POST then Reset. After a couple attempsts it then asked me to use "chkdsk" because one of the drives was flakey. After that test it would BSOD everytime at windows.

hmmmmm, ram problem????


I swapped out my 2gigs of Crucial with another 2 gigs of Crucial that I had just bought. mmmmmm. $45 never tasted so good. Still BSOD at Windows startup!!!

I then booted from CD from my WinXP pro disk and repaired XP. Seem to work.

I'm now back up and running with my spare 2gig of Ballistix I then put in my orginal 2gigs (4 gigs total) and BSOD Long story short. It looks like my original 2gigs of Crucial whet bad AND corrupted XP

I called Crucial and they RMA'd and crossed shipped free of charge.

I'm back up and running still at 3.6 OC Oh, btw, my NB heatsink was not all the way on Hmmmm, How many time have we heard this ? I thought I was going to be an Abit owner tonight

Knock one wood

 

[Triton67]

Ain't got the exact modles, but E6850 / P35 DS3R (rev 1.0)

Finished 9x400, 3600MHz with 1.36875?v BIOS, 1.33v idle, 1.296v load - 3,5 hours Prime95 v25.3 stable. Will refine some parameters (vdimm 2.0, lower mem settings since I dropped from FSB 425)

Since SpeedFan reports 15C low, I add that to core temps:
typically 44+15C = 59C load in cosy warm room, fans at lowest. Abs. max was 46+15C, 61C at 8K test....but Crysis beta GPU bench likes 8x425 better, dispite lower CPU MHz

EDIT ////

Flashed the F7 BIOS today (vcore, vdimm, 12v showing in PC health, no need to press CTRL+F1 for advanced settings) and continued oc'ing:

E6850 L723A650
9x422 (3,8GHz)
BIOS 1.456v
IDLE 1.41v
LOAD 1.376-1.392v
MCH +0,2v
VDIMM +0,3v
FSB & PCI-E stock volts
4-4-4-11 2T (4x1GB sticks)
66C max load read out during the 30min Blend test, but generally avg ~63C
played games all nite, typical max gaming cpu temp was 48-50C

 

[egrimisu]

if woth orthos the core's temp are 68, is this ok or it;s to hot?


What do you recomend shall i flash another bios?
my memory's are rated to 2.0v is the mono applying wrong v (+0.2= 2.24)??++

For now just 3 questions

 

[honolululu]

Originally posted by: egrimisu
if woth orthos the core's temp are 68, is this ok or it;s to hot?


What do you recomend shall i flash another bios?
my memory's are rated to 2.0v is the mono applying wrong v (+0.2= 2.24)??++

For now just 3 questions

I would say that's borderline safe. Some would say that's just fine, some will say too hot. I'd say if you want to run the chip for more than 3 years, get those temps down. EDIT: I just saw you have the Thermalright 120? Sounds like a bad mount. With those voltages I would think it would run cooler than that. What's the room temperature there?

I don't have the same board, but it's possible F4 might let you see actual voltages in the BIOS if you already can't in the PC HEALTH section.

These boards seem to 'know' if the RAM needs 1.8V or 2.0V and then overvolts even more from there.

I'm doing an RMA on my RAM, so I've got a slower set of pc2-5300 in right now, and at normal voltage it puts me at 1.89V. With the Mushkins in, normal voltage puts me at 2.01V or so.

__________________________
I've got a general question someone might be able to answer. I'm loving this Mushkin memory, but from the get go I had one bad stick in memest. So after getting great forum support from them, but to no avail, I finally sent them both back. I got the new pair, and it was the same thing. One stick would pass memtest alone, the other wouldn't no matter which slot, timings, dividers, stock cpu, all the choices.

I can pass memtest with the pc2-5300 set in dual channel so it seems the RAM is the problem. Could it be possible that when I first put the set it it breaks the second stick. But then if I switch them around shouldn't they both get broken? I have no idea how it might break it besides over volt. Can it overvolt a part of the RAM?

So I open a support ticket and the guy says, " I will advance you a replacement ASAP" I asked if that was like a cross ship, but he never answered about that. Does this mean I can have the four sticks in my hand to choose from? I would prefer to run the two new ones together, but if one is good and the other bad again...

I guess I should test them one at a time when they get here. Then put them both in and test.

 

[wpcoe]

OK, I actually read through this entire long thread before posting. Whew! I'm somewhat new at overclocking, so some of the more esoteric things like Vdrop and Vdroop are beyond me, but I hope I'm clear on the more mundane basics.

I'm currently overclocking to 3.0Ghz (375 x 8), which is the highest I can get and still pass MemTest and Ortho regardless of what voltages I use. I'm assuming that is a limitation of using PC2-4300 (266 Mhz) RAM. In fact, I'm pretty impressed that this Kingstom ValueRam works stably at 375.

Of course, any feedback on the previous paragraph is welcome, but I actually have questions about two other issues:

CoreTemp
From comments in the OP, and all through this thread, it seems that people using CoreTemp 0.95.4 beta don't add 15°C to the reported temp. I downloaded and installed from "CoreTemp Beta 0.95.4.zip" and it reports the Tjunction as 85°C and the Core #0/Core #1 temps read identically to those reported by SpeedFan 4.33, which everybody says to add 15°C to.

The CPU temp reported in BIOS (~49°C) is indeed about 15°C higher than what CoreTemp shows at idle (23°-27°C). When running the Orthos blend test CoreTemp reports max temps in the 44°-46°C range. Unless I add 15°C to that, that is pretty darn cool for moderate overclocking with stock Intel cooler, no?

So, why do I get the impression that everybody else's CoreTemp 0.95.4 shows Tjunction as 100°C and no need to add on a factor of 15°C?

BIOS 'upgrades'
My mobo came with the F2 BIOS on it, and it's the only one that lets me boot up! I have downloaded the F4 and F5 BIOSes from the Gigabyte site, but even in "Fail Safe" and "Optimal" configurations with no tweaking, I cannot boot off of my hard drives.

With the F4 and F5 BIOS, and only testing one RAM stick at a time, I cannot complete a MemTest series of tests.

The few times that I actually got the system to access my hard drives and try to boot Vista, it would spontaeously reboot and on the reboot run CHKDSK on one or another of the partitions. Usually it would find no errors, but other times would report errors. Even SpinRite would not launch from a floppy. I thought I had bad hard drives. Booting from the Vista DVD and trying "repair" did no good. I even re-installed Vista twice thinking my hard drives had been affected somehow (this was before I ran MemTest and SpinRite). I dropped back to F2 BIOS, and all was dandy, except I now had to reinstall Vista again...

Just for kicks, yesterday, I tried flashing both the F4 and F5 BIOSes, and still cannot boot.

FYI, my original Vista install had been with my SATA drives set to AHCI and with the current install I forgot and had them set as IDE before I installed Vista, so that appears not to be a relevant factor.

Any idea about what on earth causes my system to choke on the newer BIOSes?

 

[egrimisu]

It may be a problem if i put to much thermal paste? i put like 1.5cm on the CPU and 1.5 on the heatsink, i now raised the voltage to 1.38125v in bios and in idle CPU-z reads 1.328v and with orthos SMALL FFTs test reads 1.296v. I put the (G)MCH to default + .0
After 3 minutes of Orthos the temps where 68 and at test 3 faild to run.
What shall i do? shall i raise the cpu voltage or shall i put (G)MCH to +.1v or i shall re-paste the cpu and the heatsink with less thermal paste.IF yes what will be the amount of thermal paste that i shall use.
The rpm of my cpu fan is 1400rpm that pushes ~44.8CFM of air and the noise is 17db.(Shall i replace the fan with an thermaltake 1400rpm 16db 78CFM?)
The ambient temp of my room is ~ 24c.


Thanks in advance.

 

[biltong]

@egrimisu - I'd say that that's too much paste. Assuming that the base of the heatsink and the top of the CPU heatspreader make good contact, then you should only need to apply the paste to the CPU heatspreader.

The eaisest way I've found to do it is to apply an amount approximately the size of a grain of rice. You can then spread this over the entire surface using either a credit card, or, the method I've always found easiest is to use some plastic cling film (the stuff you use for covering food) stretched over your index finger and then spread with your finger. You should end up with a layer of paste that is only just thick enough to cover the surface, less than paper thin.

Make sure you remove all the old paste before applying the new paste. Preferably use a proprietary TIM cleaner like Akasa TIM cleaner.

You need to cover the entire surface of the CPU heat spreader.

If temperatures don't improve (or get worse), then its likely you're not getting perfect contact between the CPU heatspreader and the base of the heatsink. In that case, you may need to apply a grain of rice size amount of paste to both the CPU heatspreader and the heatsink base (and spread as usual).

I did several tests using Artic Silver 5 and Artic Cooling MX-2 and found that the AC MX-2 lowered the temperatures a couple of degrees under load, so that might help too.

Of course you could also lap the the heatsink base and the CPU heatspreader, but you'd have to be a braver man than me

It may be worth trying a different fan, but my guess is you'll see very little difference cooling wise between 2 fans of the same size running at the same speed, although the noise levels may vary.

I suspect you may need to increase the Vcore to get stable, but I think your first priority should be to get those temperatures down.

 

[mrfatboy]

Originally posted by: wpcoe
OK, I actually read through this entire long thread before posting. Whew! I'm somewhat new at overclocking, so some of the more esoteric things like Vdrop and Vdroop are beyond me, but I hope I'm clear on the more mundane basics.

I'm currently overclocking to 3.0Ghz (375 x 8), which is the highest I can get and still pass MemTest and Ortho regardless of what voltages I use. I'm assuming that is a limitation of using PC2-4300 (266 Mhz) RAM. In fact, I'm pretty impressed that this Kingstom ValueRam works stably at 375.

Of course, any feedback on the previous paragraph is welcome, but I actually have questions about two other issues:

CoreTemp
From comments in the OP, and all through this thread, it seems that people using CoreTemp 0.95.4 beta don't add 15°C to the reported temp. I downloaded and installed from "CoreTemp Beta 0.95.4.zip" and it reports the Tjunction as 85°C and the Core #0/Core #1 temps read identically to those reported by SpeedFan 4.33, which everybody says to add 15°C to.

The CPU temp reported in BIOS (~49°C) is indeed about 15°C higher than what CoreTemp shows at idle (23°-27°C). When running the Orthos blend test CoreTemp reports max temps in the 44°-46°C range. Unless I add 15°C to that, that is pretty darn cool for moderate overclocking with stock Intel cooler, no?

So, why do I get the impression that everybody else's CoreTemp 0.95.4 shows Tjunction as 100°C and no need to add on a factor of 15°C?

BIOS 'upgrades'
My mobo came with the F2 BIOS on it, and it's the only one that lets me boot up! I have downloaded the F4 and F5 BIOSes from the Gigabyte site, but even in "Fail Safe" and "Optimal" configurations with no tweaking, I cannot boot off of my hard drives.

With the F4 and F5 BIOS, and only testing one RAM stick at a time, I cannot complete a MemTest series of tests.

The few times that I actually got the system to access my hard drives and try to boot Vista, it would spontaeously reboot and on the reboot run CHKDSK on one or another of the partitions. Usually it would find no errors, but other times would report errors. Even SpinRite would not launch from a floppy. I thought I had bad hard drives. Booting from the Vista DVD and trying "repair" did no good. I even re-installed Vista twice thinking my hard drives had been affected somehow (this was before I ran MemTest and SpinRite). I dropped back to F2 BIOS, and all was dandy, except I now had to reinstall Vista again...

Just for kicks, yesterday, I tried flashing both the F4 and F5 BIOSes, and still cannot boot.

FYI, my original Vista install had been with my SATA drives set to AHCI and with the current install I forgot and had them set as IDE before I installed Vista, so that appears not to be a relevant factor.

Any idea about what on earth causes my system to choke on the newer BIOSes?

Coretemp 0.95.4 should be reporting 100c for you. Mine does for the E6750. So no need to add the 15C.

Regarding your Bios/Boot problem. I bet it's your ram. I had the same "chkdsk" problem a couple of days ago and it turned out to be bad ram. Buy 2gigs of Crucial Ballistix for $50 and I bet your problems will go away plus you will be able to overclock alot better Well worth the money.

 

[Conjugal Visit]

@mtfatboy - I happened upon one of your recent posts - forget the forum board - on your vid card, overclocking, etc. You get it worked out? I was struggling with sort of the same thing. ATITool used to control my overclock AND my fan, but not anymore. So I went to rivatuner, but was pretty frustrated configuring it, and trying to get it to control my overclock AND my fan (at specific temp thresholds.) I banged and banged on it.The most frustrating thing for me was 3D detection - getting it to implement an overclock ONLY when needed - ie. 3D apps and gaming. Truth be told, I never got it done with rivatuner. So I use them both now - rivatuner is configured to adjust my fan at specific temp thresholds, and I use ATITool to overclock and then trigger that overclock ONLY when gaming, benching, etc. Unfortunately, I have to load both of them to my tray at startup, but it's seamless now, and that's what I have 2gigs for. If you need help, you let me know.

Regards,

 

[wpcoe]

Originally posted by: mrfatboy
Coretemp 0.95.4 should be reporting 100c for you. Mine does for the E6750. So no need to add the 15C.

Regarding your Bios/Boot problem. I bet it's your ram. I had the same "chkdsk" problem a couple of days ago and it turned out to be bad ram. Buy 2gigs of Crucial Ballistix for $50 and I bet your problems will go away plus you will be able to overclock alot better Well worth the money.

Yeah, I know I should invest in faster RAM, but isn't it odd that my RAM is compatible and stable on the F2 BIOS (whether overclocked or not), but not at all on F4 or F5 (even if not overclocked)?

[NOTICE: the following was written before I was corrected by Honolululu below. I was using the wrong version of CoreTemp! I thought about re-editing the following text, but decided it would might cause more confusion...]

Here's a screen shot of CoreTemp and CPU-Z showing my E-6750 with the Tjunction as 85°C. ( www.wpcoe.com/e6750.gif ) Does yours show "Core Temp 0.95" both in the title bar and from Help > About? I expected it to say "Core Temp 0.95.4."

Based on those two oddities I'm paraniod thinking something is wonky with either my processor or the mobo.

I just now noticed in that screenshot that CPU-Z reports the Core Voltage as 1.296V, even though I have it set manually to 1.35v in BIOS ( all other voltages set to "Normal".) I had opened a large file in Photoshop and applied the Lens Correction filter to get the processor to show 3.0Ghz rather than the 2.250Ghz it shows at idle since I have EIST enabled now.

When the CPU is at idle, like right now when I only have a broswer open to write this message, the Core Voltage varies between 1.296V and 1.312V. I thought that the voltage would not vary with EIST when the processor is overclocked, only with a change-of-state to Standby with C1E enabled?

-- Bill @ getting more confused...

PS: I live in Thailand, and it's not nearly as economical to get faster RAM over here, so I'm satisfied with 3.0Ghz compromise for the moment, just puzzled why the temps show up "wrong," and why only the new BIOSes seem incompatible with PC2-4300 RAM.

 

[honolululu]

You should see Core Temp 0.95.4 in the title bar. Download here Make sure you grab the Beta version.


I just now noticed in that screenshot that CPU-Z reports the Core Voltage as 1.296V, even though I have it set manually to 1.35v in BIOS

The motherboard is somehow 'losing' (to heat probably) the voltage and it is dropping from the BIOS setting of 1.35V. This is Vdrop. When you run an intensive program the voltage to the cpu will be even more scarce, which to my understanding is Vdroop. Droop because it looks like a sag on the Vcore graph, is how I remember it, and it sort of just drops during booting never to bounce back. Of course, EIST and C1E complicate the Vdroop graph.

When the CPU is at idle, like right now when I only have a broswer open to write this message, the Core Voltage varies between 1.296V and 1.312V. I thought that the voltage would not vary with EIST when the processor is overclocked, only with a change-of-state to Standby with C1E enabled?

As I said earlier I found I could go up to 3.2 GHz OC on normal voltage and experience just what you are seeing. I was getting it to work with manual voltages with everything set to normal in BIOS. The volts were jumping from 1.35 to 1.16 or so as it went from 8x to 6x.
I think Conjugal's nice VID is so low he drops to under 1.00V on 6x multi. But I personally didn't see any effect in C1E and EIST going from F2 to F4 or F5.

Don't know about the RAM problems. Hope I helped a bit.