My first OC, help me pop my cherry

B

BassKozz

Limp Gawd
Joined
Nov 10, 2005
Messages
311
Ok guys,
This is my very first OC (on a computer at least), I've always been an advocate of leaving things at stock and keep'em cool to get the most lifetime out of them. But with the recent reviews of the Q6600 and the E8400's overclockability (word/spelling?), it would be stupid not to take advantage of this added horse power :cool:...

So tomorrow my E8400 comes in for my "Everyday Rig" (see sig), and I want to see what this sucker can do... I plan on keeping a log of my endeavor here in this thread for all to see :D

So to start off, I'll build the rig and run it at stock and run Prime95 and Orthos to get baseline temps, and from there I'll start bumping it up until it hits a level I am comfortable with (preferable keep the temps under 55c, is that a good max temp to shoot for?).

A couple of questions before I begin my quest:
  1. Is there a need for a "Burn-In" period, before I start OC'ing?
  2. (repeat above)What is a good max temp to shoot for, for a system that will be on 365/24/7 ?
  3. Any other words of wisdom?

Thanks in advance for all feedback/comments,
Wish me luck,
-BassKozz
 
For the temps, I don't like going above 60c in Prime95 with all 4 cores being loaded.

No "burn-in" is really necessary, but you might want to run some memory tests (like Memtest86+) to make sure that your memory is good before you try and overclock only to not be able to figure out why it isn't stable at all.

I usually like to pre-test the whole system before overclocking.. like installing the OS, trying out a few games.. doing some surfing, etc. Just to make sure it is fully stable at stock speeds before overclocking.

Also, check you idle/load temps before starting any overclocking. If the stock temps seem high, you will need to address that before anything else.
 
Well, my E8400 arrived at 1pm, and the build was complete at 6pm... yeah, go ahead laugh now and get it out of your system. I like to take my time and make sure I get everything just right. I meticulously ran the cabling for maximum airflow.
Unboxing the E8400:

Abit IP35-E:

Wounder why the CPU Cover says FOXCONN:confused:?

This wasn't the easiest build, especially when it came to mounting the "Arctic Cooling Freezer 7 Pro" (ACF7P)... and one of the mounting pins still isn't all the way in (i think, at least I didn't hear it click like the other 3 did). I can't tell if this is fully locked in:

here is a picture of one of the other pins that I know did mount fully (heard the click):

and one of the chipset heatsinks was too close, so I had to preform surgery on the ACF7P.

And I cut my self twice on the NSK4480's rough edges around the hd bays. :(

But it's all done...

now for some testing...
cyclone3d said:
For the temps, I don't like going above 60c in Prime95 with all 4 cores being loaded.

No "burn-in" is really necessary, but you might want to run some memory tests (like Memtest86+) to make sure that your memory is good before you try and overclock only to not be able to figure out why it isn't stable at all.

I usually like to pre-test the whole system before overclocking.. like installing the OS, trying out a few games.. doing some surfing, etc. Just to make sure it is fully stable at stock speeds before overclocking.

Also, check you idle/load temps before starting any overclocking. If the stock temps seem high, you will need to address that before anything else.
Click to expand...
I just read this post but I've already ran 5 passes of memtest86+ v1.7 w/out any errors... I thought that was sufficient enough for now... I'll re-run for longer after I OC.

I am installing WinXP now, will report back temps and results of Prime95 & Orthos testing shortly.
Question: How long should I run Prime95, & Orthos before I can take readings of temps, and ensure the rig is stable?
NKDietrich said:
Hold on my sister is out of town let me give her a call.
Click to expand...
Hehehe LOL :D
Tell her to hurry, I am bored watchin XP install.

... Will report back shortly on how she preformed (both the computer and NKDietrich's sister)...
 
Stock Settings (cpu-z):

(Click to see larger)

Ok, I am finally done testing at stock speeds and the results are in:
Idle
CoreTemp
Core #0: 38°c
Core #1: 35-37°c
Abit's EQ Utility
CPU: 22°c

(Click to see larger)
-=-
LOAD
CoreTemp
Core #0: 44-46°c
Core #1: 51-52°c
Abit's EQ Utility
CPU: 37°c

(Click to see larger)
-=-
For some (very) odd reason SpeedFan read 40°C consistently both idling and under load :confused:

So... a range of 44-52°c in CoreTemp seems alittle high for stock settings?
Or am I wrong here?
 
I am mid 40s idle and high 50s under load, which I am not happy with, so my chip appears to run hotter than others, my brother is right around the same, we both got our chips from Microcenter, same batch. Disappointing due to temps, happy with 4 ghz.
 
Ok now to start seeing what this puppy can do;
I went into the BIOS's "SoftMenu" and here are what the default stock settings are:
CPU Operating Speed: 3000(333)
External Clock: 340
Multiplier Factor: 9x
Estimated CPU Clock: 3060Mhz
DRam Speed (CPU:dRAM): Default (DDR2-816)
PCI Express Frequency: 100Mhz
Voltages
CPU Core: 1.2125V
DDR2: Auto
CPU VTT: 1.2V
ICH 1.5V: 1.5V
MCH 1.25V: 1.25V
CPU GTLREF: 63%

So for my very 1st OC (OC#1) I did the following:
Bumped the External Clock to 400Mhz (go for the gold :D)
- Which caused the Estimated Clock to change to 3.6Ghz
- Also caused the Dram Speed: Default (DDR2-960)

It wouldn't post :(
My guess is I would need to tweak Voltage settings to get this config to work, but changing voltages scares me :eek:
So I reset CMOS (jumper), and tried again, but this time I also changed the Multiplier from 9x to 8x, which adjusted the Estimated clock to 3.2Ghz ...
I am now running Memtest for a few to make sure the RAM can handle it...
One really odd thing is that Memtest is reporting the CPU as "Pentium III 3.66Ghz" :confused: What's that about? I thought the clock would read 3.2Ghz, why is it reading 3.66 ?

Will report back (with temps) after I run memtest for a few passes and throw Orthos and Prime95 at it to make sure it's stable...
 
Ok, After 1hour and 3 successful passes of memtest, I decided it was time to boot into XP, and see if it is Orthos + Prime95 stable...
But I didn't get that far... as soon as I opened up abitEQ, the case speaker started blaring like a police siren, and the voltage tab on abitEQ was red:

Item | Value | Limit | Status
DDR2 Voltage | 2.18V | 1.67V ~ 2.04V | Warning
Should I be concerned about this? I looked at the spec's for my memory on newegg, and it reads 2.2V, so 2.18V should'nt be bad, right?
  • So now what... do I need to manually set the DDR2 Voltage? If so, to what?
  • Also notice CoreTemp is reading Clock Speed as 3.6Ghz (same as what MemTest was reading), this doesn't make sense, because I set the Multiplier to 8x, which should be 3.2Ghz :confused: ???
 
BassKozz said:
Also notice CoreTemp is reading Clock Speed as 3.6Ghz (same as what MemTest was reading), this doesn't make sense, because I set the Multiplier to 8x, which should be 3.2Ghz :confused: ???
Click to expand...

Some programs always calculate the standard multiplier, trust your bios readings. I Don't see anything wrong with that mem voltage, If i were you I'd get rid of the stupid Abit utilities, they do more harm than good. If that doesn't shut it up look for some setting in the bios, something like "Memory Voltage Warning" ect. And I'd set the memory voltage manually.
 
AliP said:
Some programs always calculate the standard multiplier, trust your bios readings. I Don't see anything wrong with that mem voltage, If i were you I'd get rid of the stupid Abit utilities, they do more harm than good. If that doesn't shut it up look for some setting in the BIOS, something like "Memory Voltage Warning" ect. And I'd set the memory voltage manually.
Click to expand...
Thanks for the input AliP,
I changed the DDR2 Max limit in Abit EQ to 2.2V and the beeping stopped...
Idle and Load results are in OC#1 - 3.2Ghz (400Mhz External Clock 8x Multiplier):
Idle
CoreTemp
Core #0: 38°c
Core #1: 35-38°c
Abit's EQ Utility
CPU: 22°c

(Click to see larger)
-=-
LOAD
CoreTemp
Core #0: 46-48°c
Core #1: 52-53°c
Abit's EQ Utility
CPU: 38-39°c

(Click to see larger)

Here are the temp Differences:
Item | Stock | OC#1 | Difference
Idle CoreTemp Core#0 | 38°c | 38°c | 0
Idle CoreTemp Core#1 | 35-37°c | 35-38°c | +1
Idle AbitEQ CPU | 22°c | 22°c | 0
LOAD CoreTemp Core#0 | 44-46°c | 46-48°c | +2
LOAD CoreTemp Core#1 | 51-52°c | 52-53°c | +1
LOAD AbitEQ CPU | 37°c | 38-39°c | +1-2

Also note the following voltages have changed (BIOS still set at Default/Auto Voltages):
Item | Stock | OC#1 | Difference
CPU | 1.04 | 1.06 | +.02
DDR2 | 1.94 | 2.18 | +.24
(Readings based upon Abit EQ)

-=-

Time to go back into the BIOS and Start tweaking again to get to 3.66Ghz (my goal)...
Adjusting Voltages Scares me :eek:... I don't want to fry my board or CPU...
  • How high should I put my voltage to get to 3.66Ghz?
  • If I manually set the CPU voltage do I also have to change the CPU VTT Voltage?
  • Will I need to adjust the DDR2, NB, SB Voltages or CPU GTLREF?
  • Any advice on what other settings I should also mess with to get to 3.66Ghz (safely)?
  • And lastly, Should I go 400x9 or 450x8?

EDIT: Currently running memtest at 400x9, CPU Volt Manually set @ 1.2525V, and DDR2 Volt manually set @ 2.2V...
EDIT 2: Memtest ran for 40minutes / 2 passes / 0 errrors, so I tried to boot to XP, no-go :(... I bumped the CPU Voltage to 1.2725V, it booted into XP :), so I am now running Orthos & Prime95 tests...
Will report back temps once done testing...
 
As someone who is to be experimenting for his first time OCing, I must say that this little story is helpful thus far, though I can't contribute much of my own..!
 
Klaerth said:
As someone who is to be experimenting for his first time OCing, I must say that this little story is helpful thus far, though I can't contribute much of my own..!
Click to expand...
Glad I could help...
Stay tuned, as I completely fry my rig :eek::p
 
upping voltages shouldn't be a problem for you, but i would stay under 1.4-1.5 volts to be safe.....

if you want you can up your voltage to 1.1v or 1.2v, or 1.3, or if you need to 1.4 (probably safe max) if you want.... I have 0 experience on a e8 series, but I do have a q6600 running at 3.4ghz..
 
Ok I am done testing OC#2 - 3.6Ghz (400x9 External Clock CPU Volt @ 1.2725V DDR2 Volt @ 2.2V)...
Here are the results:
Idle
CoreTemp
Core #0: 38°c (geez, Core #0, doesn't want to budge from 38°c at Idle)
Core #1: 37-42°c
Abit's EQ Utility
CPU: 22°c

(Click to see larger)
-=-
LOAD
CoreTemp
Core #0: 54-55°c
Core #1: 58-59°c :eek: (bummer, this is above my goal threshold of 55°c :()
Abit's EQ Utility
CPU: 45°c

(Click to see larger)
-=-
Here are the temp Differences:
Item | Stock | OC#1 | OC#2 | Difference OC1 vs. OC2 | Difference OC2 vs. Stock
Idle CoreTemp Core#0 | 38°c | 38°c | 38°c | 0 | 0 (Core #0 Held Strong @ 38°c @ Idle)
Idle CoreTemp Core#1 | 35-37°c | 35-38°c | 37-42°c | +2-4 | +2-5
Idle AbitEQ CPU | 22°c | 22°c | 22°c | 0 | 0
LOAD CoreTemp Core#0 | 44-46°c | 46-48°c | 54-55°c | +8-7 | +10-9
LOAD CoreTemp Core#1 | 51-52°c | 52-53°c | 58-59°c | +6-6 | +7-7
LOAD AbitEQ CPU | 37°c | 38-39°c | 45°c | +7-6 | +8
-=-
Voltages
Item | Stock | OC#1 | OC#2 | Difference OC1 vs OC2 | Difference OC2 vs Stock
CPU | 1.04 | 1.06 | 1.14 | 0.08 | 0.1
DDR2 | 1.94 | 2.18 | 2.18 | 0 | 0.24
(Readings based upon Abit EQ)
-=-
So now I've hit my goal of 3.6Ghz, but i've broken my rule of not exceeding 55°c (hit 59°c under load):
  • Should I be concerned? Should I back it down a bit?
  • Is there anything I can do to keep 3.6Ghz but also keep it under 55°c (besides changing HSF)?
  • Also have a question in regards to my memory timings, but I've placed it under the [H] memory subforum, if anyone is interested taking a stab at it.
 
Well, I ran 5hrs of memtest w/out any problems;

But when I ran Orthos "Blend - stress CPU and RAM" overnight, I woke up to this;
:(
Now what... Do I need to tweak my memory settings;
DRAM Frequency - 480mhz
FSB:dRAM - 5:6
CL - 5
tRCD - 5
tRP - 5
tRAS - 16
CR - 2T


EDIT: Changed to 1:1 and it seems to be working fine now... I will re-visit this once I get the CPU temps under control to my liking.
 
In an attempt to get my core temps down below 55°c, I decided to crack open my case and try and re-seat my Arctic Cooling Freezer 7 Pro (ACF7P)...

First thing I had to do was remove the ACF7P, which was no easy task thanks to the cheap pins the ACF7P uses to mount itself, I also had to remove the 120mm casefan on the back so that I could fit my fat hand in there to pop the pins out:

This next picture shows how one of the pins is actually broken (luckily I was still able to re-mount):

Here is what a working pin looks like:

Some 91% Rubbing Alcohol to remove the OEM Thermal Paste:

Now to apply some TX-2:

I put my finger inside a plastic bag to apply the TX-2 to the base of the ACF7P:

Used a razor blade to apply a even thin layer on the HS of the E8400 (did i use too much?):

And the E8400 is back in it's home:

I also replaced the 120mm fan that came with the NSK4480 with a Scythe S-FLEX SFF21F, and I placed the Antec fan back in the Scythe case for safe keeping. I figure the more powerful Scythe should help pull heat from the ACF7P, and keep it cooler, thus lowering my CPU temps :fingers-crossed:



And after all that hard work...
The load temps are exactly the SAME;
LOAD
CoreTemp
Core #0: 54-55°c
Core #1: 58-59°c :eek: :(

Now what?
  • Should I just chalk it up to the fact that the ACF7P isn't a ThermalRight Ultra-120 Extreme?
  • I have a COOLER MASTER GeminII laying around, should I try that instead?
  • Could it be because I used too much TX-2? Does the TX-2 need a burn in time (300hrs) to work fully?
 
WAY too much thermal paste.

Clean both the CPU and heatsink up.

ONE small dot in the center of the CPU (about the size of a grain of rice). None on the heatsink. Thermal paste is supposed to be used to fill in microscopic crevices in the surface of the CPU and heatsink to improve heat transfer, not as another "layer" of material to transfer heat through.
 
lol good thing you posted pics, now I know how much artic silver not to apply :p. I probably would have done the same thing and emptied the whole tube on the cpu.
 
killingfield1975 said:
i wonder if the tx-2 is that much better than the artic silver. let me know what temps y ou're getting.
Click to expand...
I got it because this review: http://www.madshrimps.be/?action=getarticle&number=1&artpage=2953&articID=635
of which I probably should've read better:
Essentially two schools of thought exist. Spreading - in which an even layer is distributed over the entire IHS (Integrated Heat Spreader) surface using a flat edged card or other object. And Dollop - in which a large drop or dollop of paste is deposited onto the center of the IHS and then the heatsink is mounted. Where-as spreading can incorporate tiny air-bubbles, compressing the paste should force more air out. Here we'll look at some examples before and after.

Tuniq TX-2 recommends using the dollop method applying their paste. Its consistency is more viscous compared to most thermal pastes. In the photo above I spread TX-2 covering the entire IHS. The photo below was taken after 7-days of testing cycling temps between IDLE and LOAD. In the thumbnails below an example of Tuniq TX-2 applied using the dollop method on the Danger Den TDX.

Tuniq TX-2 retained more moisture then any other sample tested, albeit under the recommended dollop method or spread as seen above. I was most impressed with TX-2 and one reason I believe this paste is such a strong performer is in its ability to retain moisture longer then other pastes.

AC’s MX-2 is a synthetic ceramic paste which has similarities to TX-2 insofar as its high moisture content. The flow rate makes it easy to apply whether you're spreading the paste or using the dollop method. MX-2 also recommends the dollop applications method.

After seven days MX-2 did lose some moisture. I noticed several of the pastes tested had great consistency out of the syringe, but after 7-days viscosity and texture can be completely different. Note the distribution of the paste which may seem inadequate until you realize the cores beneath the IHS only account for about 70% of its total area.

I find this method of application very revealing since it matters not if you were to spread the paste over the entire area of the IHS. Spreading the paste is not going to improve the contact area between the two surfaces. Only lapping your heatsink will do this. Problem is, just about any changes you make even using an OEM thermal interface material voids your processor warranty.
Click to expand...
-=-
2uantuM said:
WAY too much thermal paste.

Clean both the CPU and heatsink up.

ONE small dot in the center of the CPU (about the size of a grain of rice). None on the heatsink. Thermal paste is supposed to be used to fill in microscopic crevices in the surface of the CPU and heatsink to improve heat transfer, not as another "layer" of material to transfer heat through.
Click to expand...
How much of a difference could it make? Do I really have to dismount the HSF Again (major pain in my ass) :confused:

-=-

I ran Prime 95 "In-place LARGE FFTs" test over night, and my temps actually came down (according to core temp)
LOAD
CoreTemp
Core #0: 52-53°c
Core #1: 57-58°c

But then again, it could have something to do with the ambient temps in my apartment in the NE dropping from 66°f to 62°f? Or could I expect the CPU temps to drop more as the TX-2 gets broken in?
 
BassKozz said:
I got it because this review: http://www.madshrimps.be/?action=getarticle&number=1&artpage=2953&articID=635
of which I probably should've read better:

-=-

How much of a difference could it make? Do I really have to dismount the HSF Again (major pain in my ass) :confused:

-=-

I ran Prime 95 "In-place LARGE FFTs" test over night, and my temps actually came down (according to core temp)
LOAD
CoreTemp
Core #0: 52-53°c
Core #1: 57-58°c

But then again, it could have something to do with the ambient temps in my apartment in the NE dropping from 66°f to 62°f? Or could I expect the CPU temps to drop more as the TX-2 gets broken in?
Click to expand...

It could make a large difference. Using too much thermal paste causes the paste as an insulator rather than helping thermal transfer.

Edit: Just looked at that review. They pretty much fail at applying paste.
 
But wouldn't the weight and the compression of applying the HSF cause the TX-2 to thin out so that it would be a smooth, flat, thin layer?
 
But this review of TX-2 @ EXTREME Overclocking says:
The first major difference to note is that the TX-2 is infinitely easier than the AS5 to spread across the CPU. It has always taken forever for me to get a nice even coat of AS5 on anything because of how thick it is. With the TX-2, I put a glob in the middle of the heat spreader, and using a razor blade spread it into a nice even coat as easy as if I was using cream cheese.
Click to expand...
 
if you wanna stick with your high temperatures, then go for it. Its common sense.
 
Ok, I went ahead and lapped my HS (Arctic Cooling Frezzer 7 Pro), it was a major pain in the ass getting the bracket off the bottom, I had to bend it a bit, but managed to do all that and get it back in place:

(Click to see larger)
I also purchased some AS5 (hopeing it will work better then the TX-2 I've been using), i placed the AS5 on my CPU as directed by the AS5's website directions:

...
And again, after all that...

Same IDLE and LOAD temps :mad:

This stinks... I give up... I guess I am going to stick with my 59°c temps @ only 3.6Ghz (20% increase over stock) :(

A couple of final questions:
  • How bad is it if I run this comp 24/7 @ these temps (59°c)?
  • Is there anything else I can do to lower my temps (aside from swapping HS or Lapping the actual CPU)?
  • Would rising the voltage help lower the temps? I know this is counter intuitive, but I can't seem to understand why my temps are so high at such a low voltage (1.2725v)
  • Any final words of advice, before I screw on the back case screws and throw this sucker in a corner?
 
Like in most of the E8400 threads, I think this is a problem with the temp reporting not working properly. This is just me, gathering info from google, but that seems to be the gereral common ground with people that have this processor. Take that for what it's worth, and by all means Google it! :)
 
baleedit said:
Like in most of the E8400 threads, I think this is a problem with the temp reporting not working properly. This is just me, gathering info from google, but that seems to be the gereral common ground with people that have this processor. Take that for what it's worth, and by all means Google it! :)
Click to expand...

Is this a CPU problem or a Motherboard problem?
Will I beable to flash the bios of my Abit IP35-E (currently using latest BIOS v15) in the future and get correct temp readings or is it an internal CPU/Intel issue?
 
baleedit said:
Like in most of the E8400 threads, I think this is a problem with the temp reporting not working properly. This is just me, gathering info from google, but that seems to be the gereral common ground with people that have this processor. Take that for what it's worth, and by all means Google it! :)
Click to expand...

How can I find out if this is a temp reporting issue or if my temps really are this high?
I really don't get these high temps because I am able to reach 3.6ghz on my server rig (Q6600 G0), and that's a 66% increase over stock, and my temps don't go above 56°c. But when I go up a measly 20% over stock on the newer E8400 proc my temps hit 59°c :confused:
Is the difference between a Thermalright Ultra-120 Extreme and an Arctic Cooling Freezer 7 Pro that drastic?

and bump these questions:
BassKozz said:
A couple of final questions:
  • How bad is it if I run this comp 24/7 @ these temps (59°c)?
  • Is there anything else I can do to lower my temps (aside from swapping HS or Lapping the actual CPU)?
  • Would rising the voltage help lower the temps? I know this is counter intuitive, but I can't seem to understand why my temps are so high at such a low voltage (1.2725v)
  • Any final words of advice, before I screw on the back case screws and throw this sucker in a corner?
Click to expand...
 
You must log in or register to reply here.
Back
Top