Is it normal that I can't get my E4300 to hit 2.9Ghz?

Boltaction

Gawd
Joined
Jun 10, 2004
Messages
939
Gigavbyte 965P-S3 rev 1, firmware F8.
C2D E4300 Batch: Q644A519
X800 GTO Fireblade
Enlight 450W power supply
Vista Business
2GB DDR2 667 CAS5

My settings are:

Vcore 1.375
VDIMM 2.1
MCH +0.1
Ram @ 1/1
Currently set @ 266FSB, CPU is running at 2.4Ghz.

Every time I try to go higher I can't boot to windows. What's wrong?
 
Super Talent, the special on hotdeals. But the ram shouldn't be the issue, since it is overvolted to 2.1 from 1.8 stock... I think I found something though.

I had fan control on in the BIOS. Turning it off my fan speed went from 800 to 1700. I am now @ 2.7 and will try for 2.9ghz.


edit: Much better temps. 67C on both cores using TAT @ stock volts, 300FSB
 
Boltaction: TAT does not report the correct cpu core temperature in all situations on all boards.

Your real core temperature is being reported by SpeedFan and if you don't trust those 56C numbers then check out the program CoreTemp and they will confirm that. You have plenty of room to overclock your system.

TAT was designed to create maximum heat for laptop testing. It's a little overkill for desktop users since there are no programs that can heat up your cpu the way tat does. Orthos is a little more practical than TAT for load testing.

333 X 9 should be simple now that your fan is working at the proper speed. 366 X 9 is likely at about 1.400 v - 1.425 volts for the CPU.
 
Unclewebb:

I tried@ 333, no go. I will attempt to raise the Vcore to 1.4 to verify.


UPDATE: boots to desktop but BSOD's on OCCT run when trying to run 333 @ 1.4. I am now checking 1.425... nope that crashes too. Core Temp showed 66C on both cores. OCCT took 30 seconds to fail.

Maybe my PCI isn't locked down? I set my PCI-E to 100mhz


:(




Stable at 2.7ghz on OCCT@ 1 hour...
 
You might want to try 2.2 volts for the memory. It might help and definitely won't hurt.

The E6300 / E6400 generally have no problem getting up to about 3200 MHz without needing any additional core voltage.

If you can, run a program like MemTest86+ overnight to make sure your memory isn't the problem.

http://www.memtest.org/

Even if your memory is new and from a brand name source you still shouldn't trust it 100% until it's been proven to be error free. I would set the cpu MHz to something near default, core voltage to ~1.325 v and raise your memory speed with a different memory divider so you can isolate the memory that way.

Run MemSet to show what memory timings you're getting.
Version 3.0: http://rv.page.cegetel.net.perso.cegetel.net/MemSet.exe
Version 3.2: http://www.overclockers.ru/download?794

In Windows I think I'd start with Orthos testing and then use OCCT.

It's probably something really minor holding you back so keep at it.
 
I think you need to play with your bios settings or get it from someone that has the same board thats got a good overclock..... setting something at enabled instead of disable sometimes stops your overclock by quite a bit....
 
I'm also using the stock HS, might that be an issue?


Update
The system posts to desktop @ 1.45V, 333FSB but goes past 70C when running OCCT. I'm beginning to consider that this isn't a "good" overclocking stepping. It was manufactured 1/30/07
 
I'm afraid a lot of people can't get > 3.0GHz out of their 4300
 
I'm afraid a lot of people can't get > 3.0GHz out of their 4300

Check my sig ;) I also know a lot who can run at 3.0-3.2 without issues :)

On topic, I guess it's a improper heatsink installation so recheck it and make sure the push pins is 100% engaged. Those are a bitch to engage fully.
 
Ram shouldn't be the issue with the e4300.

Depending on the motherboard and its vdroop an E4300 needs close to 1.425 to run @ 3000 under load. As far as your temps are concerned, are you running the stock cooler? If so then I would see you getting beyond 2.8 easily.

The new rev 3.3 DS3 is running my E4300 @ 3200 with 1.425vcore and 1.9 vdimm on the cheepest G.Skill DDR2 800 2G kit from newegg. Temps are fine on the Tuniq Tower 120.
 
Test your ram with MEMTEST to make sure there are no issues at that end.
Set the RAM to its default specs.

Update to the latest bios, I think F10 for DS3 and S3.
Make sure you have these in the MB settings:

RAM: + 1,+2 or +3 depending on ram specs
FSB: +1
MCH: +1
Some have the PCIE:+1 but mine is set to normal and don't have issues with it
Set the CPU auto adjust stuff to disable, forgot the specifics but something like C1E and thermal throttling thing for CPU.
Set the voltage to whatever highest you want to go then start with 333x9.

This method worked well for me to start with, I actually hit 3GHz at 1.33v, I think I can hit 3.2 but I'd probably have to up the vcore to 1.4.

Now after hitting my target speed and temps with no issues on OCCT and Orthos I proceeded to tweak the memory speed and did memory test again.

Specs:
[email protected] (1.37v)
Katana Cu with Shit etsu x23
DS3 rev 3.3
Cooler Master Copper NB cooler with Shin Etsu x23
SB TIM was replaced my white goop
2GB GSkill
8800GTS
250GB SATA
DVDRW SATA
Antec 900 case
Corsair 620 PSU
etc. etc.

Good Luck on your OC
 
try these settings.

[img=http://img337.imageshack.us/img337/5/occtstablewm3.th.png]

temps are a little high when stress testing , but during gaming and rendering I rarely get over 56c on stock hsf.
one other thing ,i have the same mobo and cpu and my vvoltages are set too:
v-core:1.375
ddr voltage:+0.3
mch voltage:+0.3

also my pci-e freq is set too auto , when i set it manually for some reason my cpu freq drops from 375 to 374
 
ok... so 2.25ghz (950 fsb) is as stable as i can get my e4300. i've tried to do 1000fsb, and that just freezes up during bootup. it detects speed of processor then stops right before detecting ide. i've gone up and down with voltages, i was at 1.4v the last time i tried. i'm not sure if i tried it, but i think it does the same thing at 1.45v. do the temps for 1.4 volts look right? it's hovering around 41-43C's with stock e6300 cooler. maybe i have an fsb hole on my p5nE? i got this thing from fry's in a combo deal. i tried a little oc'n /w it on my abit f-i90hd, froze, and now my abit board won't boot at all no matter if i use the e4300 or my spare celeron 356. this is really confusing me because my celeron 356 and e6300 overclocks just fine one both boards. could it be that i got a craptastic overclocking chip? stepping is: SL9TB 0649A111





steppings
error message
 
I'm also using the stock HS, might that be an issue?

Yes, I think it is.




xkon -
could it be that i got a craptastic overclocking chip?

No, at first glance I think its because you dont have a clue what you are doing, Start your own thread and put your full system specs in it and perhaps we can help you. Reset the cmos on your Abit, check the manual for instructions.
 
Yes, I think it is.




xkon -


No, at first glance I think its because you dont have a clue what you are doing, Start your own thread and put your full system specs in it and perhaps we can help you. Reset the cmos on your Abit, check the manual for instructions.

bwahaha. a little offensive at first, but partially true. :D

i've reset the cmos on the abit mobo numerous times, did a couple hard resets too (took battery out and shorted jumpers). i've taken it up to 360 without a hiccup on my e6300. i throw this e4300 in there and start off slow, but it wouldnt boot. second time around that i turn it on it goes through bootup screen then error message comes up, and i turn it off. after that fans and lights turn on, but nothing on keyboard/monitor. i switched processors to the celeron and it does the same thing, no matter how many cmos resets i do. thing is is that i'm not using the abit board anymore as it's not booting up with any cpus. the error screen i got was from the abit board. i was using a noctua nh-u9 /w the abit.

the p5n-e on the other hand boots up with it, but it's bus speed wont do anything over 238 (@ 950 rated fsb) 40 degrees Celcius. i was doing 50Mhz rated fsb jumps because that's what the bios uses to overclock with. i've tried upping the voltage to 1.45volts, but same results, i freeze in detection or in bios.

here's my new post: e4300 not overclocking well, suggestions?

arctic cooling's freezer 7 seem good for its price. if you can find one :D nanosys1 19.99 + shipping
thermaltake blue orb II: xioxide 32.49 + shipping
scythe mine: newegg 32.99 + shipping
 
What's the cheapest heatsink I can get away with?
The Intel heatsink and fan! I can get 3400 MHz out of my E6400 with it and even 3500 MHz or 3600 MHz out of it in the cool basement for some quick benches. If you're not over 3000 MHz there's usually no reason to replace it.
 
The Intel heatsink and fan! I can get 3400 MHz out of my E6400 with it and even 3500 MHz or 3600 MHz out of it in the cool basement for some quick benches. If you're not over 3000 MHz there's usually no reason to replace it.

Ok... whats too high a temp?
 
The temperature is too high when your computer becomes unstable.

The Core 2 Duo is rated to 85C before it will start to throttle itself and temporarily slow down.

I've found excellent stability up to a core temperature of about 70C to 72C. If you're not running at full MHz and full volts then you can probably go a little higher than that reliably.
 
Ram shouldn't be the issue with the e4300.

Depending on the motherboard and its vdroop an E4300 needs close to 1.425 to run @ 3000 under load. As far as your temps are concerned, are you running the stock cooler? If so then I would see you getting beyond 2.8 easily.

The new rev 3.3 DS3 is running my E4300 @ 3200 with 1.425vcore and 1.9 vdimm on the cheepest G.Skill DDR2 800 2G kit from newegg. Temps are fine on the Tuniq Tower 120.

I have a 4300 (week 41), DS3 3.3 and I can hit 3.04ghz (8 x 380) at 3.25v. I hit 3150 at 9x, but that was at 1.35v (hey, the box says max voltage is 1.35, who am I to argue with Intel). Probably could have gone higher, but I was lookinat at TAT and those temps scared me.

I may try to see what it can do at some point, but it's hard to complain about a 70% oc.

One thing I don't get is the board simply will not boot at 400mhz, even if I set the multiplier at 6. I'm not sure what the issue is. Then again, I'm not sure what some of voltages are by default (chipsets and FSB come to mind), nevermind whether I should leave hpet on or turn it off.
 
I hit 3150 at 9x, but that was at 1.35v (hey, the box says max voltage is 1.35, who am I to argue with Intel). Probably could have gone higher, but I was lookinat at TAT and those temps scared me.
.


Don't use TAT for temp monitoring, in most cases, it is wrong on the newer stuff.
 
Greetings!Do use TAT in E4300.
CoreTemp shows idle E4300 temperatures bellow ambient temperatures. Its off by ~15ºC.

Get the latest CoreTemp or SpeedFan 4.32 and you will get a accurate core temperature. TAT is off by 12 degrees, on the high scale. Me and unclewebb did some extensive testing to find out which software give a accurate reading and we decided CoreTemp is the most accurate.

It's possible to get the core temperature a few degrees below ambient since the metal has a tendency to get cooler at ambient temperatures. Core 2 Duo are running very cool at stock speed and voltage.
 
Greetings!

TAT and CoreTemp read from within the CPU die. Its impossible to be bellow ambient temperature using a Zalman 9700.

CoreTemp fails reading Allendales cores (E4300 and the new E6300/E6400).

Ambient temperature is 20ºC, MB is P5B Deluxe, BIOS is 1004 (0604 and 0910 showed the same difference).
Check it:
tempspp8.jpg



On E4300s use TAT for temperature reading.
 
CoreTemp fails reading Allendales cores (E4300 and the new E6300/E6400).
I've been preaching CoreTemp and SpeedFan v.4.32 but your results seem to confirm that there is a problem on the new E4300 Allendale chips. I read a while ago on the CoreTemp forum about this so I know the programmer of CoreTemp is looking into it.

If ambient is 20C and your computer has had time to warm up then you're right. The core temp can't be lower than that. Not even a Zalman 9700 can perform miracles!
 
I've been preaching CoreTemp and SpeedFan v.4.32 but your results seem to confirm that there is a problem on the new E4300 Allendale chips. I read a while ago on the CoreTemp forum about this so I know the programmer of CoreTemp is looking into it.

If ambient is 20C and your computer has had time to warm up then you're right. The core temp can't be lower than that. Not even a Zalman 9700 can perform miracles!

After some thinking, It may be possible that it's 10C off. Remember my testing with TAT workload and rebooting at around 75C. Maybe it's really 85C when it rebooted...
 
I've seen some references to an FSB ceiling on e4300s, at around 400 FSB. Maybe this is your problem? ... Willing to try a CPU side "pin" mod (note quotes since the pins are in the socket now of course).

short thread: http://forums.vr-zone.com/showthread.php?t=131468
the mod : http://xtremesystems.org/forums/showthread.php?t=134550

You can apparently force either 1066 boot (edit: on the below still maybe forcing 1066 could unlock the magic o/c juice even though ur already @ that point), or 1333 - assuming your mobo allows it/supports it No idea if this will help you, but I don't imagine that it could hurt - and is easily reversible if you want to remove it.

[edit] Oh, wait you're @ 266 eh? Hmm well I suppose testing a forced 1066 (or even 1333 if you're board is capable) is worth a try. You could have some problem with a screwy bootstrap change, hence the CPU and/or board having issues going further -- hence the mod..But I've not seen the issues @ your ~266, more like the previously mentioned ~400
 
Using ddr2 667 cas 5 (cheap stuff) could be limiting you.

Do you notice any stability change at a stable overclock setting if you change vdimm?

When I had some PQI ddr2 667 cas 4 sticks, I had similar problems with my e6600 (9x multiplier). It was only stable to about 2.7Ghz (9x300).

Swapped out to ddr2 800 and it does 3.2Ghz (still not great, but much better)
 
I've been preaching CoreTemp and SpeedFan v.4.32 but your results seem to confirm that there is a problem on the new E4300 Allendale chips. I read a while ago on the CoreTemp forum about this so I know the programmer of CoreTemp is looking into it.

If ambient is 20C and your computer has had time to warm up then you're right. The core temp can't be lower than that. Not even a Zalman 9700 can perform miracles!

I think it depends on the board. This room is roughly 22c. My Core Temp reads at 32/35. TAT reads 49/51. I'm using an e4300 on a DS3....week 41 if it matters.
 
I have an Allendale E6400 on the DS3, and both CoreTemp and Speedfan agree with each other, while TAT is at least 10c too high.....so at least on some Allendales, CoreTemp does work properly.
 
P.S. I assume you've already done this but be sure to disable C1E, speedstep, etc. in BIOS.

Also just to be safe go in to Power management in windows under the screensaver tab, and set it to Always On or Home/Office... ie: NOT Portable or Battery, etc.
 
I think it depends on the board. This room is roughly 22c. My Core Temp reads at 32/35.
The 30C to 35C range for an idle core temp on a C2D with 2MB cache is pretty normal.

I don't understand why but it does seem that the temp of the E4300 on some boards is too low.

I guess we need more crazy volunteers that are willing to bring their new E4300 up to 85C by reducing their cpu fan speed so we can see what's really going on. The only thing I know for 100% sure is that CoreTemp works correctly on my E6400 on my P5B Deluxe.

impar: I agree that CoreTemp / SpeedFan are wrong on your system but that doesn't mean or prove that TAT is right. My theory is that these new chips are being misread so that the thermal throttle or max temperature is being read as 100C like the laptop chips instead of 85C like the original Conroe E6300/E6400 chips. If this is what's going on then I think adding exactly 15C to your CoreTemp / SpeedFan readings will get you the "real" temperature of the cores in your cpu.

How about some load temps while running Orthos?

Your readings of 17C / 19C would then be 32C / 34C which are very realistic idle temps for your ambient temperature and the MHz and voltage you are using.
 
Greetings!
I guess we need more crazy volunteers that are willing to bring their new E4300 up to 85C by reducing their cpu fan speed so we can see what's really going on.
That was already done:
http://www.xtremesystems.org/forums/showthread.php?t=131008
The only thing I know for 100% sure is that CoreTemp works correctly on my E6400 on my P5B Deluxe.
Is yours a true Allendale 2MB native, or a Conroe with half cache disabled?
I agree that CoreTemp / SpeedFan are wrong on your system but that doesn't mean or prove that TAT is right. My theory is that these new chips are being misread so that the thermal throttle or max temperature is being read as 100C like the laptop chips instead of 85C like the original Conroe E6300/E6400 chips. If this is what's going on then I think adding exactly 15C to your CoreTemp / SpeedFan readings will get you the "real" temperature of the cores in your cpu.
...
Your readings of 17C / 19C would then be 32C / 34C which are very realistic idle temps for your ambient temperature and the MHz and voltage you are using.
That puts the temperature very close to TAT reports.
How about some load temps while running Orthos?
There:
loadtempsgs3.jpg

 
Ok, so you are saying when coretemp shows @ 70, I am hitting 85?



OK, here is my mode of "attack":

Set CPU back to default, all volts at default
put RAM @ 667, CAS 5 and 1.8, run memtest for 30 minutes
put ram @ 667 CAS 4 1.8v, 30 minutes
if stable, try 800 @ CAS 5, raise VDIMM +.2

Today I am going to isolate my ram.
 
Back
Top