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

unclewebb

[H]ard|Gawd
Joined
Jun 21, 2006
Messages
1,807
What does TAT tells you?
Here's my early Conroe E6400 with half the cache disabled on a P5B Deluxe.

tortureuc4.png


I could not get CoreTemp higher than 85C. TAT reported that the Intel thermal throttle was active on both cores which halted the temperature at 85C, the Intel maximum for a C2D.

At this point TAT was reporting 83C. From this test on my board I came to the conclusion that CoreTemp is reporting the temperature most accurately.

If the E4300 on some boards is being reported 15C too light then 70C as reported by CoreTemp would equal 85C. I guess the only way to properly test is to run TAT and CoreTemp at the same time and see when it starts to report thermal throttling.

The post on the xtremesystems forum that shows TAT reporting temperatures over 90C is definitely wrong.

And Intel thought going to an on chip digital temperature probe would make things less confusing! :rolleyes:

impar: CoreTemp reports yours at 47C under load so I believe that is really 47C +15C = 62C. TAT always reports 2C or 3C less than what CoreTemp reports on my board so that equals 62C - 3C = 59C which is your TAT reported temperature. It all sort of makes sense.
 

Boltaction

Gawd
Joined
Jun 10, 2004
Messages
939
RAM:
(TM2 function enabled, everything else set to disabled. All voltage except for RAM set to Auto. CPU running stock speed)
533@5-5-5 1.8V
memtest: pass

667@5-5-5 1.8V
memtest: FAIL
memtest @ 2.3v: FAIL

800@5-5-5 2.3V
memtest: FAIL
2.4v memtest: FAIL


Motherboard:
plans- drop the multi and ram down as low as it will go,raise MCH by +1 and try for highest FSB.
not started


CPU:
drop ram to 1/1, stock vcore and raise. 30 minutes ORTHOS per clock, raising by 50mhz FSB until error.
not started




Looks like I found the culprit. :(
 

unclewebb

[H]ard|Gawd
Joined
Jun 21, 2006
Messages
1,807
At the other extreme, with an ambient temperature of 19C, CoreTemp reports my E6400 at 24C when using default voltage, default 2133 MHz and the Intel OEM heatsink and fan.

intele6400defaultns7.png


A Zalman might reduce that a couple of degrees but it should never get cooler than ambient.

Maybe the above test is a good benchmark to see if your case is being properly cooled and your heatsink was properly installed.

5C over ambient is the number to beat at default MHz & volts! :D

Update: With the core voltage dropped down to 1.176 volts, as reported by CPUz, the core temp hit 2C above the 18C ambient. Very cool! :cool:
For this second run I also changed the multiplier from 8X down to 6X so MHz dropped from 2133 MHz down to 1600 MHz.
In my testing, MHz doesn't effect core temps nearly as much as core voltage does.

intele6400118voltscu2.png


Nice chip Intel.

Anyone can set their multi to 6X and drop their cpu voltage to 1.18 volts as reported by SpeedFan or CPUz and should be able to get their C2D down to within a couple of degrees of their ambient temperature. This should be a very easy, safe way to test and calibrate CoreTemp and SpeedFan. If during this test your E4300 readings are way below ambient then add on 15C and see if that makes sense. Most users will be a lot happier calibrating this way compared to taking their new investment to 85C and praying to God that the thermal throttling works properly.
 

AbRASiON

Limp Gawd
Joined
Jul 28, 2004
Messages
354
Just like to chime in here.

I have the E4300, SL9TB from Malaysia
It's a Q644A312 batch CPU and frankly, it's junk - I can't think of a better word for it, it's the polar opposite of the ones people have been reviewing at 3.2

I have extreme difficulty getting it to post on the DS3P 3.3 at more than 2.8ghz, real real difficulty - it's really only stable around 2.5.......

Just a warning - so if anyone else is having problems - might wanna cross reference that batch number

- Good luck.
 

impar

n00b
Joined
Sep 1, 2004
Messages
53
Greetings!

Tested:
temps095rc8.jpg



It adds 15ºC, as can be seen by SpeedFans readings and the 100-85ºC of Tjunction.
 

unclewebb

[H]ard|Gawd
Joined
Jun 21, 2006
Messages
1,807
Thanks for the update impar.

I tried CoreTemp 0.95 on my E6400 and it reports the same as what CoreTemp 0.94 reports so it looks like he fixed it for the new E4300 chips without screwing up the readings for the previous chips! That's great.

If you have a chance impar can you try my experiment to see how low your idle temps can go compared to ambient.

Lock your multi at 6X in the bios and set core voltage down to about 1.175 v to 1.200 volts.

It would be interesting to see what CoreTemp 0.95 has to say now.
 

CpuMan

Gawd
Joined
May 6, 2005
Messages
608
Thanks for the update impar.

I tried CoreTemp 0.95 on my E6400 and it reports the same as what CoreTemp 0.94 reports so it looks like he fixed it for the new E4300 chips without screwing up the readings for the previous chips! That's great.

If you have a chance impar can you try my experiment to see how low your idle temps can go compared to ambient.

Lock your multi at 6X in the bios and set core voltage down to about 1.175 v to 1.200 volts.

It would be interesting to see what CoreTemp 0.95 has to say now.

I just tried CoreTemp 0.95 on my E6400 and it reads 15 degrees higher than CoreTemp 0.94 or speedfan. Now I do not know what to believe??

 

unclewebb

[H]ard|Gawd
Joined
Jun 21, 2006
Messages
1,807
Now I do not know what to believe??
Here we go again!

Why not try my experiment and run your CPU as slow as possible with low volts and see if your core temperature gets down to ambient temperature.

Post your exact details in terms of heatsink and fan you're using, ambient temperature, MHz and core voltage as reported by SpeedFan so we can see if your new number makes sense.

Do you know if your E6400 is one of the new ones? Check out the revision with CPUz. My E6400 is an early, revision B2, and the new ones should be revision L2.
 

jbmx4life

2[H]4U
Joined
Oct 14, 2004
Messages
2,550
It seems that the 15c higher reading on CoreTemp 0.95 comes right along with the 15c higher Tjunction. That seems logical :confused:
 

CpuMan

Gawd
Joined
May 6, 2005
Messages
608
Here we go again!

Why not try my experiment and run your CPU as slow as possible with low volts and see if your core temperature gets down to ambient temperature.

Post your exact details in terms of heatsink and fan you're using, ambient temperature, MHz and core voltage as reported by SpeedFan so we can see if your new number makes sense.

Do you know if your E6400 is one of the new ones? Check out the revision with CPUz. My E6400 is an early, revision B2, and the new ones should be revision L2.

I am pretty sure that CoreTemp 0.94 is correct and 0.95 broke something.
I do have the new E6400 with the L2 stepping @ 2.66GHz as you can see in the screenshot in my above post.
My pc details are in my sig. I am using a ZALMAN CNPS7000B-ALCU.

I do not have time right now to drop my vcore voltage, I need the system up and running right now. I might mess around with it this weekend though.
 

vanilla_guerilla

Supreme [H]ardness
Joined
Oct 3, 2005
Messages
5,895
why would you want to drop your vcore voltage anyway? well this weekend when you have 10 or 12 hours to spare, go into bios and change your vcore.
 

unclewebb

[H]ard|Gawd
Joined
Jun 21, 2006
Messages
1,807
10 or 12 hours to spare? To adjust the multi and drop the core voltage?

Last time I checked it was one reboot and a couple of minor bios adjustments. Simple. Your computer is still quite usable and for most light duty things the change is barely noticeable.

The reason I suggest doing this is because it seems to be an easy way to calibrate your temperature readings. By dropping things down to low MHz and low voltage I found that CoreTemp reported about 2C great than my ambient temperature.

If CoreTemp reports something totally different than ambient temperature then it is a sign that something is not right.
 

Xilikon

[H]ard|DCer of the Year 2008
Joined
Oct 12, 2004
Messages
14,944
10 or 12 hours to spare? To adjust the multi and drop the core voltage?

Last time I checked it was one reboot and a couple of minor bios adjustments. Simple. Your computer is still quite usable and for most light duty things the change is barely noticeable.

The reason I suggest doing this is because it seems to be an easy way to calibrate your temperature readings. By dropping things down to low MHz and low voltage I found that CoreTemp reported about 2C great than my ambient temperature.

If CoreTemp reports something totally different than ambient temperature then it is a sign that something is not right.

Missed the new release with the whole basement finish job I had during the weekend. I will give a try and report the results but if 0.95 increased 15C, it means mine is running a bit hot at 75C on average.
 

impar

n00b
Joined
Sep 1, 2004
Messages
53
Greetings!
If you have a chance impar can you try my experiment to see how low your idle temps can go compared to ambient.
Lock your multi at 6X in the bios and set core voltage down to about 1.175 v to 1.200 volts.
What FSB?
 

unclewebb

[H]ard|Gawd
Joined
Jun 21, 2006
Messages
1,807
Greetings!What FSB?
I used my E6400 default of 266 MHz but for an E4300 you could use its default of 200 MHz.

The point is to try and get your chip to produce as little heat as possible. With the Intel cooler I got down to 2C above ambient temperature which makes perfect sense. I think most people should be able to get down to a similar temperature.

It would make no sense if you ended up 10C or more below ambient. Being below ambient is a sign that the temperature is not being reported correctly.

If a person was 15C or more above ambient then either they have horrible case air flow, they screwed up installing their heatsink or CoreTemp isn't reporting the correct temperature.

AMD T-type: Try using SpeedFan v.4.32. They added C2D core temperature monitoring with that version.
 

vanilla_guerilla

Supreme [H]ardness
Joined
Oct 3, 2005
Messages
5,895
10 or 12 hours to spare? To adjust the multi and drop the core voltage?

Last time I checked it was one reboot and a couple of minor bios adjustments. Simple. Your computer is still quite usable and for most light duty things the change is barely noticeable.


yea no shit. the guy said he didnt have the time to change his vcore. that was a little joke.
but thanx for the heads up re changing bios settings. (thats a little joke too)
 

CpuMan

Gawd
Joined
May 6, 2005
Messages
608
10 or 12 hours to spare? To adjust the multi and drop the core voltage?

Last time I checked it was one reboot and a couple of minor bios adjustments. Simple. Your computer is still quite usable and for most light duty things the change is barely noticeable.

The reason I suggest doing this is because it seems to be an easy way to calibrate your temperature readings. By dropping things down to low MHz and low voltage I found that CoreTemp reported about 2C great than my ambient temperature.

If CoreTemp reports something totally different than ambient temperature then it is a sign that something is not right.

LOL, I did not mean to sound rude, sorry if I did, I appreciate all you have contributed. I understand about the vcore drop and I think its a good way to check temperature sensors.
The reason I wanted to wait is I have stuff running and did not want to shut down the pc just yet, I have some stuff downloading also. When I get a chance I'll shut every thing down and do as you said.

I would be interested to know if anybody else is seeing a 15c difference between CoreTemp 0.94 and CoreTemp 0.95 on L2 steppings with the e6300 and e6400's?

Also what is Tjunction? It is 85c on CoreTemp 0.94 and 100c on CoreTemp 0.95? Also if you notice in the screenshot, CoreTemp 0.95 shows my cpu to be a "conroe-2m" but CoreTemp 0.94 shows "Allendale.
 

unclewebb

[H]ard|Gawd
Joined
Jun 21, 2006
Messages
1,807
Tjunction is the maximum rated temperature that the processor is designed to safely operate at.

All of the original B2 E6300 / E6400 cpus that I've seen have a Tjunction of 85C. When testing my cpu by slowing down the cpu fan speed it allowed the temperature to come up to 85C but thermal throttling kicked in and didn't allow it to exceed this temperature. The C2D has this feature built into it to keep it from self destructing due to heat. That's why people should have TM ( thermal monitoring ) or TM2 turned on in the bios. It's a good safety feature if the cpu fan ever fails.

I believe the C2D for many laptops has had a Tjunction rating of 100C since the beginning.

In theory a processor that can operate reliably at a higher temperature is generally considered a better grade of processor. In reality there may not be much real world difference between the 85C C2D and a 100C C2D. It might only be Intel raising the rating of these processors to better reflect their true capabilities.

My C2D was surprisingly very stable at 85C though I didn't hold the pedal to the metal for too long.
 

twogunpete

n00b
Joined
Mar 13, 2006
Messages
54
ok !would the t-junction being 100c in coretemp 0.95 mean that my e4300 is a mobile processor?
 

Xilikon

[H]ard|DCer of the Year 2008
Joined
Oct 12, 2004
Messages
14,944
I tried 0.95 last evening and it is now up by 15C and now matching TAT within 2-3 degrees. However, with the new heatsink mounting installation (ordered a set of bolt-down Thermalright retention mechanism designed for a Ultra-120 to remplace the crappy push-pin mechanism I had with my SI-128 and it fit perfectly and more reliably. One of the pins was borderline from popping out so it's not 100% fully seated), I got up to 62-63C under Folding SMP load (meaning it's approximately 68-70C under OCCT).

Upon further checking, I looked at my electronic thermostat in the room and it's currently 21C and SpeedFan 4.32 show 19C when idle with C1E enabled (6x multiplier) while CoreTemp 0.95 show 33-35C, which is more logical.

I'm now lost with which software is right :eek:
 

unclewebb

[H]ard|Gawd
Joined
Jun 21, 2006
Messages
1,807
ok !would the t-junction being 100c in coretemp 0.95 mean that my e4300 is a mobile processor?
An E4300 is not a mobile processor. The E4300 may have the same 100C t-junction as mobile processors but they're not the same processor.

Xilikon: TAT has always been 2C - 3C less than CoreTemp on my P5B Deluxe with an E6400 at full load.

If SpeedFan is showing CPU core temps below ambient then that's a good sign that it's not reporting correctly. Do my test and drop the MHz and cpu voltage and you will see your SpeedFan reported core temperature drop even further below ambient which isn't possible with a heatsink and air cooler.

I found on my board that enabling C1E in the bios made no difference for either power consumption at the plug or for reducing the idle temperature.
 

unclewebb

[H]ard|Gawd
Joined
Jun 21, 2006
Messages
1,807
CoreTemp 0.95 shows my cpu to be a "conroe-2m" but CoreTemp 0.94 shows "Allendale.
This confusion is Intel's fault. At the design level they might have Conroe and Allendale processors but at the retail level they don't seem to want consumers to know what they've got.

They are switching the manufacturing of the E6300 / E6400 from the original 4MB cache with 2MB disabled to the new generation that only has 2MB to begin with or what enthusiasts have been calling the Allendale core. If too many consumers discover that one is better than the other than they can't sell the other without having to discount them. No one wants old technology if new is better and no one wants new technology if the original E6300 / E6400 processors perform better.

Information is coded into these chips but Intel doesn't like to share too much information with individual programmers writing a utility like CoreTemp. If you're part of a large corporation and your company is willing to sign a NDA then maybe they'll talk to you.

Head to the Intel website and try to get this basic information and you'll find it's not readily available.


Searched for Allendale

Your search - Allendale - did not match any documents.
No pages were found containing "Allendale"

Suggestions:

* Make sure all words are spelled correctly.
* Try different keywords.
* Try more general keywords.
 

impar

n00b
Joined
Sep 1, 2004
Messages
53
Greetings!
If you have a chance impar can you try my experiment to see how low your idle temps can go compared to ambient.
Lock your multi at 6X in the bios and set core voltage down to about 1.175 v to 1.200 volts.
Finally remembered to do it before turning system on. :)
6x200 with 1,1v in BIOS. Ambient temperature was 17ºC.



Noticed that CoreTemp and Orthos are reading 300 FSB. Check third post here:
http://www.bleedinedgesupport.com/ocz/forum/showthread.php?t=23803
9/6*200=300
 

CpuMan

Gawd
Joined
May 6, 2005
Messages
608
This confusion is Intel's fault. At the design level they might have Conroe and Allendale processors but at the retail level they don't seem to want consumers to know what they've got.

They are switching the manufacturing of the E6300 / E6400 from the original 4MB cache with 2MB disabled to the new generation that only has 2MB to begin with or what enthusiasts have been calling the Allendale core. If too many consumers discover that one is better than the other than they can't sell the other without having to discount them. No one wants old technology if new is better and no one wants new technology if the original E6300 / E6400 processors perform better.

Information is coded into these chips but Intel doesn't like to share too much information with individual programmers writing a utility like CoreTemp. If you're part of a large corporation and your company is willing to sign a NDA then maybe they'll talk to you.

Head to the Intel website and try to get this basic information and you'll find it's not readily available.


Searched for Allendale

Your search - Allendale - did not match any documents.
No pages were found containing "Allendale"

Suggestions:

* Make sure all words are spelled correctly.
* Try different keywords.
* Try more general keywords.

Well said,

This is some of the best info so far that I have seen on the newer steppings, not enough info to write code for coretemp. Just some general information.

http://en.wikipedia.org/wiki/List_of_Intel_Core_2_microprocessors
 

unclewebb

[H]ard|Gawd
Joined
Jun 21, 2006
Messages
1,807
impar: I think your numbers make it pretty obvious that CoreTemp 0.95 is reporting your E4300 core temperature properly and SpeedFan v.4.32 is wrong.

With an ambient of 17C, it's impossible to get a core temperature of 12C with air cooling. What sort of thermal paste are you using? Applying a thin rectangular area of AS5 directly over the cores is working great for my idle temps. The thin line that ArcticSilver recommends on their website wasn't as effective for me.

CpuMan: If the wiki info is true, which I believe it is, then writing a utility that determines Allendale or Conroe is simple. The 3 known Allendale cores, E4300 and the new E6300 / E6400, are all reported as Revision L2 by CPU-z. The Revision B2 E6300 / E6400 are all built on the Conroe core. Mystery solved!
 

Marvelous

Supreme [H]ardness
Joined
Apr 26, 2003
Messages
4,117
Why doesn't INTEL just enable the cache on the E6300 and E6400 in the first place since these aren't Allendale.

These 2meg cache E6000 series are supposed to be faded it out anyway. The 4meg cache is coming and no one is going to buy the 2meg cache version of these chips at the same price point. Not to mention overclock ability of the Conroe over Allendale. They wasted time and energy building E63-6400 allendale just so they can phase them out.
 

CpuMan

Gawd
Joined
May 6, 2005
Messages
608
impar: I think your numbers make it pretty obvious that CoreTemp 0.95 is reporting your E4300 core temperature properly and SpeedFan v.4.32 is wrong.

With an ambient of 17C, it's impossible to get a core temperature of 12C with air cooling. What sort of thermal paste are you using? Applying a thin rectangular area of AS5 directly over the cores is working great for my idle temps. The thin line that ArcticSilver recommends on their website wasn't as effective for me.

CpuMan: If the wiki info is true, which I believe it is, then writing a utility that determines Allendale or Conroe is simple. The 3 known Allendale cores, E4300 and the new E6300 / E6400, are all reported as Revision L2 by CPU-z. The Revision B2 E6300 / E6400 are all built on the Conroe core. Mystery solved!

Right, I agree. All the CoreTemp author would have to do is have the program check the steppings on the 6300 and 6400's, B2 = Conroe, and L2 = Allendale, and adjust accordingly.
 

coldpower27

[H]ard|Gawd
Joined
Jan 27, 2005
Messages
1,154
Why doesn't INTEL just enable the cache on the E6300 and E6400 in the first place since these aren't Allendale.

These 2meg cache E6000 series are supposed to be faded it out anyway. The 4meg cache is coming and no one is going to buy the 2meg cache version of these chips at the same price point. Not to mention overclock ability of the Conroe over Allendale. They wasted time and energy building E63-6400 allendale just so they can phase them out.

They are doing just that, with the introduction of the E6320/E6420, the initial disabling was probably due to market segmentation more then anything else.

The Allendale core isn't just used for the E6300/E6400 alone, they are going to be used for sometime on the E4xxx line as well as the Pentium E21xx line as well. It's isn't a big deal to stop binning for E6xxx Allendale as they are phased out because the Allendale core still has a long life ahead.
 

unclewebb

[H]ard|Gawd
Joined
Jun 21, 2006
Messages
1,807
Intel took over the top end of the market last year as well as the mid range with the E6300 / E6400.

Intel just introduced the Allendale cores this year which are going to be around for a while putting the squeeze on AMD in the entry level market.
 

merlin87

Gawd
Joined
Aug 18, 2001
Messages
964
I got an offtopic question. I just got my e4300 to 3.2ghz. It was orthos stable for 9 hours and passed the occt test on an hour setting. I now notice that whenever I get my comp to 100% usage my mouse starts to jump around and randomly click on things. This only started after I got to 3.2. Is this connected to my OC? At load Im hitting about 70*c-74*c
 

Marvelous

Supreme [H]ardness
Joined
Apr 26, 2003
Messages
4,117
70C is pretty high. You want to keep it about 10c under what your temp is right now.
 
Top