Sign in with
Sign up | Sign in
Your question

A simple question about overclocking

Last response: in Overclocking
Share
May 31, 2008 10:37:32 AM

I have amd athlon64 x2 4400+ (brisbane) 2.3ghz processor.
Overclock it and whoops..... 2.6ghz (using stock cooler fan)

The question is, how further I can go overclocking my processor using STOCK fan?

Thanks.
May 31, 2008 1:55:02 PM

General tools for overclocking.
CoreTemp : http://www.alcpu.com/CoreTemp/
cpuid HW Monitor : http://www.cpuid.com/hwmonitor.php
cpuid CPU-Z : http://www.cpuid.com/cpuz.php

Prime95 : http://www.majorgeeks.com/download.php?det=4363
(make sure you enable "Advanced -> Round Off Checking")
OCCT : http://www.ocbase.com/perestroika_en/index.php?Download

MemTest86 : http://www.memtest86.com/memtest86-3.4a.iso.zip

As long as CoreTemp reports Prime95/OCCT load temperatures under 55°C-60°C and Prime95/OCCT doesn't report errors you are good.
May 31, 2008 2:20:10 PM

as long as you dont overheat it is fine.but in general.even a cheap after market cooler like Hyper TX2 is worth the investment.
Related resources
May 31, 2008 3:29:05 PM

thanks for the software.
andrius wat do u mean temp under 55-60 celcius? Is it just core 1/2 or total of them?
May 31, 2008 3:37:09 PM

the temp of all the core when under 100% load dont exceed those numbers.
May 31, 2008 7:48:08 PM

CoreTemp reports temperatures for both cores.
If you run Prime95 with 2 threads(1 per core) both cores should be below 55°C-60°C. If they go over 55°C I would reduce the overclock. If they go above 60°C you are getting close to the thermal shutdown threshold and at that point you are risking potential heat damage to the CPU.

I wouldn't run it above 55°C for prolonged periods of time but that's just me.
May 31, 2008 8:34:39 PM

The 55-60C should be read from HWMonitor where it says just "CPU Temp" or something to that effect. CoreTemp reads the temps directly from the cores, and they can run considerably warmer than the 'thermal threshold'.
June 14, 2008 1:06:22 PM

Ok... right now I just got the cheap ThermalTake DuCool cooler fan..
do the magic and..

Bus = (200 default) 244 x 11.5
Core= (2300 default) 2805.5 MHz (21.98 % overclock)
HT link = 1219.7 MHz
Core volt = 1.35 v

temp = core 1 core 2
23c 21c
(@ 1-8% load)

temp = core 1 core 2
45c 46c
(@ 100% load)


so... is this good? I think its hot. Touch the heatsink pipe and burn like hell (@ 100% load)



June 14, 2008 1:39:45 PM

Looks good if those are CoreTemp readouts. Everything below 50°C is good. Metal at ~45°C will burn you because it transfers heat better than air.
It's the same thing with a car in direct sunlight. It's only a few degrees hotter than the air surrounding it but it burns to the touch.
a b K Overclocking
June 14, 2008 5:12:44 PM

Andrius said:
General tools for overclocking.
CoreTemp : http://www.alcpu.com/CoreTemp/
cpuid HW Monitor : http://www.cpuid.com/hwmonitor.php
cpuid CPU-Z : http://www.cpuid.com/cpuz.php

Prime95 : http://www.majorgeeks.com/download.php?det=4363
(make sure you enable "Advanced -> Round Off Checking")
OCCT : http://www.ocbase.com/perestroika_en/index.php?Download

MemTest86 : http://www.memtest86.com/memtest86-3.4a.iso.zip

As long as CoreTemp reports Prime95/OCCT load temperatures under 55°C-60°C and Prime95/OCCT doesn't report errors you are good.

Drop CoreTemp. It is inaccurate by up to +/-15C or higher. Get RealTemp.
June 14, 2008 9:30:24 PM

^While I agree with that, RealTemp only works for Intel "Core" CPUs.
Since he is using an AMD it is useless.
a b K Overclocking
June 14, 2008 11:47:10 PM

^lol, I should have noticed that OP was using a X2. That's what I get when I assume every one is using a C2D these days. :lol: 
June 14, 2008 11:48:11 PM

Shadow you are like Hard core Anti-Core Temp arent you?lol
June 15, 2008 10:46:49 AM

CoreTemp == simple and easy to use (like soccer mom minivan, practical and does many things good enough)

RealTemp == advanced, highly accurate, needs calibration (like an f1/indy racecar, razor sharp at what it does but useless when buying a large widescreen TV)

At the end of the day people with AMD chips don't have a choice and people with 45nm Core2 need to figure out RealTemp. :) 
June 15, 2008 2:18:16 PM

or they can always stuck with SpeedFan 4.34(i think it is).lol:p 
June 16, 2008 1:13:08 PM

Shadow703793 said:
Drop CoreTemp. It is inaccurate by up to +/-15C or higher. Get RealTemp.


I highly disagree with this post. Here is what the makers of Core Temp say.

"Core Temp lets you monitor Intel "Core Duo", "Core Solo" (Yonah), "Core 2 Duo", "Core 2 Extreme", "Core 2 Quad", " Pentium E2000" series, "Celeron 400\500" series (Allendale, Conroe, Merom, Kentsfield, Conroe-L respectively), "Xeon 3000/3200/5100/5300" series (Woodcrest, Clovertown respectively) and all AMD K8 (AMD64) and K10 (Phenom, Opteron) series die temperature.
The temperature readings are very accurate as the data is collected from a Digital Thermal Sensor (or DTS) which is located in each individual processing core, near the hottest part. This sensor is digital, which means it doesn't rely on an external circuit located on the motherboard to report temperature, its value is stored in a special register in the processor so any software can access and read it. This eliminates any inaccuracy that can be caused by external motherboard circuits and sensors and then different types of programs trying to read those sensors.

This is how the program works:

Intel defines a certain Tjunction temperature for the processor. In the case of Yonah it is 85C° or 100C°. First of all the program reads from a Model Specific Register (or MSR), and detects the Tjunction temperature. A different MSR contains the temperature data, this data is represented as Delta in C° between current temperature and Tjunction.

So the actual temperature is calculated like this 'Core Temp = Tjunction - Delta'

The size of the data field is 7 bits. This means a Delta of 0 - 127C° can be reported in theory. But from preliminary tests, the reported temperature doesn't go below 0C°, no matter what kind of cooling was used.

AMD chips report the temperature by a special register in the CPU's NB. Core Temp reads that register and uses a formula provided by AMD to calculate the current temperature.
The formula for the K8 is: 'Core Temp = Value - 49'.
The formula for the K10* is: 'CPU Temp** = Value / 8'.

The sensor in AMD CPUs can report temperatures between -49C and 206C.

*K10 = Phenom (Agena), Opteron (Barcelona). The K10 reports a temperature value that is relative to a certain predefined value, it doesn't report the actual processor temperature! So take that into consideration.
**CPU Temp is because the Phenom\Opteron (K10) have only one sensor per package, meaning there is only one reading per processor. "

If you have a K8 or K10 processor, CoreTemp is one of the best temp monitoring solutions out there. Super lightweight, and free.
June 16, 2008 1:34:44 PM

i can see Realtemp is dominating Coretemp. i think they author of Coretemp and fil a lawsuit just like AMD!lol
June 16, 2008 1:41:44 PM

iluvgillgill said:
i can see Realtemp is dominating Coretemp. i think they author of Coretemp and fil a lawsuit just like AMD!lol


Yes.... Realtemp dominates! But it only dominates on INTEL and ONLY IF YOU SET THE CORRECT SETTINGS! If you don't set the calibration right for the Tjunction, then it's worthless.
June 16, 2008 2:35:36 PM

Verillion said:
Yes.... Realtemp dominates! But it only dominates on INTEL and ONLY IF YOU SET THE CORRECT SETTINGS! If you don't set the calibration right for the Tjunction, then it's worthless.


yep thats right!

but i still like Core temp as it display the VID value and in a smaller and more attract window then Realtemp. Realtemp seems professional BUT too dull for my liking.
June 16, 2008 3:49:17 PM

realtemp is better for 45nm intel processors, where as 65nm is fine for coretemp, but i prefer realtemp.
June 16, 2008 4:10:19 PM

CoreTemp FTW...

But that aside, we're being off-topic here. :) 
June 16, 2008 9:39:16 PM

CoreTemp has 2 general issues:
1. Core 2 (Duo and Quad, 45nm/65nm) idle temperatures are either too high or too low (if you read the documentation of RealTemp you will understand why).
2. In some cases(some CPUs) there is no data for TjunctionMAX and since the DTS only reports the distance to Tjmax the value can be off by 10°C or more (the larger the distance the more inaccurate the value).

I like CoreTemp for it's simple interface and very nice layed out display. It is however just a very nice looking toy.
June 16, 2008 11:53:49 PM

but realtemp needs calibrate which is not convenient to noobs who are just got into OC. do realtemp give VID?i cant remember now.
June 17, 2008 10:35:49 AM

Yes. You get VID (it's was in the top right corner when switching displayed data) in the previous version. Now it's under settings.

Calibration is a downside of RealTemp (it's quite complicated and requires quite a bit of time). So is the non user friendly interface. It's a great program, but the UI needs a lot of work. The simple table look of HW Monitor is much easier to read, PROChot flags should not be done with checkboxes as they make users think they can select something. A dropdown menu would be an easy improvement. Colorcoding or temperature icons would be great for an easy readout. It's quite clear from a programmers point of view that RealTemp is not meant to be user friendly.
:) 
June 17, 2008 10:22:17 PM

i think the author should allow users to customise the name(tag) for each temp. CPU-1.40? whats that some may ask! let them change it to Vcore.better no?
June 17, 2008 11:32:36 PM

Customizing labels can be a pain. I've been doing some of that for translations on an Application Specific Software project (great acronym).

I've been writing a simple testproject with complete UI customization.
The default layout is hardcoded into the application but it let's you reposition every control via a layout XML. Labels are also customizable.

Now all it needs is something to do. :) 
June 18, 2008 10:40:40 AM

pity andrius!lol

i saw your pic being use by someone else.first i thought it was you!but when i look at the name is someone else!lol
June 18, 2008 12:00:56 PM

yeah. spud uses it. I've been here for over 7 years and I never came across it. I made a new kitty picture avatar but I couldn't be bothered to upload it yet.
a b K Overclocking
June 18, 2008 3:43:35 PM

@Verillion: I know how CoreTemp works, but the fact still stands that it is inaccurate. Even CompuTronix recommends RealTemp. (Although, I still prefer to calibrate SpeedFan according to his guide, but it's been deleted since it comes +/-1C as measured by an IR thermometer. )
June 18, 2008 8:10:00 PM

Wow. A lot information I got from here. Thanks guys. I'm using CoreTemp since its perfect for amd proc. Like Andrius said; simple and easy to use ;p

To Shadow703793 : What makes u think I will post an intel related problems in amd forum? :p 
June 18, 2008 8:13:00 PM

Oh yeah.. since I oc'ed my amd athlon x2 4400+ from 2.3 ghz to 2.8 ghz, will it run as fast as stock amd athlon 64 x2 5600 2.8 ghz?
June 18, 2008 8:16:02 PM

grizz86 said:
Oh yeah.. since I oc'ed my amd athlon x2 4400+ from 2.3 ghz to 2.8 ghz, will it run as fast as stock amd athlon 64 x2 5600 2.8 ghz?

*amd athlon 64 x2 5800+ 2.8ghz
June 18, 2008 11:40:57 PM

^The best way to figure this out is to run your favourite benchmark (if you can't see a real world benefit that is). I checked my Q6600 against a QX6850 in sisoft sandra (multimedia arithmetic benchmark, mips/flops).
I don't care much for benchmarks, it's more important the system is responsive and I can't say I notice a big difference between 2.4GHz and 3.0GHz during normal work.

The time saved when building a large VHDL project makes it worthwhile however.
June 18, 2008 11:46:58 PM

benchmark to tell the difference! but maybe L2 size different? now thats size matters!LOL
June 19, 2008 9:34:22 PM

Yes, like I said. I changed mine today.
I hope this time it's a bit more unique. :) 
June 19, 2008 11:30:04 PM

so what your new one looks like?
June 20, 2008 10:47:12 AM

It hasn't been approved yet I guess and for some reason I can't even upload a different/new picture. Maybe it was too offensive...
Oh well... it's not like it matters anyway. :) 
June 20, 2008 2:36:58 PM

ah!i can see it.Must go faster aye?!lol
June 21, 2008 2:25:00 PM

It only really makes sense if you imagine the kitty with a red cape and helmet. We all know red cars are the fastest. But it is very much the facial expression of the kitty that made me pick it. :D 


June 21, 2008 2:41:31 PM

i still like your old one.
June 22, 2008 9:40:29 AM

It was better yes. More fitting to the general irony of life and this forum. Everything you need to know in one place and a search function that works as well as a blindfold.

I was however beaten to it. :( 
June 22, 2008 1:15:43 PM

i dont think search function work well though.just cant find what i want when i need it.
June 22, 2008 4:20:18 PM

Cat avatars FTW...

Although I must say nothing beats the smile on randomizer's avatar. :D 
June 22, 2008 4:33:59 PM

eat the nigiri!!!FTW!:D 
June 23, 2008 12:32:50 AM

@Evilonigiri
True. Randomizer's new signature addition does come close to it IMO (the one about the hot girlfriend).

@iluvgillgill
Searching for stuff with a blindfold usually means finding other stuff with random bodyparts painfully. That is about the same result you get when using the "search forum" function. In my very few attempts I usually ended up with 0 results. Even when searching for stuff I knew I posted myself. The irony of this forum is that none of the really cool features work. :D 
June 23, 2008 12:34:03 AM

i cant even find the stuff i posted!><

just a question how many people in here OC with speedstep TURN ON??
June 23, 2008 1:22:44 AM

I stopped counting after the sticky thread was placed.
Usually there were 3 or so a week. That's what happens when all an admin cares about is cold hard ads-shown numbers. :) 
!