Jump to content
soundx98

The ULTIMATE MBM 5 Settings/Help Thread

Recommended Posts

There's always going to be a pretty significant discrepency between the two sensors - I always go by the low one.

 

Pretty sure Dan is saying that if there is a large difference between the two core temp readings, he goes by the lower one (of the two core temp readings).

 

But so you are aware, different versions of the program will reverse the readings.

So while an earlier version of Core Temp (or Everest, or any program that measures the core temps) may read that Core 1 is higher than Core 2, a later version will show the opposite core reading higher.

 

I can't remember ever seeing Core Temp tracking both cores equally across a siginificant temp increase/decrease. (nothin' wrong with the program, all it does is do the math for us :) )

 

SmartGuardian etc. gets it's reading from a Diode and CoreTemp etc. is reading the DTS Thermal Sensor, therefore different readings. As Soundx98 has told us many times.
Danke George :D

 

So it really doesn't matter which you use, as long as you are comparing apples to apples and not apples to oranges (diode to diode or core temp to core temp, NOT diode to core temp). Verstehen?

 

Diode calibration can be off by as much as 10 degrees (I can't believe that DTS readings don't have their own set of foibles as well).

So a comment like, "my CPU runs 8 degrees cooler on xyz board compared to my abc board" don't do us much good either.

 

It's necessary to compare like values when comparing temps.

 

georgetakei-copy.jpg

Share this post


Link to post
Share on other sites

Thanks for the post Soundx98, all of which makes sense to me.

 

Not sure if its coming across clear or not, but if I have a bunch of apples all reporting Xdeg, and a bunch or oranges reporting their own Ydeg, which of the two would be more accurate? I guess I'd have to get a temp probe stuck in there somewhere to get a more accurate idea of what the actual temp is...I'm just trying to figure out which (diodes or DTS) is better to go by. Like I said, MBM gives me ~40deg while coretemp gives ~51. Can I trust that the 40 is more reliable and I can therefore o/c that much more, or is the ~51 a warning sign to find a better cooling system?

 

I used to go by MBM for everything, but with a 10+ deg difference I want to be able to stop worrying that I'm on the verge of heat damage.

 

Thanks again for all the input guys!!

Share this post


Link to post
Share on other sites

Funny thing about oranges and apples ... neither is wrong, an apple is an apple and an orange is an orange. So the word "accurate" is maybe not that meaningful. Both are probably "accurate."

 

On this form most people have traditionally gone by the sensor that MBM reads, which is the same one that SG reads I think and the same one the BIOS reads. So it will always be an ongoing discussion. It seems each seup is somewhat different.

 

I was expecting to see a bigger difference between SG and Core Temp (I don't have MBM installed yet) on my new Ultra-D/165 setup because of discussions like this. At [email protected] I am 31C in SG and high 20's in Core Temp. When I try to push it with an OC and load temps reach over 50 in SG, Core Temp is within 2 degrees +/- of SG's temps.

 

I have the same board as you with the same CPU and you have a 10 degree difference ... go figure. But there can be a big difference in steppings and who can predict the effects of individual heat sinks.

 

Maybe an average between the two sensors would be meaningful.

Share this post


Link to post
Share on other sites

I can tell you that while running Prime95/Orthos on many chips, I generally find that on my rigs (lot of variables here remember) MBM temps are generally 45-50C and Core Temp readings are between 50-55C.

I DON'T THINK YOUR TEMPS WILL EVER GET HOTTER THAN WHEN RUNNING THESE PROGRAMS.

 

If they go higher than that I back off the Vcore.

Of course both the latest Everest and NextSensor (see links in 1st post in thread) can report both values. MBM5 can as well on later boards and C2D processors, but the MBM5 date won't work on the NF4s.

The various programs do "sample" or "poll" the sensors at different time intervals but if you would rather see instantaneous changes MBM5 can be set to a shorter interval.

 

I use both Everest and NextSensor frequently as well, but MBM5 uses slightly less resources than most other programs and I like the look and feel of it.

(I like NEC drives, you may prefer Plextor)

 

Vataliy Jungle (the CBID and NextSensor dude) has an interesting program called Coretemeter that can track the Core Temps and Super I/O chip readings.

 

Coretemeter is a small and handy tool designed to construct the graph of die temperature of the AMD64 processors using readout of on-die CPU thermal sensor. The utility supports drawing the core temperature for each core of a dual-core processor. There is an option to set a rate of reading the temperature. You can select 1, 2 or 6 times per second for the rate. Besides, Coretemeter calculates an average value of the die temperature for each time period of monitoring. Also, you can define a priority of reading the temperature. During stress load on the CPU this allocates more time to the utility to construct the graph. Besides, both for single and dual core AMD64 processors Coretemeter constructs a temperature graph using readout of external hardware monitor also known as Super I/O chip. This logical addition transforms Coretemeter into the temperature comparator. Now you can examine how far the reading of hardware monitor differ from the readings of processor’s on-die thermal diode. The utility is very simple to use.

http://cbid.amdclub.ru/files/corterm.zip

 

He also has a new version of his NextSensor program out (July 11, 2007) that I'll be checking out when I get home this weekend (works on A64 only)

http://cbid.amdclub.ru/files/nxsensor.zip

(select DFI NF4 under custom, or NF590 if using that AM2 board)

 

 

While there is certainly fun and bragging rights to getting your CPU to 3GHz, I can verify that you will no difference in real world performance running at 2.7GHz.

As Angry_Games and the others have said, If you are a gamer go ahead and get yourself a kickass Video Card.

 

There is no "right" or "wrong" or "better" program for temp measurement.

I like MBM5 because it labels sensors correctly, is very easy to customize, allows easy comparison to most other users, and can log values, be used as a screensaver, works with my G15 KB, can be used with Samaurize, etc, etc, etc. (Plus, this old dawg is pretty familiar with it by now.)

 

I originally started this thread on March 6th, 2005. We had been using something similar on the DFI NF2's in 2003 when A_G was moderating the DFI forum at amdmb/pcper. So 2 and half to 4 years of geeks using MBM5 for comparisons and this thread itself has had almost a quarter of a million (229,491 as of this post). So there is a lot of opportunities to compare values with other DFI boards.

 

Just realized I've gone over 7,500 posts here at DIY-Street.

In the words of the immortal James Brown "Good God" :D

 

DFIerror.bmp

 

http://premium1.uploadit.org/soundx98//Whoop-Dee.jpg

Share this post


Link to post
Share on other sites
DFIerror.bmp

 

That is just plain funny. Many thanks for all the work on MBM and temp sensoring in general. MBM still has the perrrdiest display of them all and I'm all about cute.:)

Share this post


Link to post
Share on other sites

Looks like I'll just have to pick one and go with it...

 

Nextsensor is pretty neat:

Processor (28/37deg, diode).

Coretemp A&B (37/48deg, 31/45deg, thermistor)

Looks like all the same info in a different format. Nice though, I'm sure it will float around my rig for a while. Can't really see a way to relabel readouts though...

 

Now that issue's been resovled, I tried adding in the coretemp plugin into my MBM setup, following instructions on this page, But get a "Detected processor does not support DTS reporting. Plugin failed to initialize" error. Ever since, my chipset reads -9.51v, and DIMMv is -.031. Any idea on this? I do not have a custom voltage ini, only happened after MBM update3. I tried going back to the MBM update1, but no change. Webtemp and Nextsensor report the correct voltages though.

 

Congrats on the 7.5k mark, and the most-excellent thread!

bios_guests_brown.gif

Share this post


Link to post
Share on other sites

MBM5 Update 3 does NOT work with the DFI config file.

You'll need to uninstall MBM5, delete the registry entries (under software/Alex Van Kaam, I think) and then start all over with MBM5.

Shouldn't take very long man (I did it too)

Share this post


Link to post
Share on other sites

I hoped it wouldn't be like that soundx, ah well - shoulda asked first :-p

 

syngensmyth - looks like all the apples and oranges are talkin a different language...if apples said 45deg, and oranges said 30deg, which would be better to base an o/c off of if you're keeping an eye on temps?

Share this post


Link to post
Share on other sites
I can tell you that while running Prime95/Orthos on many chips, I generally find that on my rigs (lot of variables here remember) MBM temps are generally 45-50C and Core Temp readings are between 50-55C.

I DON'T THINK YOUR TEMPS WILL EVER GET HOTTER THAN WHEN RUNNING THESE PROGRAMS.

 

If they go higher than that I back off the Vcore.

 

I think following Soundx98's advice here is good.

 

In reality it is somewhat academic because as temps go over 50 (MBM style) stability goes out the window anyway. At least that is what has happened to me. Then you need more vcore which causes more heat:eek2::eek::sad:

 

You can easily watch them both while OC'ing.

Share this post


Link to post
Share on other sites

what i hagve to do to get the ATI temp monitoring in MBM5 after i'v uninstall my nvidia card?

 

i update from 6600Gt to X1950XT,

Share this post


Link to post
Share on other sites

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.

Guest
Reply to this topic...

×   Pasted as rich text.   Paste as plain text instead

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.


×
×
  • Create New...