[hellcamino]

Originally Posted by Neuromancer

I use HWMonitor instead of coretemp/realtemp/everest anymore. Nice and small and you can get it the same place you grab CPUz Monitors temps AND voltages of mobo, CPU, PSU video cards and just temps on harddrive


For DAAMIT users

AMD Overdrive, and I see lots of sickos running k10stat.

Software voltage monitors are unreliable, hwmonitor is great for the temps it displays and a few of it's voltage readings as long as the volt readings aren't very important.

[Neuromancer]

Software voltage monitors have come a long ways. And everytime I see someone say they are bad, someone pulls out a DMM and measuers within a few hundreths of a volt what the SW is reading.

Software temperature monitors on the other hand. WILDLY unreliable. Especially with 45nm Intel and AMD chips. 2 different softwares will show readings 20C different. Granted not many people have the equipment or patience to accurately record the temperatures of their CPUs, so seeing someone comapre them to a SW reading... I have never seen it.


The idea of software monitoring, is not to stake your life on the results. But to give you a general idea of what is actually going on in your system.

For example, seeing your gcore voltage move from 1.05 to 1.15 when you load it, you have an indication that your card has moved to 3d mode.


I have more faith in SW voltage measurements than SW temperature monitoring software anyday of the week.

[DrNip]

EVGA's Eleet is pretty much spot on except for its VTT readings.

[punx223]

very nice chuch

[hellcamino]

[QUOTE=Neuromancer;6405]Software voltage monitors have come a long ways. And everytime I see someone say they are bad, someone pulls out a DMM and measuers within a few hundreths of a volt what the SW is reading.
QUOTE]

Pick one you like and I'll tell you whether or not my PSU even passes muster...So far they all tell me it's crap and needs to be replaced as the voltages are often wrong by 50% or greater. Oddly the DVM still says it's alright but then maybe it's biased...As a matter of fact I don't have any PSU's that pass muster except when I consult my meter.

[Neuromancer]

Originally Posted by Neuromancer

Software voltage monitors have come a long ways. And everytime I see someone say they are bad, someone pulls out a DMM and measuers within a few hundreths of a volt what the SW is reading.

Software temperature monitors on the other hand. WILDLY unreliable. Especially with 45nm Intel and AMD chips. 2 different softwares will show readings 20C different. Granted not many people have the equipment or patience to accurately record the temperatures of their CPUs, so seeing someone comapre them to a SW reading... I have never seen it.


The idea of software monitoring, is not to stake your life on the results. But to give you a general idea of what is actually going on in your system.

For example, seeing your gcore voltage move from 1.05 to 1.15 when you load it, you have an indication that your card has moved to 3d mode.


I have more faith in SW voltage measurements than SW temperature monitoring software anyday of the week.

Originally Posted by hellcamino

Pick one you like and I'll tell you whether or not my PSU even passes muster...So far they all tell me it's crap and needs to be replaced as the voltages are often wrong by 50% or greater. Oddly the DVM still says it's alright but then maybe it's biased...As a matter of fact I don't have any PSU's that pass muster except when I consult my meter.

For everyone elses sake I will quote myself since you seemed vexed on one point of my post, rather than the entirety. Make your own assumptions there.

Unfortunately I tried very hard not to get annoyed by your comments (coming from a great OCer such as yourself, compared to a n00b like me) and spent a long time writing a civillized response. Only to loss power. ARGH!!!!


Damn why havent I gotten UPS by now!!!

Anywho...

Before I disagree with you further, I do want to say one thing.. your OC prowess is amazing. If I need advice on OCing or extreme cooling I will definitely assume you know what you are talking about I say what follows only to keep people from perpetuating fallacies and because, well, I am right.


You said pick one.

I pick vcore. You really actually measuring your 1.3 sw set vcore at .65 or 2volts??? 50% offset...

Now I assume what you really where talking about was PSU rail sw tools. Which I agree, can be VERY wrong. Those are not what I was talking about but lets entertain that thought... So by 50% your motherboard reads.. 12v and your dmm reads 6 or 18 volts? thats a 50% difference. Or you mean that your 12v SW reads 11.8 and your HW reads 11.9? (technically that is less than a 1% difference. pretty freaking sweet)

But lets compare that to temperature readings....

My software tells me my 550BE CPU idles at -256C and loads at 0C. Since it is actually running at 30C and 45C.. thats more than 1300% off.

Okay lets talk average which is what I was talking about anyway...

There are many people that quote coretemp at 25C only to find out they are really running at 35C with those damn e8XXX cpus, (only a 40% difference, which 10 degrees is pretty standard on adding to your SW temp readings actually), oops that would mean SW temp readings are not accurate :p) Sorry my house is running at 24C .. what fucking air cooler is so good it will take an OCed CPU and run it at ambient?

So your 50%? If real idle temp was 37.5.. you hit 50% right there. A difference of 2 degrees. (since people with temp probes are much more lacking than dmm types.. you gotta now that Sw temp readers are SO far off the mark.. its a marketing scheme (oh... not like Intel and nVidia don't play those games nudge nudge say no more)

Okay one last thing..

How do you set your voltages when you overclock?

You use the software (ahem... firmware) that comes with your mobo? Or do you have some machine that clamps down on it and you use dials and switches to set vcore, vdimm etc?

nuff said.

(PS that asus pod... does not count as HW )


EDIT: yeah I am not good at arguing, because my mind is faster than my fingers so if I left too many incomplete thoughts I will try and expound them later.

EDIT EDIT: For that matter.. HW monitor (except on vgpu) lists sensors as specs. So PEBKAC comes into play???

[Deux]

Eleet + Nhancer seem to be the only big ones missing from the list.

There is a nifty little program written by Tiborr (IIRC) called Prioaif that is really helpful for AM3 and wPrime. It is posted in the first 2K1 tweaking thread over at XS. I'll try to dig it up later today.

OCX Spi Tweaker can be helpful for 32M especially. It requires some serious trial and error to get the most out of it but in the end it's worth it.

[hellcamino]

Neuromancer I have absolutely nothing to prove to you! When I said pick one it was intended that you pick a voltage monitor specifically for monitoring psu voltage and I would test my own with it and then give you an actual reading from a multimeter unless you prefer that I use my ossciloscope...

If you are basing your observations off of a single utility and/or a single motherboard or cpu then you have some catching up to do. AMD doesn't even use a core sensor so if you are basing your reading off of that then you are setting yourself up for failure.

Math games? Why would I actually mess around like that? Seriously...what exactly would be the point?

I have seen my 12v rail as low as 2.2v recently, my 5v at 9v and my 3.3v at 1.9v all according to software! The reason they are inaccurate is due to the way they sense voltage...err don't. All of the voltage readings that you see in monitors are simply relayed from the sm bus via bios and so they reflect what is set in bios. When you see changes in windows from what you set in the bios it is due to fluctuations in system load.



Now I say again: pick a software utility for me to test PSU voltage that is accurate. When I say accurate I mean the exact same reading I get with a meter in real time.

As far as temp monitors go I'm honestly not sure, I've been told that an i7 will throttle at exactly 100c but haven't made the experiment (2Chesapeakes did though), I was also told that 100c was threshold temp for the majority of LGA775 socket quad cores and that they would shut down when they reached this temp. The highest I ever saw on my Q6600 was IIRC 119c and on my X3360 xeon I saw it at 123c, the xeon shut down immediately afterward but the Q6600 lasted a few more seconds as I was trying to kill the load and crashed in the proccess. Well I would think that the temp sensors for these cpu's would be tighter than that so I am guessing that those temp readings were higher than actual. The utility in use was core temp and it reads lower than Real temp, hwmonitor or Everest. That doesn't render them entirely useless though as the reported temps can still be used as a guide.

[Kal-EL]

Both of you can remove the f-bombs during intelligent discussions and err on the side of understanding during these intelligent discussions. You might give people the wrong idea or something.

thx- Supes

[Neuromancer]

Originally Posted by Kal-EL

Both of you can remove the f-bombs during intelligent discussions and err on the side of understanding during these intelligent discussions. You might give people the wrong idea or something.

thx- Supes

Sorry Supes,

Edited for content, inflammatory comments and spelling


HC.

Post whatever you like. I see you are still on about PSU rails. Which is fine, I specifically said HWmonitor in my first post and that it gives you an idea about voltage on your video card.

How about OCCT? Run the bench test it gives you a nice graphical display of your 12v rail.

Also out of curiosity, do your rails read from SW) that far off at stock? If not, how long do they stay "accurate" before instability kicks in? Extreme OCes or just any OC at all?


AS for my experience with PCs. No it is not based on a single system. I have 5 PCs running in my house now and I "rent" new ones every few months. (I call it renting, flip HW so often I never really "own it" lol) Benn doing this for the last 5 years, and my experience with PCs goes back to learning how to program on a TRS-80 around 1979


And if "none" of your PSUs will past muster on your board based on SW measurements. Well that could be an issue with your board then, not SW voltage monitoring as a whole. It could also be an indicator that your mobo has been damaged in some way. If for instance when it was brand new it read everything fine)