Gigabyte EasyTune

[mav777]

I have a gigabyte p55 udp4 board and easy tune wont run under xp, 64 bit.
Come up with GUI.exe with error in normal.dll
Not worried about overclocking, especially as the system is still bedding in. I need easy tune to slow down the fans running at full speed.
System is an i7 860 and runs temps in low 30s, so not hot, just noisy from full speed fans.
Speedfan is a great app, however won't slow the fans down and need easy tune.

Help!

P.S.
Wish the easy tune app / board was half as good as my old abit board for fan control.

 

[r_manic]

You've tried reinstalling the software, right?

 

[mav777]

Used the latest from their site.

Didn't work, uninstalled and installed an older version on the disk that came with the mainboard.
Same problem.

Gigabyte are taking ages. 2 Days just to ask me if i;m using a legitimate version of windows!

In a few days i'm expecting they will ask me to turn the computer off and restart!

 

[jsc]

You should be able to set smart fan profiles in the BIOS. Look in "PC Health".

 

[mav777]

JSC,
thanks for the reply.
The PC Health section has the facility to turn on / off the variable speed - which doesn't seem to work and only runs at full speed.
Also, there is the ability to alter via PWM etc
No profiles. Profiled would be great.
Gigabyte has always been poor with fan support.
They are finally back to me and suggest I flash the BIOS with a beta version.
I've provided them with good technical information about the issue, yet they provide this?
Tried the beta bios and it says "BIOS ID Check Error"
My guess is they have sent me a BIOS not compatible with my board.
Hopeless!
3 Fans at full speed. Nice and cool but loooooouuuuuuudddd!

 

[bilbat]

Assuming it's a 55-UD4P, F7 seems to be the latest BIOS (F7a beta appears to have been withdrawn, for some reason...), available here:
http://www.gigabyte-usa.com/Support/Motherboard/BIOS_DownloadFile.aspx?FileType=BIOS&FileID=15436
If you got a checksum error, it may just be a corrupted download; consider yourself lucky - I FUBAR'd the same board TWICE 'cause the first time, I refused to believe a corrupted file could have possibly PASSED checksum! Had to do the 'dualBIOS' recovery twice, restore CMOS (which, thank Murphy, I had saved to disk...), etc!

There are two things going on here, one of which 'belongs' to GB, one to MS... On the GB end, you need to use the version (but not necessarily the revision) that was shipped with your board. People get ver 3, and later think that by going to pro, or 4, they are 'upgrading'; what happens is that EasyTune has to be 'built' to access code 'stubs' in the BIOS that 'hook in' to the LPCIO (which, in your case, is a pretty-much standard [for GB] iTE 8720 - it does the fan control, sensing, most of the temp reads, and a dozen other things). It's difficult enough (and sometimes simply not successful) to keep these code interfaces 'lined up' when BIOS revisions are done; either the BIOS must remain 'aligned' at each rev (probably not an easy thing to do if the revisions get major), or, alternatively, a new rev of the same version of EasyTune has to be upgraded to match the new 'stubs' - complicated by the fact that each version 'services' a large number of boards, and all the new platforms (1156/1366) are 'works in progress', and are getting BIOS revs up the ying-yang! And, of course, all of this must be documented, which, we all know - just never gets done! If we are lucky, we get a two word description in the BIOS' changelog - but usually, NOT!

On the MS end, Xp 64 has always been sort of Microsoft's 'red-haired bastard stepson' - there's never been anything approaching even a tolerable level of support for it. The reason has a lot to do with architecture; Xp 64 was MS' first try at a 64-bit framework, which meant it was everyone else's, too. It it's heart, windoze is just an API - an application programmer's interface. When I wanna do some file access, I really don't want to know whose disk control chip you have, or whether it's IDE or SATA, or how your drive is organized in cylinders and sectors - I just pass a message to windoze telling it my needs, and it hands me back a file handle, or a diagnostic telling me why it can't. This is all 'concealed' by the HAL - the hardware abstraction layer - which is comprised of the windoze 'core', drivers, and 'ring zero' programs. Ring zero is MS' nomenclature for the place programs have to run to get 'priveleged access' directly to the hardware - and, pretty much forever, their overwhelming advice has been: don't do it! There are all kinds of restrictions on how hardware is to be handled, how exceptions have to be 'thrown', and on and on... When they did Xp 64, again, being their first foray into 64 bit, the HAL and ring zero were, well, inscrutable, and massively confusing. Lots and lots of people who were used to 'following the rules' for drivers' structure, and ring zero access found the new paradigm nearly impossible to deal with - thus, very little support - the miniscule size installed base just didn't pay off for the effort involved!

Now, MS learned a crapperload from this, too - one of the first things that was taken care of at the inception of Vista (which was designed to be an 'either or' OS - they wanted to 'capture' the high-performance market that was headed to 64 bit, with or without them...) was to restructure the drivers and ring zero so that they were more 'tractable', and to do a really thorough job of the DDK/SDKs (driver development kit/software development kit) so everyone had a well-delineated 'direction' to go in...

I'm pretty picky about what I 'allow into' ring zero on my production system... I recently wanted a program for 'ripping' DVDs to work on some hardware accelerated transcoding, and a buddy recommended SlySoft's AnyDVD - but, when it installed, it 'dumped a piece of itself' into ring zero - and that was the end of it! Uninstall, followed by a restore - I just don't allow it! There are a number of popular things that do this - lots of CD/DVD stuff, .iso image 'mouters', all verboten on a machine I depend on daily! There are advantages to ring zero - for instance, AutoCAD uses ring zero components to accellerate the hell out of its video access - and I'm willing to risk that they know what they're doing, but in general - nope... I often recommend EasyTune as a 'quick & dirty' way to test individual parameters for an overclock (or MemSet for memory tweaking), but I would never trust them for 'day to day' operation

 

[mav777]

Bilbat,

Pretty awesome post.

I can tell you one thing for certain. Gigabyte support is a joke.

It's a clown with a big red noze and oversized shoes.

I've sent them plenty of info, attachments, error dumps the works. They reply with, try a different fan this time!

Perhaps the dodgy BIOS not working was a blessing in disguise.

To slow the fans i'll buy a couple of the in line power adaptors that slow them down.

My preferred method would have been to link the cpu temp with fan speed.

I might build a new os and re-install the software. Perhaps theres a conflict i don't know about.

My xp 64 bit build is from a previous ep45 motherboard. A clean install would have been better, but we all know how much hassle it is to re-install the works!

Mav

 

[ekoostik]

bilbat,
How do you know when you install something and it drops a piece of code into ring zero?

mav77,
My Gigabyte GA-P55M-UD2 does variable CPU fan speeds so I would assume your UD4P allows it. Are you having problems with that, or is it just the case fans that are causing problems?

 

[mav777]

eko,

The mainboard should run the fans at variable speed, however only runs at full speed.
I've used the easy tune software before and was able to slow them down using this.
The fans are noctua and therefore very good and fairly quite, however they should not be running at full speed all the time.
No idea on how to slow them down. Adjusting in speedfan does squat as well. At least speedfan runs!

Mav

 

[bilbat]

The BIOS will (at least, on every one I've had experience with) control the fan on the CPU_FAN header, and you can cheat, and get all your fans buzzed (well, at least five or six) by using this splitting scheme:

Isn't mine, I stole the idea from Arctic, and they have a line of PWM fans that implement it - their website says (and I've verified by looking at the spec) that any PWM header should have at the very least, enough 'oomph' to drive five fans (I get six using the fan's max and the PWM spec's min...); there are PWM splitters/extenders at NewEgg that are pretty close, just need a couple wires truncated... The SYS_FAN2 header on most boards is also controlled (in theory, linked to the MCH, or 'system' temp sense), and in this case, it's designed to operate with either a three-pin standard, or a four-pin PWM fan - all they do is 'lock on' the on-fan PWM switches by providing a steady +5V on the 'command'pin (4), and then physically PWM the 12V main supply...

I have wanted, for a long time, to build a muti-fan, multi-temp-input controller based on something often implemented in industrial analog control. It's called PID (proportional, integral, derivative) loop control: it runs a control equation that allows variable 'weighting' of three factors - the proportional term works off the 'distance to target'. Say I'm heating a vessel with a variable-venturi gas valve; the farther the contents of the tank are from their 'set-point' temp, the more 'oomph' I goose the gas valve with. The integral term we often call 'nudge'; what it does is integrate (sum) the variation from set-point over time, the idea being that most processes will 'come to equilibrium' some distance from the set-point; by summing the small error that persists, we can 'nudge' the controlled variable (again, say the gas valve), to 'pull it in'... The third term is a little more complex - here we take the derivative of the error signal, i.e., we look at its 'rate-of-change' versus its distance from the set-point; the idea here is that when we're way way off, we want to give 'er - valve wide open; but, if we keep that up, when we do get to the set-point, we'll overshoot it a bunch, and wind up 'clamping it' closed - the temp will drop until we're considerably low, and we'll wind up oscillating around the set-point, in ever-smaller 'swings'; if we measure the rate of change and factor it for distance from set-point, we can avoid that oscillation - typically, a well-controlled process 'overshoots' once, by not a lot, and then settles out...

I bought an ESA fan controller board because I listened to nVidia's 'babble' about it, assuming that I could 'roll my own'; however, nVidia documents nothing, will not 'hand it over' and, unless you get lucky, and find the 'sectret software', won't even allow it to operate on systems without nVidia chipsets!

I can tell you one thing for certain. Gigabyte support is a joke.

...tell us about it! I have a pretty good sized collection of looney quotes sent to me, or posted, by various folks pursuing various support issues through them - some of them are simply hard to believe... I've been trying, for some time now, to do a semi-coherent job of helping out here, and I've often jokingly (or, maybe not so jokingly) suggested that they'd do better by firing their tech support, saving the cost of one dinghy PR campaign, and sending me a pallet of MOBOs, and a small monthly budget for ancillary parts - I'll beg samples myself from the memory manufacturers; not likely to be, overall, worse than what they've got now!

Another thing that I've railed about to absolutely no avail is 'model proliferation' - last time I counted, I think they (GB) had one-hundred twenty-five plus part numbers in socket 775 MOBOs alone! That means if their most expensive board is, say, four-hundred USD, and the cheapest is, again a guess, $40 - they have an entry into the marketplace every two dollars and eighty-eight cents! Is that kind of market penetration, and segmentation, really advantageous?? If you can't provide decent, detailed support for a dozen boards, how in the hell are you going to attempt it for three-hundred?