OS2World OLD-STATIC-BACKUP Forum

WebSite Information => Article Discussions => Topic started by: Andi710 on 2012.06.24, 10:08:08

Title: ACPI version 3.20.03 released
Post by: Andi710 on 2012.06.24, 10:08:08
http://www.os2world.com/content/view/21818/2/ (http://www.os2world.com/content/view/21818/2/)
QuotePosted by Martin Iturbide - Saturday, 23 June 2012


I have jus. .. .

Runs well here on a Phenom II X6. However, with SNAP video enabling 4 cores is still the limit. (6 cores only work with plain VGA, SNAP hangs at WPS boot-up)

And the reactor temperature widget for XCenter does not show anything except of "Zzz".
Does anybody have a working version that can show system/CPU temperature as measured using ACPI?

Cheers
Andreas
Title: Re: ACPI version 3.20.03 released
Post by: melf on 2012.06.24, 11:33:22
Hi Andy,
this version has much better power reduction on my system (Thinkpad T43-2668). Previous version didn't really work, while this version keep my temperature at about 55C (which is good). However resume results in a system freeze, so I have reverted back to 3.19.14 (where I got working suspend-resume) combined with DougB:s coolman, which does the power reduction work as well on this Thinkpad. I can't notice no effect of the idle time HLT, however I don't know what to look for.

When I started up my system after install of acpi 3.20.03 the config-file (in MPTN/ETC) was blank and the GUI consequently didn't work. However the temp was around 55C even then, which make me wonder if this in fact was an effect of the idle time HLT? Enabling the power management (copy conf template on acpiD.cfg and enabling) in fact didn't seem to add anything.
Title: Re: ACPI version 3.20.03 released
Post by: Blonde Guy on 2012.06.24, 18:05:31
Melf, that T43 would run cooler and suspend/resume would work better with APM than with ACPI.

My T60p runs cooler with 3.20.03 than it did with 3.20.02.

My X60 may run a bit cooler with 3.20.03, but the hangs on desktop start. These hangs it has always had on boot-up are worse with 3.20.03. I never had a hang with 3.20.02, which is great because it's the only ACPI thus far that didn't hang. Hangs are particularly bad for X60 because it then does a high temperature shutdown. For X60, it would be great if ACPI would run the machine cooler during the processing of config.sys.

I have heard that hangs on desktop start go back many years on many machines and that ACPI may only be involved by changing the timing.

The reactor widget works with both machines.
Title: Re: ACPI version 3.20.03 released
Post by: DougB on 2012.06.24, 18:31:27
QuoteDoes anybody have a working version that can show system/CPU temperature as measured using ACPI?

That works, for me, by RMB on the widget, select properties (NOT the eCenter properties), then select motherboard = Asus P5B. I suspect that that may be required on many AMD systems.

Quotethis version has much better power reduction on my system (Thinkpad T43-2668)

I haven't got around to trying it on my T43 (1871-W8M) yet, but it does much better on my Asus M3A78-EM, with quad core Phenom processor. It is also better on my Lenovo ThinkPad T510, which, contrary to instructions, I run with hyperthreading enabled (2 cores look like 4). I have had NO TROUBLE with that. I don't think it does much for eCS, but it sure helps Win7, and I don't have to mess around in the BIOS to change it.

Quotecombined with DougB:s coolman, which does the power reduction work as well on this Thinkpad

My COOLMAN program uses the SETGETSTATE program to slow the clock, which seems to do more than throttling on single processor machines, where it will work. From what I can tell, the new power manager does not use the clock slowing mechanism, and throttling doesn't work well on every machine either. My Asus M3A78-EM motherboard has a problem where it won't reboot if it is left throttled while a reboot is attempted (usually after a crash). Sometimes, it needs to be rest by "pulling the plug" for a couple of minutes, to get it to go again. Setting the min throttle settings to 100% (all of them) seems to work around that problem, and the new ACPI still drops the temperature, significantly (I assume that it is HLT that does part of that, but it is mostly turning off unneeded processors that makes the big difference on SMP machines).

Melf. You may want to play with COOLMAN to disable the throttling section, and just let it use the SETGETSTATE part, then use it with the new ACPI. I will try that on my T43, when I find the time (probably this morning).
Title: Re: ACPI version 3.20.03 released
Post by: DougB on 2012.06.24, 18:38:12
QuoteMelf, that T43 would run cooler and suspend/resume would work better with APM than with ACPI.

I don't know about Melf's machine, but I do know that my T43 (1871-W8M) won't work, at all, with APM (won't even power off). The support code is just not in the BIOS.
Title: Re: ACPI version 3.20.03 released
Post by: melf on 2012.06.24, 19:48:57
Doug: Sounds like a good idea to run Coolman, without throttling together with the new ACPI, ought to reduce power use even more. The bad side for me is that I use suspend/resume quite often, when pausing work, so 3.19.14 is the best option for me - so far.

Blond Guy: I used APM on my (2669) and I think this machine (2668) should work with APM as well. When I used APM latest it worked OK, but adding and enabling the "throttle-package" made it freeze of some reason. Also I get a better distribution of irq:s with ACPI - although I don't know how important that is.
Title: Re: ACPI version 3.20.03 released
Post by: DougB on 2012.06.24, 20:46:42
QuoteDoug: Sounds like a good idea to run Coolman, without throttling together with the new ACPI, ought to reduce power use even more.

It does. I don't remember which version I sent you (I made a lot of changes after that, trying to get my Lenovo ThinkPad T510 to play nicely). If it has the option to set the minimum throttle, just set that to 100, and it won't play with throttling (except when it first starts up).

FWIW, on my T43 (1871-W8M), with no power control, it runs at about 69C. With ACPI 3.20.03, and throttling enabled, it runs near 54C. If I slow the clock to level 3 (using the SETGETSTATE command), it runs at about 44C. Of course, battery life is extended too.

I think that the basic S/R is working with my T43, but it is difficult to tell, because a resume never turns on the screen backlight, and that makes it difficult to tell if it really came back to life, or not. The T510 appears to do the suspend okay, but I am pretty sure that resume never completes. It also does not turn on the screen backlight.
Title: Re: ACPI version 3.20.03 released
Post by: StefanZ on 2012.06.27, 19:33:27
Wow!

Power-off works now!!!  :D
Title: Re: ACPI version 3.20.03 released
Post by: Sigurd on 2012.06.30, 00:05:37
Works really great with my fanless minipc! Power off works, and HLT as well.

See picture one, application pm123 just started , all four kernel (2 kernel + 2 hyperthreading - the readme suggest to disable hyperthreading but I am still testing :-) )  working


See picture two - a few seconds later three kernel are halted as one kernel is enough by far :-)

I started then open office and the other three kernel are back in business - really great!

And the fanless pc is still not hot :-)


Some additional information for the mini pc:
- still problems with IRQ, wlan minipcie crashes...
- Super Dolby surround Sound with Uniaud
- still problems with USB, drivers are not able to switch off USB legacy mode
- but still a lot of fun :-)
Title: Re: ACPI version 3.20.03 released
Post by: rwklein on 2012.06.30, 18:36:45
Quote from: Sigurd on 2012.06.30, 00:05:37

See picture one, application pm123 just started , all four kernel (2 kernel + 2 hyperthreading - the readme suggest to disable hyperthreading but I am still testing :-) )  working


Some additional information for the mini pc:
- still problems with IRQ, wlan minipcie crashes...
- Super Dolby surround Sound with Uniaud
- still problems with USB, drivers are not able to switch off USB legacy mode
- but still a lot of fun :-)

About hyperthreading. There are bugs open in the eComStation bugtracker where a system would trap during startup with hyperthreading switched on.

About the MINI PCI stuff. I don't know if you have had time to open a bug for this. Or if this crash is caused by loading an unsupported genmac driver (don't know the details).

Roderick Klein
Mensys
Title: Re: ACPI version 3.20.03 released
Post by: rwklein on 2012.06.30, 18:41:44
Quote from: Andi710 on 2012.06.24, 10:08:08
http://www.os2world.com/content/view/21818/2/ (http://www.os2world.com/content/view/21818/2/)
QuotePosted by Martin Iturbide - Saturday, 23 June 2012


I have jus. .. .

Runs well here on a Phenom II X6. However, with SNAP video enabling 4 cores is still the limit. (6 cores only work with plain VGA, SNAP hangs at WPS boot-up)

And the reactor temperature widget for XCenter does not show anything except of "Zzz".
Does anybody have a working version that can show system/CPU temperature as measured using ACPI?

Cheers
Andreas

So what happens if you run with Panorama do the 6 cores work ? If that works then we know SNAP has an SMP defect!
We have somebody working on SNAP at the moment. Based on your report I opened a ticket:

http://svn.ecomstation.nl/snap/ticket/21

Regards,

Roderick Klein
Mensys
Title: Re: ACPI version 3.20.03 released
Post by: rwklein on 2012.06.30, 19:09:27
While suspend resume we don't support officially its still being worked and moving forward.
More and more systems are starting to work and 3.20.03 version has a new Intel ACPI code in the PSD that has fixes for suspend resume for none Windows systems (if I understood David his explanation correctly).

While I see that Melf his system does not suspend resume correctly anymore with the current version the suspend/resume is a very complex procedure. But as we do have (as far as it now seems) a good working HLT support that seems to be working on a lot of systems!

The SMP kernel never supported this. We have in the test lab also other systems see other work much better with suspend resume then ever before. While not ready for production usage never have so many systems worked with suspend resume.

To give you an idea of how systems are working now. System that never would even go to sleep now suspend and then resume and you can access the system via the kernel debugger (via the serial cable).  But things go wrong in restoring the video buffer.
Thats why with testing suspend/resume try the /VBE switch.

As time will move forward more progress is being made that will only see the success rate of suspend/resume go up.
We will never have the coverage that Windows has. But that is not that weird in it self.  Suspend resume from my point of view is screwed standard. With that I mean that a standard OEM install of a Windows 7 CD will also not work on all systems.
As most hardware vendors preload drivers for there own equipment.

So right now suspend resume is officialy not supported as more stuff is being debugged. But progress is being made.

Hope the new ACPI versions will help keep your systems cool and reduce your electricity bill :-)

Roderick Klein
Mensys
Title: Re: ACPI version 3.20.03 released
Post by: Andi710 on 2012.07.05, 23:48:42
Quote from: rwklein on 2012.06.30, 18:41:44
Quote from: Andi710 on 2012.06.24, 10:08:08
Runs well here on a Phenom II X6. However, with SNAP video enabling 4 cores is still the limit. (6 cores only work with plain VGA, SNAP hangs at WPS boot-up)


So what happens if you run with Panorama do the 6 cores work ? If that works then we know SNAP has an SMP defect!
We have somebody working on SNAP at the moment. Based on your report I opened a ticket:

http://svn.ecomstation.nl/snap/ticket/21

Roderick Klein
Mensys

Indeed, the 6 cores run fine with plain VGA. I did not try panorama though. However, I need SNAP to power my two LCDs with a large WPS...
Also, video performance with multiple cores enabled is not great.

Will comment on the SNAP ticket
Cheers
Andreas
Title: Re: ACPI version 3.20.03 released
Post by: Blonde Guy on 2012.07.06, 05:51:17
Quote from: melf on 2012.06.24, 19:48:57
Doug: Sounds like a good idea to run Coolman, without throttling together with the new ACPI, ought to reduce power use even more. The bad side for me is that I use suspend/resume quite often, when pausing work, so 3.19.14 is the best option for me - so far.

Blond Guy: I used APM on my (2669) and I think this machine (2668) should work with APM as well. When I used APM latest it worked OK, but adding and enabling the "throttle-package" made it freeze of some reason. Also I get a better distribution of irq:s with ACPI - although I don't know how important that is.

Doug's machine isn't supported by APM -- he should use ACPI. The T43 2xxx is supported by APM, and that's what makes those machines special. But at this point ACPI is so good that it could be a possibility, too. You can't really use throttling with non-ACPI systems.
Title: Re: ACPI version 3.20.03 released
Post by: Andi710 on 2012.07.08, 12:20:09
Quote from: Andi710 on 2012.06.24, 10:08:08
And the reactor temperature widget for XCenter does not show anything except of "Zzz".
Does anybody have a working version that can show system/CPU temperature as measured using ACPI?

Roderick: Should I open a ticket requesting support for my board's temperature sensors through ACPI?
Title: Re: ACPI version 3.20.03 released
Post by: DougB on 2012.07.08, 17:49:18
QuoteRoderick: Should I open a ticket requesting support for my board's temperature sensors through ACPI?

I am not Roderick, but... Did you try changing the widget motherboard setting from "ACPI standard" to "Asus P5B"? That works for my Asus M3A78-EM motherboard.

If that doesn't work, then open a ticket, but I am not sure if you should open it against ACPI, or the widget.
Title: Re: ACPI version 3.20.03 released
Post by: DougB on 2012.07.08, 19:09:50
QuoteThe T43 2xxx is supported by APM, and that's what makes those machines special

I don't know about "special". It does seem that the T43 2xxx series does support ACPI. Possibly APM too (I don't know for sure). You can use the ACPI utilities SETGETSTATE and SETGETTHRTL to determine if the machine supports those functions (you do need ACPI loaded to do the tests). If it does, you can use them to reduce power consumption, which also reduces heat output, and extends battery life.

My old Coolman prototype used both of those, when they would work, but REXX and external programs, are really too limited to be really effective. I found that SETGETSTATE did more to reduce power consumption than SETGETTHRTL does. The new ACPI power management appears to use throttling, but not the STATE settings. I am not sure exactly what throttling does, but the STATE settings reduce the machine clock speed, which lets the hardware work with less power consumption, but it is slower. The new ACPI has added a new twist, that I never thought of (probably because I was working with a single core machine), and that is to turn off all but one core of a SMP machine. Now, they have also added the use of the HLT command, to further reduce power consumption. Of course, with all of these methods, there needs to be a way to speed it up when the power is needed, and slow it again when the demand is not there. ACPI, using throttling, core reduction, and HLT, seems to work great on SMP machines, but it is far less effective on single core machines. You can effectively measure the effectiveness of these techniques by monitoring the system temperature. For example, my Asus M3A78-EM, with quad core Phenom processor, runs at about 59C (138F), with no power control. With throttling, it runs at about 55C (131F). When it uses throttling, and core reduction, it runs at about 42C (107F). Add HLT , and it runs at about 38C (100F). While the machine claims to have two STATE settings, SETGETSTATE cannot change them, for some reason (although, I haven't tried that with the new ACPI).

My IBM ThinkPad T43 (1871-W8M - single core), runs at about 72C (161F) with no power control. With throttling only, it runs at about 62C (143F). Add HLT, and it runs at about 58C (136F). Of course, core reduction is not possible with a single core machine, so that is what ACPI can do for it. Now, if I use SETGETSTATE to set it to the slowest clock speed, it runs at about 42C (107F). The battery life is extended with each step too, and, of course, the T43 doesn't cook your legs, if you use it on your lap, until you run a program that speeds it up for more than a few seconds.   :)

My Lenovo ThinkPad T510 (dual core, running with hyperthreading enabled - contrary to suggestions in the ACPI package), runs at about 62C (144F) with no power control. Throttling knocks it down to about 52C (125F), core reduction goes to about 48C (118F), and HLT goes to about 43C (109F). SETGETSTATE does not work. Note that I use hyperthreading because it makes Win7 work MUCH better. It doesn't seem to do much (good, or bad) for eCS, so I just leave it enabled.

Of course, suspend/resume is not usable with any of them. On the other hand, I have an antique IBM ThinkPad A22e that has full APM support, and it does suspend/resume, and hybernation (full power off, with very quick recovery), with no problem (the BIOS actually does it, with a little support from the OS). It does not do power reduction though. It has no support for ACPI, and it will cook your legs if you use it on your lap.
Title: Re: ACPI version 3.20.03 released
Post by: Blonde Guy on 2012.07.12, 05:11:11
Well, the T43 2xxx is like your T22, with full support for everything. IBM certified it. Sure ACPI probably also works, but not everything works as well. For example, with SpeedStep, the T43 2xxx runs about 40 C.

We need ACPI to run newer machines, and for some older machines. I'm just saying that there are a lot of T43 2xxx machines out there because they were the last machine IBM certified for OS/2.
Title: Re: ACPI version 3.20.03 released
Post by: eirik on 2012.07.12, 15:20:49
I fully agree with Blonde Guy that the emphasis should be ongetting ACPI support to newer models -- after all that is where the market is going to be.

My old T43 ran excellent without ACPI - suspend/resume worked wonders, wired and WiFi access, in brief the works.

On my 2 years old T410 I run ACPI ver. 3.20.03.  Boot time is so fast that the lack of suspend/resume does not feel like a big problem.  Fixing this issue may however also be useful in terms of getting the FnFx buttons to work as they should.  My reactor temperature is 62 degrees when I run on the net, and between 52 and 58 degrees when I run on battery (I've activated throttling).  On the hot side, but no real danger.  Throttling also extends battery time substantially, so it beats my Win7 (64).

Otherwise, the system runs wonderfully -- wired access is excellent with the E1000E driver from the MultiMac project.  The most pressing  issues as I see it are: WiFi (I tried with the help of others to get the Intel 4995 under GenMac mimic the Intel 6300 a/g/n/ but no success), the FnFX buttons (so I can use my eCS partition for presentations), and Java integration into FireFox (or any other useful browser).
Title: Re: ACPI version 3.20.03 released
Post by: diver on 2012.07.12, 18:40:44
@Eirik the Java part is being worked on. Should not really take very long till GA3 is out.
Title: Re: ACPI version 3.20.03 released
Post by: melf on 2012.07.23, 17:23:19
Found 3.20.04 available. Now both suspend and resume works again on my T43!