Viktor the issue is quality control. Major lack of. Lot of makers hack up enough that the windows machines appear to run right. Even that for the chipsets they should have longer battery life. This is why at times you get two laptops from two different makers with identical chips on the motherboard and some-cases identical motherboard and identical batteries that have different battery life.

biosbits.org I am point two is the start of this lack of good quality checking for motherboards being turned around.

Viktor is one thing you really cannot blame Microsoft completely for. Yes the defective ACPI test suite Microsoft made, Microsoft has to take responsibility for. Yes a test suite that if you passed Windows would randomly lock up and was not to ACPI spec either. Since then its basically been learnt don’t ask Microsoft to make a hardware test suite they don’t know enough. So some of the defective hardware does track to Microsoft but they did not just break Linux they broke there own OS as well. If it was conspiracy of Microsoft to destroy Linux they are highly incompetent. Linux flatly refused to boot so did everything else using the reference implementation.

Really that bad test suite from microsoft was the best thing for Linux since after that a lot of hardware makers joined the kernel.org and started running Linux as part of regular testing.

There are still things like turning on and off backlights in acpi where there is not good automated tested. To test properly you need monitoring hardware to make sure that everything in the hardware side is acting as per software directions. Shocking part is even today that does not fully exist for you x86 machines.

So on quality control we have a long way to go. Viktor sticking head in sand does not help the problem.

Items like this https://wiki.ubuntu.com/OHW/VgaSwitch
Not even Microsoft has. So there really does need to be a focus getting the hardware for OS makers made so they can check there hardware interactions and that motherboard makers can check that everything is operating correctly.

Yes the problem is lets build and pray that it works.

Yes open standard hardware would be a good idea so everyone can use the same hardware for confirming everything is working right.

biosbits.org also comes out of giving hardware makers what they want. They don’t want to have to repair a laptop or desktop after its left. So all tests have to be run in the prototype stage or not at all to get change in quality for a lot of makers.

I’m disappointed. As a long-time reader you should know by now that the evil, evil industry in concert with the evil, evil Redmond empire has rigged ACPI in order to make Linux look bad.

In short: it’s a conspiracy!

I don’t use “7″ so I have nothing to say about that. I don’t use Ubuntu on notebooks, either, so I have nothing to say about that. I do use Debian GNU/Linux on notebooks and the previous owner of one of them noticing that the battery still had lots of guts after six years, was amazed. One cannot get better battery-life than “adequate”. I use notebooks at home, not on the bus, so battery-life is not really an issue. The machines all save/restore to disk perfectly too.

One can reduce power-consumption on most systems by reducing clock-speed and so on. I have no need of that because all my systems have modest clock-speeds from the start. I have no use of ultra-high clock-speeds resulting in more wasted power/cycles. Even Beast which is insanely powerful for a personal computer only runs at 2.5gHz. All the others run from 400MHz to 2.2gHz. They are all snappy because I match the hardware to the tasks. Here, hydro-electric power is 6 cents/KWH and we heat with electricity. PC-usage is just part of the heating system in winter. Even in the summer, we use no more energy than if we had a few teenagers living at home.

Of course, caching is wonderful if you need it and a pain for stuff you don’t need. Why lash the performance of the PC to “average” usage of files instead of instantaneous usage of files? If I want caching, I will use a GNU/Linux terminal server and let the server run holding all the cached files for the whole organization which won’t need to cache anything when a user boots.

e.g. In an organization of 100 users, That other OS will seek 100X files instead of none in the case of GNU/Linux terminal servers. That’s a huge savings in wear, noise, heat, capital equipment and energy consumption. It blows people’s minds when I show them 8 year old PCs beating that other OS on new machines with a single hard drive seeking itself to death at every boot.

Is this true for all Linux Distrobutions. I think not.

Also here is the funny one that test was against desktop ubuntu not ubuntu configured for laptops. There are options you can enable that speed up windows but ruin its power usage as well.

The real answer is no where near as black and white.

Also “NVIDIA IOn graphics” this also shows problem Nvidia has been known for not providing power management options under Linux. ATI machine more interesting question.

There is a big problem with limited sample testing. Its not a valid way to make any assessment on the general status at all Phenom.

World is shades of grey. Some will run Linux better than windows some will run Windows better than Linux.

You are not providing enough data to make a ruling either way Phenom.

“Starting with Vista, Windows keeps statistics of which applications users run most, and caches”

Ubuntu preload uses statistics also can be pretold to load particular things as well at user control.

Both systems have a preload system.

Interesting enough http://behdad.org/preload.pdf Windows copied Linux here. The formulas Windows Vista and 7 are using for preload are in this document by the author of the Linux preload system.

Reason MS made theirs after the Linux one already existed.

Linux preload you can choose to exclude particular applications. This is because windows vista and 7 are a poor copy in the preload system.

Since you don’t know history of where these techs come from you did not have the right response Phenom. What is basically bow you head and walk away because windows implementation is poor copy of the Linux preload system.

Later generations of Linux preload addressed many issues the first document contain that MS copied. MS has not updated theirs from the master document so its buggy.

Key change was not preloading stuff for userspace usage while system is still starting up. Split the statistics. Start up services split from user-space applications with two statistic numbers. So after to login screen then start preloading user stuff if you are going to preload any.

kozmcrae is basically refering to the bug in MS preload implementation that the first Linux preload had but is now fixed. Linux users are use to this fixed so are annoyed by windows not being fixed.

Phenom reality MS does a really poor job coping at times.

Some of these things you should be on MS back about. Like boot up statistics and user application statistics being split for preload so system does not stall badly on start because startup services are being loaded after pre-loading some user application that cannot be run yet because you cannot log in to run it.

Starting with Vista, Windows keeps statistics of which applications users run most, and caches (“pre-loads”) these. As a result, users get better performance for their mundane tasks for the costs of a small delay for rare tasks.

Sounds pretty reasonable to me. After all, nowadays the hard disk is the performance bottleneck, and the cheap RAM makes a great cache when unused.

You wrote quite a rant, and nothing you say can change the simple fact that Windows 7 has better battery life than Ubuntu.

He can write for himself, but one of the things that really annoys me is that other OS “preloading” stuff a user may never use delaying “useful desktop”. I may intend to run certain applications but that other OS certainly would not know which, even from my history. On one day I may want to write a bunch of WordPress articles and on another to add a bunch of software to a bunch of PCs. A quite different list of applications are involved and that other OS cannot read my mind. Pre-loading from a hard drive really sucks juice. A typical hard drive uses a few watts idle and more than 10W active. That other OS certainly likes to swap on real machines I have seen in schools. The malware sucks resources too, just by having all the accessed files sniffed.

Eh, like this?
http://www.myasuseee.com/windows-7-beats-ubuntu-10-04-on-battery-life/

“No one cares. People just want them to work according to their expectations.”

Expect to purchase/download extra security software to make it safe for the Internet.

Expect it to slow down with age.

Expect it to be useless every patch Tuesday while it updates and reboots/re-reboots.

Expect the battery life to plummet because the CPU is busy doing a whole bunch of stuff and you’re not sure what exactly it’s doing.

Expect your shiny new computer to slowly become a useless piece of crap as Windows gets cluttered up with day to day normal use.

Microsoft has conditioned an entire generation of computer users about just what to expect from their computer. What a shame.