Did the Netbook Improve Windows 7's Performance? 440
Arnie87 writes "One Microsoft Way has an interesting article suggesting that the reason Microsoft is focusing so much on speed with Windows 7 is the whopping sales of netbooks. The article concludes by saying: 'If you plan on adopting Windows 7, you have the netbook to be thankful for, because Vista's successor would be a very different beast if Microsoft had less motivation to pursue performance.'"
End justifies the means (Score:3, Interesting)
It doesn't matter if I have a netbook or not, if this is true, then everyone benefits. Even the guy with a multimedia powerhouse machine will see an improvement if performance is the bottom line.
Microsoft's fascination with taking advantage of new hardware and technologies has led to a consistent decrease in performance over the years, with Vista perhaps being the most obvious and poorly received example. The tide seems to be turning, though. Symantec pulled all the stops on making the newest releases dramatically lower in memory & faster, everyone's re-writing pages so they scale properly for mobile devices, now Microsoft is paying attention too?
This is a good trend. I hope it continues.
Re:Bloat (Score:4, Interesting)
The thing about Linux is that nearly every netbook maker is developing a custom Linux distro that removes the cruft and makes it run faster.
Re:There is some bad news too (Score:5, Interesting)
10% of 14 million is still a sizable market share. There are countries all over the world that are smaller than that number that speak their own unique language. The netbook might not be most people's primary machine, but 1.4 million people who are now OK with using linux that would have blindly bought a windows PC before is a giant leap in terms of consumer penetration. 10% penetration is a number Apple's been clawing after for years .
Short answer - no (Score:5, Interesting)
Short answer -- No.
First of all, the obvious: Microsoft started working on Windows 7 late in 2006, even before Vista was released. Netbooks became popular in 2008. 2007 worldwide sales of Netbook-type machines were less than half a million.
Any self-respecting computer programmer knows what's really going on. When you spend months or years working on a major new release, you're often struggling to get the new stuff working at all. Your managers are pushing you to get the thing out the door; deadlines are looming; adding more people to the team would probably be counterproductive since they'd only slow down the people who need to be 100% focused on finishing things up.
Once you get that x.0 release out the door, you take a vacation, reintroduce yourself to your wife and kids, putter around at work for a while, and then dive back in and make your code faster, cleaner, more reliable, more useful. The x.1 release that follows ends up being the one everyone likes; people say "It's what x.0 should have been!" ... Right? That's what happens!
And that's exactly what's happening with Windows 7. This isn't a major "reinvent the wheel" release... it's all about optimization, performance, better user interfaces, and tacking on some new things that have become popular since Vista was released, like proper support for SSD drives, multi-touch, multi-core GPUs, and so on...
Who says it performs better anyway? (Score:2, Interesting)
Who says it performs better anyway? Arstechnica gives no information on what tests they ran. Windows 7 is really just Vista SP3, so I'm a bit sketical.
Re:End justifies the means (Score:4, Interesting)
Microsoft's fascination with taking advantage of new hardware and technologies has led to a consistent decrease in performance over the years, with Vista perhaps being the most obvious and poorly received example.
Oh, please. A "fascination with taking advantage of new hardware and technologies" is not why Windows has sucked on an an epic level. Windows has always been an "also ran" when it comes to adapting "new hardware and technologies". Always.
Hardware support? Even the abysmal Mac OS 9 had more leading-edge hardware support than W9x and W2K on their respective releases. Windows XP and 2k3 can, and have been, a huge pain to install if you've got SATA and/or necessary USB devices on the system. Sure, decent support is available after you're installed - but that's not due to Microsoft.
Emerging technologies? Can you name one software/OS/desktop feature which MS was first-to-market on for Windows? I seem to remember something called Cairo that was making news back in the mid-90s, which had a feature list similar to what we now know as Time Machine - on OS X. MS still hasn't come up with such a functionality. Hell, they don't even have simple search indexing working well in Vista, yet.
No, MS has been behind the curve with implementation - and well ahead of it with outright lies and broken promises ("Vista Ready", anyone?)
Historically, these are the things a new version of Windows has been certain to bring to the table:
* Slower performance
* Bigger memory footprint with little related advantage (see "slower performance").
* The first release/pre-SP will be buggy, unstable, and nearly unusable.
* A lot of stuff that's supposed to work, won't. This includes applications which are supposedly designed for said OS.
* If it's a complete lemon, they'll silently drop actual support and focus their efforts on their next release (See: ME -> 2k, Vista -> W7).
Yes, there are various other improvements to new Windows releases. But, consider: Windows still can not approximately estimate the time it will take to copy a file from one local directory to another. That's hardly a focus on new technologies.
Though, I absolutely agree with you on the whole low-end focus in the IT industry being a good deal for everyone. Now, if only we could get away from the "the browser is the OS" idea, as we're running into all sorts of the same bloat and instability we got with OSes, as browser developers re-implement containers and other OS-level features at a highly abstracted level.
Re:Why is this bad news? (Score:3, Interesting)
Well, it's bad news for a reason he didn't quite mention. The bad news is that it's likely, in part, due to how pokey Linux has gotten on the desktop in the last 5 years.
We need a serious initiative within open source to push for a feature-lock for a year or two (or even 6 months), and focus on improving the ability of OSS to run within small constraints. Granted, a lot of this is happening currently (see: Firefox), but I think a more concerted effort needs to be taken.
Re:Microsoft performance and security ! (Score:3, Interesting)
Quick Answer: "No"
I simply meant since XP was released, or more specifically, since the internet became really popular, and .Net was released, there are now hordes of craptastic applications out there.
Vista is pretty much irrelevant, although with Vista, they introduced (to the average windows user) things like Widgets, so now people are a little more familiar with running stupid little shit all the time, so maybe Microsoft realized that when people run all their craptastic software, they blame the OS, rather than the software, so they are trying to minimize that blame by partially taking responsibility for various peoples poorly coded software, and allowing for even more of it to run, or for the "normal" amount now, to run better.
And I'm not necessarily even saying that's what I think, I was only adding another possibility to the GP's comment, although that I'm sure that it does play a part, however major or minor it may be, in addition to just the general desire for a slimmer/faster/smoother OS.
Duh (Score:2, Interesting)
Intel giveth, Microsoft taketh away.
The thing that breaks this paradigm is an Intel platform that moves backwards in net performance. When the goal shifts from ever increasing net performance to performance per power it's only expected that Microsoft should miss the turn.
The question is, how did they miss being informed that the turn was coming? Did they get told and disbelieve, or were they just not told? I believe the former, not the latter.
Interesting parallel in the mac world (Score:3, Interesting)
This has a few (opposite) parallels over in the Mac world.
When the PPC platform stagnated, Mac OS releases started doing a strange thing.... they actually tended to be faster than the previous release on old hardware.
I've got a 450MHz G4 in its (mostly) original hardware configuration currently running 10.3.9. Unless I'm doing video encoding, or something else similarly processor-intensive, it certainly doesn't feel like a 10 year old machine. (The video encoding example is an interesting one, given that I used the machine 2 years ago for a large video-editing project with Final Cut Pro, and simply farmed out the rendering and encoding tasks to a more powerful machine -- FCP has remarkably modest hardware requirements)
This is all on a computer that shipped 2 years before the release of OS X. (As a random sidenote, I've also always been impressed that it could handle up to 2 GiB of RAM. That was unprecedented for its time)
Once Apple switched to Intel chips, new releases started to become progressively slower. Leopard would be an embarrassment if it weren't for the fact that Vista was even a bigger embarrassment.
IMO, the PowerPC's limitations actually drove a lot of innovation at Apple during those few years.
Comment removed (Score:5, Interesting)
Re:Or maybe you're pulling that from your ass (Score:2, Interesting)
There are lot of problems with portable applications which try to write into the directory where .exe file is installed.
Problems? Are you installing this app in a directory where the user actually has the write-access they need?
If you need to put the EXE in a subfolder under %ProgramFiles%, then just modify the ACLs on the folder to let users write to it. This has all worked fine since at least NT3.x, and continues to work fine in Vista and Win7.
Oh, you're probably talking about when the user is logged in as an admin (i.e. they have all the rights they need) but has a non-elevated token. You're right, that is stupid. But you know what? Expecting your users to be logged in with admin rights is pretty stupid too.
In summary, Vista may suck, but your app sucked first.
But IS Windows 7 faster? (Score:3, Interesting)
I really want to like Windows 7.
On one hand, I hate Microsoft as much as the next guy.
On the other hand, I use their software everyday at work. And if the market leader massively improves their (somewhat crappy) software, it forces the competitors to get better too.
For instance, IE8 seems much faster and better than IE7 (and of course IE6). This will hopefully wake up Mozilla and force them to improve on Firefox.
Regarding Windows 7, I can see that the memory footprint is lower, and that's a good thing.
=====> But it still _feels_ much slower than XP in everyday use! =====
I am talking about the little things that make up the experience of responsiveness. It just takes a noticeable amount of milliseconds more when I click on an icon, until the OS reacts. Opening a new browser window just have that extra lag. Copying files feels slower. Etc.
At first, I sort of liked Windows 7 and ran it for a couple of weeks. Then I booted back into XP (not a fresh installation of XP, mind you). I was depressed by how much snappier XP feels. I was hoping to have a good reason to ditch XP.
Makers of desktop operating systems should focus intensively on responsiveness. The OS should react as fast as possible on any user request, regardless of whatever else it is doing.
It's fair enough that some heavy calculation takes longer time if you have some other heavy job or service running, but the initial latency from any user request until you get some sort of reaction should be as low as possible. And XP is much better in this regard than Windows 7 or Vista (and also faster than all Linux distros I've been running).
To use an analogy from network land: I would much rather have 10 ms ping times and 1 Mbps than 1000 ms ping times and 100 Mbps.
Re:But IS Windows 7 faster? (Score:3, Interesting)
This is a modest Athlon 64 X2 2200 Mhz with 1GB of RAM. I realize that modern systems have more power than this, especially more RAM, but on the other hand it's not like Windows 7 is accessing the page file all the time. In fact, when I have done this kind of test, there's no hard disk activity at all, and plenty of free physical memory.
Also, we're talking about netbooks here which are even slower than this system. And while XP shines on this system, Windows 7 and Vista does not.
My XP does not get slower with time, as you suggest. But then, I'm fairly conscious about what software I install and which services are running etc.
Also, remember that I am comparing a newly installed Windows 7 with an "old" XP install, so even if it _is_ slower than a newly installed XP, the newly installed Windows 7 is still even slower, much more so in fact.
Linux Netbooks at IT Show, Singapore, Mar 2009 (Score:2, Interesting)
These are the netbook manufacturers I've collected brochures from:
These were advertised prices. Actual prices were about S$50-150 lower.
None of the vendors advertised the availability of GNU Linux. Asus is the only one that advertised the GNU Linux option. They also put up Eee PCs and Eee Boxes with GNU Linux on display, hooking up to big LCD screens for passer-bys to see.
This happened at the IT Show event, 12th-15th March 2009, Singapore.
So, no, I don't think your "nearly every netbook maker is developing a custom Linux distro that removes the cruft and makes it run faster" statement appears to be true, for the moment.
Re:thank the netbook? (Score:3, Interesting)
Now, however, I still do a lot of work on machines with roughly similar specs to the machine I was using in 2001. Every computer I've owned was fast enough that some tasks would not benefit from a faster machine. The only programs I regularly run that are CPU-limited enough for me to notice are gcc and pdflatex. Video playback is slow on some older machines, but modern systems offload this to a GPU or DSP.
That's not to say I haven't bought new computers, but since around 2003 I've found portability much more important than raw speed. Speed is nice, but it's no longer the driving factor in my purchases because 90% of the time there is no perceptable difference between a 1GHz Celeron M and a 2.16GHz Core 2 Duo (I use both on a daily basis).
I tried running Windows 3.11 on my Pentium, and it was much faster than NT 4, but I didn't run it very often because NT gave me features that were worth the overhead. Even if NT 4 had used 100% of the CPU and RAM of the machine that ran 3.11, I would not have minded because it would have been under 10% of my new computer's power.
This is no longer the case. People are going from 2GHz desktops to 2GHz laptops rather than to quad 3GHz desktops and they expect the next version of their OS to run well on the new computer.
Re:Or maybe you're pulling that from your ass (Score:4, Interesting)
You're wrong. The only thing that matters at all with any product anywhere is wrong with Vista; my customers don't want it and won't buy it. Therefore, it is a terrible product that I don't carry and have no interest in.
Now, with Windows 7 my customers have been showing some cautious optimism, so we'll see how it goes.
Re:Or maybe you're pulling that from your ass (Score:3, Interesting)
While I agree with most of your points, I have to take issue with no. 7:
7) This is linux and not windows. Bash is much more powerful than the pitiful shell windows provides.
It's rather newish (2006) but IMO PowerShell generally blows bash and all other Unix shells out of the water. Arguably, PowerShell is much better for Windows, as more APIs in Windows are object-oriented and thus fit better with PowerShell.
Take a look at my sig. It's a one-line, slashdot sig fitting (OP has a point: If you are handling files with spaces in them, *many* scripts will break down due to the fact that *nix shells pipes are text-only and that many tools by default parse using whitespace as delimiters. In a shell with object-oriented or structured pipes this will not happen.
Sorry for chiming in, but it is kinda my pet topic at the moment.
Re:Interesting parallel in the mac world (Score:3, Interesting)
Once Apple switched to Intel chips, new releases started to become progressively slower. Leopard would be an embarrassment if it weren't for the fact that Vista was even a bigger embarrassment.
What are you talking about? If you've got an Intel-based Mac, Leopard is actually faster. The kernel handles SMP much, much better, and many of the things like Spotlight received serious optimization -- try using a Tiger-based Mac and a Leopard-based one side-by-side under load and you'll see a difference.