I've been running Windows Vista Ultimate on my Acer Aspire One netbook (with 1.5GB of RAM, and a 30GB hard drive) for a while now, without any problems or performance issues. I have the full Aero Glass experience, and I didn't need to do any performance tweaking or fiddling with services. I even made a few very crappy videos to show it all off. Apparently, Steven Sinofsky thinks Vista - and therefore, Windows 7 - can run just fine on a netbook too, and that's why he demonstrated Windows 7 running on a netbook this morning during the Windows 7 keynote. In an interview with Ars he gave a little more details.
Post a Comment
Member since:
2005-07-13
"If you're buying a netbook, avoid the SSD models. They might be cheaper, but you are going to get extremely frustrated with the SSD."
If you can show me a laptop with an SSD HDD in it that is CHEAPER than a good old fashiond moving parts drive, then i want you to help me pick out my next notebook because i have yet to see that option make one cheaper 
.
now if you are reffereing to notebooks that ONLY offer that and are cheap or intended for developing countries, then i understand.
Member since:2006-02-05
Member since:2005-11-11
the faster SSDs in the 900 and 901 are not bad at all, at least with Xandros on them. Most MLC SSDs pretty much suck though.
I've also got Windows installed on an SDHC card to use with my 900*. Using EWF**, no changes are made to the SD card (until/unless I choose to commit all changes, they are stored in RAM) so any write penalties are thrown out the window. This could of course be done on the SSD too.
I think it's pretty neat. Don't like how a change has affected things? Reboot without commit. Satisfied with the current state? Commit and write it all out at once instead of in many speed sapping random writes.
I wish I could do this in Linux (maybe someone knows how). There's tmpfs of course (though AFAIK they start off emtpy, so it'd be hard to mount / as one 
and I don't know of any easy way to commit changes). Maybe one can do a unionfs with the fs where changes are recorded being a tmpfs? Again though, how to commit any changes if I decide they are beneficial?
ANYWAY flash based works fine for me with Linux and Windows. None of the stuttering people can see with slower SSDs or no tweaks to prevent random writes all over the place. Out of 4 laptops in the house I've had to order 3 new hard drives in the last year and a half, so I rather appreciate not having a mechanical drive that can get banged around.
*For Starcraft, which can't quite run at full speed in wine on the 900
**Enhanced Write Filter: http://support.microsoft.com/kb/814257
***EDIT*** beaten by Thom
Edited 2008-10-29 08:56 UTC
Member since:2008-10-23
Starcraft is the top of the platinum list in wine app db:
http://appdb.winehq.org/
It means that it runs flawlessly. Is there really a performance hit on your 900? There shouldn't. On my desktop at least, it runs just as fast or faster than on Windows.
Anyway, puppy linux does just what you describe. By default, it does commit your changes every half an hour to the pup_save file. You can have several pup_save files and use them to create several branches where you commit some different changes.
Edited 2008-10-29 09:58 UTC
Member since:2005-11-11
Yes, there is a performance hit. Each frame displayed has to be translated from the 8 bits (256 colors) it runs at to whatever bit depth your x server is using (16 or 24 most likely - unfortunately 8 bit doesn't work).
This is done on the GPU where possible, using shaders. Otherwise it's done on the CPU. This is slow. My old Athlon 1400 couldn't do it fast enough until I applied the DirectDrawRenderer opengl registry key which let it use the GPU. Unfortunately the eee's integrated intel decelerator can't do it, and the CPU can't do it fast enough.
That platinum rating means everything works, and it does (aside from some battlenet menu issues, which the platinum rater overlooked). Most modern computers will have either a sufficient GPU or CPU to get Starcraft running in wine with no slowdown, but the eee has neither. It's close but ~70% doesn't do it (it runs with no errors, just slow). Perhaps wine will implement a DIB engine someday and then it will work at full speed even on the eee. I'll keep checking.
***
Thanks for the info on Puppy Linux. That's close to what I want, but it is sort of missing the essential ingredient of not writing to the SSD at all unless I want. Unless the pup_save file can be stored on a tmpfs and the changes written to / when desired? I guess that could do it. I figured it was possible (someone must have done something like that) so thanks for the heads up. Next I'll want to know if it can be done with Debian
Edited 2008-10-29 12:53 UTC
Member since:2008-10-23
Does the performance hit apply if you use a 8bit X server? If not, why not spawn a new X server with a 8bit layout for Starcraft?
Edit: You can tell puppylinux not to save to the pup_file at all, to save only at shutdown, to save periodically or directly. Bu default, it saves periodically, but just change the PUPMODE environment variable and you have what you want.
It is certainly possible on debian, but you have to do all the work they have done on puppy yourself I'm afraid. It involves using unionfs, tmpfs and squashfs.
Edited 2008-10-29 13:22 UTC
Member since:2008-04-28
I've been running Starcraft through wine on a Pentium 3 Compaq laptop with a junk video card and 768mb ram for months now. That's without tweaking anything either. I've played some LAN matches with my room mate but not battle.net. I'd like a desk instead of my couch if I were to play competitively.
Are netbooks really THAT slow that a P3 can out perform it?
Member since:2005-07-06
. I don't know...the Acer Inspire One mentioned in the article perhaps? The SSD version is quite a bit cheaper than the mechanical HDD version. Sure the SSD version has 8GB of storage as opposed to 120 or 160 GB, but it is cheaper.
Member since:
2005-10-12
"If you're buying a netbook, avoid the SSD models. They might be cheaper, but you are going to get extremely frustrated with the SSD."
Can't comment on the SSD, but I'm really happy I went for the HDD. Currently multibooting WinXP, Linpus, and the latest development versions of Ubuntu and Fedora. My only complaint is that the Jmicron multicard reader drivers don't work with the latest Linux kernels.
Member since:
2007-03-20
I've had my Aspire One for a quite a while now and the amount of tweaks required to get decent performance out if it are quite high. These devices are still new and various operating systems weren't really designed with SSD in mind, the performance is awful until you tweak it.
This is one thing Microsoft should be taking advantage with Windows 7, trying to optimise everything as much as possible for SSD. Telling users to avoid SSD netbooks isn't exactly the best strategy. Eventually the cheaper netbooks will probably end up with 1-2GB of RAM and enough disk space to take the newer versions of Windows.
Member since:
2007-02-17
FTA: "We do a lot of really customer focused things, like we have a gigabyte and a half of printer drivers."
Wow. 1.5 GB, just for printer drivers.
http://www.linuxfromscratch.org/blfs/view/svn/pst/cups.html
"Download size: 3.6 MB
Estimated disk space required: 62 MB "
Do we have an outrageous discrepancy here, or what?
Use a filesystem for the SSD that caches small writes.
http://lwn.net/Articles/275706/
http://notechie.com/linux-2627-management-ubifs-for-netbook/
http://twinturbo.org/linux/linux-kernel-2627-officially-released/
* write-back support - this dramatically improves the throughput of the file system comparing to JFFS2, which is write-through;
* fast mount time - unlike JFFS2, UBIFS does not have to scan whole media when mounting, it takes milliseconds for UBIFS to mount the media, and this does not depend on flash size; however, UBI initialization time depends on flash size and has to be taken into account (see here for more details);
* tolerance to unclean reboots - UBIFS is a journaling file system and it tolerates sudden crashes and unclean reboots; UBIFS just replays the journal and recovers from the unclean reboot; mount time is a little bit slower in this case, because of the need to replay the journal, but UBIFS does not need to scan whole media, so it anyway takes fractions of a second;
* fast I/O - even with write-back disabled (e.g., if UBIFS is mounted with -o sync mount flag) UBIFS shows good performance which is close to JFFS2 performance; bear in mind, it is extremely difficult to compete with JFFS2 in synchronous I/O, because JFFS2 does not maintain indexing data structures on flash, so it does not have the maintenance overhead, while UBIFS does have it; this is because of the way UBIFS commits the journal - it does not move the data physically from one place to another but instead, it just adds corresponding information to the file system index and picks different eraseblocks for the new journal (i.e., UBIFS has sort of âwanderingâ journal);
* on-the-flight compression - the data is stored in compressed form on the flash media, which makes it possible to put considerably more data to the flash than if the data was not compressed; this is very similar to what JFFS2 has; UBIFS also allows to switch the compression on/off on per-inode basis, which is very flexible; for example, one may switch the compression off by default and enable it only for certain files which are supposed to compress well; or one may switch compression on by default but disable it for supposedly uncompressible data like multimedia files; at the moment UBIFS supports only Zlib and LZO compressors and it is not difficult to add more"
Write-back support is apparently what you want.
This means that if you have a netbook with SSD flash only, you don't have to make it vulnerable by installing XP Home, you don't have to buy a hard disk drive to install Vista on it, and you don't have to wait for Windows 7 ...
... just get a Linux distribution that includes the 2.6.27 kernel, and use the ubifs filesystem for the root partition on the SSD.
The only problem will be getting grub or some other bootloader to load it ... sigh!
Edited 2008-10-29 03:26 UTC
Member since:2006-09-22
Member since:2007-02-17
If you just cut the power on any OS, data in transit to open files will be lost. ubifs is no different.
You would lose the data being written in "small writes" to an SSD under the current method in any event.
On a normal shutdown, the cache would of course be flushed to SSD no matter how little data was in it.
Member since:2005-06-29
...and Windows can do the exact same thing. And you don't even need a special filesystem for it, so no re-installation required. It's called EWBF or EWF.
http://www.aspireoneuser.com/forum/viewtopic.php?f=6&t=833&sid=1e66...
http://www.aspireoneuser.com/forum/viewtopic.php?f=6&t=179&sid=1e66...
Now you again.
Member since:2007-02-17
...and Windows can do the exact same thing. And you don't even need a special filesystem for it, so no re-installation required. It's called EWBF or EWF.
http://www.aspireoneuser.com/forum/viewtopic.php?f=6&t=833&...
http://www.aspireoneuser.com/forum/viewtopic.php?f=6&t=179&...
Now you again. "
Good. Solutions abound, for more than one OS. Excellent.
So now perhaps you can drop the pretense about why you had to install Windows on your Acer Aspire One, and just admit that you are slave to the lock-in.
Member since:2006-11-24
without knowing about these drivers in particular, the drivers are of course not 1.5 GB. You have the same on Mac for example and the reason is not the driver in the true sense but rather localization, help, etc -- and if you were honest, Linux drivers are never that fully featured as MS. For example, if I am thinking of my old Creative Sound under XP, true, that was a big download but you got all kind of fancy stuff like tools, etc, whereas for Linux, you only got barely the sound output. I can do drivers small at 3.6 MB. But honestly, for the past 10 years, there have only been times where I had wished my Linux drivers were more like on Windows - and never the other way round.
So it is not that black and white as you make it out to be. Having said that, I don't know what is in these W7 drivers - but so don't you. An MS driver offering the same as a Linux driver is not much different from a Linux driver.
Member since:2007-02-17
I don't think so.
A particular driver in Linux will accomodate a great many devices. Most often, but by no means always, the range of devices covered by a single driver covers a single "chip" rather than any one given manufacturer. Video card drivers address the GPU, not the manufacturer of the card. This becomes even more apparent though when you look at network interfaces including wireless and modems. Sometimes a single driver will cover two or more similar "chips" from different manufacturers ... mouse drivers are a good example here.
In the case of printer drivers for Linux ... applications "write" either postscript or PDF. From an applications point of view those are the two different "types" of printer you can have.
The process from application to printer is convoluted, but essentially CUPS is a set of filters that convert the print data into the final language/format of the printer.
http://en.wikipedia.org/wiki/CUPS#Filtering_process
Foomatic is one such filter:
http://en.wikipedia.org/wiki/Foomatic
I'm pretty sure that Gutenprint is another:
http://en.wikipedia.org/wiki/Gutenprint
Each of the filters covers a very large range of printers.
Most often, the only thing that is required to accomodate a new printer model (and sometimes even a new printer type) is a simple PPD file.
http://en.wikipedia.org/wiki/PostScript_Printer_Description
PPD files apply even to non-postscript printers (such as cheap inkjets) becasue CUPS itself is the Postscript interpreter.
When I connect a new network printer on Linux, all I generally need to do is select the PPD file from a long, long list contained within CUPS. If the version of CUPS installed is older, I can often get a newer printer working by downloading a PPD file for it from the net.
http://www.linuxfoundation.org/en/OpenPrinting/Database/DatabaseInt...
You can even describe a new printer and the database can often make a new PPD file for you where none existed before.
http://www.openprinting.org/edit_printer.cgi?newentry=1
When I install a new network printer on Windows ... Windows attempts to download executable dlls from the printer server to the new client! (This BTW means that a Windows print server can ONLY serve Windows clients).
The fact that each printer type seems to use a specific set of dll files as its printer driver, IMO, easily accounts for the 1.5 GB of printer drivers needed by Windows, versus the 60-odd megabytes needed by Linux.
Edited 2008-10-30 01:40 UTC
Member since:
2006-03-26
Member since:2006-02-05
Member since:
2007-12-16
"Sinofsky also explained that Windows 7 will allow you to remove more components post-installation, so that you can save diskspace."
It doesn't matter what can be removed post-installation, because it still isn't gonna git the 8 GB flash disk, if the trash needs to be installed at some point (in the beginning) anyways.
Member since:
2008-10-29