I've got it on two machines -- this one (a Pentium II 266 -- I think), with 176 meg RAM, and a AMD K6-2-3D running at somewhere around 290mhz at home, with 64 meg RAM. On here Win95, Win98 and Windows 3.11 all ran like a charm. Never got Win98 working with the drivers so video was kind of pokey, but I can play back that southpark starwars trailer under Win95 using VMWare without any glitches or audio dropouts. I boot them with 64 meg RAM.
At home with only 64 meg RAM, I boot them with 32, and its pokey because of the godawful large amount of RAM GNOME takes up. If I log in to an account I setup to load Windows directly, it runs like a charm too... well except for the annoying tendancy Linux has to hardlock on heavy drive access... four months and I still haven't tracked that bugger down, and that's a Linux thing not a VMware thing.
Concerning the price for VMWare, sending e-mails can't hurt. I suggested a WMWare "lite" version for people who just wanted to be able to boot Windows, and they said they made note of the suggestion. If they keep getting it ("I'd rather buy a new PC than spend $300 on the software, but $150 might be reasonable") then they might consider a price change or a "lite" version.
Internet Explorer is available for several Unix platforms. I'd like to see it available for Linux. As much as I dislike Microsoft and the poor quality of their software, I dislike the poor quality of Netscape's more -- and the lack of options for web browsers with any sort of modern standards.
Netscape is a lousy package under Linux. I've never had a version that felt stable. Currently ps shows netscape gobbling up almost 100 meg of RAM on here, after only a day's work. Netscape also seems to bloat X's memory,although not being an X-programmer I don't really understand how that could happen.
On all three machines I regularly run Netscape on I have a mysterious problem with it spawning error windows until the system crashes...
Gotta figure IE for Linux is a smaller porting project than Word, given the HP and Solaris versions.
What's everyone's opinions? I wish Mozilla was ready... hopefully it won't be as bad as 4.0/4.5.
Except that VNC is slugish on > 640X480 displays
on
The $299 PC
·
· Score: 1
1600x1200? What color depth? 24 bit would mean the screen is 1600x1200x24 is a hair under 6 meg.
Of course that would be sluggish. But the few applications most people who run Linux need to be able to use under Windows don't need screens that big. For a couple years I've used VNC as a replacement for a KVM switch. I usually configure the Windows machine at 800x600x16bit, works like a charm, even over a 10mb network. Since the advent of the Mac server, I've been using that as well -- its slow, but its nice to be able to get to a mac from my desktop as well. I'm sure once the mac server matures some it'll perform as well as the linux and windows ones.
My point is that on my machine at work (Pentium II 350, 176 meg RAM, ATI Rage Pro) I get pretty decent performance out of VMware's product. Its noticeably slower than Windows 95 or 98 on its own, particularly in a window as opposed to full screen (with the upgraded DGA XFree86). Video display is kind of pokey, and so is the hardware. I'm sure I'd get just as much performance to my linux desktop using VNC on a $299 machine.
VMware is a good product at $300 for commercial use -- if I have to test a product against Windows 95, 98, linux, etc... its a great system. $300 instead of dedicated testing machines? Can't beat that.
But $300 to simply run Windows, its a better deal to spend the extra $$$ to buy a second PC. (Although in my case I'd just stick to running Win98 on my 586/133 and using VNC as I am right now...)
This makes it all the more disappointing that VMWare is planning on selling their software at $299. Why spend $299 for their stuff when $299 will buy one of these? Throw in a $20 network card, run VNC and you get the same result...
I wouldn't want a $300 PC, but maybe the fact that they exist might entice VMWare to sell their software at a more reasonable $150 or something.
This is quite a bit of code... Something new? Probably not, but its pretty slick none the less. Very quick on here, Win98 and Win95 installed without a hitch, both seem to work fine, all the apps I've tried have worked.
There's some wierd little bugs with the mouse cursor in X when running their device drivers under Windows... my netscape window is overlapping the VMware one right now, if I put the mouse in the window its on the windows desktop until I pull it completely out of the partially hidden VMWare window, then it appears on top of netscape.
Everything else seems to work... $300 is kind of pricey considering I can buy a new PC for a bit more than that, but if I can get it to run Cakewalk, I might even spend $300...
I bet they'd sell a lot of them to people who want to run Linux primarily but need Win95 for some various things, if it was a bit cheaper. $150 seems more reasonable given the $399 cost of those new E-Machine PC's.
I thought they closed the downtown one entirely... As long as one is open, I don't even care.:)
Aw man, now I really want one. *grrr* Too bad I'm goin snowboarding this weekend or a road trip would be in order.:)
I ran into someone in switzerland once who knew what Nicks was though -- you you never know. And anyone who doesn't know, should know. Stack of nicks plates, big bottle of Devils Springs, pile of misidentified DEC boards... its the makings of... um... well the old jake and elwood actually.
I asked them about fedex overnight once... they looked at me like I was smokin' crack.
But you're right, 24 hours is too long. (Although a mistimed nicks run during a blizard turned a 6 1/2 hour each way nicks run with far too many people jammed into my car into a 20 hour roundtrip fiasco a few years ago... but I got my plate...)
SGI's NUMA architecture means data can be pumped *much* more quickly between nodes, 100-1000 times as fast. Network-based Linux clustering is useful only for calculations that are fairly self-contained and don't need a lot of data to process.
What I think would be more interesting, given SGI's leanings towards supporting Linux on their MIPS and Intel platforms, is if they eventually tweak the multiprocessing in the kernel to support NUMA style multiprocessing and I can throw Linux on an Origin server. Or maybe better yet a NUMA-architecture Intel machine (i'm not really up on floating point speed comparisions between newer MIPS and newer Intel chips). Since they've dropped real PC-compatibility on their new Intel machines, that sort of a shift is a lot less painful than the initial dropping of support for DOS/16 bit apps.
So SGI doesn't get hurt by Linux. Linux *can't* really compete with a Cray at any real-world tasks (not yet...). And SGI is in a *real* good spot to be the ones selling the Linux-compatible hardware that actually could. In which case, why would they care? Their profit may be lower on a $500k Cray-comparible NUMA linux system than on the Cray, but I'd bet they'd sell enough more of them to make up the difference.
That's a point, but I doubt that's the cause in this case. I suspect its something else. If I was rendering images (as opposed to large amounts of calculations that deal with the results of other sets of calculations), then I'd just split the image up into 36 chunks and have one processor blast through each chunk, and stick them all back together at the end. My assumption is that's how this test works because they mentioned how a few scanlines were dropped when one node went off line. So its not true PVM-style clustering like Beowolf.
In this case, 18 times the processors should give 18 times the speed -- unless the test really isn't processor bound. I'm not sure what it would be bound by, however... lousy implementation? I/O? Network?
If the test isn't really processor bound then the comparison to the cray is meaningless, because there's something wrong with the way the software is coded to work on parallel machines, I'd think.
I disagree that the $12k cost means it was a cluster of Pentium II's. I've bought a couple Pentium II systems in that range, its easy to get up there when you add a lot of RAM, lot of harddrive space, etc. Using name parts jacks the price up a lot. (ie, VA's selling systems with Intel boards rather than supermicro or some other lower-cost company...)
On a side note, I remember reading a year or two ago that someone was working on a networking layer that allowed IP and other protocols to be routed between cluster machines over a 40MB/sec SCSI bus. Anyone know if that ever got to completion? A four-fold jump in network speed would make quite a difference to I/O bound applications. (And SCSI cards are a lot cheaper than Gigabit ethernet or other real high-speed networking technologies...)
I think what's interesting about this is what it didn't say -- that the old record of 9 seconds was on a Dual Pentium II. So having *18 times* the number of processors only got it three times the performance...
Any comments on this? Obviously a dual Pentium II is pretty damn good at this too, being only 1/3 the speed of a $5.5 million supercomputer. Anyone have any idea why adding so many processors to the Linux cluster would improve results so little?
We had a discussion about this back in January on the autolinux mailing list (www.bangsplat.org/autolinux).
The best way to handle it is to mount everything you can read-only. I've been experimenting with mounting everything read only and then mounting a ramdisk as an overlay filesystem allowing the system to be written to, as long as the files you write to don't need to be saved.
Another way is to mount a ramdisk as a partition, copy off a read only partition, and periodically mount the original partion read/write, sync the two, and remount it read only.
You're very right. I had to remove 1.0 and go back to 0.99.7 to get three of my systems back to a usable state.
I think its a shame that the GNOME team shot for feature expansion not stability for 1.0. It seems like a poor and careless way to handle the public side of software development. They obviously have every right to code it any way they wish, but this sort of a fiasco just hurts their reputation. Yesterday would've been a big day for KDE if KDE was actually a useful environment.
Lets hope 1.1 or 1.0.1 or some announcement saying "just kidding, *this* is actually 1.0" comes out sometime soon.
For all those who say a stable 1.0 is impossible..
on
GNOME 1.0 Released
·
· Score: 1
Don't compare GNOME 1.0 to KDE 1.0.
Compare GNOME 1.0 to GNOME 0.99.7 or 99.8.
Then you'll see how much of a piece of crap 1.0 is. Usually software works better when you get off of beta, not worse. Unfortunately when the eyes of the world are on the GNOME team, they've botched things by releasing the least stable version they've had in months.
Has anyone noticed that 1.0 is *much* less stable than 0.99.7? Its the worst I've seen GNOME in six months! At first I thought it was just my PC at work -- a mishmash of RedHat 5.0, 5.1, 5.2 with multiple GTK versions, half compiled, half RPMS, etc.
But here at home on a clean 5.2 install -- as well as a 5.1 install that hasn't had any GNOME software (only KDE) installed on it, things that used to work perfectly don't work any more. The control panel doesn't function in any usable form. Most of the components of it crash randomly, the Enlightenment one isn't even listed any more.
The menu editor crashes and half the time causes the panel to crash to. In *no* cases has making a change caused the panel menus to actually change, which has worked flawlessly the last five or six versions I've ran.
Half the games seem to crash or not function properly. GNOME sounds seem to cause the system to crash randomly. (Aparently because of wav files missing from what it expects to have, and its not handling the missing file cleanly)
Now I've got to work myself back through downgrading to a version that actually is usable.
This is really pathetic. I think its terrible that such a lousy product would be announced to the public when so many eyes are on Linux this week, and the software works far worse than the multitudes of betas that thousands of people have happily been using over the last few months.
If there's SRPMS on there, you could just download all of them and rebuild PPC versions of the RPMS.
It'd take quite a while -- Gnome isn't quick to compile -- but it should work. I squeaked in a moment or two after this article was posted here and got right into the server, but I just grabbed the binary RPMS not the source ones so I don't know if the source ones are there.
Which makes sense, I've never been able to get RedHat to run on an 8meg system, regardless of what their minimum requirements say.
I use RedHat on my alpha and 8meg+ intel systems, Debian on anything 8 meg or less.
RedHat could learn a bit from Debian in the installer allowing you to mount an existing swap partition first step of the install. In 8 meg, I've found fdisk and the other partitioning program won't load... at least in the two or three I've tried.
First one on the last week or two to send in something about it. Hasn't there been a half-dozen postings about the Empeg on here in the last six month?
Slashdot Mirror
I've got it on two machines -- this one (a Pentium II 266 -- I think), with 176 meg RAM, and a AMD K6-2-3D running at somewhere around 290mhz at home, with 64 meg RAM. On here Win95, Win98 and Windows 3.11 all ran like a charm. Never got Win98 working with the drivers so video was kind of pokey, but I can play back that southpark starwars trailer under Win95 using VMWare without any glitches or audio dropouts. I boot them with 64 meg RAM.
At home with only 64 meg RAM, I boot them with 32, and its pokey because of the godawful large amount of RAM GNOME takes up. If I log in to an account I setup to load Windows directly, it runs like a charm too... well except for the annoying tendancy Linux has to hardlock on heavy drive access... four months and I still haven't tracked that bugger down, and that's a Linux thing not a VMware thing.
Concerning the price for VMWare, sending e-mails can't hurt. I suggested a WMWare "lite" version for people who just wanted to be able to boot Windows, and they said they made note of the suggestion. If they keep getting it ("I'd rather buy a new PC than spend $300 on the software, but $150 might be reasonable") then they might consider a price change or a "lite" version.
Internet Explorer is available for several Unix platforms. I'd like to see it available for Linux. As much as I dislike Microsoft and the poor quality of their software, I dislike the poor quality of Netscape's more -- and the lack of options for web browsers with any sort of modern standards.
Netscape is a lousy package under Linux. I've never had a version that felt stable. Currently ps shows netscape gobbling up almost 100 meg of RAM on here, after only a day's work. Netscape also seems to bloat X's memory,although not being an X-programmer I don't really understand how that could happen.
On all three machines I regularly run Netscape on I have a mysterious problem with it spawning error windows until the system crashes...
Gotta figure IE for Linux is a smaller porting project than Word, given the HP and Solaris versions.
What's everyone's opinions? I wish Mozilla was ready... hopefully it won't be as bad as 4.0/4.5.
1600x1200? What color depth? 24 bit would mean the screen is 1600x1200x24 is a hair under 6 meg.
Of course that would be sluggish. But the few applications most people who run Linux need to be able to use under Windows don't need screens that big. For a couple years I've used VNC as a replacement for a KVM switch. I usually configure the Windows machine at 800x600x16bit, works like a charm, even over a 10mb network. Since the advent of the Mac server, I've been using that as well -- its slow, but its nice to be able to get to a mac from my desktop as well. I'm sure once the mac server matures some it'll perform as well as the linux and windows ones.
My point is that on my machine at work (Pentium II 350, 176 meg RAM, ATI Rage Pro) I get pretty decent performance out of VMware's product. Its noticeably slower than Windows 95 or 98 on its own, particularly in a window as opposed to full screen (with the upgraded DGA XFree86). Video display is kind of pokey, and so is the hardware. I'm sure I'd get just as much performance to my linux desktop using VNC on a $299 machine.
VMware is a good product at $300 for commercial use -- if I have to test a product against Windows 95, 98, linux, etc... its a great system. $300 instead of dedicated testing machines? Can't beat that.
But $300 to simply run Windows, its a better deal to spend the extra $$$ to buy a second PC. (Although in my case I'd just stick to running Win98 on my 586/133 and using VNC as I am right now...)
They pretty much explicitly say it doesn't. The two higher priced ones do. Its just like the $399 eMachine doesn't have one either. The $499 one does.
This makes it all the more disappointing that VMWare is planning on selling their software at $299. Why spend $299 for their stuff when $299 will buy one of these? Throw in a $20 network card, run VNC and you get the same result...
I wouldn't want a $300 PC, but maybe the fact that they exist might entice VMWare to sell their software at a more reasonable $150 or something.
Damn, beaten to the punch. Oh well I won't be the first, but I'll be one of the few with full-time internet in my car. Wish I had that kinda $$$. :)
http://www.bangsplat.org/autolinux
This is quite a bit of code... Something new? Probably not, but its pretty slick none the less. Very quick on here, Win98 and Win95 installed without a hitch, both seem to work fine, all the apps I've tried have worked.
There's some wierd little bugs with the mouse cursor in X when running their device drivers under Windows... my netscape window is overlapping the VMware one right now, if I put the mouse in the window its on the windows desktop until I pull it completely out of the partially hidden VMWare window, then it appears on top of netscape.
Everything else seems to work... $300 is kind of pricey considering I can buy a new PC for a bit more than that, but if I can get it to run Cakewalk, I might even spend $300...
I bet they'd sell a lot of them to people who want to run Linux primarily but need Win95 for some various things, if it was a bit cheaper. $150 seems more reasonable given the $399 cost of those new E-Machine PC's.
I thought they closed the downtown one entirely... As long as one is open, I don't even care. :)
:)
Aw man, now I really want one. *grrr* Too bad I'm goin snowboarding this weekend or a road trip would be in order.
I ran into someone in switzerland once who knew what Nicks was though -- you you never know. And anyone who doesn't know, should know. Stack of nicks plates, big bottle of Devils Springs, pile of misidentified DEC boards... its the makings of... um... well the old jake and elwood actually.
Ah the good ol' days...
There's a good slashdot question:
"Nicks?"
1) Yes
2) No
3) Huh?
:)
I asked them about fedex overnight once... they looked at me like I was smokin' crack.
But you're right, 24 hours is too long. (Although a mistimed nicks run during a blizard turned a 6 1/2 hour each way nicks run with far too many people jammed into my car into a 20 hour roundtrip fiasco a few years ago... but I got my plate...)
Uuuhhhh... that's all I needed, a reminder of CSH -- just enough to trigger a craving for a Nicks plate... damn seven hour drives.
*grumble*
Ah, that clears it up. Its always good to hear the real deal from the source. :)
I doubt it...
SGI's NUMA architecture means data can be pumped *much* more quickly between nodes, 100-1000 times as fast. Network-based Linux clustering is useful only for calculations that are fairly self-contained and don't need a lot of data to process.
What I think would be more interesting, given SGI's leanings towards supporting Linux on their MIPS and Intel platforms, is if they eventually tweak the multiprocessing in the kernel to support NUMA style multiprocessing and I can throw Linux on an Origin server. Or maybe better yet a NUMA-architecture Intel machine (i'm not really up on floating point speed comparisions between newer MIPS and newer Intel chips). Since they've dropped real PC-compatibility on their new Intel machines, that sort of a shift is a lot less painful than the initial dropping of support for DOS/16 bit apps.
So SGI doesn't get hurt by Linux. Linux *can't* really compete with a Cray at any real-world tasks (not yet...). And SGI is in a *real* good spot to be the ones selling the Linux-compatible hardware that actually could. In which case, why would they care? Their profit may be lower on a $500k Cray-comparible NUMA linux system than on the Cray, but I'd bet they'd sell enough more of them to make up the difference.
Time will tell.
That's a point, but I doubt that's the cause in this case. I suspect its something else. If I was rendering images (as opposed to large amounts of calculations that deal with the results of other sets of calculations), then I'd just split the image up into 36 chunks and have one processor blast through each chunk, and stick them all back together at the end. My assumption is that's how this test works because they mentioned how a few scanlines were dropped when one node went off line. So its not true PVM-style clustering like Beowolf.
In this case, 18 times the processors should give 18 times the speed -- unless the test really isn't processor bound. I'm not sure what it would be bound by, however... lousy implementation? I/O? Network?
If the test isn't really processor bound then the comparison to the cray is meaningless, because there's something wrong with the way the software is coded to work on parallel machines, I'd think.
I disagree that the $12k cost means it was a cluster of Pentium II's. I've bought a couple Pentium II systems in that range, its easy to get up there when you add a lot of RAM, lot of harddrive space, etc. Using name parts jacks the price up a lot. (ie, VA's selling systems with Intel boards rather than supermicro or some other lower-cost company...)
On a side note, I remember reading a year or two ago that someone was working on a networking layer that allowed IP and other protocols to be routed between cluster machines over a 40MB/sec SCSI bus. Anyone know if that ever got to completion? A four-fold jump in network speed would make quite a difference to I/O bound applications. (And SCSI cards are a lot cheaper than Gigabit ethernet or other real high-speed networking technologies...)
I think what's interesting about this is what it didn't say -- that the old record of 9 seconds was on a Dual Pentium II. So having *18 times* the number of processors only got it three times the performance...
Any comments on this? Obviously a dual Pentium II is pretty damn good at this too, being only 1/3 the speed of a $5.5 million supercomputer. Anyone have any idea why adding so many processors to the Linux cluster would improve results so little?
I was just happy I could pull down the OS selection and pick "none".
Anyone know if the onboard SCSI in the Dell 2300 servers is supported by Linux? Its it a 7880? I wish their site had more info on it.
I wonder if they're going to have an ASCII-art logo... what if my cdplayer application isn't graphical?
CDDB isn't that complex, and what its doing isn't that complex. I assume they don't have a patent on it, someone should just duplicate the service...
We had a discussion about this back in January on the autolinux mailing list (www.bangsplat.org/autolinux).
The best way to handle it is to mount everything you can read-only. I've been experimenting with mounting everything read only and then mounting a ramdisk as an overlay filesystem allowing the system to be written to, as long as the files you write to don't need to be saved.
Another way is to mount a ramdisk as a partition, copy off a read only partition, and periodically mount the original partion read/write, sync the two, and remount it read only.
There was quite a bit of fun to be had at Web98 last year in Boston...
What we really need is a good non-corporate linux expo in Boston or NYC. (Non-corporate meaning cool like Web98 not sucky like InternetWorld)
Of course there was better free stuff at InternetWorld.
Okay, good non-corporate linux expo with lots of free stuff and killer parties. That's the ticket. In Boston. Or NYC.
You're very right. I had to remove 1.0 and go back to 0.99.7 to get three of my systems back to a usable state.
I think its a shame that the GNOME team shot for feature expansion not stability for 1.0. It seems like a poor and careless way to handle the public side of software development. They obviously have every right to code it any way they wish, but this sort of a fiasco just hurts their reputation. Yesterday would've been a big day for KDE if KDE was actually a useful environment.
Lets hope 1.1 or 1.0.1 or some announcement saying "just kidding, *this* is actually 1.0" comes out sometime soon.
Don't compare GNOME 1.0 to KDE 1.0.
Compare GNOME 1.0 to GNOME 0.99.7 or 99.8.
Then you'll see how much of a piece of crap 1.0 is. Usually software works better when you get off of beta, not worse. Unfortunately when the eyes of the world are on the GNOME team, they've botched things by releasing the least stable version they've had in months.
Has anyone noticed that 1.0 is *much* less stable than 0.99.7? Its the worst I've seen GNOME in six months! At first I thought it was just my PC at work -- a mishmash of RedHat 5.0, 5.1, 5.2 with multiple GTK versions, half compiled, half RPMS, etc.
But here at home on a clean 5.2 install -- as well as a 5.1 install that hasn't had any GNOME software (only KDE) installed on it, things that used to work perfectly don't work any more. The control panel doesn't function in any usable form. Most of the components of it crash randomly, the Enlightenment one isn't even listed any more.
The menu editor crashes and half the time causes the panel to crash to. In *no* cases has making a change caused the panel menus to actually change, which has worked flawlessly the last five or six versions I've ran.
Half the games seem to crash or not function properly. GNOME sounds seem to cause the system to crash randomly. (Aparently because of wav files missing from what it expects to have, and its not handling the missing file cleanly)
Now I've got to work myself back through downgrading to a version that actually is usable.
This is really pathetic. I think its terrible that such a lousy product would be announced to the public when so many eyes are on Linux this week, and the software works far worse than the multitudes of betas that thousands of people have happily been using over the last few months.
If there's SRPMS on there, you could just download all of them and rebuild PPC versions of the RPMS.
It'd take quite a while -- Gnome isn't quick to compile -- but it should work. I squeaked in a moment or two after this article was posted here and got right into the server, but I just grabbed the binary RPMS not the source ones so I don't know if the source ones are there.
Which makes sense, I've never been able to get RedHat to run on an 8meg system, regardless of what their minimum requirements say.
I use RedHat on my alpha and 8meg+ intel systems, Debian on anything 8 meg or less.
RedHat could learn a bit from Debian in the installer allowing you to mount an existing swap partition first step of the install. In 8 meg, I've found fdisk and the other partitioning program won't load... at least in the two or three I've tried.
First one on the last week or two to send in something about it. Hasn't there been a half-dozen postings about the Empeg on here in the last six month?
And I thought it was something new...
www.bangsplat.org/autolinux
for anyone interested...