Try going to a college in the cornfields, where there are cows nearby. Trust me when I say that 5 miles is not too far away from 50 cows with far less than 30000 tons of manure among them.
Unfortunately, gravity did it for me. The problem is that my portable CD player had no mounting kit. (Hell, if wanted to spend money on a portable cd player and then more on a springy attachment device, I could have bought an in-dash cd player). Thus it slid around a bit, tugging on the cable occasionally.
Okay, the REAL problem was that the cassette adapters were from Radio Shack, and should have been sold as the $2 pieces of s*** they were, not extortion to the tune of $20.
I've owned several cassette adapters, and they all went bad. The main problem is the cheap quality of the audio cable they always use, so they break at the plug. The quality itself was great -- better than a tape since it's coming from a CD player and you don't have to worry about the quality of metal on the tape.
The point is, an FM adapter may be slightly more durable since there's less fiddling with the cabling. Less fiddling itself is a bonus, except that now you have to mess with the radio station.
The AlphaServers will run Linux and Compaq will sell you one, but no buyers yet among these big machines.
I'm pretty sure there are smaller ES45 (or ES50 or whatever) linux clusters around, in the low hundreds of nodes instead of thousands. Also, I thought Sandia's CPlant was running linux on some of these guys and it's sitting at #30. The TOP500 list doesn't say, and I've even used it and still don't remember....:)
I'll agree with that, but it's poor for other reasons as well. First, having used it, I'm convinced of the futility of synthetic benchmarks, since I don't think it belongs in the #4 spot. Pure processor count is the only way it's up that high. IO performance is also pitiful.
And if you want to look at the data output from one of their large simulations, you ideally would like to do it in parallel on the machine it was run on (i.e. Red itself). It doesn't support X11 so you can't run a graphical app at all and get a window to display to your desktop. It doesn't support sockets, so you can't even run a parallel visualization app and send data over sockets to a workstation with graphics hardware. The only way to get data off that machine is by sucking the terabytes of data you created off their slow filesystem.
10. IBM SP Power3 375 MHz 16 way
Deutscher Wetterdienst
Again, I'm sure it's AIX.
All Unix. No, no linux on there yet, but Pacific Northwest will be right up there near the top, and Lawrence Livermore is also probably getting a linux cluster of almost that size pretty soon. That will make two in the top few slots.
Unless there is something you really want to include that is just too big to be downloaded easily by 56kers. Unfortunately, many people still seem to be limited in this regard, and even I don't really want to download hundreds of megs on a cable modem.
Apart from that, anyone who does open source or other types of coding already has net access, so it is not worth even a few bucks. For just source code and maybe some binaries, just make sure you have a good web or ftp site available.
I have dozens of CDROMs that came with books that I have never touched. The one exception was one that had a 50MB source tar that I didn't feel like waiting to download from a slow site.
DivX encoding: 30% (note that the program is highly optimized, by Intel themselves, for the P4. How many programmers have an Intel engineer handy?)
IIRC, it is a standard pass through the current Intel optimizing compiler, available for a couple hundred dollars as a plugin to MSVC. Not exactly hand coding by an Intel engineer, although it was originally done in one night by an Intel employee.
Thinking vs Planning (was Re:FIREFOX LIVES !!!!!)
on
Think And Click
·
· Score: 1
Its esay , sometimes to think about something and not pysically execute that movement. WHAT HAPPENS when you JUST THINK and it happens, I could think of all kinds of scenarios this would be MUY bad, remember the end of the first ghostbusters movie:)
There is a real distinction here between "thinking about doing it" and "plan on doing it". You can think about doing lots of things, but this doesn't mean you will. This is important, because it actually applies here.
When you are about to move your mouse, some pre-initiation of the neurons that do the moving actually happens. This is happening about when your muscles start to contract, but before any actual movement happens.
In other words, you have to actually make a conscious effort to pretend to command your arm to move. It does not make your arm move, but it's a lot more like making the cursor move instead of thinking about it moving. This is the right behavior for such a system, IMHO.
It's a little pricey (though well within your range -- about $200 US), but they make one that handles the high resolutions as well as both USB and PS2 keyboard and mouse. It even switches audio if you want it.
I use one at home and at work, and I have only tried it up to 1920x1440, but it handles it with a bare minimum of ghosting. The most important thing is the cables. Belkin also sells those, some good gold plated ones, for not a ridiculous amount. That's critical to preventing ghosting.
Assuming you go with something that can run either windows or linux as an OS:
Citrix has an ICA client for both Windows and for Unix. I have mainly used the one for Linux, and as far as I can tell, it works exactly as well as the one for Windows. I doubt there is much difference between the price for either.
If you want to use Windows as a host OS, that is a decent chunk of money right there. You can certainly run X11 apps to displayed to windows using StarNet's X-Win32, Hummingbird's eXceed, or WRQ Reflection X. The last one I know supports accelerated OpenGL and I don't think the other two do. If that's important for your type of work, take that into consideration.
If you really want good remote X11 support, though, do not use a Windows X server. Speaking from personal experience, they all have flaws. Use Linux as host OS, get the Citrix ICA Linux client, and just run X remotely when you need to. Best price/performance for sure. Plus, it gets the worker bees accustomed to the Linux desktop while using the Windows one. Makes it easier for the bait'n'switch later.;)
I work for one of the large US national laboratories, and we have been doing large-scale parallel CFD-style codes for years. One billion elements is not uncommon anymore. These are codes designed to run on 100 to 10,000 processors.
We have traditionally gone with IBM hardware (RS6000) running AIX. I personally hate AIX for its crappy development environment (shared libraries in particular are completely broken). However, IBM has done a decent job with inter-processor communication and IO.
We also have some 32-128 processor SGI Onyx (Origin 2k?) systems running IRIX. I personally love them, but we do *not* typically use them for CFD, but for visualization of results (yes, in parallel). Part of their beauty is that they are SMP machines, so codes which don't use MPI but do use threads work great on them. The IBMs, by contrast, are 4-way (or more recently 16-way) parallel per node, but you are forced to so cross-box communication a la MPI for more than that.
However, we are looking into a 60-teraflop machine right now. It could be a big IBM, but at the current hardware speeds, that's about 50000 processors (!). One realistic option is a massive Intel/Linux (Lintel?:) cluster.
The current problem is that no P4 systems are more than 2-way SMP, so the interprocessor communication will bog you down much faster. You can do something like 4-way PIII's cheaper at the sacrifice of single-processor speed. It depends on how much the code you use depends on the interprocessor communication vs pure processor speed.
You can use something like a Quadrics interconnect to get really good inter-box communication speed, but those only scale to 128 nodes without extra work, and even then I only know they support 256 nodes.
This sounds like you are doing something on a much smaller scale for sure, but if your Fluent code is very parallizable, go with the cheap Linux cluster. If you need interprocessor speed, you can do the same as long as you get a good interconnect. Except for the problems of the more do-it yourself style support, it will cost you far less for the same number of FLOPS.
If it is possible to get a larger SMP machine for your money and Fluent is only threaded, go the SGI route, or maybe a single node 16-way IBM, to get the most processors within a node that you can.
It's not specifically "intelligence" per se, but more the difference in knowledge. I have a hard time explaining Winblows to people who have never used a computer before. This is even true for my parents, who have been using computers for almost as long as I have (considering they bought me my first one at age 7), but who still don't understand many of the things going on under the covers. It took a while, but I think I got them to the point where they can install a CDROM drive with only phone-based tech support from me.:)
The real problem is the frame of reference. Could you explain a mouse to a complete tech newbie if you used one for twenty years? You wouldn't realize that even if they figure out how the mouse connects to the cursor, they might not know what the "buttons" look like, that "icons" require a double-click to activate instead of a single click, etc.
This also points to one strength of Windows -- a consistent user interface. If buttons look different between GTK, Qt, even Motif and Athena widget sets, can you imagine how confusing this is for newbies?
With that exception, however, Linux is not really much harder to learn that Windows and the problem is that it is typically the very experienced user who is trying to teach it, not just the most intelligent ones.
The big question -- did you rip to uncompressed (.wav) files before encoding? And if so, did it have the aforementioned errors?
Make sure to try CD-ripping software with error correction.
I can't speak much for Linux (I only know of grip), but under Winblows try Exact Audio Copy (EAC). It's got great error correction settings and you can rip in essentially a paranoid mode if necessary.
If that doesn't help, try a different encoder. LAME comes highly recommended both from professional trials and from personal experience. Use the latest beta -- it's better than the "stable" release.
(Okay, so that's two words....)
Seriously, they have just recently become available to the masses, they are still fighting a format war, and the blanks are far too pricey.
However, give 'em a few months. I would bet the price will drop from $500-$600 to closer to $300 in the first couple months of the new year, the DVD+RW vs DVD-RW formats will have some semblance of a winner, and the blanks should drop to $5 or so. It's just too early to make a reasonable purchase of one.
1. Box in a Box. This is a cardinal rule of packaging.
My brother packed his computer in a box in a box. UPS shipped it. The box arrived in one piece.... with a big gaping hole.... and no inner box.
They never found his computer. We think someone at UPS said "Oh, look! A computer in a box with no label. Guess it's mine now."
Glad it was insured for $2k, but such a loss of data.... Sad.
Ah yes, the Dell Inspiron 8k series. I've got a PIII 1GHz, 512MB PC133 RAM, GeForce2(Go), and a 15" 1600x1200 screen.
Plus, a builtin DVD/CDRW, 2 USB, IEEE1394, SVideo and normal TV out, 10/100 Ethernet, 40G HD and a decent sound card.
With the CD builtin, the second media bay holds a second battery. Plenty of time to watch Braveheart on a plane as long as you don't need the floppy drive.
As for gaming, the GF2Go is nearly as fast as a normal GF2, and it smokes the hell out of my home system right now. 100FPS Quaking. Just hook it up to a real monitor (which it will drive at 2000x1400 if your monitor will handle it), and suddenly a glorious dual-boot Win2k/Linux 7lb wonder machine. Who needs a desktop with this thing?
Can't wait to get my hands on this GF3Go/NV17M thingy....:)
I think Trolltech has actually done a very good job keeping binary compatibility between versions, at least in the 2.x series. They happen to do so by a method where they create much of their private class data as classes defined in the SOURCE file, not the header file.
This is actually somewhat unfortunate for developers, because it makes it nearly impossible to subclass some of their builtin widgets without copying large sections of their source code. To me that defeats some of the purpose of writing it in C++.
However, there are many obselete functions left over from *old* versions of Qt that had to go away. No one wants to support those kinds of things. And without breaking at least binary compatibility, you can't rip them out of the class definitions and the virtual function tables.
"Recode"? It for the most part requires only a re-link. Thus the "breaks binary compatibility" but "almost complete source compatibility". Changes to any source code which uses Qt should be minor to zero.
Seriously? I thought that was all weapons design codes and input decks, and those had all been Secret Restricted Data since some presidental act decades ago.
There are about 100 million "pixels" in the human eye, so if a display took up your entire visual field it would need to be 10000x10000 pixels.
However, there is a much higher density of receptors at the fovea (center of focus), so if the eye is allowed to refocus on different areas of the display (you don't always fix your eyes in the center of your display, do you?), this is more like, assuming a 22" monitor two feet away, about 1000 pixels per inch, or 1 megapixel per square inch of display.
Check out the Oracle at Gathering of Developers. It has many questions in a similar vein and answers from those actually in the industry, and it has a lot of good info. As they say, it is a "moderated public forum geared toward striving game developers".
Slashdot Mirror
Try going to a college in the cornfields, where there are cows nearby. Trust me when I say that 5 miles is not too far away from 50 cows with far less than 30000 tons of manure among them.
Turkey manure must be on the mild side.
Unfortunately, gravity did it for me. The problem is that my portable CD player had no mounting kit. (Hell, if wanted to spend money on a portable cd player and then more on a springy attachment device, I could have bought an in-dash cd player). Thus it slid around a bit, tugging on the cable occasionally.
Okay, the REAL problem was that the cassette adapters were from Radio Shack, and should have been sold as the $2 pieces of s*** they were, not extortion to the tune of $20.
They're both cheap kludges!
I've owned several cassette adapters, and they all went bad. The main problem is the cheap quality of the audio cable they always use, so they break at the plug. The quality itself was great -- better than a tape since it's coming from a CD player and you don't have to worry about the quality of metal on the tape.
The point is, an FM adapter may be slightly more durable since there's less fiddling with the cabling. Less fiddling itself is a bonus, except that now you have to mess with the radio station.
The AlphaServers will run Linux and Compaq will sell you one, but no buyers yet among these big machines.
:)
I'm pretty sure there are smaller ES45 (or ES50 or whatever) linux clusters around, in the low hundreds of nodes instead of thousands. Also, I thought Sandia's CPlant was running linux on some of these guys and it's sitting at #30. The TOP500 list doesn't say, and I've even used it and still don't remember....
I'll agree with that, but it's poor for other reasons as well. First, having used it, I'm convinced of the futility of synthetic benchmarks, since I don't think it belongs in the #4 spot. Pure processor count is the only way it's up that high. IO performance is also pitiful.
And if you want to look at the data output from one of their large simulations, you ideally would like to do it in parallel on the machine it was run on (i.e. Red itself). It doesn't support X11 so you can't run a graphical app at all and get a window to display to your desktop. It doesn't support sockets, so you can't even run a parallel visualization app and send data over sockets to a workstation with graphics hardware. The only way to get data off that machine is by sucking the terabytes of data you created off their slow filesystem.
And that is why I think it is a poor OS.
1. IBM ASCI White,SP Power3 375 MHz
;)
Lawrence Livermore National Laboratory
It runs AIX.
2. Compaq AlphaServer SC ES45/1GHz
Pittsburgh Supercomputing Center
Haven't used it, but I'm guessing Tru64.
3. IBM SP Power3 375 MHz 16 way
NERSC/LBNL
Once again, AIX.
4. Intel ASCI Red
Sandia National Labs
A poor home-grown OS (no offence) called Cougar or TFlops which doesn't even support X11 or sockets.
5. IBM ASCI Blue-Pacific SST,IBM SP 604e
Lawrence Livermore National Laboratory
Can you say AIX?
6. Compaq AlphaServer SC ES45/1GHz
Los Alamos National Laboratory
I assume Tru64.
7. Hitachi SR8000/MPP
University of Tokyo
No idea. Sorry.
8. SGI ASCI Blue Mountain
Los Alamos National Laboratory
IRIX.
9. IB SP Power3 375 MHz
Naval Oceanographic Office
Don't know for sure, but you can bet it's AIX.
10. IBM SP Power3 375 MHz 16 way
Deutscher Wetterdienst
Again, I'm sure it's AIX.
All Unix. No, no linux on there yet, but Pacific Northwest will be right up there near the top, and Lawrence Livermore is also probably getting a linux cluster of almost that size pretty soon. That will make two in the top few slots.
No Windows on these puppies!
Unless there is something you really want to include that is just too big to be downloaded easily by 56kers. Unfortunately, many people still seem to be limited in this regard, and even I don't really want to download hundreds of megs on a cable modem.
Apart from that, anyone who does open source or other types of coding already has net access, so it is not worth even a few bucks. For just source code and maybe some binaries, just make sure you have a good web or ftp site available.
I have dozens of CDROMs that came with books that I have never touched. The one exception was one that had a 50MB source tar that I didn't feel like waiting to download from a slow site.
DivX encoding: 30% (note that the program is highly optimized, by Intel themselves, for the P4. How many programmers have an Intel engineer handy?)
IIRC, it is a standard pass through the current Intel optimizing compiler, available for a couple hundred dollars as a plugin to MSVC. Not exactly hand coding by an Intel engineer, although it was originally done in one night by an Intel employee.
Its esay , sometimes to think about something and not pysically execute that movement. WHAT HAPPENS when you JUST THINK and it happens, I could think of all kinds of scenarios this would be MUY bad, remember the end of the first ghostbusters movie :)
There is a real distinction here between "thinking about doing it" and "plan on doing it". You can think about doing lots of things, but this doesn't mean you will. This is important, because it actually applies here.
When you are about to move your mouse, some pre-initiation of the neurons that do the moving actually happens. This is happening about when your muscles start to contract, but before any actual movement happens.
In other words, you have to actually make a conscious effort to pretend to command your arm to move. It does not make your arm move, but it's a lot more like making the cursor move instead of thinking about it moving. This is the right behavior for such a system, IMHO.
It's a little pricey (though well within your range -- about $200 US), but they make one that handles the high resolutions as well as both USB and PS2 keyboard and mouse. It even switches audio if you want it.
I use one at home and at work, and I have only tried it up to 1920x1440, but it handles it with a bare minimum of ghosting. The most important thing is the cables. Belkin also sells those, some good gold plated ones, for not a ridiculous amount. That's critical to preventing ghosting.
Honda is going to start selling a Civic hybrid this spring. 50 mpg, normal size Civic. See here for the press release and for the pictures.
I personally would find this a huge step up from the Insight, and probably the Prius as well.
Assuming you go with something that can run either windows or linux as an OS:
;)
Citrix has an ICA client for both Windows and for Unix. I have mainly used the one for Linux, and as far as I can tell, it works exactly as well as the one for Windows. I doubt there is much difference between the price for either.
If you want to use Windows as a host OS, that is a decent chunk of money right there. You can certainly run X11 apps to displayed to windows using StarNet's X-Win32, Hummingbird's eXceed, or WRQ Reflection X. The last one I know supports accelerated OpenGL and I don't think the other two do. If that's important for your type of work, take that into consideration.
If you really want good remote X11 support, though, do not use a Windows X server. Speaking from personal experience, they all have flaws. Use Linux as host OS, get the Citrix ICA Linux client, and just run X remotely when you need to. Best price/performance for sure. Plus, it gets the worker bees accustomed to the Linux desktop while using the Windows one. Makes it easier for the bait'n'switch later.
NCD's terminals (used to) run in the $1000s. Ones we use here at work that support just X11 were about $6K new.
I work for one of the large US national laboratories, and we have been doing large-scale parallel CFD-style codes for years. One billion elements is not uncommon anymore. These are codes designed to run on 100 to 10,000 processors.
:) cluster.
We have traditionally gone with IBM hardware (RS6000) running AIX. I personally hate AIX for its crappy development environment (shared libraries in particular are completely broken). However, IBM has done a decent job with inter-processor communication and IO.
We also have some 32-128 processor SGI Onyx (Origin 2k?) systems running IRIX. I personally love them, but we do *not* typically use them for CFD, but for visualization of results (yes, in parallel). Part of their beauty is that they are SMP machines, so codes which don't use MPI but do use threads work great on them. The IBMs, by contrast, are 4-way (or more recently 16-way) parallel per node, but you are forced to so cross-box communication a la MPI for more than that.
However, we are looking into a 60-teraflop machine right now. It could be a big IBM, but at the current hardware speeds, that's about 50000 processors (!). One realistic option is a massive Intel/Linux (Lintel?
The current problem is that no P4 systems are more than 2-way SMP, so the interprocessor communication will bog you down much faster. You can do something like 4-way PIII's cheaper at the sacrifice of single-processor speed. It depends on how much the code you use depends on the interprocessor communication vs pure processor speed.
You can use something like a Quadrics interconnect to get really good inter-box communication speed, but those only scale to 128 nodes without extra work, and even then I only know they support 256 nodes.
This sounds like you are doing something on a much smaller scale for sure, but if your Fluent code is very parallizable, go with the cheap Linux cluster. If you need interprocessor speed, you can do the same as long as you get a good interconnect. Except for the problems of the more do-it yourself style support, it will cost you far less for the same number of FLOPS.
If it is possible to get a larger SMP machine for your money and Fluent is only threaded, go the SGI route, or maybe a single node 16-way IBM, to get the most processors within a node that you can.
Agreed, completely.
:)
It's not specifically "intelligence" per se, but more the difference in knowledge. I have a hard time explaining Winblows to people who have never used a computer before. This is even true for my parents, who have been using computers for almost as long as I have (considering they bought me my first one at age 7), but who still don't understand many of the things going on under the covers. It took a while, but I think I got them to the point where they can install a CDROM drive with only phone-based tech support from me.
The real problem is the frame of reference. Could you explain a mouse to a complete tech newbie if you used one for twenty years? You wouldn't realize that even if they figure out how the mouse connects to the cursor, they might not know what the "buttons" look like, that "icons" require a double-click to activate instead of a single click, etc.
This also points to one strength of Windows -- a consistent user interface. If buttons look different between GTK, Qt, even Motif and Athena widget sets, can you imagine how confusing this is for newbies?
With that exception, however, Linux is not really much harder to learn that Windows and the problem is that it is typically the very experienced user who is trying to teach it, not just the most intelligent ones.
The big question -- did you rip to uncompressed (.wav) files before encoding? And if so, did it have the aforementioned errors?
Make sure to try CD-ripping software with error correction.
I can't speak much for Linux (I only know of grip), but under Winblows try Exact Audio Copy (EAC). It's got great error correction settings and you can rip in essentially a paranoid mode if necessary.
If that doesn't help, try a different encoder. LAME comes highly recommended both from professional trials and from personal experience. Use the latest beta -- it's better than the "stable" release.
(Okay, so that's two words....)
Seriously, they have just recently become available to the masses, they are still fighting a format war, and the blanks are far too pricey.
However, give 'em a few months. I would bet the price will drop from $500-$600 to closer to $300 in the first couple months of the new year, the DVD+RW vs DVD-RW formats will have some semblance of a winner, and the blanks should drop to $5 or so. It's just too early to make a reasonable purchase of one.
1. Box in a Box. This is a cardinal rule of packaging.
My brother packed his computer in a box in a box. UPS shipped it. The box arrived in one piece.... with a big gaping hole.... and no inner box.
They never found his computer. We think someone at UPS said "Oh, look! A computer in a box with no label. Guess it's mine now."
Glad it was insured for $2k, but such a loss of data.... Sad.
Ah yes, the Dell Inspiron 8k series. I've got a PIII 1GHz, 512MB PC133 RAM, GeForce2(Go), and a 15" 1600x1200 screen. :)
Plus, a builtin DVD/CDRW, 2 USB, IEEE1394, SVideo and normal TV out, 10/100 Ethernet, 40G HD and a decent sound card.
With the CD builtin, the second media bay holds a second battery. Plenty of time to watch Braveheart on a plane as long as you don't need the floppy drive.
As for gaming, the GF2Go is nearly as fast as a normal GF2, and it smokes the hell out of my home system right now. 100FPS Quaking. Just hook it up to a real monitor (which it will drive at 2000x1400 if your monitor will handle it), and suddenly a glorious dual-boot Win2k/Linux 7lb wonder machine. Who needs a desktop with this thing?
Can't wait to get my hands on this GF3Go/NV17M thingy....
I think Trolltech has actually done a very good job keeping binary compatibility between versions, at least in the 2.x series. They happen to do so by a method where they create much of their private class data as classes defined in the SOURCE file, not the header file.
This is actually somewhat unfortunate for developers, because it makes it nearly impossible to subclass some of their builtin widgets without copying large sections of their source code. To me that defeats some of the purpose of writing it in C++.
However, there are many obselete functions left over from *old* versions of Qt that had to go away. No one wants to support those kinds of things. And without breaking at least binary compatibility, you can't rip them out of the class definitions and the virtual function tables.
"Recode"? It for the most part requires only a re-link. Thus the "breaks binary compatibility" but "almost complete source compatibility". Changes to any source code which uses Qt should be minor to zero.
Seriously? I thought that was all weapons design codes and input decks, and those had all been Secret Restricted Data since some presidental act decades ago.
Information is not classified after it has been born unclassified. It must be born classified as part of a classified project.
I also doubt the judge possesses the clearance required to evaluate it himself, so no one may be able to evaluate it's accuracy.
There are about 100 million "pixels" in the human eye, so if a display took up your entire visual field it would need to be 10000x10000 pixels.
However, there is a much higher density of receptors at the fovea (center of focus), so if the eye is allowed to refocus on different areas of the display (you don't always fix your eyes in the center of your display, do you?), this is more like, assuming a 22" monitor two feet away, about 1000 pixels per inch, or 1 megapixel per square inch of display.
Check out the Oracle at Gathering of Developers. It has many questions in a similar vein and answers from those actually in the industry, and it has a lot of good info. As they say, it is a "moderated public forum geared toward striving game developers".