if the reason Sony choose linux as a dev environment was because they had to much trouble with their engineers playing computer games, and since they realized there are so few for linux, they wouldn't have quite as large a problem. Of course, with more games coming, and wine, that may not solve anything.
Last fall, my roommate had to go to Microsoft for a week. He came back, and described all the people he met there as extremely arrogant, and they all fully believed in and supported MS, but not because the disliked the alternatives and competition, but because they didn't know any better. In short, all the people who worked there were just there trying to make as much money as they could, as fast as possible. They all had the attitude that anything MS did was correct because they are successful, so therefore whatever they do must be right. It was nice to finally see someone print this. It IS true.
I know you guys are in San Jose and busy, but I've submitted this article twice, once on Saturday, and once on Sunday. What's taken so long for it to be posted?
Why are people comparing the G3 and G4 and P3 to the Merced and Elbrus chips? There's no comparison to be made. When/if they arrive, they will be in an enterprise/supercomputing league, not the desktop league. This means, they will kill the desktop style processors.
I'm sick of seeing all these reviews and statements about the over-hype of the P3. The only hype I've seen is generated by the press saying how it's overhyped, or how it's not that great. Anyone who believed that the P3 would greatly enhance 3D didn' do their research on the instructions that have been added. The added instructions to help with 3D to some extent, but they are more geared toward speeding up specific algorithms, like the kind used in decoding and encoding video and voice recognition. In short, the P3 is offering the same type of idea as multiple processors: your computer won't necessarily be faster, but it will be able to handle more compicated calculations at one time. Thus, since the overall speed increase is minimal and mostly due to clock speed increase, and since NOBODY seems to develop the sorts of multimedia apps that would take advantage of these for linux, no linux user would benefit from a P3 over a P2. End of story.
If I could get dual k6-3s....
on
K6-3 on Monday
·
· Score: 1
I'd love to be able to run dual K6-3s. They'll be so cheap... it would be the best deal. I know some people are ready to get dual K7s if K6-3s won't do SMP (which I'm sure they won't). However, I'm fairly sure that at least until next year around this time, K7 processors and motherboards are just going to be way to expensive for most people. Sigh.
is that an eastern company will pick up and make this chip if western companies don't. Someone like Hitachi. I think they'd love to get the jump on something so high-end. Well, if the chip is actually producable, and the results are real.
I want FAST. So hard drive size seems to be growing exponentially. So what? I personally don't need that much space. That's not to say that I never will, of course. But I'm much, much more interested in speed than size. So you've got a dual Xeon P3 450 system. Or a Quad. You've still got a bottleneck on your harddrive. Any time you do anything disk intensive (say, compiling the latest kernel), away all your clockcycles go. People always seem to be so worried about memory accesses and small, slow caches bottlenecking processors. But those type of memory accesses are still way faster than harddrive access. 30 megs/sec is about what the fastest drives throughput. I really wish we could get something faster. That's why I spent $200 on a 7200 RPM UW SCSI drive (4.3) gig, instead of the same amount on a 10 GB UDMA 5400 RPM IDE drive. I want SPEED!!!!!!
Oddly enough, just the other day I attended a talk on patents in the US as they apply to engineering and software. Here are some quick facts:
1) In the US, proof of original concept is required. That is, the first to think of and implement (and have proof of this), can patent something. So, if you can prove that you implemented something before someone else who has a patent on something, you can fight it.
2) Software is patentable... but only the interface, which is why, I suppose, MS was able to get the patent described above.
3) Algorithms are patentable, but only in some sort of interface. So, a pure algorithm cannot be patented, but use of that algorithm in a program with a interface can be patented.
4) Patents now last 20 years from the time of filing.
Now, everytime I see one of these crazy patents, it makes me mad. I think, "How the hell could they allow these things to be patented?" Well, I was told that the patent office is understaffed, and that they are looking to hire more examiners. The problem is, the salary is only about $27000 US/year. My guess is these people don't know very much about computers.
The other thing is, yes, it is irritating to see these patents, but the thing is, they go away. And once they do, they become public domain.
Quake3 will have threading--> Good for SMP
on
Parallel Mesa
·
· Score: 1
Carmack had trouble trying to have one processor do opengl calls and the other processor do everything else... there was just too much overhead. What he said in his.plan was that he would just go back to natively threading it, because doing that he saw a fairly linear increase in performance. So, Q3, should have a threaded executable that will run faster on SMP machines, as well as a regular executable for UP machines.
1)Most of the posts here seem to be addressing which chip is "fastest." We've got people calling x86 CISC outdated (probably), and saying that G3s'4s are faster because they are RISCian. These arguments just don't hold up. Why? Because the G3/4 design is RISCian, but the instruction set is not. The same holds true with x86, be it Intel or AMD. Check out this article at Ars-Technica for greater explanation of RISC vs. CISC.
2)It's not just about the speed!!! If it was, why not argue that someone buy and SGI MIPS box and get all the great FPU performance? Because it costs too much. Sure, at clock speed, the G3/4 kicks x86 (at least integer wise, not sure about fpu). AMD and Cyrix also beat Intel at clock speed. But Intel just ups MHz. Now, the G3/4, probably not only beat Intel chips at clock speed, but at one or two higher clock speeds as well. Here's the thing: you can still get a faster x86 system for much cheaper (note: since we're just talking about speed here, we don't care about friendliness. That's the usual explanation for Mac prices, that you pay for the easiness and friendliness of the system). In october, I put together a Dual P2-350 system, with UW SCSI card, and UW SCSI hard drive, and 64 MB RAM, Matrox G200 vidcard, SB AWE 64, PCI ethernet, and a 12x SCSI cdrom, for $1000. No monitor. But with a monitor, that would have cost the same as an iMac. And there's just no way that the 233 MHz G3 running on a 66MHz bus was gonna beat that system.
So, when the K7, G4, and P3 are all out later this year, if I (a power user), feel like upgrading for more speed, it will most likely be to another x86 chip because speed/$ is much greater.
NOTE: I'm no x86 fanatic (I'm the guy who posted about cheap alphas yesterday), and this discussion does not apply to normal computer users, because to them speed is not all that is important, and they don't understand what makes a computer fast. Fast is not what sells, marketing and gimmicks sell.
I work at the National Center fro Supercomputing Applications at the University of Illinois at Urbana/Champaign, and we use all the latest mass storage technologies (DLT drives, TLM Robots, Tape Silos, etc), and we've also had do to a migration from older media. And it takes time. First of all, our migrations weren't from such old media, which meant that they held more than NASA's tapes. So, we had less tapes to deal with, and they transferred faster to DLT. NASA has so many (relatively) low-capacity tapes that read slowly, it would take a huge amount of time to do anything. It doesn't matter how fast the medium you're copying to can write at, in this case the bottleneck is reading the old media. Not to mention the fact that tape drives are relatively unreliable. That is, they tend to break every few months when you use them 24hours/day, 7days/week. And we are talking about huge amounts of data... I know at NCSA I once had a user request the deletion of a 100GB file that was tarred and gz'ed. Optical drives would be great, but they don't hold enough compared to tapes.
a unix doesn't HAVE to have remote login. You can shut off all remote accesses except http, if you want. That's what makes unix unix: virtually unlimited choices of what you want. I'd be interested in an finding out if an NT web server is more secure than a unix-based web server with all remote logins turned off.
As for inexperienced sysadmins... well, if you use linux, I'd think that with the money saved you could get yourself a more experienced sysadmin, which would be better in the long-run anyway. As you said, NT looks pretty good "on the surface".
why firewire is said to be so good? I can understand that it's much better (faster) than USB, and would make an excellent replacement for serial ports, but I've seen articles that say there will be firewire harddrives, etc. Isn't it too slow? The max hard drive data transfers right now are 80MB/sec (on a U2W scsi card and drive). Now, if firewore (or this new non-Apple implementation of it) allow 100MB/sec, than I don't see how it will be enough if you've got a hard drive + cdrom + peripherals.
This really has little to do with the OS. It's just an application to download that was compromised with a trojan on one server. This is the equivalent of someone hacking and distributing a copy of WinZip for Windows that would do something similar. The only difference is that with open source, the trojan is caught fairly quickly by people who go over the source. In Windows, you never know exactly what you're running... like BackOrifice.
Slashdot Mirror
if the reason Sony choose linux as a dev environment was because they had to much trouble with their engineers playing computer games, and since they realized there are so few for linux, they wouldn't have quite as large a problem. Of course, with more games coming, and wine, that may not solve anything.
Last fall, my roommate had to go to Microsoft for a week. He came back, and described all the people he met there as extremely arrogant, and they all fully believed in and supported MS, but not because the disliked the alternatives and competition, but because they didn't know any better. In short, all the people who worked there were just there trying to make as much money as they could, as fast as possible. They all had the attitude that anything MS did was correct because they are successful, so therefore whatever they do must be right. It was nice to finally see someone print this. It IS true.
Time to upgrade. Glad I'm back on T1.
I know you guys are in San Jose and busy, but I've submitted this article twice, once on Saturday, and once on Sunday. What's taken so long for it to be posted?
No, SMP was available in the 2.0.x kernels. It just didn't work nearly as well (could lock up the system/worse performance).
Why are people comparing the G3 and G4 and P3 to the Merced and Elbrus chips? There's no comparison to be made. When/if they arrive, they will be in an enterprise/supercomputing league, not the desktop league. This means, they will kill the desktop style processors.
I'm sick of seeing all these reviews and statements about the over-hype of the P3. The only hype I've seen is generated by the press saying how it's overhyped, or how it's not that great. Anyone who believed that the P3 would greatly enhance 3D didn' do their research on the instructions that have been added. The added instructions to help with 3D to some extent, but they are more geared toward speeding up specific algorithms, like the kind used in decoding and encoding video and voice recognition. In short, the P3 is offering the same type of idea as multiple processors: your computer won't necessarily be faster, but it will be able to handle more compicated calculations at one time. Thus, since the overall speed increase is minimal and mostly due to clock speed increase, and since NOBODY seems to develop the sorts of multimedia apps that would take advantage of these for linux, no linux user would benefit from a P3 over a P2. End of story.
I'd love to be able to run dual K6-3s. They'll be so cheap... it would be the best deal. I know some people are ready to get dual K7s if K6-3s won't do SMP (which I'm sure they won't). However, I'm fairly sure that at least until next year around this time, K7 processors and motherboards are just going to be way to expensive for most people. Sigh.
is that an eastern company will pick up and make this chip if western companies don't. Someone like Hitachi. I think they'd love to get the jump on something so high-end. Well, if the chip is actually producable, and the results are real.
I want FAST. So hard drive size seems to be growing exponentially. So what? I personally don't need that much space. That's not to say that I never will, of course. But I'm much, much more interested in speed than size. So you've got a dual Xeon P3 450 system. Or a Quad. You've still got a bottleneck on your harddrive. Any time you do anything disk intensive (say, compiling the latest kernel), away all your clockcycles go. People always seem to be so worried about memory accesses and small, slow caches bottlenecking processors. But those type of memory accesses are still way faster than harddrive access. 30 megs/sec is about what the fastest drives throughput. I really wish we could get something faster. That's why I spent $200 on a 7200 RPM UW SCSI drive (4.3) gig, instead of the same amount on a 10 GB UDMA 5400 RPM IDE drive. I want SPEED!!!!!!
Oddly enough, just the other day I attended a talk on patents in the US as they apply to engineering and software. Here are some quick facts:
1) In the US, proof of original concept is required. That is, the first to think of and implement (and have proof of this), can patent something. So, if you can prove that you implemented something before someone else who has a patent on something, you can fight it.
2) Software is patentable... but only the interface, which is why, I suppose, MS was able to get the patent described above.
3) Algorithms are patentable, but only in some sort of interface. So, a pure algorithm cannot be patented, but use of that algorithm in a program with a interface can be patented.
4) Patents now last 20 years from the time of filing.
Now, everytime I see one of these crazy patents, it makes me mad. I think, "How the hell could they allow these things to be patented?" Well, I was told that the patent office is understaffed, and that they are looking to hire more examiners. The problem is, the salary is only about $27000 US/year. My guess is these people don't know very much about computers.
The other thing is, yes, it is irritating to see these patents, but the thing is, they go away. And once they do, they become public domain.
Carmack had trouble trying to have one processor do opengl calls and the other processor do everything else... there was just too much overhead. What he said in his .plan was that he would just go back to natively threading it, because doing that he saw a fairly linear increase in performance. So, Q3, should have a threaded executable that will run faster on SMP machines, as well as a regular executable for UP machines.
1)Most of the posts here seem to be addressing which chip is "fastest." We've got people calling x86 CISC outdated (probably), and saying that G3s'4s are faster because they are RISCian. These arguments just don't hold up. Why? Because the G3/4 design is RISCian, but the instruction set is not. The same holds true with x86, be it Intel or AMD. Check out this article at Ars-Technica for greater explanation of RISC vs. CISC.
2)It's not just about the speed!!! If it was, why not argue that someone buy and SGI MIPS box and get all the great FPU performance? Because it costs too much. Sure, at clock speed, the G3/4 kicks x86 (at least integer wise, not sure about fpu). AMD and Cyrix also beat Intel at clock speed. But Intel just ups MHz. Now, the G3/4, probably not only beat Intel chips at clock speed, but at one or two higher clock speeds as well. Here's the thing: you can still get a faster x86 system for much cheaper (note: since we're just talking about speed here, we don't care about friendliness. That's the usual explanation for Mac prices, that you pay for the easiness and friendliness of the system). In october, I put together a Dual P2-350 system, with UW SCSI card, and UW SCSI hard drive, and 64 MB RAM, Matrox G200 vidcard, SB AWE 64, PCI ethernet, and a 12x SCSI cdrom, for $1000. No monitor. But with a monitor, that would have cost the same as an iMac. And there's just no way that the 233 MHz G3 running on a 66MHz bus was gonna beat that system.
So, when the K7, G4, and P3 are all out later this year, if I (a power user), feel like upgrading for more speed, it will most likely be to another x86 chip because speed/$ is much greater.
NOTE: I'm no x86 fanatic (I'm the guy who posted about cheap alphas yesterday), and this discussion does not apply to normal computer users, because to them speed is not all that is important, and they don't understand what makes a computer fast. Fast is not what sells, marketing and gimmicks sell.
I work at the National Center fro Supercomputing Applications at the University of Illinois at Urbana/Champaign, and we use all the latest mass storage technologies (DLT drives, TLM Robots, Tape Silos, etc), and we've also had do to a migration from older media. And it takes time. First of all, our migrations weren't from such old media, which meant that they held more than NASA's tapes. So, we had less tapes to deal with, and they transferred faster to DLT. NASA has so many (relatively) low-capacity tapes that read slowly, it would take a huge amount of time to do anything. It doesn't matter how fast the medium you're copying to can write at, in this case the bottleneck is reading the old media. Not to mention the fact that tape drives are relatively unreliable. That is, they tend to break every few months when you use them 24hours/day, 7days/week. And we are talking about huge amounts of data... I know at NCSA I once had a user request the deletion of a 100GB file that was tarred and gz'ed. Optical drives would be great, but they don't hold enough compared to tapes.
a unix doesn't HAVE to have remote login. You can shut off all remote accesses except http, if you want. That's what makes unix unix: virtually unlimited choices of what you want. I'd be interested in an finding out if an NT web server is more secure than a unix-based web server with all remote logins turned off.
As for inexperienced sysadmins... well, if you use linux, I'd think that with the money saved you could get yourself a more experienced sysadmin, which would be better in the long-run anyway. As you said, NT looks pretty good "on the surface".
why firewire is said to be so good? I can understand that it's much better (faster) than USB, and would make an excellent replacement for serial ports, but I've seen articles that say there will be firewire harddrives, etc. Isn't it too slow? The max hard drive data transfers right now are 80MB/sec (on a U2W scsi card and drive). Now, if firewore (or this new non-Apple implementation of it) allow 100MB/sec, than I don't see how it will be enough if you've got a hard drive + cdrom + peripherals.
This really has little to do with the OS. It's just an application to download that was compromised with a trojan on one server. This is the equivalent of someone hacking and distributing a copy of WinZip for Windows that would do something similar. The only difference is that with open source, the trojan is caught fairly quickly by people who go over the source. In Windows, you never know exactly what you're running... like BackOrifice.
Actually, Chrome has now been un-cancelled.
Title says it all. Read the article. This was only included on the PC, not PSX.