In independent reports made by WHO and the Red Cross, nationwide rate of Thyroid Cancer increased by 2400% from 1985 to 1992 in Belarus alone. And that's just Thyroid cancer, and not other variants.
Except that theory takes a severe beating. Compare entries 4, 5 and 6 here at Top 500, especially the number of CPU's. Notably, entries 4 and 6 use the same kind of high-speed interconnects between the nodes, so the difference can't be blamed on that.
The problem with the AMD approach is that you get the NUMA drawbacks not only against other nodes, but internally on the node. If the data CPU 1 needs isn't in it's own memory banks, it's got to request them from CPU 2, 3 or 4, with a latency penalty(Letting the memory controllers read/write freely from each other's banks doesn't sound like a good idea, really). It works well with databases, serving websites, file servers etc, where the processes don't need to share memory, talk to each other a lot etc, but for physics and chemistry simulations etc, you get some penalties
Application speed can be significant, it all depends on what kind of app... A simple low-life editor such as Vi will run fast on anything beyond a 286, and will be I/O-limited when you work with larger files. OpenSSH, OpenSSL, rendering programs, graphics programs etc tend to enjoy a greater benefit from optimizations for your architecture, especially if you enable the use of MMX/3DNOW/SSE and similar extensions. Even though GCC is not auto-vectorizing, those programs do gain a noticeable boost from enabling those extensions.
There are many legitimate reasons, such as information exchange, handling of investigations under the protocols for such, and FBI has lots of offices around the world. The FBI agents were kicked out for doing illegal stuff.
Actually, several FBI agents have been seized and more or less thrown out of the country by Swedish police, for failing to comply with Swedish law and international treaties.
The Swedish Stridsfordon 90(Combat Vehicle 90 for foreigners) is available in several variants, with the most heavily armed and armoured being used in Sweden, the StrF9040, and moving on to the 9040C variant. The weapon system is a Bofors 40mm automatic cannon that can use several kinds of ammo, the 4 primary being HE, APDS(Tungsten Penetrator), Programmable HE rounds for Anti-air use, and APDS-I(Incendiary package added to the Tungsten Penetrator).
Two variants being tested are Stridsfordon 90 AMOS and Stridsfordon 90120. The AMOS variant is a joint development with Finland, and it's a StrF90 chassis with an AMOS turret. Dual 120mm mortars, a total of 8 rounds/minute, capable of firing Strix rounds(Fire-and-forget IR-guided anti-tank mortar rounds) as well as ordinary HE rounds.
The Strf90120 has a turret armed with a 120mm smoothbore cannon.
The Friend Computer gives you a mission to kill a Commie Mutant Traitor in the form of one of the programmers with UV clearance(The player chars are nowhere near that, of course), and remains secluded in his UV clearance sector manse. When the characters eventually talk to the programmer, they find out that the computer has really been reprogrammed by Commie Mutant Traitors(And he has real proof that there has really been tampering, and that he's not responsible). The characters must reprogram the computer to fix the problem. Have fun =)
Or when you let all the players have Machine Empathy.... Because the computer has recently been upgraded to a biocomputer....
Congrats. The Debian guy in my class tried to do that and ended up having to fix even more stuff than the usual Gentoo install has to do.... Was kinda fun hearing him first tease us Gentoo users, and then just go "shut the hell up" whenever we asked him how the compile was going. =)
I don't see how it would cripple creativity and advancement? Did it hinder Linus et al from making Linux? No. Did it hinder Theo de Radt from making OpenBSD? No. Did it hinder a lot of other projects? No. Innovation is creating something entirely new. Improving a product... Well, if you had just tried to go about that instead, I'd maybe have agreed with you. But if you need someone elses kernel etc to make something, the odds are that your idea wasn't innovative at all.
And as for POSIX? Personally, I think it's way too much cruft and bloat.
As for competition, if everyone agrees on doing soemthing in just one way, you won't have any advancement of technology, really, since specs commitees etc will bog down quite a lot
Let us take an example from the CPU market... If we'd only go with IA32 or x86-64 as a standard, we'd be locked into that(Never mind that, for economical reasons, we already are, and many Linux proponents want that to happen, judging by comments on Slashdot etc... So much for advancement of technology...).
As for the freedom aspect... Microsoft has excercised their freedom to say no. Sure, RMS, ESR and a whole bunch of other fanatics might not like that someone says no to them, but freedom isn't just you getting things from someone because you want it, freedom is also for the other part to say no if they don't want to.
I never said that either. Both have their inherent security flaws. It's just that among open source projects, there's quite a lot of blind trust in that every participant is clean.
Well, can you trust the contributors? Can you trust the entire core team? What if someone participates with a fake identity also, and uses the legit and the fake identity to insert exploits, with the legit ID saying "I've checked out his patch, it seems ok", occasionally fixing a trivial error etc. Maybe not very likely, but it's definitely possible.
The whole "Many eyes makes the problem shallow" only works if everyone is equally skilled, and hopefully as skilled as the potential exploit creator. There's also the fact that the more people that become involved, the more things tend to screw up, with people not doing things because they think that someone else will look at it.
First of all, PCI Express is an implementation of parts of the 3GIO specifications(For example, PCI Express only allows for implementation on a PCB, i.e backplane, while the full 3GIO supports linking systems via fiber-optic cable).
As for your comments about PCI-X, not all implementations of PCI-X supports that, and you're still dependent upon transfers between various units, and RAM etc. DMA channels don't matter much when you've still only got that one physical conncetion and it's half-duplex. And using UDP is a hit-or-miss proposition. Any packet loss on the way, and the demo won't look quite as good
Improvements tend to come from working on other peoples projects. Innovation has that annoying tendency to come from the people who actually choose to work from scratch.
Case in point: Marketing aside, Xerox, Apple, MS etc etc has done far more innovation than the Open Source/Free Software Movement has done(Hell, the GNU project started because they made their own implementations of already existing tools). There's nothing in *BSD, Linux, XFree86, Mozilla, KDE, Gnome, GIMP etc that can be considered innovative, just implementations of what has been done in earlier research projects or commercial offerings.
Having recently done some research on various high-end bus and network technologies, I'd say go with PCI Express instead. PCI-X still has CPU issues, especially if you combine it with TCP. PCI-X also has the problem of being half-duplex.
Most impressive of the bunch though was Infiniband and 3GIO, which can be implemented from the north bridge straight out to devices several km away, via fiber-optic cables.
Listening too much to the community can be very detrimental to the development process, however, since you might lose sight of your original ideas and planning, and wind up in an unstructured mess of widely disparate ideas.
""Geek and proud!" As proof, I offer not merely the fact that I prefer digital watches, but that I set them to 24-hour time."
Since when did that become a matter of geekhood? Military all over the world has used it for quite a while, and the 24h-system has been standard in Sweden for at least 80 years.
So? Slashdot links to stuff from pro-Linux biased people all the time, without any such disclaimers either. The really annoying part is that noone whines about biased PoV then....
Because, assuming basic competency, if you write it from scratch, it does _exactly_ what you want it to do, it's developed just the way you want it, and with the documentation that you want and need.
With Open Source, you can easily end up spending more time modifying it to your specific needs than it takes to write it from scratch, especially if the source you're looking at is using a method for doing things that is not suitable for your problems. Especially since lack of or inferior documentation is all too common among many projects.
"On the other hand, I recieve many, MANY spam e-mails. I think anyone with an e-mail address has experienced lots of spam."
Unfortunately not true for me. None of my yahoo adresses have ever been spammed, much less my ISP-supplied email, or my own email account on a box. I got spam on an email adress I had over a couple of years, on a box owned by an IRC buddy, but it wasn't even 3/week
Slashdot Mirror
In independent reports made by WHO and the Red Cross, nationwide rate of Thyroid Cancer increased by 2400% from 1985 to 1992 in Belarus alone. And that's just Thyroid cancer, and not other variants.
Except that theory takes a severe beating. Compare entries 4, 5 and 6 here at Top 500, especially the number of CPU's. Notably, entries 4 and 6 use the same kind of high-speed interconnects between the nodes, so the difference can't be blamed on that.
The problem with the AMD approach is that you get the NUMA drawbacks not only against other nodes, but internally on the node. If the data CPU 1 needs isn't in it's own memory banks, it's got to request them from CPU 2, 3 or 4, with a latency penalty(Letting the memory controllers read/write freely from each other's banks doesn't sound like a good idea, really). It works well with databases, serving websites, file servers etc, where the processes don't need to share memory, talk to each other a lot etc, but for physics and chemistry simulations etc, you get some penalties
Application speed can be significant, it all depends on what kind of app... A simple low-life editor such as Vi will run fast on anything beyond a 286, and will be I/O-limited when you work with larger files. OpenSSH, OpenSSL, rendering programs, graphics programs etc tend to enjoy a greater benefit from optimizations for your architecture, especially if you enable the use of MMX/3DNOW/SSE and similar extensions. Even though GCC is not auto-vectorizing, those programs do gain a noticeable boost from enabling those extensions.
There are many legitimate reasons, such as information exchange, handling of investigations under the protocols for such, and FBI has lots of offices around the world. The FBI agents were kicked out for doing illegal stuff.
Actually, several FBI agents have been seized and more or less thrown out of the country by Swedish police, for failing to comply with Swedish law and international treaties.
The Swedish Stridsfordon 90(Combat Vehicle 90 for foreigners) is available in several variants, with the most heavily armed and armoured being used in Sweden, the StrF9040, and moving on to the 9040C variant. The weapon system is a Bofors 40mm automatic cannon that can use several kinds of ammo, the 4 primary being HE, APDS(Tungsten Penetrator), Programmable HE rounds for Anti-air use, and APDS-I(Incendiary package added to the Tungsten Penetrator).
Two variants being tested are Stridsfordon 90 AMOS and Stridsfordon 90120. The AMOS variant is a joint development with Finland, and it's a StrF90 chassis with an AMOS turret. Dual 120mm mortars, a total of 8 rounds/minute, capable of firing Strix rounds(Fire-and-forget IR-guided anti-tank mortar rounds) as well as ordinary HE rounds.
The Strf90120 has a turret armed with a 120mm smoothbore cannon.
The Friend Computer gives you a mission to kill a Commie Mutant Traitor in the form of one of the programmers with UV clearance(The player chars are nowhere near that, of course), and remains secluded in his UV clearance sector manse. When the characters eventually talk to the programmer, they find out that the computer has really been reprogrammed by Commie Mutant Traitors(And he has real proof that there has really been tampering, and that he's not responsible). The characters must reprogram the computer to fix the problem. Have fun =)
Or when you let all the players have Machine Empathy.... Because the computer has recently been upgraded to a biocomputer....
Congrats. The Debian guy in my class tried to do that and ended up having to fix even more stuff than the usual Gentoo install has to do.... Was kinda fun hearing him first tease us Gentoo users, and then just go "shut the hell up" whenever we asked him how the compile was going. =)
I don't see how it would cripple creativity and advancement? Did it hinder Linus et al from making Linux? No. Did it hinder Theo de Radt from making OpenBSD? No. Did it hinder a lot of other projects? No. Innovation is creating something entirely new. Improving a product... Well, if you had just tried to go about that instead, I'd maybe have agreed with you. But if you need someone elses kernel etc to make something, the odds are that your idea wasn't innovative at all.
And as for POSIX? Personally, I think it's way too much cruft and bloat.
As for competition, if everyone agrees on doing soemthing in just one way, you won't have any advancement of technology, really, since specs commitees etc will bog down quite a lot
Let us take an example from the CPU market... If we'd only go with IA32 or x86-64 as a standard, we'd be locked into that(Never mind that, for economical reasons, we already are, and many Linux proponents want that to happen, judging by comments on Slashdot etc... So much for advancement of technology...).
As for the freedom aspect... Microsoft has excercised their freedom to say no. Sure, RMS, ESR and a whole bunch of other fanatics might not like that someone says no to them, but freedom isn't just you getting things from someone because you want it, freedom is also for the other part to say no if they don't want to.
I never said that either. Both have their inherent security flaws. It's just that among open source projects, there's quite a lot of blind trust in that every participant is clean.
More likely, _if_ you find the malicious code.
Well, can you trust the contributors? Can you trust the entire core team? What if someone participates with a fake identity also, and uses the legit and the fake identity to insert exploits, with the legit ID saying "I've checked out his patch, it seems ok", occasionally fixing a trivial error etc. Maybe not very likely, but it's definitely possible.
The whole "Many eyes makes the problem shallow" only works if everyone is equally skilled, and hopefully as skilled as the potential exploit creator. There's also the fact that the more people that become involved, the more things tend to screw up, with people not doing things because they think that someone else will look at it.
First of all, PCI Express is an implementation of parts of the 3GIO specifications(For example, PCI Express only allows for implementation on a PCB, i.e backplane, while the full 3GIO supports linking systems via fiber-optic cable).
As for your comments about PCI-X, not all implementations of PCI-X supports that, and you're still dependent upon transfers between various units, and RAM etc. DMA channels don't matter much when you've still only got that one physical conncetion and it's half-duplex. And using UDP is a hit-or-miss proposition. Any packet loss on the way, and the demo won't look quite as good
Improvements tend to come from working on other peoples projects. Innovation has that annoying tendency to come from the people who actually choose to work from scratch.
Case in point: Marketing aside, Xerox, Apple, MS etc etc has done far more innovation than the Open Source/Free Software Movement has done(Hell, the GNU project started because they made their own implementations of already existing tools). There's nothing in *BSD, Linux, XFree86, Mozilla, KDE, Gnome, GIMP etc that can be considered innovative, just implementations of what has been done in earlier research projects or commercial offerings.
Having recently done some research on various high-end bus and network technologies, I'd say go with PCI Express instead. PCI-X still has CPU issues, especially if you combine it with TCP.
PCI-X also has the problem of being half-duplex.
Most impressive of the bunch though was Infiniband and 3GIO, which can be implemented from the north bridge straight out to devices several km away, via fiber-optic cables.
Why not? Age of consent is 15 in Sweden
Listening too much to the community can be very detrimental to the development process, however, since you might lose sight of your original ideas and planning, and wind up in an unstructured mess of widely disparate ideas.
yet it's still sluggish compared to Opera, and uses more memory on my box here in class.
Why not something really worthwhile, such as donating to the Red Cross, or the World Wildlife Foundation, or Greenpeace or similar?
""Geek and proud!" As proof, I offer not merely the fact that I prefer digital watches, but that I set them to 24-hour time."
Since when did that become a matter of geekhood? Military all over the world has used it for quite a while, and the 24h-system has been standard in Sweden for at least 80 years.
I'm just wondering wether anyone has attempted to invert the iso image....
So? Slashdot links to stuff from pro-Linux biased people all the time, without any such disclaimers either. The really annoying part is that noone whines about biased PoV then....
Because, assuming basic competency, if you write it from scratch, it does _exactly_ what you want it to do, it's developed just the way you want it, and with the documentation that you want and need.
With Open Source, you can easily end up spending more time modifying it to your specific needs than it takes to write it from scratch, especially if the source you're looking at is using a method for doing things that is not suitable for your problems. Especially since lack of or inferior documentation is all too common among many projects.
I do use my accounts, among other things when I sign up for downloads etc.
I'm just careful about what sites I download from, etc. I'm a member of a couple of web communities, and no spam to my publicly listed accounts.
"On the other hand, I recieve many, MANY spam e-mails. I think anyone with an e-mail address has experienced lots of spam."
Unfortunately not true for me. None of my yahoo adresses have ever been spammed, much less my ISP-supplied email, or my own email account on a box. I got spam on an email adress I had over a couple of years, on a box owned by an IRC buddy, but it wasn't even 3/week