The XBox PowerPC processors are not designed for general purpose work. They're in-order, they have a long pipline like the Pentium 4 and they have crappy branch predition. Just being PowerPC doesn't make them good at every.
I noticed that the Star Destroyers and fighters shown in that scene were different from the ones we'd seen in the rest of the movie. The Star Destroyers were white instead of white and orange or whatever, and IIRC they were structurally different as well; they didn't seem to have the docking bay doors on top like the earlier ones. The fighters were the first ones in the whole movie that had the distinctive TIE roar.
With entirely new classes of ships in service, it seems likely that the scene took place years after the rest of the events of the movie. No way to tell how many.
Also, the construction of the second death star may have begun at any time. When something takes that long to construct, it's common (for us anyway) to build many in parallel, and as others have said the better experience and industrial consolodation of the Empire may have reduced the construction time of the second one.
"But the benefits of this technology are not to extend processor life, or primarily to decrease power consumption."
Electromigration is greater at higher voltages, so I would say going from 1.4 volts to 1.1 volts would increase the processor life. It's not a primary motivation, but it certainly doesn't hurt when deciding to buy the thing.
Heat, power consumption, and noise are all directly related, so an attempt to improve one necessarily improves the others.
OpenBSD is biglock; that is, only one of the processors can be in the kernel at any given time, the entire thing is protected by one big lock. The other processor can be running something in userspace no problem.
This isn't a problem when using a limited number of CPUs (say 4 or less) because the probability of contention is low enough that it's not a big hit on performance for most tasks. For example, Linux 2.4.x and FreeBSD 4.x were biglock and people seemed pretty happy with their performance. People even complain about performance regressions compared to the older versions.
Linux and FreeBSD have moved on to finer grained locking. It's an advantage on systems with more CPUs, but the advantage is questionable on 2-4 processors because there's a lot of overhead, it's a HUGE amount of work, and it takes a long time to get right.
The post you replied to might have been joking and it was certainly wrong.
A number of OpenBSD developers will be speaking at the CUUG meeting on tuesday the 24th. It's extremely interesting to see them discuss the stuff they do, and it's a good opportunity to ask questions.
"I wonder if the copyrights holders associations would have less of a problem with Peer to Peer networking if the source and destination of the files (and who placed the files in the p2p network in the first place) were clearly recorded.
After all, If you "only read Playboy for the articles", you should not object to getting a copy with all of the pictures removed.
If you want to protect BitTorrent and other kinds of P2P, Make them be responsible citizens."You assume innocent people should
That's idiotic. Additional records would make no difference because BitTorrent makes no attempt to conceal the participants now. When you connect to the tracker, you request a list of IP addresses and it gives them to you. That's as transparent as you can get.
The only reason that hasn't stopped people from using it is that the RIAA and MPAA can't sue people fast enough. They already have the identities of more people than they can sue. Only when they started suing the trackers and.torrent file repositories did they get results, and it's easy to find those regardless of what records they keep.
That whole "if you're innocent you have nothing to hide" attitude is offensive anywhere it happens. It is not the responsibility of the individual to record their activities in order to later prove that they didn't do anything wrong. It never has been.
The only reason the RIAA and the MPAA have a problem with BitTorrent is that it's so mind-bogglingly efficient. It's capable of distributing multi-gigabyte files to an arbitrary number of people with nothing more than broadband Internet access. It's completely content agnostic, it just happens to be a lot better at what it does than any of the alternatives.
I don't see this as a positive thing. I see this as further evidence that Apple doesn't spend enough on QA.
This will be about the tenth time my iBook has had something wrong with it. In this case the inconvenience is very small because I can continue to use the computer until the replacement arrives, but Apple's consistent failure to do good QA guarantees that this will continue to happen in the future.
They're not any worse than other OEMs, but they're not any better.
"Running a CPU at lower than its rated speed is just plain silly.... So you see, there is no point to what you're doing, and I think its stupid:P"
Only if the assumptions you made were correct. Given the information you seem to have, your conclusions are reasonable. But you're missing something.
Newer desktop Athlon64s have laptop-like power-saving features. Basically, the chip I have (Athlon64 3000+) can run at 1.0 ghz or 1.8 ghz (changes between 1.1v and 1.4v). This can be changed dynamically at any time. I'm not doing anything at all weird, it's supported in the chipset, the CPU, and the OS without any modifications by me apart from installing appropriate software.
So it's only underclocked when it's idle, and it's handled automatically so I can't tell the differnce except for the benefits.
And here I got one of those new Athlon64s because I wanted to be able to underclock it.
It reduces stability, it decreases component life, and it increases power usage and heat. If you want to do it, I'm not going to stop you, but I'm not going to complain if Intel and AMD come up with a way to effectively prevent it, and I'm still going to think it's stupid.
"But PF and OpenSSH were/started/ at a hackathon and are now open projects and open for much discussion. My point in my original post was not taken in the right way. I meant to say that if Theo wants everyone to just code and not discuss their methods, then he'll end up with a million different implementations of something each with its own set of problems."
As I understand it, they've already worked out how things are going to go well before anyone steps on a plane. They've done most of the design work.
As I interpret the "shut up and hack" quote, they mean by the time everyone arrives in Calgary it's time to implement what they've discussed, not rehash design decisions that have been talked to death.
Re:Where did the devil put the .iso images...
on
OpenBSD 3.7 Released
·
· Score: 1, Informative
"Buy the official CDs and support the project, roll up your sleeves and make your own or use another OS. It's a free world."
It's easier to just do an FTP install.
Well, once the mirrors calm down in a few days anyway.
My setup: I use OpenBSD 3.6 (no upgrade for me yet, mirrors are totally) as a firewall and server (NFS, web, SSH, DHCP, that sort of thing). I use Debian-testing as a desktop on a different machine.
The documentation is second to none. That includes all the Linuxes I've tried as well as the BSDs. The fact that it's actually worth reading the docs means you do it early and often, which is nice.
Installing things out of ports is about as easy as any of the other good package managers I've used on Linux. The one thing to be aware of is that pkg_add can take a filename argument in the form of a URL to the FTP site. It's better to do that because it can resolve dependencies from the FTP site instead of you having to download them manually.
DO NOT ROLL A CUSTOM KERNEL. The generic one supports virtually everything that has been tested properly. If you have a problem with generic, report the bug because they'll want to fix it.
The firewall, PF, rocks. It's easy to set up and very powerful.
Overall, the differences between distros are as big as the differences with OpenBSD. I had to go through a lot of distros before I found the one I was happy with, and if you can wrap your brain around, say, both Slackware and Suse, you'll be fine.
I don't use OpenBSD as a desktop system so I don't know how well this works, but it might be worth a shot to set up Linux binary emulation and just using the Linux binary from mozilla.org. That might work better anyway for flash and stuff.
I use Debian-testing as a desktop and I avoid apt for the small number of things that I want to be 100% up to date with, including Firefox.
I prefer to code C on OpenBSD because the man pages for the C library calls are better. I can port it to Linux later if necessary. I don't use OpenBSD as a desktop OS, so I don't know what graphics support is like. They seem to be running a recent X.org, so it can't be that bad...
I like OpenBSD better in general, but Debian is a lot less work as a desktop OS and I'm lazy. OpenBSD runs my firewall/server and Debian-testing my desktop.
The documentation is MUCH better than any Linux I've used, which may be an edge if you're starting out.
"You don't see OpenBSD winning any awards for number of installations, do ya?"
OpenBSD, no. It's a niche OS. It doesn't surprise me that firewall machines are outnumbered by other machines. That doesn't say anything about the quality of OpenBSD code.
The fact that PF (written at a hackathon) has displaced IPF and IPFW as the BSD firewall of choice speaks to the quality of OpenBSD code.
The fact that OpenSSH is by far the dominant SSH implementation speaks to the quality of OpenBSD code. 90% of all SSH servers are OpenSSH, and a good chunk of the rest are re-branded OpenSSH servers.
" while Darwin may be based heavily on BSD, are they actually contributing anything back?"
Sometimes yes, sometimes no.
KHTML is huge, and the Apple changes are huge. They've simply forked the Safari version from the KDE version because there's too many changes happening too quickly to keep the two in sync. Their contributions back are in a form that is largely useless to KDE (a big blob, with no way to tell what a change fixes and no way to tell what it depends on).
OTOH, they submit reasonable patches back to FreeBSD, as their changes are relatively small, and are generally in response to specific bugs rather than large changes in functionality. In these situations, a patch that is accepted upstream is one more patch that you don't have to maintain yourself, so even though the BSD license doesn't force them to release the code they do it anyway.
"I disagree. Debian especially is very focussed on quality, and each package having a very good ratio of maintainers."
I use Debian-testing as a desktop system, I agree with your statement, but Debian has plenty of its own problems. For example, the stable release schedule is pretty tragic.
Conversely, 3.7 is the first OpenBSD release in years that broke the 6-month cycle, and it's less than 3 weeks late. It's been years since there was a Debian-stable release, and after the new one is released Real Soon Now there's no reason to believe the subsequent one will be relased any sooner.
"The free BSDs are almost as distro-like as something like Debian, in that they write and maintain a small core set of software, and then package, test, and include 3rd party (free) software with changes to config management and packaging to fit their system, and often a few of their own changes to the code."
That's not correct. Most of the OpenBSD userspace is maintained directly by the OpenBSD team. They have ports to make it convenient to install 3rd party software, but the base system is mostly privately maintained software.
Slashdot Mirror
The XBox PowerPC processors are not designed for general purpose work. They're in-order, they have a long pipline like the Pentium 4 and they have crappy branch predition. Just being PowerPC doesn't make them good at every.
I noticed that the Star Destroyers and fighters shown in that scene were different from the ones we'd seen in the rest of the movie. The Star Destroyers were white instead of white and orange or whatever, and IIRC they were structurally different as well; they didn't seem to have the docking bay doors on top like the earlier ones. The fighters were the first ones in the whole movie that had the distinctive TIE roar.
With entirely new classes of ships in service, it seems likely that the scene took place years after the rest of the events of the movie. No way to tell how many.
Also, the construction of the second death star may have begun at any time. When something takes that long to construct, it's common (for us anyway) to build many in parallel, and as others have said the better experience and industrial consolodation of the Empire may have reduced the construction time of the second one.
I think he's thinking of someone that has theoretically not seen any of the movies, and is watching them all at once.
Of course, you'd have to be pretty determined or pretty young to not know the important parts by now.
Isn't there an orwellian definition of non-toxic that means "It won't leak, so it's not toxic."?
That bubble burst pretty quick.
"But the benefits of this technology are not to extend processor life, or primarily to decrease power consumption."
Electromigration is greater at higher voltages, so I would say going from 1.4 volts to 1.1 volts would increase the processor life. It's not a primary motivation, but it certainly doesn't hurt when deciding to buy the thing.
Heat, power consumption, and noise are all directly related, so an attempt to improve one necessarily improves the others.
OpenBSD is biglock; that is, only one of the processors can be in the kernel at any given time, the entire thing is protected by one big lock. The other processor can be running something in userspace no problem.
This isn't a problem when using a limited number of CPUs (say 4 or less) because the probability of contention is low enough that it's not a big hit on performance for most tasks. For example, Linux 2.4.x and FreeBSD 4.x were biglock and people seemed pretty happy with their performance. People even complain about performance regressions compared to the older versions.
Linux and FreeBSD have moved on to finer grained locking. It's an advantage on systems with more CPUs, but the advantage is questionable on 2-4 processors because there's a lot of overhead, it's a HUGE amount of work, and it takes a long time to get right.
The post you replied to might have been joking and it was certainly wrong.
"When I asked a simple question regarding OpenSSH, he proceeded to pick a fight over something extremely inconsequential in way it was phrased."
Funny, RMS did exactly the same thing last week for me...
A number of OpenBSD developers will be speaking at the CUUG meeting on tuesday the 24th. It's extremely interesting to see them discuss the stuff they do, and it's a good opportunity to ask questions.
"I wonder if the copyrights holders associations would have less of a problem with Peer to Peer networking if the source and destination of the files (and who placed the files in the p2p network in the first place) were clearly recorded.
.torrent file repositories did they get results, and it's easy to find those regardless of what records they keep.
After all, If you "only read Playboy for the articles", you should not object to getting a copy with all of the pictures removed.
If you want to protect BitTorrent and other kinds of P2P, Make them be responsible citizens."You assume innocent people should
That's idiotic. Additional records would make no difference because BitTorrent makes no attempt to conceal the participants now. When you connect to the tracker, you request a list of IP addresses and it gives them to you. That's as transparent as you can get.
The only reason that hasn't stopped people from using it is that the RIAA and MPAA can't sue people fast enough. They already have the identities of more people than they can sue. Only when they started suing the trackers and
That whole "if you're innocent you have nothing to hide" attitude is offensive anywhere it happens. It is not the responsibility of the individual to record their activities in order to later prove that they didn't do anything wrong. It never has been.
The only reason the RIAA and the MPAA have a problem with BitTorrent is that it's so mind-bogglingly efficient. It's capable of distributing multi-gigabyte files to an arbitrary number of people with nothing more than broadband Internet access. It's completely content agnostic, it just happens to be a lot better at what it does than any of the alternatives.
I don't see this as a positive thing. I see this as further evidence that Apple doesn't spend enough on QA.
This will be about the tenth time my iBook has had something wrong with it. In this case the inconvenience is very small because I can continue to use the computer until the replacement arrives, but Apple's consistent failure to do good QA guarantees that this will continue to happen in the future.
They're not any worse than other OEMs, but they're not any better.
Makes you wonder why they don't spend more on QA upfront.
"Running a CPU at lower than its rated speed is just plain silly. ... :P"
So you see, there is no point to what you're doing, and I think its stupid
Only if the assumptions you made were correct. Given the information you seem to have, your conclusions are reasonable. But you're missing something.
Newer desktop Athlon64s have laptop-like power-saving features. Basically, the chip I have (Athlon64 3000+) can run at 1.0 ghz or 1.8 ghz (changes between 1.1v and 1.4v). This can be changed dynamically at any time. I'm not doing anything at all weird, it's supported in the chipset, the CPU, and the OS without any modifications by me apart from installing appropriate software.
So it's only underclocked when it's idle, and it's handled automatically so I can't tell the differnce except for the benefits.
And here I got one of those new Athlon64s because I wanted to be able to underclock it.
It reduces stability, it decreases component life, and it increases power usage and heat. If you want to do it, I'm not going to stop you, but I'm not going to complain if Intel and AMD come up with a way to effectively prevent it, and I'm still going to think it's stupid.
"People claiming BitTorrent is "perfectly legitimate""
The OpenBSD mirrors are completely hosed today because 3.7 has just been released, so I downloaded it with BitTorrent.
If that is other than "perfectly legitimate" in any way, please let me know so I don't do it again.
I'm on Shaw Cable (Calgary). I can load the page just fine.
"But PF and OpenSSH were /started/ at a hackathon and are now open projects and open for much discussion. My point in my original post was not taken in the right way. I meant to say that if Theo wants everyone to just code and not discuss their methods, then he'll end up with a million different implementations of something each with its own set of problems."
As I understand it, they've already worked out how things are going to go well before anyone steps on a plane. They've done most of the design work.
As I interpret the "shut up and hack" quote, they mean by the time everyone arrives in Calgary it's time to implement what they've discussed, not rehash design decisions that have been talked to death.
"Buy the official CDs and support the project, roll up your sleeves and make your own or use another OS. It's a free world."
It's easier to just do an FTP install.
Well, once the mirrors calm down in a few days anyway.
My setup: I use OpenBSD 3.6 (no upgrade for me yet, mirrors are totally) as a firewall and server (NFS, web, SSH, DHCP, that sort of thing). I use Debian-testing as a desktop on a different machine.
The documentation is second to none. That includes all the Linuxes I've tried as well as the BSDs. The fact that it's actually worth reading the docs means you do it early and often, which is nice.
Installing things out of ports is about as easy as any of the other good package managers I've used on Linux. The one thing to be aware of is that pkg_add can take a filename argument in the form of a URL to the FTP site. It's better to do that because it can resolve dependencies from the FTP site instead of you having to download them manually.
DO NOT ROLL A CUSTOM KERNEL. The generic one supports virtually everything that has been tested properly. If you have a problem with generic, report the bug because they'll want to fix it.
The firewall, PF, rocks. It's easy to set up and very powerful.
Overall, the differences between distros are as big as the differences with OpenBSD. I had to go through a lot of distros before I found the one I was happy with, and if you can wrap your brain around, say, both Slackware and Suse, you'll be fine.
I don't use OpenBSD as a desktop system so I don't know how well this works, but it might be worth a shot to set up Linux binary emulation and just using the Linux binary from mozilla.org. That might work better anyway for flash and stuff.
I use Debian-testing as a desktop and I avoid apt for the small number of things that I want to be 100% up to date with, including Firefox.
I prefer to code C on OpenBSD because the man pages for the C library calls are better. I can port it to Linux later if necessary. I don't use OpenBSD as a desktop OS, so I don't know what graphics support is like. They seem to be running a recent X.org, so it can't be that bad...
I like OpenBSD better in general, but Debian is a lot less work as a desktop OS and I'm lazy. OpenBSD runs my firewall/server and Debian-testing my desktop.
The documentation is MUCH better than any Linux I've used, which may be an edge if you're starting out.
"You don't see OpenBSD winning any awards for number of installations, do ya?"
OpenBSD, no. It's a niche OS. It doesn't surprise me that firewall machines are outnumbered by other machines. That doesn't say anything about the quality of OpenBSD code.
The fact that PF (written at a hackathon) has displaced IPF and IPFW as the BSD firewall of choice speaks to the quality of OpenBSD code.
The fact that OpenSSH is by far the dominant SSH implementation speaks to the quality of OpenBSD code. 90% of all SSH servers are OpenSSH, and a good chunk of the rest are re-branded OpenSSH servers.
" while Darwin may be based heavily on BSD, are they actually contributing anything back?"
Sometimes yes, sometimes no.
KHTML is huge, and the Apple changes are huge. They've simply forked the Safari version from the KDE version because there's too many changes happening too quickly to keep the two in sync. Their contributions back are in a form that is largely useless to KDE (a big blob, with no way to tell what a change fixes and no way to tell what it depends on).
OTOH, they submit reasonable patches back to FreeBSD, as their changes are relatively small, and are generally in response to specific bugs rather than large changes in functionality. In these situations, a patch that is accepted upstream is one more patch that you don't have to maintain yourself, so even though the BSD license doesn't force them to release the code they do it anyway.
*looks out window*
We're in the mountains?
*squints*
Well, I can see some mountains. They look pretty far away though.
"I disagree. Debian especially is very focussed on quality, and each package having a very good ratio of maintainers."
I use Debian-testing as a desktop system, I agree with your statement, but Debian has plenty of its own problems. For example, the stable release schedule is pretty tragic.
Conversely, 3.7 is the first OpenBSD release in years that broke the 6-month cycle, and it's less than 3 weeks late. It's been years since there was a Debian-stable release, and after the new one is released Real Soon Now there's no reason to believe the subsequent one will be relased any sooner.
"The free BSDs are almost as distro-like as something like Debian, in that they write and maintain a small core set of software, and then package, test, and include 3rd party (free) software with changes to config management and packaging to fit their system, and often a few of their own changes to the code."
That's not correct. Most of the OpenBSD userspace is maintained directly by the OpenBSD team. They have ports to make it convenient to install 3rd party software, but the base system is mostly privately maintained software.