no one came down with it. it was only harmful to monkeys. i did hear a romour that there were other tropical diseases that some of the workers caught from the monkeys, but that was unrelated.
i live in and grew-up in reston, btw. the monkey house (which was torn down and a new building is just now being completed on that lot) was across the street from a mcdonalds, a pizza hut, and a taco bell. a few friends of mine in high school actually broke into the monkey house, after it was boarded up, and grabbed some souveniers and took some pictures. and no, they didn't catch ebola or anything. i believe the monkey house was disinfected by sealing the building and boiling off formaldehyde vapor.
Tricks such as "register
renaming", "out of order execution", and "predication".
In other words, if the programmer won't fix the code, the chip will do it for him
First he's complaining the intel's cpus place too much demand on compilers, now he's complaining that the chip is optimizing instructions. I guess intel can't do ANYTHING right! Also, 1) afaik, there is no ISA that allows for register rename hints, 2) out of order execution is useful for doing hit-over-miss. Compilers can't predicted cache misses.
Quote:
The PowerPC G3 and G4 chips use much the same tricks (after all, all these silicon engineers went to the same schools and read
the same technical papers) which is why the G3 runs faster than a similarly clocked 603 or 604 chip
G3s are NOT faster than 603es at the same clock speed. G3s use 603 based cores.
Quote:
The theory with RISC processors, which has long since proven to be bullshit, was that by making the instructions simpler the chip
could be clocked at a higher clock speed. But this in turn only made up for the fact that more instructions were now required to
implement any particular algorithm, and worse, the code grew bigger and thus used up more memory. In reality a RISC processor is
no more or less powerful than a CISC processor.
Funny that he mentions code bloat on risc machines, but not the massive code bloat that ia-64 brings. Anyone who thinks risc and cisc are the same needs to go learn some x86 assembly and then some alpha assembly. Compare the clean alpha ISA to all the silly x86 instructions that only work on certain registers. Also, much of the code bloat from a risc ISA is due to the instructions all being 32bits long. If all x86 instructions were 32bit aligned, they'd take up more space, too. I really doubt this guy has that much assembly experience, since he seems to think that cisc instructions (eg, x86 mov) actually do more work per-issue, than risc instructions. yes, mov can do more than stq (it can both load and store), but how often does mov actually do both a load and a store (eg, mov addr1,addr2) at the same time? the real difference between risc and cisc is that risc instructions are more specific and cisc ISAs place more of a burden on the cpu instead of the compiler.
the *only* types that're expanded on 64bit arches are pointers. ints are still 32bits!! yes, the pointers take up twice as much space, but regular ints do NOT. that is why you can't store a pointer in an int on 64bit cpus (although, you shouldn't ever do that anyway).
It will be a trivial task to port solaris to MIPS. (snip) Linux on MIPS and ultrasparc is just reverse engineered drivers and hacks.
you need to get a clue. let me guess how many OS you've ported to different archs: none. anyone who thinks porting an OS to a different arch is "trivial" has no idea what they're talking about. also, i'd love to have you explain why linux on mips and sparc64 are "hacks" and let me know what reverse engineering took place.
there several reasons why using a ramdisk instead of a disk-backed fs is a bad idea. eg, no nonvolatile backing store (unless you run a backup script every so often, but that's much worst than most filesystems' automatic syncing). also, it's permanently wired in-core which means non-optimal usage of ram. why waste space on files that are rarely accessed?
you don't buy cds, but you do buy dvds? that seems to imply that you believe that it's better for a company to try to restrict free speech (decss source) than to protect their own copyrights.
another thing: a lot of you kiddies jump all over anyone who violates the gpl without giving them a second to explain and correct the violation, while at the same time you violate licenses on music and then complain when the riaa, etc. try to enforce their licenses. in other words, you believe in copyrights only when it suits you. you sure don't like it when people violate the gpl, but you don't mind posting mp3s of copyrighted music on napster.
if you look at all the other software compaq has available for free (as in beer) download, it probably has exactly the same license. someone at compaq who's not familiar with the gpl more than likely added that standard license, not know any better. wouldn't an email to compaq explaining the situation do more good than posting this on slashdot? does every company that makes an honest mistaken need to be dragged into the town square?
altivec has kernel support and motorola has patches for gcc (though i don't know if they've been intergrated yet or if the altivec C extentions will be allowed).
as for smp and altivec, saving altivec contexts is no more difficult than saving fpu contexts on an smp system.
Though, it all seems pointless to me anyway. People by Apples so they can use MacOS.
not everyone buys a mac for mac os. just because you do, doesn't mean everyone else on the planet does.
Why spend all that money and then waste it by running software that wasn't designed for it?
for the same reason people buy x86 boxes and run non-windows OSes on them. it almost sounds as if you're questioning why someone wouldn't want arch and platform dependant software.
It'd be like getting a proprietary SGI (O2, Indy, etc) and running Linux on it instead of IRIX. The user experience would be gone. So it is with Apples.
that's retarded statement. is the "user experience" lost when someone runs mac os on an old mac clone? or solaris on an x86?
drivers are not written in asm. drivers for the vast majority of OSes are written in c, obj-c, or c++. the only asm they contain is in asm macros or calls to inlined functions written in asm, often only one instruction (such as a memory barrier op).
no, it only means a speed up when using 64bit intergers. the disadvantage (unless you need 64bits of address space) is 8 byte pointers. 64bit gprs are nice, but currently there isn't much need for them (in the consumer mainstream, anyway).
the only advantage other than more address space, is that 64bit interger ops will be atomic. that is, single instruction, instead of having to concatenate two 32bit values.
on a related note, linux (i mean the native monolithic kernel, not mk) can now run on nubus PDM machines ([678]100s) thanks to Takashi Oe. his has patches are available here.
1) linux runs perfectly fine on 601s and definetly runs on 7200s
2) mklinux (the kernel+vmlinux) runs perfectly fine on other dists, except for the fact that it's 2.0.x. only a couple things break. i have personally used debian under mklinux.
3) http://ppc.linux.or.jp/~toe/kernel monolithic linux on nubus pmacs
please don't mod that up, it's completely wrong. open firmware needs some extra.note sections (which netbsd has had for a while now) to boot elf kernels. this has ABSOLUTELY NOTHING to do with userland. this ONLY applies to the kernel image.
pure water is not conductive. only the contaminates and minerals in water are.
Re:Yeah, but when are we gonna effectively use it?
on
Maxtor's 80GB Drive
·
· Score: 1
1394 isn't 400 megabytes a second it's 400 megaBITS a second. that's 50 megabytes per second. you can easily max that out with a couple newer drives on the bus.
Your right. Wow. What brilliant insight. You must NOT be stupid and uninformed because you said "~advocate what would cause more corruption" which.. really makes no sense, but your still smarter than everyone who disagrees with you.
what exactly didn't you understand?
Lets put you in charge. We wouldn't want Americans voting in an election, because they may be uninformed.
sorry, i'm not running for office. i didn't say americans shouldn't vote. i DID say that i don't want the stupid ignorant ones voting.
Jackass.
your mom
Yep, this is a flame but I'm not baiting you. Your really a jackass.
you really advocate limiting voting and decision making power into the hands of/fewer/ people?
yes. i'd like uninformed mindless idiots to stop voting (regardless of whether or not they'd vote on my side). wouldn't you prefer people understand the issues they're voting on? if not, then why even bother to vote? just flip a coin.
How few?
i'd like a lot of people to stop voting; i get the impression that there are lots of ignorant, stupid voters out there.
Limited by education? Limited by taking a test? Limited by what?
currently they're limited by their own laziness and unwillingness to get off their asses and to the polls. personally, i'd like to see less of these go-vote movements (eg, rock the vote), and more movements to encourage people to educate themselves on issues before voting on them. if you don't understand what you're voting on, you shouldn't be voting!
Some places have very limited democracies. They're called/dictatorships/. We dont like those.
neither do i.
Whats going on is a recurrence of what has happened when a new "class" has been empowered to vote in this country - the status quo is being challenge and is scared.
what status quo are you refering to? democrats? republicans? both? if this new class is lazy ignorant people, i'd prefer that they don't vote.
Online voting now better enables the people that were too lazy to go out and vote, but cared enough to compose rants on the internet.
if someone doesn't care enough to even get off their ass to vote, do you really WANT them to?? i don't want every internet rant writer voting, either (assuming they fall into the ignorant or too lazy to goto the polls catagories).
you're also forgetting the online voting is EXTREMELY risky and VERY likely to be tampered with. it is IMPOSSIBLE for online voting to reach the level of security a physical polling center has.
It gives bite to our bark.
only if in their ignorance and laziness they happen to be on your side.
Arizona had its largest voter turnout in recent state history with internet voting. This scares the hell out of the politicians and those in control - they have an entirely new set of issues to deal with, because there are an entirely new demographic voting.
you're stuck in that mind-set of "more is good". voting is (or should be) about quality of voters, not quantity (though lots of quality voters isn't bad).
A very good example of increasing voter turnout causing unexpected results - Jesse Ventura - this is a state that had motor-voter registration, and Jesse managed to mobilize a large amount of disenchanted voters with common sense and straight talk.
as i've said, laziness and ignorance don't make for good voters. if someone doesn't want to vote, i say good!
/That/ is what this is about - losing control and predictability.
huh? you'd rather have choas and random election results than informed voters? what's wrong with predictability?
Slashdot Mirror
no one came down with it. it was only harmful to monkeys. i did hear a romour that there were other tropical diseases that some of the workers caught from the monkeys, but that was unrelated.
i live in and grew-up in reston, btw. the monkey house (which was torn down and a new building is just now being completed on that lot) was across the street from a mcdonalds, a pizza hut, and a taco bell. a few friends of mine in high school actually broke into the monkey house, after it was boarded up, and grabbed some souveniers and took some pictures. and no, they didn't catch ebola or anything. i believe the monkey house was disinfected by sealing the building and boiling off formaldehyde vapor.
Tricks such as "register
renaming", "out of order execution", and "predication".
In other words, if the programmer won't fix the code, the chip will do it for him
First he's complaining the intel's cpus place too much demand on compilers, now he's complaining that the chip is optimizing instructions. I guess intel can't do ANYTHING right! Also, 1) afaik, there is no ISA that allows for register rename hints, 2) out of order execution is useful for doing hit-over-miss. Compilers can't predicted cache misses.
Quote:
The PowerPC G3 and G4 chips use much the same tricks (after all, all these silicon engineers went to the same schools and read
the same technical papers) which is why the G3 runs faster than a similarly clocked 603 or 604 chip
G3s are NOT faster than 603es at the same clock speed. G3s use 603 based cores.
The theory with RISC processors, which has long since proven to be bullshit, was that by making the instructions simpler the chip could be clocked at a higher clock speed. But this in turn only made up for the fact that more instructions were now required to implement any particular algorithm, and worse, the code grew bigger and thus used up more memory. In reality a RISC processor is no more or less powerful than a CISC processor.
Funny that he mentions code bloat on risc machines, but not the massive code bloat that ia-64 brings. Anyone who thinks risc and cisc are the same needs to go learn some x86 assembly and then some alpha assembly. Compare the clean alpha ISA to all the silly x86 instructions that only work on certain registers. Also, much of the code bloat from a risc ISA is due to the instructions all being 32bits long. If all x86 instructions were 32bit aligned, they'd take up more space, too. I really doubt this guy has that much assembly experience, since he seems to think that cisc instructions (eg, x86 mov) actually do more work per-issue, than risc instructions. yes, mov can do more than stq (it can both load and store), but how often does mov actually do both a load and a store (eg, mov addr1,addr2) at the same time? the real difference between risc and cisc is that risc instructions are more specific and cisc ISAs place more of a burden on the cpu instead of the compiler.
the *only* types that're expanded on 64bit arches are pointers. ints are still 32bits!! yes, the pointers take up twice as much space, but regular ints do NOT. that is why you can't store a pointer in an int on 64bit cpus (although, you shouldn't ever do that anyway).
you need to get a clue. let me guess how many OS you've ported to different archs: none. anyone who thinks porting an OS to a different arch is "trivial" has no idea what they're talking about. also, i'd love to have you explain why linux on mips and sparc64 are "hacks" and let me know what reverse engineering took place.
there several reasons why using a ramdisk instead of a disk-backed fs is a bad idea. eg, no nonvolatile backing store (unless you run a backup script every so often, but that's much worst than most filesystems' automatic syncing). also, it's permanently wired in-core which means non-optimal usage of ram. why waste space on files that are rarely accessed?
in short: don't do it
you don't buy cds, but you do buy dvds? that seems to imply that you believe that it's better for a company to try to restrict free speech (decss source) than to protect their own copyrights.
another thing: a lot of you kiddies jump all over anyone who violates the gpl without giving them a second to explain and correct the violation, while at the same time you violate licenses on music and then complain when the riaa, etc. try to enforce their licenses. in other words, you believe in copyrights only when it suits you. you sure don't like it when people violate the gpl, but you don't mind posting mp3s of copyrighted music on napster.
if you look at all the other software compaq has available for free (as in beer) download, it probably has exactly the same license. someone at compaq who's not familiar with the gpl more than likely added that standard license, not know any better. wouldn't an email to compaq explaining the situation do more good than posting this on slashdot? does every company that makes an honest mistaken need to be dragged into the town square?
as for smp and altivec, saving altivec contexts is no more difficult than saving fpu contexts on an smp system.
Though, it all seems pointless to me anyway. People by Apples so they can use MacOS.
not everyone buys a mac for mac os. just because you do, doesn't mean everyone else on the planet does.
Why spend all that money and then waste it by running software that wasn't designed for it?
for the same reason people buy x86 boxes and run non-windows OSes on them. it almost sounds as if you're questioning why someone wouldn't want arch and platform dependant software.
It'd be like getting a proprietary SGI (O2, Indy, etc) and running Linux on it instead of IRIX. The user experience would be gone. So it is with Apples. that's retarded statement. is the "user experience" lost when someone runs mac os on an old mac clone? or solaris on an x86?
drivers are not written in asm. drivers for the vast majority of OSes are written in c, obj-c, or c++. the only asm they contain is in asm macros or calls to inlined functions written in asm, often only one instruction (such as a memory barrier op).
*sigh* there is no 2gig limit, it's 4gigs. if you're going to complain about something, at least get the facts straight first.
no, it only means a speed up when using 64bit intergers. the disadvantage (unless you need 64bits of address space) is 8 byte pointers. 64bit gprs are nice, but currently there isn't much need for them (in the consumer mainstream, anyway).
the only advantage other than more address space, is that 64bit interger ops will be atomic. that is, single instruction, instead of having to concatenate two 32bit values.
alpha is definetly NOT a VLIW cpu. also, VLIW is nothing new, intel had a VLIW chip out about ten years ago, which never caught on.
btw, there's a reason no one else is making VLIW cpus. it's because it's an inherently flawed idea.
there's one problem with that:
i s, can natively execute x86 code.
wine == wine is not an emulator
btw, merced/itanium/pentium4/whatever-this-weeks-name-
linux on alpha has always been 64bit. you're getting confused by intel's (piece of crap) ia64.
check out the openstep web site. it's suppost to open "sometime june 1997". heh, and i thought *i* was lazy when it came to writing wed content.
on a related note, linux (i mean the native monolithic kernel, not mk) can now run on nubus PDM machines ([678]100s) thanks to Takashi Oe. his has patches are available here.
yes, most of the ppc dists are redhat based.
debian runs on several archs, including alpha, mips, sparc, ppc, and m68k.
1) linux runs perfectly fine on 601s and definetly runs on 7200s
2) mklinux (the kernel+vmlinux) runs perfectly fine on other dists, except for the fact that it's 2.0.x. only a couple things break. i have personally used debian under mklinux.
3) http://ppc.linux.or.jp/~toe/kernel
monolithic linux on nubus pmacs
please don't mod that up, it's completely wrong. .note sections (which netbsd has had for a while now) to boot elf kernels. this has ABSOLUTELY NOTHING to do with userland. this ONLY applies to the kernel image.
open firmware needs some extra
pure water is not conductive. only the contaminates and minerals in water are.
1394 isn't 400 megabytes a second it's 400 megaBITS a second. that's 50 megabytes per second. you can easily max that out with a couple newer drives on the bus.
what exactly didn't you understand?
Lets put you in charge. We wouldn't want Americans voting in an election, because they may be uninformed.
sorry, i'm not running for office. i didn't say americans shouldn't vote. i DID say that i don't want the stupid ignorant ones voting.
Jackass.
your mom
Yep, this is a flame but I'm not baiting you. Your really a jackass.
ha!
yes. i'd like uninformed mindless idiots to stop voting (regardless of whether or not they'd vote on my side). wouldn't you prefer people understand the issues they're voting on? if not, then why even bother to vote? just flip a coin.
How few?
i'd like a lot of people to stop voting; i get the impression that there are lots of ignorant, stupid voters out there.
Limited by education? Limited by taking a test? Limited by what?
currently they're limited by their own laziness and unwillingness to get off their asses and to the polls. personally, i'd like to see less of these go-vote movements (eg, rock the vote), and more movements to encourage people to educate themselves on issues before voting on them. if you don't understand what you're voting on, you shouldn't be voting!
Some places have very limited democracies. They're called /dictatorships/. We dont like those.
neither do i.
Whats going on is a recurrence of what has happened when a new "class" has been empowered to vote in this country - the status quo is being challenge and is scared.
what status quo are you refering to? democrats? republicans? both? if this new class is lazy ignorant people, i'd prefer that they don't vote.
Online voting now better enables the people that were too lazy to go out and vote, but cared enough to compose rants on the internet.
if someone doesn't care enough to even get off their ass to vote, do you really WANT them to?? i don't want every internet rant writer voting, either (assuming they fall into the ignorant or too lazy to goto the polls catagories).
you're also forgetting the online voting is EXTREMELY risky and VERY likely to be tampered with. it is IMPOSSIBLE for online voting to reach the level of security a physical polling center has.
It gives bite to our bark.
only if in their ignorance and laziness they happen to be on your side.
Arizona had its largest voter turnout in recent state history with internet voting. This scares the hell out of the politicians and those in control - they have an entirely new set of issues to deal with, because there are an entirely new demographic voting.
you're stuck in that mind-set of "more is good". voting is (or should be) about quality of voters, not quantity (though lots of quality voters isn't bad).
A very good example of increasing voter turnout causing unexpected results - Jesse Ventura - this is a state that had motor-voter registration, and Jesse managed to mobilize a large amount of disenchanted voters with common sense and straight talk.
as i've said, laziness and ignorance don't make for good voters. if someone doesn't want to vote, i say good!
huh? you'd rather have choas and random election results than informed voters? what's wrong with predictability?