sawtooth support probably isn't too far off. it should be much easier to support new macs, now that there's a unified mobo design (ie, same chips different form factors). i've never used an ATI card, so i don't know how good the support is on PPC (or on x86, for that matter). i don't think there's 3D support (though i could be completely wrong about that), but regardless, ATI's new commitment to linux can only help.
as for altivec, since mac os X uses gcc, apple has a motive for adding vector support. the general impression i get (read: rumor) is that apple or moto will release a patch for gcc 2.95 in the not too distant future. i'm not a gcc guy, but i would guess it would be pretty simple to add the asm support for the vector instructions, and not too difficult to add the C extentions (just new data types?).
one of the nice things about PPCs, is the hint bit on branch instructions. it allows you to optimize for the common case. eg, a while(i 255) loop, the common case is i = 255. i don't know if x86 supports branch hints, but i suspect other RISC chips such as sparc, mips, or alpha do.
years ago, i bought specture(sp) vr and a paperback copy of snow crash was inside. i'm glad it was bundled and i'm glad i read it. it's definetly a good book and is pretty humurous at times, especially in the beginning.
bottom line: if you're looking for a good book to read, this is it.
it seems to me, that adding pipelines (moderate space cost) and execution units (higher space cost) would be bring more performance than higher clock rates. that is, if the same effort to increase clock speed was put into superscalar expansion, the pay off would be greater, provided there's enough space.
also, it seems to me, multiple short pipelines would yield higher preformance than fewer high clocked, deeper pipelines.
i believe the reason intel goes the faster deeper route (compared to slower, wider) simply cuz:
1) it's cheaper to deepen pipelines and it isn't too hard to get a good enough signal to noise ratio for higher clock speeds (is that even an issue?)
2) marketing. this is the obvious one. they can say "our chips are fast! more MHz than our competitors" and the general public doesn't know any better.
ps, please correct me on anything, i'm just guessing at some of this stuff
so you just moderated it down cuz you don't like linux? there was nothing "out of line" in that post. i was pointing out that SCO had basically said linux sucks, yet they now seem to think otherwise (unless, of course, they're just in it for the money, trying to cash in on linux hype, which wouldn't surprise me)
here's a good macworld opinion article, that sums up what most people probably think about this 50mhz reduction thing; they should have delayed the 500mhz models and left everything else the same. apple should have absorbed the lower yields/price increases, instead of passing their problem with moto unto the consumers.
as far as i can tell, it's not a bug. what happened is motorola over estimated their yields for 500mhz chips (ie, they're very low). i believe they're moving to a smaller process to fix the problem; from a.15 micron process to.13.
notice how they use "linux" where it should say "redhat". for example, not all linux dists ship with ftp servers and other nonsense enabled by default.
they're trying to discredit linux as a whole. MS has deep pockets for advertising, and naturally, zdnet wants to play nice with them ("look, we love windows, linux sucks! give us more ad dollars!)
this whole zdnet thing is right up there with the mindcraft "study". they're both biased and totatly unscientific, making any and all results worthless and bogus.
zdnet is helping MS fight linux with FUD. unfortunely, most people who read zdnet believe what they say (people believe things unless they know of a reason not to).
what i would love to see, is redhat spend some of those intel, etc. dollars on debunking these rediculous(sp), unprofessional, FUD stunts from zdnet, mindcraft, and everyone else who's in MS's pocket.
what we need is a central PR site that debunks each and everyone of these feascos(big sp), point by point, and also shows linux advantages. (ie, don't just debunk, but show linux vs. NT features, performance, etc). the site could also run benchmarks. i'd like to see redhat and some others get together and fund a site like this.
I'd like to see him use this service; ie, put his money where his mouth is. Bill, if it's secure (as I'm sure you'll claim it is), then why not? When security problems are found, you can bet his "e-wallet" is the first they'll go for.
chesse crackers are ok, but i prefer peanut butter. wow, dvorak really is a big man to take on the cracker industry. i mean, they must be huge; just think of how many crackers are sold in one day, alone! you go boy!
Yes, you're right. Alpha Microprocessors are a high-performance way to go, but they are really expensive
no, they're not
IBM's PowerPC Open Platform hasn't launched yet and the website is rather small at the moment, but it looks interesting. Is it possible to escape from these old x86 times?
yes, buy an alpha or a mac. the sawtooth mobo's are much better, anyway, and they'll be shipping very soon (much sooner than anyone will have an open PPC board shipping or even in production).
when will the madness end?! i can't believe AMD would want to extend such a horrible arch to 64bits. even intel isn't that dumb! i have no doubt that people will buy this, however, just as i have no doubt that people will continue to buy inferior products from MS.
hopefully x86 will die in the next few years (of course, it'll still be around in the embedded world).
or use them for near-line backups. that's kind of what i do. i have a 4gig cuda in my old box at home and i back it up onto two 1gig hawks. hard drives are often cheaper than tape, unless you have massive amounts of disk space; then tape is cheaper. when i first got a 1gig hawk years ago, i bought another one just to back it up. that effectively doubled the cost of my storage, but it was still a lot cheaper than buying even a QIC tape drive.
i can't believe so many people thought the jesux thing was legit. it was very obvious that it was a joke. i'm not surprised that the media fell for it, though.
That's true, but in Mach 2.5, at least, i believe the BSD API is an intergral(sp) component; not run as a server process like in Mach 3.0 (that is, 2.5 less of a microkernel than 3.0 is). Mach drivers (3.0, at least) are very BSD and, iirc, one or more of mklinux's drivers was actually taken straight from FreeBSD.
Slashdot Mirror
sawtooth support probably isn't too far off. it should be much easier to support new macs, now that there's a unified mobo design (ie, same chips different form factors). i've never used an ATI card, so i don't know how good the support is on PPC (or on x86, for that matter). i don't think there's 3D support (though i could be completely wrong about that), but regardless, ATI's new commitment to linux can only help.
as for altivec, since mac os X uses gcc, apple has a motive for adding vector support. the general impression i get (read: rumor) is that apple or moto will release a patch for gcc 2.95 in the not too distant future. i'm not a gcc guy, but i would guess it would be pretty simple to add the asm support for the vector instructions, and not too difficult to add the C extentions (just new data types?).
one of the nice things about PPCs, is the hint bit on branch instructions. it allows you to optimize for the common case. eg, a while(i 255) loop, the common case is i = 255. i don't know if x86 supports branch hints, but i suspect other RISC chips such as sparc, mips, or alpha do.
years ago, i bought specture(sp) vr and a paperback copy of snow crash was inside. i'm glad it was bundled and i'm glad i read it. it's definetly a good book and is pretty humurous at times, especially in the beginning.
bottom line: if you're looking for a good book to read, this is it.
it seems to me, that adding pipelines (moderate space cost) and execution units (higher space cost) would be bring more performance than higher clock rates. that is, if the same effort to increase clock speed was put into superscalar expansion, the pay off would be greater, provided there's enough space.
also, it seems to me, multiple short pipelines would yield higher preformance than fewer high clocked, deeper pipelines.
i believe the reason intel goes the faster deeper route (compared to slower, wider) simply cuz:
1) it's cheaper to deepen pipelines and it isn't too hard to get a good enough signal to noise ratio for higher clock speeds (is that even an issue?)
2) marketing. this is the obvious one. they can say "our chips are fast! more MHz than our competitors" and the general public doesn't know any better.
ps, please correct me on anything, i'm just guessing at some of this stuff
of course they want it cut! i have to say, it really doesn't surprise me. what do you expect from MS? they'll do anything to get things their way.
so you just moderated it down cuz you don't like linux? there was nothing "out of line" in that post. i was pointing out that SCO had basically said linux sucks, yet they now seem to think otherwise (unless, of course, they're just in it for the money, trying to cash in on linux hype, which wouldn't surprise me)
anyone know if the various mac resellers are going to change their prices to match?
wtf?! why was that moderated down as a troll? that is a VERY legit question. *sigh* some of these moderates have no clue.
that's not why they did it. they're just using that as an excuse to make customers think they're still getting as good a deal.
here's a good macworld opinion article, that sums up what most people probably think about this 50mhz reduction thing; they should have delayed the 500mhz models and left everything else the same. apple should have absorbed the lower yields/price increases, instead of passing their problem with moto unto the consumers.
hey SCO, i thought you said linux sucked. were you lying then or are you lying now? i'd really like to see you explain this one.
as far as i can tell, it's not a bug. what happened is motorola over estimated their yields for 500mhz chips (ie, they're very low). i believe they're moving to a smaller process to fix the problem; from a .15 micron process to .13.
encrypted IP over IP
"linux" does not have services enabled by default (it can't, it's a kernel). redhat, however, does ship with unneeded stuff enabled.
notice how they use "linux" where it should say "redhat". for example, not all linux dists ship with ftp servers and other nonsense enabled by default.
they're trying to discredit linux as a whole. MS has deep pockets for advertising, and naturally, zdnet wants to play nice with them ("look, we love windows, linux sucks! give us more ad dollars!)
this whole zdnet thing is right up there with the mindcraft "study". they're both biased and totatly unscientific, making any and all results worthless and bogus.
zdnet is helping MS fight linux with FUD. unfortunely, most people who read zdnet believe what they say (people believe things unless they know of a reason not to).
what i would love to see, is redhat spend some of those intel, etc. dollars on debunking these rediculous(sp), unprofessional, FUD stunts from zdnet, mindcraft, and everyone else who's in MS's pocket.
what we need is a central PR site that debunks each and everyone of these feascos(big sp), point by point, and also shows linux advantages. (ie, don't just debunk, but show linux vs. NT features, performance, etc). the site could also run benchmarks. i'd like to see redhat and some others get together and fund a site like this.
I'd like to see him use this service; ie, put his money where his mouth is. Bill, if it's secure (as I'm sure you'll claim it is), then why not? When security problems are found, you can bet his "e-wallet" is the first they'll go for.
minor correction: G4s => 450MHz are
eh? linux has had SMP for a while, and USB as well. USB has worked at least since the first imacs came out.
chesse crackers are ok, but i prefer peanut butter. wow, dvorak really is a big man to take on the cracker industry. i mean, they must be huge; just think of how many crackers are sold in one day, alone! you go boy!
no, they use an AMD K6-2
no, they're not
IBM's PowerPC Open Platform hasn't launched yet and the website is rather small at the moment, but it looks interesting. Is it possible to escape from these old x86 times?
yes, buy an alpha or a mac. the sawtooth mobo's are much better, anyway, and they'll be shipping very soon (much sooner than anyone will have an open PPC board shipping or even in production).
when will the madness end?! i can't believe AMD would want to extend such a horrible arch to 64bits. even intel isn't that dumb! i have no doubt that people will buy this, however, just as i have no doubt that people will continue to buy inferior products from MS.
hopefully x86 will die in the next few years (of course, it'll still be around in the embedded world).
or use them for near-line backups. that's kind of what i do. i have a 4gig cuda in my old box at home and i back it up onto two 1gig hawks. hard drives are often cheaper than tape, unless you have massive amounts of disk space; then tape is cheaper. when i first got a 1gig hawk years ago, i bought another one just to back it up. that effectively doubled the cost of my storage, but it was still a lot cheaper than buying even a QIC tape drive.
i can't believe so many people thought the jesux thing was legit. it was very obvious that it was a joke. i'm not surprised that the media fell for it, though.
That's true, but in Mach 2.5, at least, i believe the BSD API is an intergral(sp) component; not run as a server process like in Mach 3.0 (that is, 2.5 less of a microkernel than 3.0 is). Mach drivers (3.0, at least) are very BSD and, iirc, one or more of mklinux's drivers was actually taken straight from FreeBSD.