Slashdot Mirror
Compaq: Alpha is Better Than IA-64
Compaq released a document (it's in PDF format) that states that their Alpha is better then IA-64 (Intel next generation Itanium Processor). The document compares Alpha (and future generations of Alpha) against the IA-64 (I hate this "Itanium" name - where do they get these names anyway?). Certainly worth a read. What do you think, folks?
I still want a 64bit CPU
first post
alpha has been 64-bit for a while ... Itanium only now .. Hello Unix vs Microsoft ..
Alpha: tru64 or Linux. As much as I love *NIX, eliminating themselves from most of the market share, they can have the best chip ever imagined, but Intel will still whip their butts. Can we say Macintosh?!?!
I think the real advantage is when operating systems become 64 bit, windows probably own't be for a long while, but is linux going to be 64 bit when it is ready on x86 architectures?
This Wiki Feeds You TV and Anime - vidwiki.org
Unfortunately it doesn't matter. The sheep will keep buying Wintel PCs. In this industry "good enough" has always been, well, good enough.
Just goes to show what bad marketing can do. Tell the average Joe about the Alpha and he's like "Alpha what?" And, yeah, I want one too. :)
My dual p2 32bit procs are ok, but 64bit whoohoo!
-- to code or not to code
I am not an expert in CPUs and I haven't read a basic CPU architecture schematic since the original Pentium came out. Therefore I cannot judge the merits of this document.
I would like to point out that this document is from Compaq, so we must suspect that the document was written with a Pro-Alpha slant to begin with. Its like Intel coming out with a paper debating the merits of the Pentium III vs. the Athlon Processor.
Manung
Oops. My fault. Fixed. Thanks
Hetz (Heunique)
The alpha processors are not changing their niche in the computer market. They are ripping fast - and Dec first and now Compaq plays to the supercomputing crowd. The XP series motherboards and 21264 chips simply rip any other motherboard/chipset out there.
However, they cost too much for anyone except a supercomputing hound. If Compaq would drop Dec's insanely idiotic OS and component licensing scheme and aid linux on alphas, they might stand a chance of making a LOT of money selling hardware. As is, people buy ten times more alphas one chip generation late and run linux instead of OSF.
Anyone interested should see the linux alpha compilers available. cc is a small improvement, and ForTran is a LARGE improvement.
http://www.unix.digital.com/linux/software.htm
But still, Itanium will come out, and an Itanium box will offer slightly less than half the floating point speed, and it will cost about 1/4th of the fast alpha box from Compaq. And the alpha motherboards will still make it tough to support third party peripherals. And Itanium will dominate the 64 bit market. And Alpha will own the supercomputing market.
Everything I've read so far, even withstanding Compaq's propaganda, indicates that the Alpha will be
faster
cheaper
*and* more powerful than Merced.
Not to mention that the Alpha, anyways, is proven technology.
Now write this down 100 times definitely definitely definitely definitely definitely ...
The PDF format is hardly Compaq's. If anything, it's Adobe's. Plus, the poster confused "than" with "then", /again/. :P
To the editors: your English is as bad as your Perl. Please go back to grade school.
It's going to be tough for Digital to edge into Intel's market, mainly because nearly all consumers have been brainwashed to look for the "Intel Inside" Logo.
"Excuse me sir, is this an Itanium?"
"No, Ma'am. This is an Alpha processor by Digital corporation."
"Well Shit, I've never heard of THEM. Where are your Itanium machines?"
Not only that, but Alphas have never really been geared toward the general consumer. Most have been high-end server machines. Also, as far as I know, Alpha won't run x86 code because it uses a different architecture. (Please correct me if I'm wrong.)
"Alpha, huh?
"No Ma'am, this machine runs a Unix variant, and has a different architecture than Intel processors."
"Well Shit, I NEED those programs. Where are your Itanium machines?"
-- Give him Head? Be a Beacon?
-- Give him Head? Be a Beacon? :P)
(If you can't figure out how to E-Mail me, Don't.
They both look like good CPUs so who cares which one is the "best". They'll both be good for most things so buy the one you like best.
Sig is taking a break!
The best feature of the Alpha is that they are available. Or are they?
I've wanted an Alpha for a while now because (for various geeky reasons: fun, supposed speed, fun, assembly programming, and fun) but I've never been able to find a reasonably priced machine (even for auction) OR good instructions on how to build them.
If Compaq were smart (note the use of a counterfactual conditional) they'd hype Linux on Alpha like all get out. What better way to screw MS than to give geeks hardware that Windows can't touch (anymore)?
But does Compaq want to screw MS? If they're smart they do: Compaq produces an ostensibly competing OS.
---
Linux MAPI Server!
http://www.openone.com/software/MailOne/
(Exchange Migration HOWTO coming soon)
The coolest thing about the paper is the details, such as they are, about the EV8. It's essentially four CPUs slammed together on the same chip talking to one another, and if the OS supports it you get four separate threads running at once.
This has the potential, along with a big cache, to really boost the performance of a box, as well as drop the price per bang down. SMP circuitry's not cheap or simple, and definitely non-trivial to design. But with the EV8, it's all been done for you...
Are you daft? Why are they sheep when they buy the best performing hardware that allows them to run the OS that allows them to run the applications they want to run? To any but the pathetic, die-hard nerds out there the OS doesn't matter much, and the hardware matters only so far as its price and bang for the buck. The applications matter. And gee, guess who the vendors are writing all their software for? Sheep? Nah. You're just a stupid freaking nerd with no grasp of reality.
So Alpha is fast and well designed and beats the shit out of x86 and IA64... What else is new? It's also irrelevant to a large majority of computing enthusiasts because the systems are really really expensive relative to the x86 boxes we can all go out and buy at Comp-U-Planet.
I know a lot of people who would absolutely adore having an Alpha box, but they're just so expensive... We have a variety of free high-quality OSes working on Alpha and we've got millions of people who are now re/entering the land of *nix. Put 1 and 1 together and you get a large potential market for low-end, moderate-cost alpha boxes... My question is, where can we find them and what's holding up the market from bringing them to us at a sane price? Are we not looking hard enough or are they not there?
-troll taker
truly, is s/w being written because the OS is popular, or is the OS popular because the s/w is being written?
I am, therefore you think.
"...better then"? How hard is it to write decent English? It's as if folks here _want_ to sound like ignorant geeks. Ach, our educational system is a mess...
I'm also not an OS (or arch) expert, BUT I think another advantage is data bus (and register) size. You've already mentioned the address bus with your 4GB physical RAM, but you've neglected to mention that you can get (and use) a full 64 bit word from that address with a 64 bit OS.
---
Linux MAPI Server!
http://www.openone.com/software/MailOne/
(Exchange Migration HOWTO coming soon)
There's no doubt that IA64 and its succesors will be hugely succesful due to this fact alone. Even though linux and other OSs are available for other archictures, lets face it, Intel is still the only game in town for almost all of us.
Alphas certainly deserve the market for high-end 64bit PCs, as it is a proven architecture that has been around since 1991 (or so). Even though the Merced is brand new, Intel will market it enough to dominate. I bet the Merced will be even more expensive than the Alpha! (Secret hope: there will be some hideous bug in Merced - like the original P5 and Chipzilla will fall flat on its face!)
To begin with, this is only the first version of the Itanium line of chips. For all we know, by the time the chip ships in SOHO computers, things will change. Also, the IA64 has one major advantage over the Alpha, the IA64 can run 32 bit programs (albeit at a considerable performance hit). For the first few years of the new chips, 32 bit software is what allot of people will use, so most computer manufacturers will only sell the IA64... Look, I am not anti-alpha, nor anti anyone else, I like competition, it lowers prices, but a realist must take all such releases with a grain of salt (Compaq is pushing its AMD line over its Intel line of machines, maybe they are trying to help AMD out, who knows). Either way, I look forward to owning a 64 bit computer someday.
Why is it that people always hear what I say, and not what I mean?
How do they come up with these processor names, you ask? An astute question, one that requires some of Intel and AMD's most closely-kept company secrets. A friend of mine who used to work for Intel managed to smuggle the following Perl script out, shortly before he was fired. Here it is:
./pnames.pl ./pnames.pl ./pnames.pl ./pnames.pl ./pnames.pl ./pnames-pl
#!/bin/perl
# Copyright (C) 1997 Intel Corporation
# This is a proprietary Intel perl script.
@prefix = ( "Pent", "It", "Max", "Ath", "Cort", "Trit" );
@suffix = ( "ium", "alon", "ex", "anium", "oricon", "agon",
"on", "eres", "obos", "ymede", "itan", "erion" );
@tag = ( "II", "III", "IV", "Pro", "MMX", "Deluxe" );
srand;
printf( "%s%s %s\n", $prefix[rand 6], $suffix[rand 12], $tag[rand 6] );
So if we run this script, we can see where the names come from:
sg1 237%
Cortium II
sg1 238%
Pentalon IV
sg1 239%
Penteres III
sg1 240%
Athalon Pro
sg1 241%
Pentitan II
sg1 242%
Maxymede MMX
Please show discretion when you refer this script to others. It is, after all, an Intel proprietary secret and should therefore only be shared with others on a "need-to-know" basis.
We're going down, in a spiral to the ground
While I don't know enough about chip design (and also don't really care) to judge the chips on their technical merits, the bottom line seems to be that it doesn't matter how good a chip you make if you can't ship/sell it in decent volumes. We all know that the early Intel chips were pretty much garbage, yet Intel today is the king of the chip world. Why? Because most (99+%) of the machines sold feature intel compatible chips.
As long as you can't go to an average computer store and pick up a PPC, Alpha or Sparc chip and build your own computer from it, the general population will not even know they exist. Don't get me wrong: I would like it if all of a sudden the availability of these chips were equal to the Intel chips, but that's just not the reality of the marketplace. With the switch to the 64-bit architecture there may be an opening in the market which will allow these chips to become a more available product in the eyes of the average consumer. But as long as the Intel/MS duopoly (which is showing signs of fracturing) is as dominant as it is now, that's just not going to happen.
I'll run an alpha KDE, sure. That's one thing. But trust your system to an alpha cpu?? I'll get one maybe after a few months of at least going Beta.
;)
Oh,
There are many organizations that still use VMS for their continuous computing appliations. Alpha is idea for that application.
Of course, there will be IA64 desktop boxes, and some people will buy them, but largely it will be wasted power.
Your general point is still valid though - the bottom line is that software out there today is compiled against x86, and this huge installed base makes it fairly easy for Intel to continue dominating the market (as long as their chips run x86 code faster than a comeptitors).
Ok, I'll try.
:-)
Macin... Linux.
Ma... Linux
Linux.
Linux.
Nope, it just doesn't seem to come out.
(the irony is I'm posting this from a mac)
"If one is really a superior person, the fact is likely to leak out without too much assistance" -- John Andrew Holmes
-----------
"You can't shake the Devil's hand and say you're only kidding."
It's the format for Word users who don't want to learn how to do HTML! Weenies.
Not able to find them at auction? *blink, blink* I have an alpha 500 sitting in a friend's basement that we went halfsies on 3 weeks ago and got for ~$850(US) on ubid. *shrugs* (sorry, this was offtopic, but the nice toys are available, or at least were before xmas ;-)
Rafe
V^^^^V
Rafe
Opinions expressed by the author may not actually exist in the wild.
For a while I was worried that when Compaq took over DEC they would just bury Alpha and forget about it.
It's nice to see them pushing it a little more heavily...I always thought DEC did a terrible job of marketing Alpha....
Looks like Compaq has a chip on its shoulder...
Ok. They're fast and lagging one generation, they are said to be affordable. Is it a recommended choice for someone only running linux? I mean, do I get twice the performance for the same overall price?
greetings,
Reinout
Reinout van Rees
because of a snowball effect that started because of a number of factors including marketing, timing, the push for homogenuity and standardization in client side computing..
the answer is obvious:
both.
simplification is stupid
I firmly want to believe that the Alpha is a better processor, but thinking so because of the contents of a marketing release from a manufacturer is prolly not the greatest of ideas.
:)
Sure, the Alpha has cool technology, and it always has, but "better" also include price-performance, something which Intel has historically been better-than-average at (until the advent of AMD, and in comparison to Motorolla/Sun/HP). There are signs that the Alpha may actually become affordable in the near future, and no one knows what Intel will do with pricing.
On a side note, there is a pretty interesting article on news.com about the past year's troubles for Compaq.. It blames most of the issues on executive malfeasanace.
Go Alpha!
--
blue
i browse at -1 because they're funnier than you are.
Why does everyone assume more is better? I compile and run lots of code on 64 bit MIPS powered SGIs. I can use the 64 bit compiler or the 32 bit one. I have tried several tests and the 64 bit stuff is never any faster. The only advantage I'm aware of is a larger address space, but this is only rarely helpful to me. Are that many of you guys running apps that bennefit from a 64 bit address space?
I have no idea where they came up with itanium, but they got pentium from latin (the root word means fifth). Now, we all know why the pentium pro was called the pentium pro, because they din't want to call it sextium
What I find more interesting is the fact that Compaq is *now* trying to promote Alpha's worth to industry. Digital seemed to have a cool road-map for the Alpha all layed out and as far as I can tell Compaq has killed all that with the reduction of R&D and fabs involved in the Alpha. As it stands, the raw statistics show a definate advantage to the PPC as now being the best processor for a number crunching Linux box.
Btw, speaking of Intel & Compaq, what the heck is going on with StrongArm?! I remember the road-map for it also being pritty cool stuff and then it fell off the face of the earth.
(smile) overhead tells the story and (imo) it will
take 5 years to eliminate 16/32 from ia64.
There's no way Compaq is going to release a comparison that doesn't make the Alpha look stunning and the IA64 like a 386SX if they can't help it.
There's no way anyone can say Alpha is better than IA64 without some solid benchmarking, and this PDF file only quotes benchmark data for the Pentium and PentiumPro, not the IA64.
Until a Third Party with No Vested Interests (aka TPwNVI[tm]) can independently compare Alpha with IA64 (with widely recognized benchmarks and other objective tests), take anything you hear from Compaq and Intel with a grain of salt.
_______
computers://use.urls. People use Networds.
So despite being the best and being on the market since 1990 thy're essentially impossible to get a hold of for the average home user.
One would think that Compaq/DEC would take advantage of IA/64's long "time to market" to crank out cheap Alpha chips *now* for use with Linux and *BSD. Having a large installed base might now be essential but have some kind of installed base is ...
To my mind DEC and now Compaq haven't shown a real commitment to the hardware that would make consumers confident. They say everything's still a "go" despite lagging MS interest in Alpha (the only other architecture NT runs on) but can money go where mouth is soon please?
The Alpha 21264's SPECint95 and SPECfp95 results _now_ are higher than the projected results for the as yet unreleased IA-64 processor. Unfortunately, the Alpha's suffered from poor marketing and and business decisions.
The big thing here is that even if Alpha can best an itanium, Dec/Compaq have pissed away their lead. Realistically how much faster than a P3 or G4 is the fastest alpha right now? (parts you can get, not the 900Mhz 264 that has been promised for about 18months now) On the order of 30% or so? There are specialized apps where you can get a more substantial improvment but in terms of general performance you're looking at something much closer to 30 to 40%. The x86 parts are much cheaper, the motherboards are plentiful, the software is plentiful. It doesn't take a genius to see where this is going. A few years ago, alpha had bragging rights and something to stand behind but now I can go buy a g4 power mac that runs cool and it can nearly perform as well as an $8000 alpha station. All Intel has to do is come close to alpha performance, which they will do by default, and have a good cost point and the 64bit market will be theirs for the taking.
True, Alpha and Merced will be the things that every one wants, but it is just not practical. Also, Im not a big linux user, but isnt the best practical chip for linux right now a PPC or G3/G4???
Lets remember that Crusoe will be announced in about 3 weeks.
my $2.e-2
You can find inexpensive Alphas. www.dcginc.com sells complete Alpha systems for $1800-$5500 and bare bone systems for much less. Alphas CAN run 32 bit code under NT using the FX32! Emulator.
Hey guys do you want alot more people to start using Linux on Alphas? Start doing some software porting!!! You guys claim you support RHL for the DS10 Webbrick? Start making some Alpha RPMs and posting them on rpmfind.net or some Compaq site! It's really annoying not being able to use Netscape, StarOffice, ssh, MySQL, Oracle, or the countless other number of apps and services available on the x86 side. And please make tar/gz'ed binaries available if not RPMs.
Sanity.html - Error 404 not found
Let moderators make their own decicions, most of them are not idiots you know.
Hmm. Does Compaq know what an order of magnitude is?
MS didn't kill NT on Alpha, Compaq did. Turns out nobody wanted to run a 32bit OS on 64 bit hardware. :)
--GnrcMan--
Hahahaha! Everyone on Outlook!! It's a "standard" and from now on you'll be able use it on any computer!
See subject
I'm going to start putting together a new linux box sometime in the next year, and am still open to the choice Intel vs Alpha. I want the box to be really spanking fast, of course, and it sounds like alpha is *still* going to be outperforming Intel even with the Merced (which is disappointing in a way, because I thought the whole point was that Intel were finally going to have a chip that could compete with alpha on grunt speed). But what worries me is: you can only buy Quake 3 in binary form, and I think I saw somewhere on the quake website that they're *only* releasing it for Intel. Plea: *please* release Quake3 for Alpha ...!
-------------
ans =
NaN
If they want some exposure or to compete for the desktop market (Which is where the money is) they need to slash the price on the chips and sell them near cost. Sure they take a hit for R&D but the volume of sales should go up. If they don't have a good strong presence by the time IA64 hits, they may as well close up their doors and go home.
I'm trying to teach myself to set people on fire with my mind... Is it hot in here?
Sounds like the big house, or maybe the big screw...or both.
LOL, what an ignoramus! Let's slam YOUR grammar, dude. 'It's' is a contraction for 'it is', while 'its' is a possessive pronoun. As for the article, it was so boring and the responses here so predictable (Intel sux; but has better marketing) that we're left picking on the grammar of peasants like you who can't get it right EVEN WHEN YOU THINK ABOUT IT! Har!
That darned closed-source programming gets people into bad habits..
/bin/perl -w
First of all with current versions of Perl the srand call is not needed.
Secondly I would recommend using qw() because it is more legible for lists.
Thirdly a little information hiding works well. There is no need to have to synchronize the length of the list with the argument to rand.
And -w is always worthwhile
So rewritten we get
#!
@prefix = qw(Pent It Max Ath Cort Trit);
@suffix = qw(ium alon ex anium oricon agon on eres obos ymede itan erion);
@tag = qw(II III IV Pro MMS Deluxe);
printf ("%s%s %s\n", &rand_elt(@prefix), &rand_elt(@suffix), &rand_elt(@tag));
sub rand_elt {
return $_[rand(scalar @_)];
}
Not that it matters in this case, but good habits are good habits...
:-P
Cheers,
Ben
PS To get the code to look like code use the TT tag, and to get indents use . Warning, IE may mess up the indented space on a cut-and-paste...
My usual seat in the cluetrain is at A HREF="http://pub4.ezboard.com/biwethey.ht
You could also say it's the "pretty" version of postscript...
I know it's asking a lot, but i'd really like to see some tried and true benchmarks of both chips...
Not some "how fast they serve web pages" or how fast compiling a new kernel would run.
I want to see some mad scientist go in and dissect each chip and tell me that the Alpha can do X better, while the Merced (I refuse to use the I word) can do Y better...
I really don't trust a release from an obviously biased party. Of course, M$ would tout that NT outperforms Linux as a Webserver... And of course Compaq would do the same for their Alpha's.
Like it's been said previously, the Compaq corporate PR people can leave out the bad stuff if they want.
What we need is the Ted Nugent of Hardware Benchmarks...
"Fuzzy Wuzzy was a bear, Fuzzy Wuzzy had no hair... Fuzzy Wuzzy wasn't fuzzy was he?"
While I'm not a programmer and therefore can't really contribute much to the cause, I still wonder why there has been no effort to make an opensource emulator... x86 to Alpha... maybe x86 to SPARC... x86 to PowerPC.... and two way, as well... though i suspect the main use would be to run x86 software on other platforms... Wouldn't that be great? And why is there no effort that I can find?
Apologies for potential off-topic.
:)
Let's port this to all other languages like LISP et al.
I will do the easy one and port it to C.
// Copyright (C) 1997 Intel Corporation
// This is a proprietary Intel C program.
static const char * prefix [ ] =
{ "Pent", "It", "Max", "Ath", "Cort", "Trit" };
static const char * suffix [ ] =
{ "ium", "alon", "ex", "anium", "oricon", "agon", "on", "eres", obos", "ymede", "itan", "erion" };
static const char * tag [ ] =
{ "II", "III", "IV", "Pro", "MMX", "Deluxe" };
int
main ( void )
{
srand ( 0 );
printf ( "%s%s %s\n",
prefix [ rand ( ) % ( sizeof ( prefix ) / sizeof ( prefix [ 0 ] ) ],
suffix [ rand ( ) % ( sizeof ( suffix ) / sizeof ( suffix [ 0 ] ) ],
tag [ rand ( ) % ( sizeof ( tag ) / sizeof ( tag [ 0 ] ) ] );
return 0;
}
P.S. Yeah, I know, I should write a perl to C printer, but then the post would be too long.
For I am not master coder yet who can code a super short compressed one-line self-compiling compiler to fit as a post.
Any challenger care to respond with one?
P.P.S. Back to doing some real coding.
Corrinne Yu
3D Game Engine Programmer
3D Realms/Apogee
Corrinne Yu
3D Game Engine Programmer
I'm convinced that the IA64 has many neat features, and I have no doubt that Intel can answer many of these criticisms with their own analysis.
That being said, the SMT idea is really neat.
There are low-cost Alpha boxes out there: the Multia. DEC sold these as "Universal Desktop Boxes" running Linux in the mid-90's. Unfortunately, they discontinued them in 1995. You can still occasionally find them for sale on the web, though. I bought one on onsale.com last year, and I won (most of) another one in a raffle from thelinuxstore.com earlier this month. They have an appealingly compact "pizza-box" desktop tower design that looks like a fat closed laptop with integral ide, video, and sound support. They also use laptop parts (PC and PCI-bus cards), making them easy to customize. Unfortunately, the Multia CPUs are quite slow by today's standards - perhaps roughly equivalent to a Pentium 100 (just a guess) for the most common version with a "low-cost" 166MHz 21066 Alpha CPU.
You can find more info by searching the web; www.viking.org/lca.html is also a good starting-point.
If only COMPAQ would come out with an updated version with a faster CPU...
- Tim
ya know what, i puled out my first year latin book, and check up, i was wrong in the first place, but for the second processor (pentium pro), i techincly an right.
Both The cardnal and hte Ordinal Number's root word is sex for six.
And since this is like the name of something, it would be a Ordinal, it would be sextus, to bad sextus is indeclinable.
BTW, hex is used in science (is it from greek?), so thats probaly where pent came from.
Fifth in latin is Quintus.
Sixth in latin is sextus
btw, for you AC who called me Yuore ignurant.
you can ii ad hade.
Yummy quadwords.
:)
This sounds like some potential for floating point to integer optimization hacks, depending on the integer/register instructions.
Fun. Fun. Yummy.
Even yummier super large register files, such that those useless register pragmas actually work in compilers.
Corrinne Yu
3D Game Engine Programmer
3D Realms/Apogee
Corrinne Yu
3D Game Engine Programmer
I'm sure everyone noticed the portion of the doc that mentioned that proper use of the IA64 platform could inflate the size of binaries by almost 33%
33% ??
I guess Intel really is a 'Microsoft Strategic Partner! They're helping them with code bloat!!
.sig: Now legally binding!
Just a small nit: Simultaneous multithreading is hardly "new" technology -- new to Alpha, perhaps, but it was introduced in '94-'95 as a slight wrinkle on the 15 year old hardware multithreading concept.
I typically declare an ARRAY_SIZE function-like macro in my project header to avoid some of the code duplication you've got above.
:-) But for applications where good pseudorandom numbers are needed, it's best to use random() or some other facility (but there again, you break strict ISO C portability since random() is not a standard library function.)
#define ARRAY_SIZE(x) ( sizeof( x ) / sizeof( x[0] ) )
This makes it less awkward to use (and more intuitive for somebody reading your code who might be a bit C-handicapped and doesn't understand what you're trying to do.)
Oh, and why srand( 0 );? Why not use the current timestamp as a seed, or better yet, some manipulation of the current timestamp and the current process ID? (This of course, requires you to stray from strictly-portable C code, because it assumes you're on a platform that has process IDs and has integral time_t values.)
Finally, rand() and srand() are typically horribly inadequate in C (and presumably, in Perl as well.) Now obviously, for this joke example, it doesn't really matter.
We're going down, in a spiral to the ground
While a few of the points they are making about the IA-64 architecture are valid, there is a significant amount of fluff and propaganda in this piece. There are many mentions of 'insignificant performance gain' or 'does not justify the cost' which from my research seem unwarranted.
The thing to remember is that Intel is not stupid. They think they can make this work, and they are one of the few companies with enough resources to make it work. I wouldn't bet outright on intel, but I do think they know what they are doing and wont end up with a useless product.
-dennis towne
Alter Aeon Multiclass MUD - http://www.alteraeon.com
Alpha suffers from what I call the "Apple problem" - it's expensive, not easily available, and it's a lock-in to a single supplier.
To make it competitive, Compaq should drop the price and license any patents in a way that allows binary-compatible clones (without necessarily releasing patents on how to implement the chip). This allows a group of cheapskate cloners to provide a low-speed compatible entry path, while Compaq continues to trade on speed.
The authors make a big stink about the fact that IA64, by it's VLIW (or EPIC) nature, cannot dynamically reschedule instructions. Here's the three main points against IA64 in the paper:
Branch Prediction
The Compaq slant
Since all the scheduling is done at compile time, IA64 has (supposedly) no way to accurately predict which instructions to fetch during branches. As stated in the paper, compilers can't do a good job of this prediction, and in most processors special counters are used to keep track of how often branches go a certain way, to make them fetch the (hopefully) correct instructions the next time around.
How IA64 handles this
VLIW processors don't do branch prediciton, because they essentially process both branches simultaneously, and drop the results from whichever branch doesn't get executed. Perhaps I'm missing something in the author's argument, but it seems like he forgot this.
Register Renaming
The compaq slant
The author states that the large number of registers in IA64 would be better served with an out-of-order (dynamic reschduling) processor, but presents no arguments as to why this might be true.
The IA64 side
Frankly, this looks to me like a non-issue. I can't even see why the author brought it up. I would personally think that compile-time scheduling would have a much better shot at using the large number of registers effectively, since it can look further ahead in the instruction stream.
Dynamic scheduling + memory latency
The compaq slant
When a cache miss occurs, dynamically scheduling processors can simply keep feeding unrelated instructions through unused functional units (FU's) while those long LOADs and STOREs are taking place. IA64, by its very nature, cannot do this, since all the instructions in a particular "packet" must feed through simultaneously. Therefore, the processor stalls for however long it takes to go to memory.
How does IA64 fix this?
This is a sticky issue. Depending upon the particular program, cache size and structure, and various other factors, this may or may not come up. What the author does is present a worst-case scenario, and prove (rightly) that IA64 doesn't handle it as well. There is a good chance that this would be caught at compile time, and the compiler would change the cache structure to compensate. If not, dynamic recompilation can be performed, using cache miss information to restructure the code.
Anyway, that's just my 2 cents. Personally, I feel that things could go either way. VLIW gives you a big performance boost on the front end, with (ideally) a lower cycle time, more FU's and a larger register space. How IA64 will match up, no one really knows yet.
...detract from the discussion. Someone cares which CPU is "best" because they have to make purchasing decisions. And the new processors will be "just fine" for your OSS Quake cheating, etc, but if you want a computer to do something more productive (read: profit-generating), which processor is best or a better value (depending on your budget :) will be the one you buy. And that is very, very important, so it does matter.
At the least, this is a forum designed to promote open discussion of issues... if you don't take a stance on either side, what's the point? You're not running for the presidency... go ahead and say what you think instead of trying to please both sides.
Acording to insiders HP has been running the IA64 show for quite a while now, due to the inabilatys of Intel's engineers. The Itanium should come with a big "HP Inside" sticker.
I don't know about you guys, but I'm waiting for my Transmeta CPU! Just hope they don't name it something stupid like Javalon with GGX technology.
Be carefull, this is a joke.
Your Momma's so fat she makes emacs look like nano!
I can't help get the feeling that the chip get it's name from the Curator of the Planitarium in South Park.
"welcome to the Plani-arium"
Vs
"Our new chip will be called: I-anium!"
The secret of success is honesty and fair dealing. If you can fake those, you've got it made. (Marx)
Lessee, my buddy here tells me the prototype alpha came in around '90...and it's been 64 bit from the git-go. We're talking about a not-yet-release, brand new chip, vs. one that's been in production for nearly ten years.
For that matter, the InfoMagic CD that I picked up in '96 or '97 had a version of linux for the Alpha, and there's a current version, as far as I know...*and* Compaq has dropped all support of NT on Alphas (there was a slashdot item on that, months ago), and has announced support for linux.
Oh, yeh, and I b'lieve that you can get Alpha processor-based systems for the price of Xeons, even from sytem retailers *other* than Compaq.
Hell, where I work, the city's been running on Alphas since '95 or '96...and the new generation system that we've just been testing...well, when I reconfigure DEC UNIX, and it's about to rebuild the kernel, it gives a message that it may take 15 minutes...and then takes (I timed this) a minute and a half! (Old message, obviously)
So, tell me again, why is it that Merced is "new and revolutionary"?
mark
The first implementation of the architecture described in the report will not be available until late 2001 -- 1.5 to 2 years away. Of course, this is assuming Compaq delivers on time...and we all know the lousy track record for delivering new Alphas on time. Further, it will be manufactured on a .18 process which will be ancient history at the end of 2001. Why is Digital -- pardon me -- Compaq always a generation behind on process technology?
Does anyone actually think that Compaq would release a study that claimed IA-64 was superior to Alpha?
C'mon folks, this is as shocking as walking out into the rain and discovering that you are wet...
It's good to see Compaq finally starting some pro Alpha marketing!! Keep it coming!!!
If you take a look at those Linux benchmarks from that guy in Germany (search for the old Byte Benchmarks), you'll see that the Alpha is surprisingly poor in 32-bit mode. The Alpha's had a hard time beating the Intel chips, when the Intel CPU's were at a lower clockspeed. The Alpha's main strengths were DEC's own libraries. You compare Alpha and Pentium under the same version of Linux, and the Alphas come out to be a dog. All of this is in 32-bit mode, of course. But still, Compaq has a serious battle if they're going to try to beat Intel. I wish them the best of luck.
You can find inexpensive Alphas. www.dcginc.com sells complete Alpha systems for $1800-$5500 and bare bone systems for much less. Alphas CAN run 32 bit code under NT using the FX32! Emulator.
32 bit x86 code no less... Also, there is support for 32 bit x86 Linux binaries available (in Linux of course.) How well it actually works is best left for someone else to answer. I'm suprised that so many people thought there was no x86 emulation available.
Of course, the emulation isn't quite as important under Linux as it is under Windows since most software for Linux is open source and able to be compiled natively. Note that I am NOT implying that it's always as easy as simply recompiling the source...
BTW, doesn't seem like a great idea to go with an Alpha/NT combo these days anyway. Microsoft ceased development of NT5/Win2k/whatever for the Alpha. Presumably because they need to focus on rigging it to work with the IA64 first. I wonder if Windows for the IA64 will end up being enough 64 bit code to call it a 64 bit OS and as much of the old 32 bit code as they can get away running under emulation. Any guesses?
numb
moderators don't like being called idiots I guess...
idiots
What makes them sheep is why they "choose" and OS, application, or computer. They do what everybody else does. In almost all cases it has nothing to do with "bang for the buck", or what application is best. If that were true, MS would have went bust years ago. For example, why do MS IE users outnumber Netscape users? Because IE is better ? Nope (allthough it is IMO), because it's what came configured on there PC when they bought it at CompUSA. This, near absolute, control that MS has gives them the resources to outlast everyone else and eventually produce decent products. People as a whole are sheep, and it follows that they buy computers, OSs, and applications, etc. like sheep (if sheep bought computers of course).
heunique is the guy who posted the damn article, why are you moderating his post "offtopic"? You just wait for the meta-moderation to rain down on you like a tidal wave, you just wait...
Just wait for Odium and Itrogen!
fish and pipes
...is available here
--GnrcMan--
Not available?
Sure, the chips are expensive, but what Alpha processors need is marketing not necessarily pricing.
And 99% of computers are shipping with Intel processors? Guess you missed AMD having the majority of computer sales a couple months ago (at over 40%) or the iMac's surprising success.
- Michael T. Babcock (Yes, I blog)
This white paper is interesting, if non-objective. In my opinion, the authors are insufficiently careful to distinguish between irreducible architectural advantages and disadvantages and the (temporary) advandates and disadvantages resulting from current implementation decisions. They are also a little slippery about identifying which features are already present in Alpha implementations and which are not yet delivered (e.g. SMT).
The implementation of simultaneous multithreading is something I very much would like to see. I'm impressed that they're able to do it as simply as this paper seems to imply.
One Alpha advantage (one that I think falls in the irreducible category) that I've never seen Digital/Compaq play up is the angle of binary compatibility of the Alpha instruction stream across different implementaions of Alpha. A binary executable that the compiler has tuned/targeted to a specific implementation of Alpha will still run, perhaps not quite optimally, on a later implementation.
Out-of-order execution is key, here. Because the programmer (or compiler) have to be explicit (with memory barrier instructions) about dependencies that might otherwise be hidden, the instruction stream in the binary executable file documents an idealized instruction execution order -- but any execution order that achieves the same result is also acceptable.
More outstanding data fetches, larger out-of-order instruction queue and wider simultaneous issue all work together to transparently make the old code work better. I haven't seen where increasing the VLIW bundle from 3 instructions to 6 instructions, for instance, would be as transparent -- so there's a much stronger need to recompile and maintain separate binaries targeting the various implementations of IA64.
This is NOT the same old Pentium vs Alpha debate. The "of course the Alpha is faster, but to run any software, I have to emulate the 32 bit pentium." Everything I've read about Itanium says that it's pure 64 bit, that it breaks backwards compatibility from earlier Intel chips. What this means is that Intel will have to emulate, too. This is why the K8 has an advantage, they're keeping backwards compatibility with older intel chips. Ah well, just my .02.
If you do a lot of very CPU intensive tasks, the alpha is quite a bit faster than a comparable x86 box. ...) is x86 only. If you need any of those, An alpha is not the right choice for you.
Other stuff (disk I/O, etc) is not faster than x86, and some hardware (e.g. many recent 3D graphics boards) can't be used in alphas.
Also, you should be aware of the fact that most closed-source Linux software (StarOffice, Netscape, Civ3,
This message is provided under the terms outlined at http://www.bero.org/terms.html
I work on Tru64 Unix, and I can tell you most assuredly that Tru64 and Alpha are both available today, and both are thoroughly 64-bit!
I sure hope the anonymous Intel employee that wrote that script is reading this! :-)
:-)
I think that you can arrange that...
I will take your word on both his gender and Perl expertise as well.
Cheers,
Ben
My usual seat in the cluetrain is at A HREF="http://pub4.ezboard.com/biwethey.ht
Intel today is the king of the chip world. Why?
Because when IBM was looking for a CPU for their PC, Motorola couldn't guarantee they could make enough 68K's.
It's too bad, because the 8086 set PC software back 5 years.
(defvar prefix '("Pent" "It" "Max" "Ath" "Cort" "Trit"))
(defvar suffix '("ium" "alon" "ex" "anium" "oricon" "agon" "on" "eres" obos" "ymede" "itan" "erion"))
(defvar tag '("II" "III" "IV" "Pro" "MMX" "Deluxe"))
(defun random-element (list)
(nth (length list) list))
(defun generate-processor ()
(concatenate 'string
(random-element prefix) (random-element suffix) " " (random-element tag)))
Hmm.. Perhaps we need a new function in Emacs: M-x intel-chip
The Alpha, MIPS, and PowerPC architectures date from the era when the goal was one instruction per clock cycle and a nice, simple CPU with a fast clock. They're a good fit to that model. But we're beyond that now; all the major CPUs are superscalar, with elaborate internal scheduling and parallel execution. With all the scheduling machinery, the programmer-visible instruction set matters less, which is why Intel has been able to wring so much performance out of an instruction set that dates from 1971. With superscalar CPU technology, key RISC issues such as how many programmer-visible registers there are matters much less.
Very Long Instruction Word architectures, like IA-64, require very smart compilers. Which don't exist. One of the compiler groups spoke at Stanford last year, and admitted they were having major problems. All that compile-time scheduling is hard, especially since the optimal instruction layout depends on the relative speeds of different parts of the processor. Early compilers weren't expected to optimize well. IA-64 looked more like an attempt at product differentiation in response to AMD than real progress.
It's worth noting that the Pentium Pro/II/III have a 48-bit segmented addressing mode, allowing physical memory beyond 4GB. Nobody uses this yet, but it's in there. It would be a coup for some Linux vendor to support this, allowing Linux PC-type machines bigger than 4GB. There'd still be a per-process limit below 4GB, but the entire machine could be bigger.
I know some people in academia working on alternative architectures, but nothing looks really promising. You need at least a 2X gain to justify changing instruction sets, and nothing on the horizon provides that.
Incidentally, a machine that executes Java bytecodes isn't the answer. It's hard to make a fast stack machine; too much implicit sequentiality.
Is it any wonder that high-karma people and frequent visitors to Slashdot aren't given moderator access (much, if at all)? I for one haven't had moderator points in about 3 months now, and I feel that the decision to give points to people with low karma or people that don't visit the site often is just *fucking stupid*. Look at the kind of trailer trash we have moderating now: these people don't even read the fucking moderator guidelines. If they did, there'd be a hell of a lot less downward-moderated articles and a hell of a lot more +4s and +5s.
Instead we have trained monkies (and that's probably insulting the trained monkies of the planet) moderating articles here. I fully expect this to be moderated down as "flamebait" or "offtopic", too. Have fun, monkey boy. Yeah, moderator, I'm talking to you...
It's going to be tough for Digital to edge into Intel's market, mainly because nearly all consumers have been brainwashed to look for the "Intel Inside" Logo.
I seriously doubt a consumer is going to want an Itanium. Or even an Alpha. These chips are designed as server and technical computing workhorses.
Like with the Alpha, all the operating systems and applications will need to be ported to the new IA-64 architecture to see any useful speed gain. All reports indicate that the on-board x86 compatibility is dog slow, with no appreciable performance gain over Pentium or Athalon chips. Why should gran'ma buy a $5000 Itanium box when the $999 iMac will run rings around it when running Quicken or MS Office?
Then there is the issue of native software: Linux, and NetBSD are gimmies. HP-UX is going to be forced marched to IA-64 (HP originally developed EPIC for the HP9000). IRIX and SCO are "definite maybes".
Sun and Microsoft, on the other hand, will probably port their OS to the platform in hopes of killing it. Microsoft had ports of NT on x86, PowerPC, MiPS and Alpha. Only x86 remains. Like with the older RISC architectures, MS will port and support the platform for a little while, but won't port it's applications, and won't promote their OS on anything other than x86. This way, Microsoft can keep control of their hardware market, and deny competitors popular support for their primary platform. And, when the market drops out, MS can quietly discontinue NT for IA-64, and place the blame squarely on Intel; just as they've blamed Compaq, Apple, and SGI for the failure of NT on RISC. Sun has a cross-platform strategy with similar goals: get them hooked on Solaris, and then entice them over to SPARC, where the applications are.
MS likes x86 becuase it -owns- x86. Linux will always be an also-ran on x86: merely a "Hobbyist's OS". The blind loyalty to intel and x86 I find expressed here is disconcerting. The only thing that will allow Linux to overcome proprietary systems is -ubiquity-, and that means cross-platform parity. Use the fastest and the best when available. That, more often than not, means Alpha.
SoupIsGood Food
Go to www.dcginc.com and check their prices. A bare bones alpha system with 2 meg of cache costs the same price as a single xeon cpu with 2 meg of cache. Not to mention the models with 4 meg of cache.
See: http://www.tux.org/~mayer/linux/bmark.html
If you use the same O.S., the same C compiler and libraries, and JUST change the CPU, the Pentium outdoes the Alpha.
Also, I was comparing the Celeron 300A to the Alphas (500-600 or so). The C300A overclocked to 504 MHz quite nicely. And beat the Alpha. It was quite a lot of bang for the buck. And still is.
Not that I'm an Intel fan. But with these results, I'd bet the K7 500 OC'd to 750+ will blow away the Alpha's too. Hmm - from a quick look at the nbench benchmarks, it looks like it does.
I too hope the Alpha's are long lived. But right now, I'm afraid they just don't have the power.
It's very simple really. Check out this article in Salon for details.
cjs
The world's most portable OS: http://www.netbsd.org.
Dude! That was, like, so lame!
i don't see any alpha mobos or cpus in frys or on pricewatch. or any other 64 bit cpu.
how exactly is that "competition?"
Treatment, not tyranny. End the drug war and free our American POWs.
See my user info for links.
5 years?
the 386 came out in 1985 i believe (1986 at the latest), and 14 years later, windows 9x is still largely *16* bit.
i wouldnt guess that 32 bit desktop software will die until at least 2010...
smash
I run: Windows, OS X, Linux, FreeBSD. Just because you have a hammer, doesn't mean everything is a nail.
Most people running Linux on sparc, alpha, etc. do not give a rats ass about all that commercial software you just listed. They use Linux because it works, not because it 'might' support some convoluted fancy commercial packages.
Dear Lord!
OK, Netscape, SO, and Oracle are binary-only products, true. However, as someone pointed out, there is 32-bit x86 binary emulation under Linux, so you can run these if you really have to.
MySQL and ssh can both be compiled. True, you do have to be one level above "retarded aardvark" to compile MySQL, but not ssh!
./configure
make
make install
Works on every Linux for me, PPC, x86, and alpha. Anyway, how can you possibly trust someone else to make a "safe" binary of something so critical as SSH? Plus, if you live in the US, you have to use the RSAREF version, so you need to compile anyway...
A bit of effort, please...
WMBC freeform/independent online radio.
You mean they've made a comeback?
(defun intel-names ()
;)
"Function to generate random names for Intel's next processor:"
(let ((prefix '("Pent" "It" "Max" "Ath" "Cort" "Trit"))
(suffix '("ium" "alon" "ex" "anium" "oricon" "on" "eres" "obos"
"ymede" "itan" "erion"))
(tag '("II" "III" "IV" "Pro" "MMX" "Deluxe")))
(concatenate 'string
(nth (random (length prefix)) prefix)
(nth (random (length suffix)) suffix)
" " (nth (random (length tag)) tag))))
;Hey, you knew *someone* was going to do it.
-- You can actually change my mind with a good argument.
Sorry, I must have misheard you. Did you just call it limited? I honestly can't believe that- go check out the W3C's Style Sheet Reference and come back and tell me that again. Style sheets offer so much more customizability and layout freedom than standard HTML, but they rejuvenate an admittedly limited standard... and Mozilla is the way to go for the correct rendering of them!
</plug>
Umm.. GCC is horrible at optimizing for the Alpha. Egcs does much better, but still not as good as the digital compiler.
On some database code I wrote a while back, I got 6x speedup going from gcc to the dec compiler.
Usually I'm silent when people moderate my comments up or down but C'Mon!
.ps version and it's marked as Overrated. Don't you think the link might be useful for some who aren't blessed with .pdf viewers?
I post a link to a
--GnrcMan--
Just what I wanted to ask, actually. Like most /.'ers, I didn't need a big technical document to see that Alpha's kick butt.
See Motorola's microprocessor info at http://mot-sps.com/
At 450Mhz, their top-of-the-line MPC7400 (aka G4) produces 825MIPS. This is comparable to my AMD K7 @ 700Mhz or a PIII/800.
PowerPCs are at the heart of (all?) IBM RS/6000s and SP (super^H^H^H^H^Hbig computers).
Unlike Intel, I don't think microprocessors are a do-or-die prospect for Motorola's business (at least for consumer markets).
So, where do they fit?
Linux (2.2.x) is still unstable on alpha. It crashes when you push it hard (ie. high scsi load on sym875 + multiple UW disks, 533 EV56, KDE 1.1.2). The bad thing is that it locks up solid. I reset and fsck at least 5 times a day. Windows NT on the same machine is rock solid. Linux on alpha is fast, it crashes fast too. I wonder if FreeBSD is more stable...
PA-8500 beats 21264@667 MHz. Yep, must be some fluke that those stupid HP engineers (if you could apply the term to them, ha!) could actually have a higher performing chip _TODAY_ than the Alpha. And the fact that HP and INTEL would pick that EPIC crap for their next architecture and sink billions of $$$$$ into it just proves that they're actually fucking morons. Won't the egg be on their face! Well, will it? NOT. Get a clue people, Compaq has to spread FUD to keep that dying architecture alive.
Doesn't anyone else remember reading a long while ago that the windows MS would release for the IA64 would be like Windows 95 in a sense? Win 95 is a "32 bit OS" so they say, yet it still has a good portion of 16 bit coding inside. What I read about many times before is MS would not change NT from it's current state (leave the OS 32 bit), but build an "emulator" that would run as a layer between NT and the hardware that would make the hardware believe NT was 64 bit. You've seen how this worked with 95.. I wonder what will happen with NT 64?
Get your facts correct. X86 instruction execution is done on the chip _without_ software emulation (i.e. done in hardware only) using the same execution units the native IA-64 instruction use.
I heartily endorsed your idea till I remembered that Jobs did use the 68K for the Mac, and he independently managed to set personal computer software back, ironically even further by coming up with a machine good enough to last longer, but with software so broken it disallowed multitasking for more than a decade! Why is it that people like Gates and Jobs wind up with the "supervisor bit", and are even lauded for technical wizardry?
You mean the POWER3 and soon-to-be-released POWER4? They, along with the Alphas compete in the same space that Itanium is aimed at.
My memory is fuzzy, but if I remember correctly a 200mhz POWER 3 has twice the floating point performance of a 600mhz Pentium 3. The POWER 3 is shipping at 600 mhz these days.
I'm not really a hardware guy, but from the technical documents I've seen about Merced/Itanium, I think it will be a flop. Itanium won't make an apreciable dent in the sales of Alpha or POWER systems.
Intel primarily targets their CPUs at the Windows market (Wintel and all that), but since there won't be a 64-bit M$ OS until 2002, who cares? Itanium will only be good for Linux and some washed up other Unices. May as well stick with alphas, which at least have run Linux for a while, as well as having a mature commercial Unix (Tru64).
about why Alpha boxen are so expensive...I remember seeing complete, reasonably equipped ~433MHz Alpha boxen for ~$1500 shortly before Compaq bought Digital...now, mysteriously, the prices have skyrocketed...makes a person wonder...
Way of the Mac? So they've come back with the two best sellings computers of all time? Get your fucking facts straight you moron.
Following the suggestion, here are a few ports of the above program to some popular languages (substitute underscores for spaces when obvious):
:P
* Scheme
(let ((rand-elt
___________(lambda (l)
________________(list-ref l (round (rand (length l))))))
______(prefix '(Pent It Max Ath Cort Trit))
______(suffix '(ium alon ex anium oricon agon))
______(tag '(II III IV Pro MMX Deluxe)))
_____(begin
__________(display (rand-elt prefix))
__________(display (rand-elt suffix))
__________(display (rand-elt tag))
__________(newline)))
* Python
def rand_elt(list):
____list[int(rand(len(list)))]
prefix = ["Pent", "It", "Max", "Ath", "Cort", "Trit"]
suffix = ["ium", "alon", "ex", "anium" "oricon", "agon"]
tag = ["II", "III", "IV", "Pro", "MMX", "Deluxe"]
s = rand_elt(prefix) + ' ' + rand_elt(suffix) + ' ' + rand_elt(tag) + '\n'
print s
That's all for now... I seem to have run out of creativity
To the editors: your English is as bad as your Perl. Please go back to grade school.
Stop ranting since you haven't shown any real proof. Take a look at the SPEC95 stats which I found on compaqs and hp's sites. SPECint95 SPECfp95 PA-8500/440MHz 33 53 Alpha 21264/667mhz 44 66 Also, I'd bet that Alpha systems cost three times less. HP hardware is rediculously expensive. Compaq isn't much cheaper either, but you can buy Alpha based systems from independent integrators also.
srand()
split("Pent It Max Ath Cort Trit", PRE)
split("ium alon ex anium oricon agon on eres obos ymede itan erion", SUF)
split("II III IV Pro MMX Deluxe", T)
b=rand()*100; c=rand()*100; d=rand()*100
CONVFMT = "%2i"
a=b ""
x=c ""
y=d ""
printf "%s%s %s\n", PRE[a%6 + 1], SUF[x%12 + 1], T[y%6 + 1]
}
--
"One World, one Web, one Program" - Microsoft promotional ad
"Remember when the U.S. had a drug problem, and then we declared a War On Drugs, and now you can't buy drugs anymore?"
The performance of HP's latest PA-RISC blows both IA-64 and Compaq 21264 out of the water, especially on floating point.
I happen to know someone that works for compaq/ digital, and I have been told a few things: The original OS that was to be the sequel to VMS was written by a Digital programmer, who was told that they really didn't need an updated version of VMS. He went somewhere else, and it's said that the initials of that system are "one better" than VMS- namely, Windows NT. Second, as the leading (lately second) seller of PC's, Compaq is tied to M$. In spite of this, the development of Alpha's continues, seemingly against Compaq's own visions and goals. Why? Linux, of course- while their proprietary OS reigns on them right now, that may change in the near future.... Just what I heard.
Coders need more than
Ok, I can understand why PC and maybe workstation users might not see the full benefit of CDE, but the fact is that in an enterprise situation, CDE is the way to go. It is the most modern interface out there, and the most advanced. You simply are not going to run an enterprise network with a desktop like KDE or GNOME.
Perhaps by "modern" you mean it should have translucent windows and you should be able to make it look like a computer from Star Trek. I'm afraid that doesn't cut it. Any modern user interface should have enterprise-wide application, object, and media integration across platforms and networks. Any modern user interface should be designed to maximise collaborative value and productivity. Any modern user interface should be light on the network--NOT heavy on it with flashy graphics and useless decorations.
The CDE is perfect for anyone who actually wants to get the job done. It is the most "modern" of any interface out there.
LOL...COMPAQ cracks me up...So when they discover something better than the "vaunted Alpha chip" then waht...and if it happens to be an Intel chip...are they going to ea their words? Anyone own a Compaq...I'm ashamed to say my mother owns one. So far, we have been unable to install anything not blessed by the proprietary demons of hell. So simply put..I guess I will never buy a COMPAQ POS so their opinion, even based upon supposed facts, can simply be placed in the trash next to the AOL discs.
First: ;-)
I visited that site and there are NO 21264's! The fastest they have listed was a 600MHz 21164. A 500MHz 21264 is the same as a 1GHz 21164 (assuming someone had a 21164 that was clocked that high), but as I said there are no 21264's... Perhaps you should read all of the results before commenting that ANY single processor is faster then a single processor 21264!! (sorry for the ranting
Second:
The GNU C, C++, and fortrain compilers are VERY unoptimized on Alpha comparied to x86. I suggest you use the Tru64 UNIX compilers to compile the benchmark app and run it on a 21264 (whatever speed, 500MHz+) and see how much faster it really is.
LONG LIVE ALPHA!!!
These support the 21264 at just over 30 specint95 and 50 specfp95:
http://ftp.digital.com/pub/Digital/info/semiconduc tor/literature/21264pb.pdf
http://www.digital.com/alphaserver/technology/ev6c hip.pdf
And this one shows the 21264@700Mhz at 39.1 specint95 and 68.1 specfp95, it's a brand spanking new GS60E 6/700.
http://www.digital.com/info/alphaserver/gs140/gs14 0_spec95.html
And, of course, the official spec web site with audited spec numbers:
http://www.spec.org/osg/cpu95/results/cint95.html and http://www.spec.org/osg/cpu95/results/cfp95.html
One of the most important things a 64 bit system gives you is a larger virtual address space. As well as the ability to use more than 4 gigs of ram which is VERY important for servers.
Further more, Alpha is NOT available from only a single source.
www.alpha-processor.com
Samsung
But help me out here, I thought they were going to use the SAME BUS and CONNECTOR as the Athlon chips? Wasn't there a DEC and AMD agreement like 30 Billion years ago about this? Rewrite the BIOS for an Athlon board (or just go without one) and boot your favorite OS on a commodity motherboard.
The Samsung chips are supposed to be around 250 in lots of 1000. Thats bloody good. But then again, Athlon is almost as good as Alpha in terms of architecture so I just AMD to go 64bit soon.
Subscribe to the "comp.sys.sun.wanted" and "comp.sys.sun.hardware" newsgroups, and then post a message asking for anyone with some 64bit hardware to reply to you if they have something for sale. You'd be surprised how quickly people respond, and how MANY.
It's way better than eBay, and people in these newsgroups are very knowledgeable about the hardware they're selling. I've had nothing but the best of service dealing with these people.
Good luck!
If you just save html, you'll have to grab a clump of files and maybe have name collisions unless you dedicate a directory (or several. for refs like ../images/whatever.gif).
And even then, you're likely to get a mix of absolute and relative URLs to trigger annoying dialup attempts while you are trying to browse the stuff offline -- unless you have a utility to do the move and URL cleanup. But from my version (4.5) of Netscape, print to pdfwriter driver is the quickest way to grab what I see. Plus the header line tells me the URL etc. Downside is the links are not live in the .pdf. I guess that would require special printer escapes to pass through.
Not everyone who posts here has that info (grammar/sp) cached... some ppl aren't fluent in english either... yet many still have valuable input.
i wish there were a grammarfish...
If and are allowed, I fail to see why and should be any different. It would sure make it a hell of a lot easier to post pre-formatted code without having to worry about doing a lot of manual formatting.
numerous posts alluded to supposedly innacurate or dumb info on the compaq paper.... I'd just like to say my thing, which is that this paper means something (possibly many things) besides the tech info. Some of you will sigh *obviously* when you read this, but: This paper states not only what is written therein, but also, that Compaq is willing to say it to the world. Out loud, to the world.. they are playing FUD fighter. This paper is anti-FUD (the result of smashing a PR particle into a SPIN particle under high energy conditions) aimed at the same people Intel aims its "Our chip is the best because everything is already running intel, and why would it be unless Intel was the best anyway? Empirical proof right?"
Okay, you can say duh now.
"Duuuuh!"
Anyway, alphas are better than ia64's because ia64's don't exist (yet). *cough*fdivbug*cough*intelsQAsuxass*cough*
PS I love the modern processor war. Check it: a laaaarge portion of the computer industry today is aimed at providing the most power possible to apps created in a high level language in order to provide functionality to the ignorant end user. These apps take huge hits on performance in order to get to market before someone else can throw up a slightly suckier product one month earlier. Now, imagine the open market machine N years from now: ridiculously powerful. Anyone wanting to use the machine itself (not some stupid M$ Word bullshit) will have the power of the very gods in her/his hands. That means that the code-capable human N years from now will be that many zillions of times more capable of processing information than the non coder of the future... best is the irony, that the non coders of today are fueling that push! Ha! My grandkids will rule this earth... from their private luxury residences on mars. Talk about telecommuting!!! Oops, i ranted.
there are three kinds of lies... lies, damn lies and benchmarks. if you want a real comparison, you'll have to get the machines yourself, tune your code etc.
I'll gladly use an Alpha system as long as I can buy the motherboard myself and build my own system. It's a great chip. I use Linux and I can run any open source Linux program on an Alpha. If it's not built into Red Hat for Alpha, just download, type "./configure", type "make", and lastly type "make install". Voila, the Alpha runs all of my software at breakneck speed - and at a good price too.
I use Linux and I can run any open source Linux program on an Alpha. If it's not built into Red Hat for Alpha, just download, type "./configure", type "make", and lastly type "make install".
Voila, the Alpha runs all of my software at breakneck speed - and at a good price too.
buy.allamerican.com/alphapro.html
I can't figure out who makes the Alpha MB. It looks just like an AMD part. Even uses the AMD 751 Northbridge and a ALI southbridge.
BTW, OSF UNIX is ancient. It it now called Tru64 Unix, and is rather different than all old OSF. In fact, I'm suprised that you still run OSF: OSF -> Digital Unix -> Tru64 Unix
I appologize if this is redundant, and I forget how to preserve indentation.
#!/bin/bash
prefix="Pent It Max Ath Cort Trit"
suffix="ium alon ex anium oricon agon on eres obos ymede itan erion"
taggix="II III IV Pro MMX Deluxe"
function randmember {
members=($*)
idx=$[ $RANDOM % ${#members[*]} ]
echo ${members[$idx]}
}
printf "%s%s %s\n" `randmember "$prefix"` \
`randmember "$suffix"` \
`randmember "$taggix"`
Some people get the idea that he is referring to the Itanium as a 64 bit x86 chip. However, he is probably referring to the AMD K8, a 64 bit x86 chip, as I doubt getting Linux to run on Itanium was a simple hack.
A deep unwavering belief is a sure sign you're missing something...
Some people get the idea that he is referring to the Itanium as a 64 bit x86 chip. However, he is probably referring to the AMD K8, a 64 bit x86 chip, as I doubt getting Linux to run on Itanium was a simple hack. (Although I hate the concept of porting an OS from 32 to 64 bit be considered a "simple hack" under any circumstances. Doing stuff like that gets you stuff like X.
A deep unwavering belief is a sure sign you're missing something...
Some people get the idea that he is referring to the Itanium as a 64 bit x86 chip. However, he is probably referring to the AMD K8, a 64 bit x86 chip, as I doubt getting Linux to run on Itanium was a simple hack. (Although I hate the concept of porting an OS from 32 to 64 bit be considered a "simple hack" under any circumstances. Doing stuff like that gets you stuff like X.)
A deep unwavering belief is a sure sign you're missing something...
Because they make MONEY! Billions and billions and billions and billions....
Actually, the POWER architechture is at the heart of those machines. PowerPC is a significantly lower performance proc. As I remember it an article posted on Slashdot about the HP PA-RISC stated that PowerPC was more or less unworthy of RISC status in that its FP performance is little better than x86. Even theoretically an 800MHz PIII is only slightly (20%) slower than a 500MHz G4 and in reality its probably much less with the PIII pulling far ahead. This comes from the fact that AltiVec is 128 bit and the G4 as two fp pipes. So you have 4floats X two pipes X 500MHz=4 gigaflops. Intel is 64 bit so 2 floats X two pipes X 800 MHz =3.2 gigaflops.
A deep unwavering belief is a sure sign you're missing something...
You appear to be playing buzzword bingo, but I'll bite - how does CDE achieve this??
Any modern user interface should be designed to maximise collaborative value and productivity.
Wow, you really are going for the buzzword bingo trophy, but unfortunately, this statement is meaningless. As in, "having no meaning".
You think 16 bit will ever disppear?
Look at the PIII. These chips still have 8 bit (8088) backwards compatability!!
Once Intel put an instruction and hardware support into the chip they will never get rid of it no matter how useless and outdated it is.
End dual-measurement, let's finish going metric!
http://gometric.us
I hate lisp. The indentation blows.
actually the IA-64 architecture doesnt execute legacy code (8 16 32), when it receives legacy code it switches to a second proc that is loaded on the back (think overdrive here) of the real proc. In the instruction set there is an interrupt that the OS sends the processor to make the switch., plus IA-32 is 3 layer (8 16 32) already...
PowerPCs are at the heart of (all?) IBM RS/6000s and SP (super^H^H^H^H^Hbig computers).
Not quite - the lower-end RS/6000 systems are built around PPC processors, but the mid-to-high-end RS6k's are built around IBM's POWER-series processors (with which, I believe, the PPC shares some features and lineage). The POWER CPUs are significantly meatier than the PPC.
Sam: "That was needlessly cryptic."
Max: "I'd be peeing my pants if I wore any!"
HP and IBM have been able to get very close to Alpha performance with much lower clocked chips several times already. Give or take the same process they can never achieve the same clock rates as Alpha though, because Alpha is much simpler. They get their performance from being sophisticated arhictecturally, whereas Alpha is classical barebones RISC. All this tells me that Alpha is going to crash and burn, no matter how much the first iteration of I-64 sucks. You can only raise the clock frequency that much...
You may also want to check out 64-bit R10K-based SGI Indigo2 systems. I recently bought a 195 MHz (MIPS R10000) Indigo2 Max Impact... not only are its SPEC marks better than those of the Sun Ultra 1, but the graphics options are much better...
XL24 = redraws feel much faster than Creator 2D
High Impact = slightly faster than Creator 3D
Max Impact = 2x as fast as High
Factor in the much faster busses, cheaper but still fast ram, fast textured Real OpenGL, and the fact that Indigo2s are going cheap these days and you can't go wrong.
Insert IRIX & MIPSpro compilers rants here
the person you replied to said "OS and component licensing scheme", not "OS" as you quoted.
But, frankly, I can't see how some obsolete RISC architecture can hope to compete with all the new VLIW and EPIC architectures from Intel, Sun, and others.
OK, I'm almost certain no one will either read this or care because I'm so late in posting, but you know what? Compaq's writers really don't know what they are talking about. It had to be some marketing guy sitting up there taking bits of things he read and pasting them together and then filling in the holes with what he or she understood to be correct... The biggest problem I had with the article is that they kept saying that the Alpha could do this: Fire out-of-order and Complete (or retire) in-order... This is completely impossible. OK, well you could do it, but it would be a waste of time, money, and speed. Alphas DO NOT complete instructions in program order. The Fire and Complete OUT-OF-ORDER. How does that work you ask? Well instructions that are independant in the same block of code really don't care when the instructions around them are executed, so Alpha fires them whenever is most convinient. But they can't just go back and say: "Well, this one should complete later to maintain program order" because that defeats the entire purpose of superscalar architecture. Program order is only maintained for dependant blocks of instructions (idealy).
Alpha is a superscalar RISC processor. IA-64 is an EPIC processor. EPIC is not VLIW. The difference between the two is simple yet subtle. VLIW is an architechture where instructions are taken and placed into Very Long Instruction Words (hence the acronym). This long instructions called a multiop is what is executed. A Multi-op is made up of usually something like 4 regular instructions that use different resources. The main focus of VLIW is reasource allocation.
EPIC, however, is different in that, instead of sending a multi-op to the processor for execution, a block of instructions which are garaunteed to be independant are sent to the processor. It is up to the processor to decide what gets executed when because it doesn't matter, as the block of instructions is garuanteed to be independant. EPIC focuses on instruction indepedance (and therefore instruction level parallelism is of utmost import).
Alpha, since it is a superscalar RISC processor, makes use of dynamic analysis of the entirety of it's executing code. It must since the complier doen't have the ability to communicate to the processor like it does in EPIC and in VLIW. So, in summary, VLIW can be done using "dumb" processors which focus on executing what it is recieves from the instruction cache VERY quickly; Alpha uses a very complicated piece of hardware that can discover the most efficient way to execute code, but is restricted in speed by size, heat, and complexity; EPIC defines a midpoint between the two, in that it optimizes the architecture to make close to maximum use of current complier technology, and it employs many of the dynamic aspects of a straigh superscalar processor to iron out what the complier couldn't predict.
Unfortunatly, Intel are stupid, and began doing things like adding tons of cycletime shortening hardware to actually predict results of calculations and other unnecessary garbage. They are idiots, and don't know what to do with themselves once they have finally found an efficient way of doing something. In truth, HP came up with most of the good stuff anyway... In my (qualified) opinion, Alpha can be a better processor than Intel will produce, simply because the designers are more practical. But they need to find a new tech writer who actually understands that about which he or she is speaking. (NO FOCI processors exist, and alpha is a FOCO processor) (FOCI: fire out-of-order complete in-order, and you figure out the second one)
Star Bridge Systems(www.starbridgesystems.com) has available super computers for as low as 90,000$, and more powerfull cpmuters at up to 1Petaflop. This is not up to critisicm, they are rated at sustained performance, not peak performance, and are commerically available NOW. There are several articles pointing out that they are not hype, but in fact are already being looked into by the militray, and communications companies for terabyte switching. I bring this up, because they plan on releasing personal computers around ~2001 according to a press release, those personal computers would operate 33,000 times faster then a 500MHz Pentium II, and cost around 1,500$. The 21264 came to late, and the athlon wont even have there 8MB cache models available, most likely. -Adam Scislowicz(core@triton.net) I hate to see technology slowed because people are so slow to let go of old habits, being static ASIC technology.
you are probably one of the few people here who actually read the thing, your two comments are quite interesting.
Power and PowerPC are about to merge in the POWER4 implementation.
So the line should read LD8 t1=[a1]
It is only typo. The point is still valid.
That is offset addressing modes of even limited displacements are very useful. And alpha does this with smaller instructions than IA64. So the alpha code has less instuctions to exe, and has smaller instuctions. This is signifigant.
Also, what really is ever new in computer arch. It is new To have a 1+ GHz cpu on one chip with SMT. What is new about EPIC as well. The same thing. Nobody has mas produced a single chip implementaion at that speed.
What other errors do you see? They are there, but Intel has error in their prop as well.
Up until a few weeks or so ago, the fastest SP were built using plain old 332 MHz 604+. If you look at the top500 super computer web page, all of the IBM silver 332 are based on this 32 bit 604+. They do have an IBM designed memory subsystem with larger busses feeding a fast l2.
The new systems are POWER3 (630) based.