Kuro5hin.org: technology and culture, from the trenches
create account | help/FAQ | contact | links | search | IRC | site news
[ Everything | Diaries | Technology | Science | Culture | Politics | Media | News | Internet | Op-Ed | Fiction | Meta | MLP ]
We need your support: buy an ad | premium membership

[P]
Alpha vs IA_64

By pulsar in News
Wed Mar 15, 2000 at 10:16:33 AM EST
Tags: Hardware (all tags)
Hardware

Compaq has put up an article comparing their Alpha with Intel's IA_64. [editor's note, by rusty] Of course, Compaq sells Alphas now, so you can imagine what their conclusions are. Still, it's a fairly in-depth piece, with a lot of detail for you hardware junkies.


Sponsors

Voxel dot net
o Managed Hosting
o VoxCAST Content Delivery
o Raw Infrastructure

Login

Related Links
o article
o Also by pulsar


Display: Sort:
Alpha vs IA_64 | 11 comments (11 topical, editorial, 0 hidden)
It'd be nice to see an in-depth tec... (2.00 / 1) (#3)
by fluffy grue on Tue Mar 14, 2000 at 02:16:47 PM EST

fluffy grue voted 1 on this story.

It'd be nice to see an in-depth technical piece comparing two pieces of hardware, but this is still somewhat lacking. For the part where it is actually making a comparison, it's quite unobjective. For starters, IA-64 is not VLIW, but a sort of RISC-CISC-VLIW fusion called EPIC. It still has a scheduler, it's just that a lot of the burden of scheduling algorithms are largely offloaded onto the compiler. There's still a lot of run-time scheduling going on, though there's no longer any out-of-order execution.

In addition, their predication vs. speculation example is a little biased. The point to predication isn't the ability to combine comparisons into as few operations as possible, but to remove jumps for small conditional blocks. With speculation, the Alpha still needs to perform a costly jump, whereas with IA-64's predication, the instructions within the conditional block are simply ignored if the condition is unmet. When you've got a deep pipeline, basically converting instructions into NOPs is a hell of a lot more efficient than having to jump. Compaq's example certainly doesn't reflect that.

Their benchmark numbers are definitely biased, given that they don't even SHOW IA-64's performance. Rather Orwellian, if you ask me. Also, comparing roadmaps isn't exactly a fair comparison of underlying technologies.

Basically, what VLIW and EPIC do is offload the expensive, convoluted and highly-bug-prone scheduler circuitry into software. Underneath every modern RISC and CISC CPU is basically a VLIW core. The AMD K7 is a perfect example of this; all that its scheduler does is convert CISC into very fine-grained VLIW. The Transmeta Crusoe is an even better example: the scheduler is completely implemented in software. Technologically, the core of the Crusoe isn't really that different from the post-scheduler pre-commit core of the K7, PowerPC, P6, or Alpha.

In any case, the extent to which this article is a comparison between the Alpha and IA-64 is very minimal; basically they make some false claims about IA-64, show their lack of knowledge in what the IA-64's predication actually does, and then go on to make random musings about their architecture which are irrelevant to the "debate."

Don't get me wrong, I like Alphas. I'm also not exactly head-over-heels with the IA-64 architecture. My favorite emerging CPU architecture right now is the Crusoe, but not because of the company or people behind it but because they've finally figured out that hardware-based instruction scheduling is limited and needlessly costly, not to mention that having a 24GHz CPU is pointless when there's so much more useful and interesting stuff to be done with a "lowly" P2-500-class CPU in embedded applications. However, I won't stand by and agree with arguments which aren't based in reality.
--
"Is not a quine" is not a quine.
I have a master's degree in science!

[ Hug Your Trikuare ]

Re: It'd be nice to see an in-depth tec... (none / 0) (#8)
by pulsar on Wed Mar 15, 2000 at 12:50:55 PM EST

Of course the Q's info is biased, just as much as it would be if Intel had written it! :) My suggestion to you would be to test drive the systems yourself. Q has testdrive systems available, both Alpha's and x86's, and several different OS'!

[ Parent ]
It would be good if you did a littl... (none / 0) (#2)
by Strider on Tue Mar 14, 2000 at 03:23:24 PM EST

Strider voted -1 on this story.

It would be good if you did a little more write-up on the link. Maybe give a brief summary of who comes out on top or something...
---
"it's like having gravity suddenly replaced by cheez-whiz" - rusty

Re: It would be good if you did a littl... (none / 0) (#7)
by pulsar on Wed Mar 15, 2000 at 12:12:53 PM EST

It would be good if you did a little more write-up on the link. Maybe give a brief summary of who comes out on top or something...

Noted! I'll be sure to say a bit more next time...

[ Parent ]
On the one hand, I've wanted an Alp... (none / 0) (#1)
by Demona on Wed Mar 15, 2000 at 08:35:41 AM EST

Demona voted 0 on this story.

On the one hand, I've wanted an Alpha ever since they first came out. On the other, I've learned never to trust hardware reviews from the people that make it. Decisions, decisions...

Anyone know if anything's come yet out of the free/open PPC motherboard spec IBM released?

Re: On the one hand, I've wanted an Alp... (5.00 / 1) (#9)
by Anonymous Hero on Wed Mar 15, 2000 at 01:13:13 PM EST

There was a problem with a bug in the prototypes and they are being reworked now. IBM already uses powerpc in 64 bit configurations(modified) so a 64 bit powerpc chip should be simple. They also scream and I heard IBM was really pushing the limits of these past MOT and Apple. Cheap, top of the line boards, cheap fast processors, made for OS operating systems.
IBM's page
more info

Joe

[ Parent ]

Awesome Icons... (none / 0) (#4)
by shepd on Wed Mar 15, 2000 at 10:28:21 AM EST

A C64 keyboard just seems like exactly the right thing for computer hardware.
:-)


Kudos to Driph (none / 0) (#5)
by rusty on Wed Mar 15, 2000 at 10:38:23 AM EST

Your fellow reader Driph made all the new icons. He's cool. :-)

____
Not the real rusty
[ Parent ]
Irony of irony. (none / 0) (#6)
by Inoshiro on Wed Mar 15, 2000 at 10:40:23 AM EST

The Compaq slide which describes how things a different is very, very slanted.

First off, they seem to imply that the CISC Alpha is better because it takes less instructions, even though those cmovereqs might take 100+ clock cycles.. I've seen plently of examples where large chunks of simple instructions can go faster than small chunks of complex instructions.

Secondly, what the hell kind of code is that that they are showing? If a person wrote that, they should be fired! It's evil to manipulate pointers in such a fashion.

Sigh. You could put an AMD K7 right next to the 21264 and see that, on average, programs are about 4 to 5% smaller because the K7 is CISC, and the Alpha is RISC. Want to save disk space? Buy CISC! :-) (it does add up)

I've never understood why people get their various underthings in a knot over different architectures, unless they think the same way this Alpha "brochure" seems to imply -- that is anything that is shown to be good by the company selling the product is good. Witness the Apple "RISC is better," followed by the Apple Zealotry of "CISC SUX0RS!!!!!" .. sigh..

All CPUs tend to be good at different things. RISCs are good at having medium sized code with fairly atomic control of the processor if you use ASM/compilers. CISCs are good at having small sized code with fairly complex instructions available for the ASM people/compilers (allowing things like libcs to be really small compared to other platforms, as it's mostly in fast hardware). VLIW is basically "everything is an instruction" and leaves most of the scheduling to the compiler, making ASM programming of it more of an art. I think code size is the largest on VLIW, but haven't (yet) had a chance to play with one. They all go faster every 18months, and all perform very equialently (except for the G4 which is behind in the Mhz market at the moment ;)



--
[ イノシロ ]
Re: Alpha vs IA_64 (none / 0) (#10)
by FlinkDelDinky on Thu Mar 16, 2000 at 03:14:11 AM EST

I think the Alpha vs. IA_64 has some interest but it's probably the wrong
comparison to make at this point.  I don't think IA_64 will be "compitetive"
until its second generation.

However, IBM's RS/6000/PPC/G4 stuff might be a truer (somebody tell me how to
spell that properly) comparison.  

I'm no intel fan, but I'm finding Intel's troubles with AMD Athalon (and
perhaps Sledgehammer), Rambus, and anything else I don't know about, well...I
don't know how I feel.	A little disturbed maybe but I don't know why. 
Strange.  

And I don't think Intel is going to just bounce back and be a BMF'er again. 
There's just to much competition inside and outside of IA-32.  IA_64 runs
Linux, BFD!  All new high end chips run Linux (except I'm not sure about knew
SPARCs).

PS. Rusty, can you put up a small tutorial on HTML article formatting codes and
then place a link to it right next to the "Allowed HTML:" .

Something like:
<p> = paragraph
<b> = bold
<?> = unbold
etc.


Re: Alpha vs IA_64 (none / 0) (#11)
by Anonymous Hero on Fri Mar 17, 2000 at 09:13:16 AM EST

Just in case some of your readers don't already know it, I'd like to point out
that the Usenet group comp.arch is one of the best Usenet groups around (not to
say that there isn't the occasional off-topic thread to kill); anyone who's
interested in processor architectures and issues of computer system design
would do well to read it.   Many frequent contributors of the group do design
processors or computer systems for a living, so they definitely know what they
are talking about.

This particular Alpha vs. IA-64 article was discussed in comp.arch when the
article was published in December; I seem to recall that the comp.arch people
didn't think the article quite as partial to Compaq's interests as some of the
commentators here suggest.

han


Alpha vs IA_64 | 11 comments (11 topical, 0 editorial, 0 hidden)
Display: Sort:

kuro5hin.org

[XML]
All trademarks and copyrights on this page are owned by their respective companies. The Rest 2000 - Present Kuro5hin.org Inc.
See our legalese page for copyright policies. Please also read our Privacy Policy.
Kuro5hin.org is powered by Free Software, including Apache, Perl, and Linux, The Scoop Engine that runs this site is freely available, under the terms of the GPL.
Need some help? Email help@kuro5hin.org.
My heart's the long stairs.

Powered by Scoop create account | help/FAQ | mission | links | search | IRC | YOU choose the stories!