Slashdot Mirror
1GHz Alphas
RelliK writes "news.com has a story about
1GHz Alphas demonstrated at PC Expo. They'll be available by
mid-2000. In the mean time, they'll start shipping 750 MHz
Alphas 21264 in July. " MMMmmm... alphariffic.
← Back to Stories (view on slashdot.org)
not really..you can have trace arrays and slop like that. it depends on how much debugging capability you can/want to build into the chip.
Alpha has been making extremely fast chips for years. Look up their old roadmaps and check the clock speeds.
Not that clock speed means a whole lot. However, Alpha's spec numbers have been impressive too.
ps: they were running the control workstation in GNOME.
As ROOT.
Have fun.
But, the more relevant question is what process
.18 micron process.
this 1Ghz Alpha is made with, and what would its
power dissipation be.
I would harzard to guess (well not really a guess) that this chip dissipates ~50W.
Not bad.
Considering that at 800mhz merced will be
a 100w chip. in a true
merced 800mhz -> 45 specint 70 specfp 100w
alpha 264 1GHz -> 60 specint 120 specfp 50w
That would be totally different then :-)
On the subject of Alphas in retail outlets, it would be darned nice if the availablity of Slot B boards made this feasible. I've met some people who work for retail chains who would like to offer Alphas but don't know how to go about it and are somewhat frightened by the cost of motherboards.
I browse with my threshold at 2 so I can't read my own comments :-)
Just a bit of Clarification.
The Slot B motherboard UP2000 ( I don't know
about UP1000) like the alpha DP264 Dual board
do not use 200mhz ev6 connections.
They use a * 333Mhz * connection to the ev6 chipset and an arrayed channel to memory.
I Think you need to visit the following sites.
www.dcginc.com or
www.microway.com or
www.alphalinux.com and go to vendors.
you can find decently priced alpha's
from $1500-$15000
What I want to know is how fast
is an alpha EV6 at 750mhz is at running
FX!32. Could it run most x86 programs
faster than a 500 mhz PIII?
API -> you need more benchmarks. with FX!32
Anybody with some data points on a EV6?
Dude, drop the pipe! The most stable version
:)
of NT is the one that runs on Alpha. The Alpha Win32 compiler is light years ahead of the x86.
Is running an x86 app via FX!32 gonna cost you some speed? Sure, but you can afford it when your 400MHz 21164 is significantly faster than a PII 450.
Finally, I would imagine that the 64 bit version is unstable, as only Dave Cutler's team is running it. (That means it's still in development
Do anal-retentive people hyphenate 'anal retentive'?
Saw it today... Nice little booth that Alpha Processor had there.
:)
More interesting IMHO was the dual 750mhz system they had running Linux. 8mb L2 cache ram on those suckers. They didn't seem to have any example of how fast it was, but it was impressive. I'm not sure why they were demoing the 1gig Alpha but didn't go all out and show a dual 1ghz system. Maybe a motherboard issue? Anyone know for sure?
More impressive was some of the serious hardware that Compaq was showing, but the people they had there seemed clueless about which worked with Linux and which didn't.
I was really disappointed at the Linux Pavilion though, seemed like half-baked attempts at doing something both by RedHat and Caldera. Liked what I saw at Cygnus's booth though. Liked that the BRU folks were actually giving away knickknacks, unlike practically everyone else there.
Anyone going to PC Expo -- take a gander at the 64" HDTV at the Panasonic booth. Hmmmm... I want one.
Volia! 100% CPU utilization. Moving the mouse really fast (last time I tried) also would put the CPU to 100% effectively locking everyone out.
Compaq was demoing a (I believe) 8 system dual-processor Beowulf cluster. I didn't get a good look at it, as I was drooling over the Alpha mainframes. :)
:)
I think I spotted one or two other companies demonstrating Beowulf clusters, and I remember spotting POV running on what looked like Linux above a stack of a half-dozen machines in one of the booths (might've been Compaq's) which I assume was showing the clustering too.
Corel... er... HCC... er... rebel.com had the Netwinder RM there, which was cool too... They're not doing a very good job differentiating themselves from Corel, sitting in the middle of Corel's booth.
Win2000 isn't 64 bit. 64bit Windows is on Microsoft's road map (msdn.microsoft.com - there's a whitepaper and development guide).
Also, Windows 2000 is one of the least bloated OSs i've ever seen, that can do what it does.
At Lawrence Tech, we run NT on Alphas (in the one lab that is open least often). Really nice display set ups for AutoCAD. But I think the only reason we use Alphas is because Digital donated all the computer equipment, including our VAX clusters.
On a side note, I rather like the VAX/VMS. The only time it's crashed recently was when the power went out, think the drop took a lightning bolt; you how much they tell us students.
And you have Linux on the brain...he can put anything he pleases on an Alpha. Who made you the spokes person for the Linux community. You don't speak for me. I don't hate anyone because of the operating system they use. However I didn't find your signature amusing "check out Canadians getting killed." What is that supposed to mean? Didn't have the time to download the plugin. You might think about re-phrasing it.
proudly Canadian eh!
#include "mysig.h"
Liquid nitrogen is as cheap as gasoline has anyone
:)
ever tried to cool a cpu with it? That is a bit
extreme. I would worry about it cracking the
plastic peices (and maybe the cpu.) But you could
put a heater on to make it a bit warmer than 77K.
If liquid nitrogen is to much, how about dry ice?
If I find an old fridge, maybe I'll stick my puter
in it
Quake 1 Linux/Alpha
Quake 2 NT/Alpha
Quake 3 no word yet
can find them at www.alphalinux.org and cdrom.com
The PowerPC boards for the Amiga aren't very good. They are made by a small company that had to design its own memory controller for the 680x0 and PPC processors. On their low budget, they came up with something that has very poor performance and does not support any level 2 cache. A PowerPC-only system would probably be faster, even though it would have to emulate 680x0 code. There is now a working 680x0 emulator for PowerPC Amigas, but it doesn't have that much use yet as there are no PowerPC-only systems available.
Do not use an apostrophe to indicate the plural form.
What a great warning! this should be on everything. Cars, hand tools, vegetables... you name it.
Probally nothing, Cray is owned by SGI and they use MIPS.
-- Bryan
*yawn* really? And exactly when did you see this, and what were the testing specs? Was the G4 running OS 8, X or Linux? Was the PIII running 95, 98, Linux or NT? What service packs/releases/updates were on either machine?
For that matter, how many times was whatever test run to get an average?
Ranting random stats... the way of the insecure.
-Markvs
46. The Hobo smiles, his eyes glaze over, and he burps. "Beware the man who has lived longer than the Wasteland."
It'd also be kind of pointless - probably cheaper to actually just have 2 boxes.
How about cray's use alpha chips, homey.
No, I'm not the original AC on this thread.
/dev/arse and let people have a little fun? Use your almighty Moderator status to promote the good posts rather than slam the ones you don't understand.
Hey Moderator, the original post was a JOKE. The response was an explanation of that joke. "Giga" in the U.S. indicates a larger number than it does in England. It was actually pretty witty. It was not an "England-is-better" flamebait.
Why don't you pull the serious-stick out of your
If you are talking about speed, the Alpha has ALWAYS beaten any x86 chip. It's more than just clock speed. To short answer is that RISC chips are faster and better designed than CISC and Alpha's are one of the best RISC based chips out there in terms of pure performance.
Hmmm, isn't this how an episode of STTNG went? Systems so complex they produce unexpected results? I think the enterpise ended up creating a new life form.
And just think, that was from a system with touch screens (eww) and people that went "woosh" and pushed doors open for you. ;)
Hmmm, but what life form would a OC'ed Alpha produce? The Enterprise produced a little ball of energy thing if I remember correctly...
check out kryotech's 600a overclocked to 767 -- http://www.kryotech.com/dindex.asp
Agreed. The common consumer will hopefully have more purchasing power for their buck when these companies finally colide in an undoubtably heated price-war. Mind you, this won happen for a few years, but I can already smell the smoke from here.
....Bacon!
...signing off
The competition between AMD and Intel has been well honestly, less than intense. I'm hoping that by the time that Intel realizes that its the underdog in this battle some REAL price juggling will occur.
Mmm smells like
From the Canadian outbacke
Blackfire
I expect that the K7 has a good chance at grabbing SMP server marketshare from intel. The EV6 bus use a crossbar switch design for full processor clock speed communication between multiple processors providing much better performance than intel's model which requires that the processors communicate through the much slower 100MHz memory bus.
I browse with my threshold at 2 so I can't read my own comments :-)
Who runs NT on Alpha? Well, probably people who need to run Alpha native NT apps...things like Lightwave come to mind.
NT/Alpha is 32bit top to bottom. Your not even running the kernel in 64bit mode. We won't see 64bit Windows until some time after Win2000 is released to the market.
I just put together a system (brand new parts) based on a Digital 164SX motherboard (533 MHz 21164PC CPU, 1MB L2 cache). The costings were as follows. These are in Australian dollars and probably should be halved for $US as our dollar is a lot weaker and prices (here in Perth) are somewhat higher.
CPU + MB $500
256MB RAM $360
8.4GB Quantum 7200 RPM IDE HDD $240
ATX midi case $70
NE2000 network card $36
S3 Virge DX card $70
Floppy disk drive $30
yeah, I have an assortment of cables etc but, as you can see, I've built a killer system that will munch any Wintel box for about $AUD 1200, or, $600US.
btw, it runs Debian "slink".
-t.
"lowering both cost and price"
it's an interesting story, but whoever wrote the article is an idiot.
In Capitalist America, bank robs you!
Umm.. the 64Bit version hasn't been released yet.. linux is still a 32 Bit OS.. so your argument about a 64Bit chip running a 32Bit os is kinda self contridicting..
I know someone who's running fluid dynamics simulations on two machines. One alpha (21164) and one dual PII.
Both machines are about a year old. The price tags where somewhere around $70K for the alpha and $3K for the PII.
When running two simulations on the PII and one on the alpha (one sim. per cpu), the PII is marginally slower than the alpha.
Ok, this is just one case. And I haven't seen the code and don't know which compiler options where used. Still, let's assume the code could be tuned to run twice as fast on the alpha, it's still one hell of a price/performance difference !
These results are far away from the SPECint/SPECfp numbers, but they are real world results achieved by real world users. I think that's what counts.
My point is, that although the alphas are really nice machines, and a new fast alpha will be faster than a new fast intel based machine, Compaq will have to get those price tags right.
I'd definitely get an alpha as my next box, if only I could afford it. A dual EV6, mmmmmmm... But there is no way in hell I can ever get that kind of money. I could buy a farm of PIIs for those money instead.
Linus himself is now working on a quad intel machine, instead of the dual alpha he had earlier. He said something like; the alpha is nice, but the intel machine is simply faster.
It would be great to see cheap alphas below 600 MHz, and really expensive ones above. That way us ordinary people could get our dirty hands on affordable alphas, and the companies that need the higher speeds can fund compaq by buying the high end ones. Much like intel is doing today.
Wow, I never realized that the Guinness Book of World Records was doing marketing for Intel. Come on, "fastest PC microchip available in retail stores". What kind of record is that?
http://www.spec.org/osg/cpu95/CINT95/
http://www.spec.org/osg/cpu95/CFP95/
You can also argue about whether your normal applications are compiled with all the nifty compiler switches the benchmarked programs are. Again this doesn't negate the fact they are real programs.
SPEC does a reasonable just in trying to make an "apples to apples" comparison under adverse conditions. Running your own specific applications as a benchmark is best. But as a reasonable substitute SPEC isn't nowhere near in the same class as FLOPS, MIPS, or "Seive".
The marketing folks bend the rules as to what are Specmark numbers ( running on a simulator doesn't officially count). There suppose to be real machines.
"they have 1 GHz chips running"
Hmmm, Intel and AMD have both showed off 1GHz systems, but if you look closly they are both existing chips (In AMD's case an almost final K7, err Athlon or whatever) with massive cooling systems. Alpha will have a 1GHz chip that dosen't require freon or liquid nitrogen or things like that. Running, yes. For sale soon for the general public without a refrigerator, no.
-----
I think what you mean is that there is more to a microporessrors performance than it's clock speed...and yes clock speed is a significant detail to be considered.
--
Linux is 64-bit on 64-bit architectures such
as the Alpha or the UltraSPARC.
Hahaha! Good one.
The K6-2 was supposed to be a Pentium II Killer... Never happened. It is common in AMD's history to announce chips that will make Intel's look like turtles, but then, when the real thing comes out, it is a disapointment.
Remember that K6-2 that was supposed to beat all Pii? Never happened. Right.
> A final note, a problem you get with high-speed
> processors is that they become nice little
> microwave transmittors (at 700-800+ Mhz, I think
> it was) and so you really need to reduce the
> power (the PA-RISC 8500 consumes 85W @ 440Mhz)
> and up the shielding when clocking at this rate,
> because otherwise you'd get a REAL pizza
> cooker/toaster in your computer...
I wouldn't worry too much about the frequency emitted by these chips (at least not for health reasons). There are a lot of signals in that region, and we live in an ocean of EM waves that are real close to the frequency of a Microwave (for instance most radar and microwave wireless communications are not too far off).
The real danger with a Microwave is not in the range of frequencies, but in a very specific frequency, which just happens to resonate with water molecules, causing vibration and heat (and yes, this is a little oversimplified). Deviate just a little from that frequency, and there is no effect on water molecules at all, and no known effect on people.
High frequencies causing interference with other devices, however, are possible, so shielding is important.
--
Do you think that it would be possible to run a dual k7/alpha system? this way the Alpha could be set as "master" cpu and use the k7 as an x86 co-processor??? It would beat emulating the x86 code and be a seriously sick system, hell, this would make sense even from a marketing standpoint. " We provide the fastest x86 chips in our system to help you migrate to our faster Alpha technology.." What do you think? or have I been smoking too much?
If your application is memory-intensive, then you should really be drooling over the modern alphas. They have 200MHz point-to-point 128-bit-wide busses to main memory, and 256-bit-wide bus to L2 cache. That rocks all over the PC's 64-bit-wide 100MHz busses. (And yes, I know that the main memory itself is only PC133 SDRAMs, but the L3 cache runs at the full 200MHz). Your memory-intensive application also probably thrashes the PII's 64-entry data TLB, especially if you're using hash tables or large trees with random access properties. Every TLB miss generates one or two (depending on whether your page directory is cached) superfluous memory accesses to look up your page mapping metadata. The 21264 has 128 TLB records and larger memory pages. You should get one and see how much better your application performs.
-
-- Guges --
-
Prototypes of the Cray 4 ran at over 1GHz with a sub-nanosecond clock using Galium Arsenide chips.
This was around 1994/1995. Nice of Compaq to
catch up...5 years later.
-Kevin
We have had an opportunity to benchmark HP's new N4000 server with 8 440 MHz PA-8500 CPUs in them. The results in a nutshell:
The PA-8500 inside that server lives up to what its SPECfp promises. It is 1.5-2x faster than SGI Origin2000 with 250 MHz MIPS R10000. That's with computational chemistry / molecular modelling codes crunching real world data sets up to 4 way parallel.
So, in contrast to what the above poster guessed, SPECfp does not seem to be skewed when compared to real world performance. I'd like to see how the machine scales up to 8way. Soon we'll know that too.
jeez, stick your head out from under that rock once in a while. Linux has been 64bit on the Alpha's and Ultra's for a *long* time now.
I occassionally hear of folks with cache busting fp code complain about SpecFP. However, I don't often hear about them complaining that Spec wouldn't accept their program/data set for the next version of the SpecFP. It seems most of these folks either don't want their code public or have non portable code (perhaps tuned to specific compiler) or don't really care about anything but their own usage.
If you go back several years (e.g., 94-95), some of the SpecFP probably were cache busting benchmarks. That's why they get revised every so often. Since benchmarks are used for marketing, the systems are queaked to boost the numbers (in addition to it becoming more cheaper/practical to do so). Thankfully, there are companies like SGI and IBM that tweak on more than just Spec benchmarks.
"All telling" benchmarks are nirvana. However, for those who don't have the time/money/opportunity/influence to run their unique/critical applications on a wide variety of boxes, the Spec benchmarks are significantly better than MIPS, FLOPS, MHz, and Sieve as indicators. You still need some insight into architecture though ( to spot the VW Bug with the Forumla 1 powerplant crammed into the back. That doesn't make it a F1 race car).
To say that the Spec numbers are completely useless is as equally fallacious as saying they completely informative.
Anyone know any other suppliers in the UK at a similar (better?) price?
d or_systems.asp
http://www.alpha-processor.com/where-to-buy/ven
Deleted
Damn!
My question is,
Alpha has always lunged ahead of the Pack with its chips.
But I wonder how this will affect the current Intel slowdown - i.e., they have 1 GHz chips running, but they're not releasing them, slowly ramping up to them.
Is this a threat to the Xeon?
It's a thankless job, but I've got a lot of Karma to burn off
Wow. 1ghz of power. Imagine playing Quake on one of those. You know what is really funny, I bet NT could still slow one down severly.
-Clump
... by the end of the year? Does that sound like a threat to the Xeon?
Of course, if Alpha were an English company, they would have reached 1Ghz *years* ago. It's true!
Its about time Alpha realized that to be #1 you have to play the "mhz game". They are by far the fastest processor on the market especially w/ the 21264 chipset. Combined w/ FX32 the alpha is the fastest desktop/server computer on the planet.
http://www.dolex.org
You're kidding...right?
--
Aaron Gaudio
"The fool finds ignorance all around him.
"Every man is a mob, a chain gang of idiots." - Jonathan Nolan, Memento Mori
"...beowulf..."
/* run away */
Save the whales. Feed the hungry. Free the mallocs.
The last time I read the Guiness Book of World Records (in 1998), the Intel Pentium II processor was the fastest computer processor (and the Alpha was knocked off that title). I guess that Intel held the title for this year as well. Now, I want to see the Alpha come out with that title once again!
It's finaly (more or less) been confirmed that the 21264 Alphas and the K7s will be interchangeable. Now all we need are mobo's with BIOSes that can support both chips. And no ads in them either please!
--
Aaron Gaudio
"The fool finds ignorance all around him.
"Every man is a mob, a chain gang of idiots." - Jonathan Nolan, Memento Mori
Although I prefer your interpretation, I don't think it was what the poster intended. See here.
Bearing in mind that the 600 Mhz 21264 on .35u draws 109W??
Doesn't the Guiness Book of World Records take paid submissions now? Somthing along the lines of putting in any particular world record an advertiser might think up.
This is an interesting innovation - at those speeds, you can cook your food with the radiated RF energy, the dissipated heat, or both. Finally, a computer that's *really* an appliance!
I can see it now, the new CPSC/FCC/DOE microwave PC warning label:
WARNING - Do not remove this tag under penalty of law!
(This isn't a matress or pillow, we mean it.)
DANGER! Microwave Microprocessor Unit! Do not ever, ever open the case of this computer!
RF Radiation Hazard inside. Opening this computer will let cancer-causing microwave
frequency photons jump out and eat their way through your retinas on their way
to your brain, where they may impair your judgement in selecting an operating system.
- You've been warned.
- Factory sealed for your protection.
- This computer contains no user-servicable parts.
- Like you could get it apart anyway, since Compaq uses these goofy screws...
- Do NOT warm strawberry Pop-Tarts in the Zip disk drive slot.
- Coffee cups on the CD-ROM drive "cup holder" may be heated, but drive
- Digital/Compaq is NOT responsible for funny little fractal patterns on your CD's.
- Discontinue use if rash, irritation, redness, or swelling develops.
- Do not use an apostrophe to indicate the plural form.
Thank you.must be closed when no coffee cup is present, or it's your retinas, baby.
Legal Department, Digital Equipment Division of Compaq Computer Corporation, Houston, Texas.
"The future's good and the present is nothing to sneeze at." - Roblimo's last
It seems like a nice chip, but sheesh...what kind of name is Alpha? Feh. I say they call it the "Alphalon" or the "Alphium". That's better, yeah.
--John Riney
jwriney@awod.com
Anyone ever heard about a PowerPC G4? heh Fuck intel I have seen 350 Mhz prototypes of this chip on an average run 30 % faster than a 500 PIII. Now bear in mind this chip will ship this fall and run not only on desktops but in laptops as well. It also incorporates 'Altivec' in one model which accelerates A/V by up to 900%. The new-age RISC chip is not Merced/Coppermine nor Alpha.
Soon we'll be requiring licenses from people like the FCC to own a computer due to the amount of RF and other sorts of radiation they'll be giving off. It'll be unlikely we'll be allowed to build our own machines as well. Oh well.
You know, as these chips start to get faster and faster, they should design some functional cases/motherboards. I want to be able to cook ramen on my P3 and fry eggs with my SMP alpha.
TAOS is what you need. It has some nifty tricks (binaries are in "bytecode" and translated into different machinecodes during loading) but I'm too tired to write them down.
I don't know if this project/OS is still alive and I don't have any links, sorry!
Unfortunately MHz is just as useless a measure of performance as a MIP or a FLOP is. Worse, the SPEC numbers are largely touted as "ok", which they really are not, as they give no real hint of how the machine performs when your problem size changes.
After all, none of my codes live in cache no matter how hard I try, the data sets just overflow. Hence knowing that an Alpha can run circles around my SGI when running inside cache really doesn't help me at all (which is what SpecFP largely measures). What I need to know is how my application scales on the machine in question, by running successively larger models over and over.
I fell for the lovely alpha specfps once. Beautiful numbers, awesome performance. As long as your code lives in cache. The moment you go outside the L3 cache, you start sucking rocks. The problem is that the nature of my code is quite memory intensive. I access memory in regular repeatable patterns, but I don't always hit cache.
When this happens, you have to touch the memory bus. That is where the current alpha based systems start to seriously suck wind (even compared to PCs!!!). The 100 MHz memory bus is operating at 1/6th the speed of the CPU (or so). Every cache miss is at least 6 clock cycles for the first words from the cache line, and more typically about 100 clock cycles before the cache line is ready to use (mem->L3->L2->L1->CPU). This hurts real bad. So it doesn't matter how fast the CPU core is, if it cannot be fed at full speed.
Never again will I fall for the MHz/SpecFP sirens song. They are completely useless IMO for predicting performance, even if you assume large error bars. A great example is if you compare the latest Sun Sparc systems to the SGI systems. Using real codes with real sized models, the Sun systems wheeze and moan. The SGI simply powers through the calculations. But the Sun chip is "faster". Yet it is 2-3x slower. Some of this could be the sorry state of fortran on Sun, but generally no one in their right mind uses Sun for scientific computations anyway, so this is moot.
The only thing that matters in the end is the performance of your code on real sized models. Don't ever believe the MHz/MIPS/SpecXY numbers to be generally valid or representative of what you might see. They are about as useful as counting transistors, using pin counts, or other pseudo-metrics. This is why the SGI and IBM systems are the real hands down winners for scientific computation as they kick ass here, and Sun/HP are the hands down winners for database and integer based stuff, though Linux is largely taking this db stuff over (on the low end).
What marketing Braniac thought "Athlon" would be a good name? Makes me think of athlete's foot (not a good thing) or an anorexic-looking marathon runner (not quite a good association for something you want to be speedy). The phonemes are all wrong for this.
:)
What AMD should shoot for are phonemes that connote strength and speed. Vehicle manufacturers are all over this -- witness "Altima" (connotations of extreme height, or peak of achievement), "Lexus" (height of snobbishness; more money than thou), "Volvo" (Latin for 'I go' -- very forceful), and even the new Suzuki "Hiyabusa" (they say it means "Falcon" but I say it has more to do with easy sex or goosing someone...
I propose something radical: a consistent name. K7 was fine by me. It's got a sexy consonant (just like M, Q, R, V, X, Z -- also great for car nomenclature), and a single digit that tells me the generation of the product. What could be better?
jon
[with way too much time on my hands...]
I think not...(*poof*)
I think there are a couple of PowerPC boards for the Amiga that does this. They call a separate library with separate Powerpcinstructions.
I don't do this myself (I have more or less abandoned my Amiga) but I know there exists a lot of these boards.
- Compaq DS20 1077 MB/s
- HP_N4000 760.0 MB/s
- SGI_Octane_300 375.3 MB/s
Mind you, the big SGI's do scale astoundingly; its not clear that Decpaq (or anybody else) has anything that competes with a 128 processor Origin_2000 yet. Wildfire will supposedly blow our collective minds, but as yet is only running (I think) 8 processors in the labs.Your friendly neighbourhood alpha booster.
One major question about the chip is how are they going to manage to keep the chip cool so it doesn't melt your computer and still keep it economical
Wow! Liquid gas sounds really funky. I gotta get me some of that... :-)
"The invisible and the non-existent look very much alike." -- Delos B. McKown
Some high-end SPECint/SPECfp results:
PA-RISC 8500 @ 440Mhz 34/51.4
21264 Alpha @ 500Mhz 27.3/57.7
Some notes: AMD's K7, even though it has a better FP unit than P-II will 'only' get SPECfp of about 20 at 600Mhz. (don't have published info, so making guess based on that it's about 30-40% faster than P-II). Sun's next gen chip (UltraSparc-III) will apparantly get SPECint/fp of 35+/60+ at 600Mhz (no actual results yet) - it is supposedly being publicly shown at the DAC (Design Automation Conference) now, but won't ship in volume until end of the year.
Anyone know how much the 8500 costs? It has 1.5Mb of level 1 cache - it has 150M transistors, to the Alpha 21264's 'mere' 15M. It must cost loads... This cache probably skews the SPEC results quite considerably when comparing to 'real world' cases - the SPEC marks scale pretty well with cache size I've heard... I also see that the 8500 doesn't seem to scale at all well at SPECfp as you add extra processors, compared to the other chips.
Real world usage can vary immensly from the SPEC values, depending on what you're doing. I have friends who've compared various machines for high end computations (fluid dynamics) and they found the SGIs ran/scaled the best, even though they didn't have the best SPECfp results for a single chip - it's their massive data buses that do the trick. Actually, the PA-RISC 8500 doesn't have a complete Fortran compiler yet... Most people I know consider the Alpha to be let down by it's IO/bus data-rate,etc. Yes, it's better than PC, but it's not much compared to the other high-end RISC guys, especially SGI, though I expect this difference to change...
A final note, a problem you get with high-speed processors is that they become nice little microwave transmittors (at 700-800+ Mhz, I think it was) and so you really need to reduce the power (the PA-RISC 8500 consumes 85W @ 440Mhz) and up the shielding when clocking at this rate, because otherwise you'd get a REAL pizza cooker/toaster in your computer...
I don't know where they got that, maybe DEC forgot to submit their benchmarks for consideration. The 1998 Alpha 21164a benchmarks beat the pants off the P2 and even today's P3 550. For that matter some of the SPARC and PA-RISC processors are way faster than intel.
I browse with my threshold at 2 so I can't read my own comments :-)
All Alpha NT apps are 32bit. FX!32 is not for 32bit apps its for NT x86 native apps. If your running NT on an Alpha you should be running the Alpha native version, not the intel x86 version, that would be foolish. SQL, Exchange and most of the other important server side apps are available in Alpha native versions.
I browse with my threshold at 2 so I can't read my own comments :-)
There's some great lines in this article. The authors should in PR or marketing.
...that did not require a special cryogenic equipment to keep it from overheating.
That guy with the mineral oil cooling system is going to be disappointed. Then again, knowing this crowd somebody going to want to overclock a 1 Ghz chip. Just make sure you use special cryogenic equipment, non-special cryogenic equipment won't work.
and puts them into a fungible package.
Okay, raise your hands. Who didn't have to look up fungible in the dictionary.
Compaq, which inherited the Alpha design team when it acquired Digital...
Oh oh, who died? I don't know about you, but if my company was bought out, I would not want to be called an inherited employee.
While universally lauded for its number-crunching prowess, the Alpha chip...
Ooooohhh, prowess. I bet that gets the chicks everytime.
I'm just being silly here today because, My new program...It's Alive! It Walks! It core dumps!
See www.alphalinux.org
--Peter
www.alphalinux.org
Kryotech is gonna sell a 1 Ghz AMD K7 by like this fall....
http://www.kryotech.com/sgindex.asp
JC over on JC's News managed to get spec marks on a K7-550 with 1/3 speed L2 (the shipping version will have 1/2 speed L2). Check out the numbers and other good stuff yourself, but it scores 25.1/22.5 on SPECint/fp. As you have noted, though, real world performance doesn't always scale with SPEC -- although it isn't totally out of touch with reality. :) And yes, the microwave frequency is right up there around 1GHz ... I think the K7 will be in the 60W range, which is certainly better than my microwave (~800W, if I read the back of it correctly). I wouldn't run a bare processor on my desk and stare at it, though.
Strange, IMHO Alpha's have currently the best name. Better then "Pentium", "Merced", "Athlon" (Erg!) or "PowerPC" (G3 - "GeeThree").
:-)
StrongARM is quite cool as well... and "SPARC" sounds very nerdic!
That was the fastest MAINSTREAM CPU, not CPU in general. The actual quote was "fastest PC microchip available in retail stores."
fall off, and it would rust, and ultimately the Japanese would buy it and fix it up and sell it, and then buy the factory they started making the horrible things in in the first place.
they are 2"x4" BEFORE finishing. After finishing they end up at 1.5"x3.5"
This is just like burger places that claim "half pound of beef" or "quarter pounder with cheese" it's half a pound, or a quarter pound BEFORE cooking it. Or to use the lumber metaphor, before finishing
Well, that answered half of a longstanding question: The Alpha will be interchangable with the K7 on SlotB motherboards.
But what about heterogeneous multiprocessing on 2+ processor systems? Can I have a K7 AND an Alpha on one dual-slot board? Obviously one would have to enjoy kernel hacking (and probably BIOS hacking), but this has been a fixation of mine ever since I played with Rainbow 100 (z-80 + 8088 in the same box) and Apple II (6502 with a z-80 card). I even upgraded my Kaypro II with a 8088 daughterboard when I was in school.
Imagine the performance gains when you can predictively send ops to a processor with an architecture best suited for the operation. Sick and twisted, I know.
I think not...(*poof*)
The J5000 is "only" $20000 for the minimal dual CPU configuration, though I remember the dual-CPU cards for the V-Class machines being something like $32000. Maybe they just don't sell many V-Class boxes, whaddya think? ;-)
Doing a little bit of guesstimating and hand-waving we could arrive at... oh, about $6000 to $8000 depending on who you believe.
(ie. I don't recall seeing any PA-8500s at auction recently... doesn't seem like an eBay kinda item)
Remember that what's inside of you doesn't matter because nobody can see it.
No, it's not Xeon that's being threatened by the Alpha; that chip, or at least its architecture, being Pentium-compatable, has a long and successful life ahead of it. However, Intel should well be worried about the Merced. Even if the Merced does keep up with the speed of the Alpha, Alpha systems are bound to be much more stable and less buggy simply because the Alpha's had more than half a decade of real-world use under NT and Unix. The main incentive for going with Intel at the moment is that you stay compatable; if you're going to take the hit of moving to a new processor, that advantage is lost.
cjs
The world's most portable OS: http://www.netbsd.org.
I recall an earlier article on shashdot that talked about 600MHz Alphaz being overclocked close to 1GHz with some funky liquid gas apparatus. I wonder if this tech could boost the 1GHz Alpha to maybe 1.5GHz or so?
What's really scary about these blindingly fast CPUs is that no one can ever hook up probes to the gates to actually see what's going on, for the probe hookup would introduce enoumous capacitance and crash the CPU. It's the uncertainty principle applied to CPUs. You can never observe what's going on, because the act of observing the CPU, changes its operation. It's all math under the ceramic case. You just trust the equations and test it like a black box.
Last week The Register posted a roadmap (leaked by Compaq, apparently) of projected SPECint95 scores for Alpha, Coppermine, and Xeon, and it don't look good for Intel. Merced rates a sorry little x at the end of the chart (turning in a SPECint95 of 46 to the 1Ghz 21264's 60); while rumor has it Willamette (Intel's all new 32-bitter) will clean up, it doesn't make the chart since it's not scheduled to arrive until Q3 next year--and that's without factoring in Intel's rather unimpressive punctuality of late. And while the SPECs weren't posted, let's not forget that Alphas have traditionally humiliated Intel chips on floating point performance. Now, take the K7^H^HAthlon (*yech*! Who came up with that one?)--with its available 2MB L2 cache and 8-way (or was it 16-way?) SMP--apparently beating Xeon performance for a whole lot less on the low end, and the new 21264 Alphas creaming them on the high end...add 64-bit Win2K for the Alpha...and finally the fact that Intel's IA-64 architecture is by all accounts late and underwhelming (at least until McKinley shows up in 2001 or 2002)...and suddenly Intel's server CPU cash cows look like dubious purchases for at least the next 2 years. As for the desktop market...the Athlon (gawd that's an awful name!) has to be the CPU of choice for cheap workstations and expensive gaming boxes; the K6 is already the fastest chip for non-FPU apps. The only reason I can think of to get an Intel CPU would be a Celeron for some cheap gaming--they still overclock a bit, but only if you're willing to run your bus at an unorthodox speed. Meanwhile, Intel's having a hell of a time getting memory companies to make Rambus DRAM that works; until then, Camino, the chipset for their new .18 micron PIII's with 256k full-speed cache, is dead in the water--it's already been pushed back until November. Oh, and nevermind that (according to Tom's Hardware at least) RDRAM's benefit over PC133 looks to be negligable. The longer they spend fitzing around with RDRAM, the longer their stuck with a 100 Mhz bus, while the K7 (so sue me) and Alpha get 200Mhz. All of which has me wondering...in a couple of months, after the K7 becomes cheap enough to get into budget boxes...what possible reason would any computer buyer anywhere have to buy an Intel CPU (at least until Willamette)?? Oh, almost forgot. They make the internet go faster.