Follow Slashdot blog updates by subscribing to our blog RSS feed

 


Forgot your password?
Close
typodupeerror
×
Hardware

Alpha 21264 And Athlon 850 Review 70

Posted by emmett from the zoom dept.
Arg! writes: "Hi. There's a comparison review of an Alpha DS10 with a 466 MHz chip running AlphaLinux up at this link on Ace's Hardware. The system is compared against an 850 MHz Athlon box (also running Linux) and there are a lot of different benchmarks also showing some memory comparisons with some other PC chips, like the Celerons, as well as some Sun Ultra benchmarks. The config tested was $5500, but a base DS10 is supposed to be around $3500 ... maybe not too bad for a nice 64-bit box. ;)"
This discussion has been archived. No new comments can be posted.

Alpha 21264 And Athlon 850 Review More

Alpha 21264 And Athlon 850 Review

Comments Filter:
  • An excellent point. Sort of like 286/386 days with Harris & AMD outdoing Intel on their own masks.

  • DS-10 clusters (Score:1)

    by Anonymous Coward
    Alpha is a quite impressive architecture. Check out what you can do with the low-profile version of the DS-10:

    ds-10l [digital.com]

    40 of those boxes in a single 19" rack makes for a pretty nice cluster.

  • The article states that gcc was designed with the x86 architecture in mind, and no optimizations are available for RISC chips. This is exactly the opposite to the way the compiler has evolved - gcc was designed with RISC optimizations in mind, and not too many x86-specific features. That is why egcc and pgcc came into being - to add pentium-specific optimizations (egcc and gcc merged, and IIRC include some of the stuff from pgcc).

    I wonder if the reason the benchmarks did not compile right was that the compiler itself was not built for the Alpha, or if the Alpha optimizations are different from the other RISC chips because of the 64-bitness of the architecture. Anyone with gcc on Tru64 tried a comparison between the compilers? It is hard to believe that gcc cannot take advantage of the FP registers...

  • rumour has it the 21264 could hit 1.6GHz!!!


    The 21264 won't ever see 1.6GHz, or even 1GHz. The now shipping 21264A (EV6.7) and upcoming 21264B (EV6.8) will see those numbers, though. The EV6.8 will be 0.18u with copper interconnects, and should arrive before the end of the year. Of couse, the EV6 was supposed to have scaled to 1GHz by last year, and we're just now at 733MHz. Now that Compaq has finally settled most of the confusion caused by the DEC buyout, things should get back on track very quickly.


    And to those who think the Alpha is a dead architecture, you've got another thing coming. Sure, its more expensive than an Athlon system, but it's not meant to compete in that market. Plus, no PC can offer the sheer memory throughput or clustering capabilities that are offered by VMS and Tru64 Unix on Alpha (and no, beowulf doesn't even come close).

    --

  • If I remember properly, Compaq was giving out free
    copies of the DEC C Compiler for Linux a while back,
    I don't know if they still are. I have the set and
    they scream on my 533 21164a.
  • For most of the pci-based alphas, most of the
    major manufacturers are supported in XFree86 4.0,
    though there are a few buggy drivers, (*cough* mga
    *cough*) this has nothing to do with whether the
    machine will accept it or not. the boards run
    whatever 'vga bios' and then run their own firmware
    underneath it. (and for the guy three posts down,
    alphas do NOT run OpenFirmware, they run SRM, ARC,
    or AlphaBios)

    -c
  • besides competition is good - since when have you, as a /. reader, been against competition ?

    i'm not against competition, but in a competition, there has to be winners and losers. sad fact is, MIPS, Alpha, SPARC etc etc are losing, or have already lost, to the x86 juggernaut. i don't care if an Athlon is a series of ugly x86 kludges piled upon a RISC core...the Athlon system cost $2,000, and Alpha system cost $3,500, and there wasn't $1500 worth of extra performance between the two systems.

    As sad as it is, the Alpha loses.
  • I recall from many years ago that the byte benchmark sucked when you failed to compile the benchmark using proper byte alignment. This was exposed when gcc suffered compared to ms compiler on windows I believe. The deal is, the cache works twice as hard to feed the cpu if words are not aligned on cache line boundaries, wich might be 64 or 128 bytes or something. The gcc compiler did much better when alignment was fixed. This may be more of a problem since the floating point numbers may actually be double precision by default (?) I wonder if this benchmark aligned properly for gcc (or for the compaq compiler for that matter).
  • The problem is that artifical benchmarks are dumb. The only thing that really matters is application performance. If the applications you use run faster on one CPU than another, voila, there you have your benchmark. Of course, good artifical benchmarks are possible if they emulate application code (like the 3D 2000 max or whatever benchmark that uses actual game engine pieces to test performance.) Cross platform testing is a bit hard. You'd have to test app that were equally optimized for both platforms. It shouldn't be too hard, however, with good optimizing compiles. Take GIMP and compile it using the Intel x86 compiler (even more optimizing than the Microsoft compiler) and test GIMP compiled with the Compaq Alpha compiler.
  • Hey Jurg,

    Alphas are far from dead. Use the right compiler (Compaq - and it is Free) and a 21264/750/8MB and the Athlon gets pasted to the road under the performance of the Alpha. (can you say 'Score of 74 in SpecFP95'?)

    Oh and by the way, you can buy a dual Alpha. You can NOT buy a dual Athlon.

    Not enough for you? Wait a few months and the .18 micron chips Alpha chips will be out running @ 1.2, 1.6 GHZ - umm thats GIGAHERTZ.

    If you are serious about your need for fast float you will pay for it. And it is paying me back in spades.

    Ever have a buggy AMD chip? They are not fun. I am in the process of replacing dozens of 166s and 200s which all have the paged memory bug above 32MB. (Replacing with Intel Pentium MMX 233s and clocking them where I can - MB dependant.)

    Need to move some data around? The alpha 21264 will do it heaps faster than the Athlon. 64bit vs 32 bit. Need a SuperComputer? #5 on the orb is an Alpha/Linux Cluster and it fuckin rocks. Cya, Dan

  • Use the right compiler (Compaq - and it is Free) and a 21264/750/8MB and the Athlon gets pasted to the road under the performance of the Alpha.

    Well its not that surprising...8MB of L2 cache?!?! how much is that going to cost?

    Oh and by the way, you can buy a dual Alpha. You can NOT buy a dual Athlon.

    maybe not this month...the Athlon itself is SMP capable, and the next generation AMD chipset will support SMP.

    there's no doubt that a 21264/750/8MB will be faster than an Athlon. It just gonna cost a shitload more, and i don't believe that it offers a shitload worth of extra performance. the intense competition in the x86 market is driving along the kludgy x86 architecture to higher levels of performance at an ever increasing rate, and at cheaper and cheaper prices. The Alpha (and other traditional RISC chips) are going to have an even harder time than they are now trying to compete, simply because of their high cost. MIPS is pretty much out of the game already.
  • One thing to remember is that no matter how good GCC is, Compaq's compiler is *amazingly* good.

    --GnrcMan--
  • Trying to categorize RNA or DNA or protein sequences using suffix arrays with dynamic programming, including sometimes large datasets.

    A substantial portion of the gene sequencing being done now is on the Alpha. This is the sort of thing that the Alpha really excels at.

    --GnrcMan--
  • There's a post at Kyles HardOCP [hardocp.com] regarding the 'demise' of the Athlon, due to the fact that they're coming out with the socket A processor. What's the general consensus?

    I'm sorry. What I meant to say was 'please excuse me.'
    what came out of my mouth was 'Move or I'll kill you!'

  • Alpha is certainly not dead. (Score:3)

    by be-fan ( 61476 ) on Saturday May 06, 2000 @11:48AM (#1088105)
    I disagree with some people when they say that Alpha is a dying platform. In reality, it is far from dead. Aside from the hideously expensive POWER3 CPU, the Alpha is the fastest CPU out there. With the release of the EV6.7 and EV6.8, they may again capture the fastest proc known to man crown again. A 1 GHz Alpha is not that far off, Compaq demoed air cooled 1GHz Alphas a few months ago. In fact, Alpha may be the only competition left for Merced (I refuse to use that Power Ranger-esque name Itanium!) in the mid to upper range market. There was an article a while ago (on Tom's Hardware I think) cogigating on whether or not Alpha would still be faster than Merced. They cited various points about how plain RISC may beat out VLIW if implemented properly (like in the Alpha) and how the DEC guys have quite a leg up in clock speed (at least at that time.) Judging by when Merced will be out the 1GHz + Alphas should be out, and should provide quite a strong competition since
    A)Alphas could just be faster, and
    B) Alphas have much more market penetration and a larger support network built up. With Intel's power, that will change, but Alpha will have quite a head start on Intel.
    Then there is the fact that Alpha is aimed at a much broader spectrum, and Intel might just not be able to take the market with Itanium.

  • Itanium vs. Alpha (Score:4)

    by be-fan ( 61476 ) on Saturday May 06, 2000 @11:53AM (#1088106)
    The compiler situation presented by the article hilights a key advantage that Compaq has over Itanium. The compiler needed to take advantage of Alpha is already freely available, while (if intel keeps with it's current practice of charging for it's compiler) one would have to pony up quite a bit for Intel's compiler. The compiler is critical because Itanium (based on VLIW, which places all code ordering work on the compiler) is heavily dependant on a good compiler for any semblance of performance. Sure GCC compiles Itanium code, but if it can barely optimize x86, do you really think it has any hope of producing good Itanium code? This might not be important on Windows, but as OSs like Linux and BeOS, which depend on the freely available gcc to compile many of the applications, the availibilty of a good free compiler is very important.
  • the right compiler is sold by compaq(/digital)
  • You can get a $2999 Alpha by visiting http://www.alpha-processor.com and signing up for the development platform.

    This is a 600MHz/2MB cache Alpha with AGP. We've been running banner ads on Slashdot and Freshmeat for a few weeks now.

    disclosure: I work for API
  • Almost any card that is supported in XFree will work. The major exception is the Nvidia cards. They make BIOS calls that are not supported by the x86 BIOS emulator.

    The API UP1000 is the only Alpha currently availabe with AGP.

    I work for API
  • Just one thing to consider, though, is the incremental cost of increasing processor speed. Much of that price is in the cost of the MB, RAM, and other peripherals. I'd wager the 21264 MB they used can support much higher clock-speed processors. So, to increase that baby from 466 to 733 or so may only be a price difference of a few hundred dollars. Now the alpha will really outperform the Athlon. Too bad you can't get the Athlon in higher speeds (yet...)
  • it's egcs (pronounced egcs), not egcc. egcs was not created to add x86 optimizing, it was forked cuz the egcs people felt the gcc maintainers were too slow at intergrating patches (fwih). pgcc was started (i believe) for the soul purpose of optimizing for x86 cpus, at the expense of other archs.

    the compiler had to be built for alpha, or it won't produce alpha binaries. also, the 64bit nature of alphas just means gprs are 64bits wide, as are virtual addresses. trust me, gcc uses the fpu on alphas. if it didn't, fp ops would have to be emulated, which is EXTREMELY slow.

    i'd guess scheduling is the major reason ccc outpreforms gcc. please correct me, if i'm wrong about gcc having ev6-specific scheduling.

  • I'm pretty psyched. I just picked up an Alpha 21164LX Motherboard + 533MHz 21164 CPU for $200 2 weeks ago. It's a really cool system. I just got MILO up this morning, and I hope to have linux (Either RedHat or Debian) on it soon. My first foray with the Alpha!!!
  • The config tested was $5500, but a base DS10 is supposed to be around $3500 ... maybe not too bad for a nice 64-bit box. ;)"

    Uh, $3500 is too much for a basic box

    For about $2500 you can have SUN Ultra5/366Mhz UltrasparcIIi/128MB ram/8GB ide-7200rpm disk with a 17 inch monitor.
    The same combo will cost you only $1300 if you are a student.

    The advantage of SUN's is that you can have Solaris on it (which I really like) and it has better application support than Linux on non-intel platforms (Linux on non-x86 platforms lacks most of the commercial/binary-only software you can have on intel)

    The only thing that I see as a big disadvantage about SUN workstations is that the SUN compiler still has to be purchased (for like $3000), otherwise you are stuck with gcc again which just like in Alphas' case does not optimize very well.

  • They should have used pgcc, egcc with better flags, or VC++. What flags did they use?

    The comparison is totally bogus. I like Alphas, but price/performance is not there.
    bytemark for unix:
    http://www.math.vanderbilt.edu/~mayer/linux/bmark. html

    Test system
    Athlon 700, Asus K7M, 384 MB PC133,
    Cheetah ST318203LW
    Not overclocked :-)

    flags:
    -s -static -O3 -fomit-frame-pointer -Wall -mpentiumpro -march=pentiu mpro -malign-functions=4 -funroll-loops -fexpensive-optimizations -malign-double -fschedule-insns2 -mwide-multiply

    Reading specs from /usr/lib/gcc-lib/i586-mandrake-linux/2.95.2/specs gcc version 2.95.2 19991024 (release)

    (cut and paste this table to a fixed font editor/term program, sorry but pre doesn't seem to work)



    You can see that these scores are better. Dunno about the neural net weirdness - NN blows up and makes the FP index way off. I would expect a KX133 Athlon 850 to outperform my 700. And I'm not using any of the high speed math libraries which would help.

    BYTEmark* Native Mode Benchmark ver. 2 (10/95)
    Index-split by Andrew D. Balsa (11/97)
    Linux/Unix* port by Uwe F. Mayer (12/96,11/97)

    TEST : Iterations/sec. : Old Index : New Index
    : : Pentium 90* : AMD K6/233*
    --------------------:------------------:-------- -----:------------
    NUMERIC SORT : 528.24 : 13.55 : 4.45
    STRING SORT : 51.58 : 23.05 : 3.57
    BITFIELD : 1.3486e+08 : 23.13 : 4.83
    FP EMULATION : 50.599 : 24.28 : 5.60
    FOURIER : 8949.6 : 10.18 : 5.72
    ASSIGNMENT : 5.0644 : 19.27 : 5.00
    IDEA : 1059.3 : 16.20 : 4.81
    HUFFMAN : 410.32 : 11.38 : 3.63
    NEURAL NET : 15241 : 24483.02 : 10298.45
    LU DECOMPOSITION : 354.4 : 18.36 : 13.26
    ==========================ORIGINAL BYTEMARK RESULTS==========================
    INTEGER INDEX : 18.045
    FLOATING-POINT INDEX: 166.002
    Baseline (MSDOS*) : Pentium* 90, 256 KB L2-cache, Watcom* compiler 10.0
    ==============================LINUX DATA BELOW===============================
    C compiler : gcc version 2.95.2 19991024 (release)
    libc : unknown version
    MEMORY INDEX : 4.417
    INTEGER INDEX : 4.569
    FLOATING-POINT INDEX: 92.072
    Baseline (LINUX) : AMD K6/233*, 512 KB L2-cache, gcc 2.7.2.3, libc-5.4.38
    * Trademarks are property of their respective holder.


    Anyone have a KX133 Athlon 850? Please post your scores.

    -pitmaster
  • Can you name an editor like emacs which was there before emacs? Stefan

  • The DS10 is not the smallest Alpha you can get. There is also a DS10L [digital.com] which is a 1U machine. You can also get this preloaded with Linux.

    The Alpha is not limited to 2-way SMP. The Alpha can also be put into much larger SMP configurations. The AlphaServer ES40 is 4-way, the GS60 is 8-way, and the GS140 can have up to 14 Alpha processors in it. There is even a new SC that can have 64-512 processors! These are SMP machines, not Beowulf clusters. The Athlon is not even 2-way capable yet.

    There are numerous supercomputer Beowulf clusters running Alphas. There are many graphics rendering companies running clusters of Alphas to render the graphics for movies.

    The Alpha is far from dead. The comparison was a past-generation 21264 EV6 466MHz vs a new Athlon. The newest Alphas are the 21264 EV67 which perform about 50% better than the EV6. They also have higher speeds up to 733MHz. The Alpha is also slated to replace the old MIPS processors in the Tandem systems. Tandem systems run about 90%+ of the world's financial systems.

    BTW, the Athlon has features that were licensed from the Alpha, such as the 200MHz frontside bus.

    The Alpha is expensive, but it is worth it on the high end. It is not designed to be a home system, although I would like to have one!

  • Interesting... They compare the lowest end 21264 to the highest end K7. Looks real fair to me! NOT! ;-)
  • <i>It seems to me that a multi-processor athalon would be the way to go.</i>

    ...and where were you planning to buy that?

    <i>The reality of precompiled binaries means that there is a real benefit in sticking with a x86 compatible architecture.</i>

    Not everyone runs debian/redhat/others with precompiles.
    Furthermore, some apps really benefit from being compiled at the user's preferences, so that one can optimize for that specific machine and to specific preferences.

    <i>I guess for more specialized applications where either only a small set of programs need be run or custom programs need be run (i.e. scientific applications) the alpha would be ideal. </i>

    Could you imagine a beowulf cluster of these?

    (sorry... couldn't help myself :)

  • Multias (Score:3)

    by Elbereth ( 58257 ) on Saturday May 06, 2000 @09:39AM (#1088119) Journal
    Of course, if you want an entry level 64 bit Alpha workstation, you can get a Multia (or UDB, as they're sometimes called) on ebay for around $75-$150 (depending on how decked-out it is).

    Many cheap Multias come with a soldered 166 MHz CPU, no floppy drive, and no hard drive. The more expensive ones (~$25 extra) have a socketed 166 or 233 MHz CPU, a floppy drive, and sometimes even a small hard drive (400 or 500MB).

    These Multias run Alpha Linux just fine. They're around the speed of a Pentium 100 at integer computations, and a Pentium 200 at floating point.

    You also might want to look at the 21164A, which is quite cheap these days. You can get a 533 MHz CPU, motherboard, and perhaps even a DIMM for around $500. This will be about the same speed (or a little faster) than a 450 MHz Pentium III at integer ops.

  • The nice thing about the mismatch is it really highlights the power of the alpha when fed properly optimised executables.

    As the article shows, using the C|O|M|P|A|Q compiler, the Alpha still defeats the Athlon at floating point work. Pretty amazing stuff.

    As a long term fan of the alpha, I am so pleased to see the architecture entering the consciousness of the enthusiasts. Perhaps someone can lend these guys an XP1000 (XP stands for "Extreme Performance") or even an ES40, just for fun.

    Course it would be more fun if they had started pumping out 0.18 micron Alphas so they could return to the days of eye-popping clock speeds - rumour has it the 21264 could hit 1.6GHz!!!

    Just don't mention the cost - always the weak spot of the Alpha.

  • Older systems are much cheaper... (Score:3)

    by seebs ( 15766 ) on Saturday May 06, 2000 @09:41AM (#1088121) Homepage
    I have a 21164 500Mhz box with 256MB of memory, total cost was around $1k. This was when memory was expensive, too...

    Not a bad system at all. Nowhere near as fast as the 21264's, but much, much, cheaper. It would have been pretty usable with less memory, too...
  • As far as I know, there is still no motherboard to fully support AMD's 200 MHz bus...if I am correct, I think its going to be amazing how fast the athlon will fly when it does have the full motherboard support...

    "spare the lachrymosity when the fulminations have inveighed"
  • <I>The nice thing about the mismatch is it really highlights the power of the alpha when fed properly optimised executables. </I>

    For more info about why a good compiler is a lot more critical to RISC-machines, read <A HREF="http://www.realworldtech.com/page.cfm?Articl eID=RWT021300000000&PageNum=1">this</A> piece about CISC/RISC architectures.

  • What can I say? (Score:3)

    by Anonymous Coward on Saturday May 06, 2000 @09:42AM (#1088124)
    The MHz's must be smaller on an Athlon than on an Alpha. This is similar to the US gallon vs the UK gallon problem. It's time we switch to metric units when reporting computer speeds.
  • What's REALLY funny is that when you use the correct compiler (hint: it ain't GCC) the Alpha wins on floating-point!

  • That the DEC compilers for AlphaLinux were that good. Wow. If only they would merge their optimizations into GCC... Incidentally, I just noticed that DEC has available a new version of em86 that should work on the 2.2 kernels, and that they also have been kind enough to make a nice RPM of netscape for DUNIX with the needed DUNIX libraries that Linux needs to run it.
  • The article said that you need a special graphics card for an alpha. Can you use all your other pci cards on an alpha?
  • The article said that you need a special graphics card for an alpha. Can you use all your other pci cards on an alpha?

    Yup... I'm using an Alpha right now with a Matrox Millenium II PCI.
    --
  • I know very little about gcc, but I do know that the PPC also has a large number of registers compared to x86 chips. Does anyone know if ppclinux is at a similar disadvantage?

    (I hope this isn't too off topic...)

  • I would like to see a Alpha put up against intel processors, like the p3 1 ghz. It would be an interesting battle, especially because in my opinon, Alpha's are a whole ton better than p3's. As someone above commented, the Alpha is doing to be *direct* competition against the intel merced, especially seeing as how as far as i know, the merced is going to need completely new software/operating systems to run, because of its being native 64 bit. The alpha already has software, operating systems, and had already demoed at 1 ghz, which intel does nto believe to reach until 3rd or 4th quarter 2001. Could it be that intel will at last loose it's foothold in the processor market, with AMD taking away market share in the main consumer market, and compaq/dec working on the workstation market......

  • Re:Motherboard Support (Score:3)

    by Elbereth ( 58257 ) on Saturday May 06, 2000 @10:06AM (#1088131) Journal
    Every AMD 750 and Via KX133 motherboard supports the Athlon's 100 MHz DDR front side bus. DDR stands for "double data rate", which is chip-speak for "it transfers data effectively twice as fast". If you take an electrical engineering class (microprocessors, for example), you'll learn that data is transmitted on the bus only at certain times (defined by the clock).

    For example, consider that the bus is on a 24 hour clock. The Pentium II with its 100 MHz bus transfers data at 12:00 noon. The Athlon, with it's 100 MHz DDR bus, transfers data at both 12:00 noon and 12:00 midnight.

    I hope that explains it.
  • Just a few comments of my own:

    * It has one PCI slot, used by a SCSI controller

    Some systems have a PCI slot it depends on if the Multia came with an internal SCSI 3.5" hard drive from the factory. If it had a hard drive the riser card doesn't have a PCI slot. If it didn't come with a hard drive it has a PCI slot. This is at least how I understand it.

    * The sound and video are integrated on the motherboard

    The line out sound is okay,it's not great though. Watch out for the air raid siren like internal speaker. Yes the internal speaker is really that annoying. It may be a good idea to take the speaker out. Of course the speaker wires are part of the power connector. The video for all intents and purposes is crap. It is an 8 plane TGA controller. It only supports 8-bit color, at up to 1280x1024 though. It may be up to 1600x1200 too I can't remember what I managed to push that poor old 15" Dell monitor up to. The Multia does support some standard intel PC video cards, if a bit poorly. As of two years ago, which isn't _that_ long ago for Multias the only card that works in above 8 bits and doesn't lock the machine up is the S3 968 chipset. Good luck finding it.

    * There is room for only a 2.5" laptop hard drive
    This is an IDE hard drive mind you and the connector is a 44-pin laptop connector. The Multia comes with a 44-pin cable from the factory, However, finding a cable is actually not as hard as it once was because the I-opener uses the same connector, but I believe the multia's connector is not mirrored like the I-opener's connector.

    * The power supply is insanely under-powered
    I can't argue there make sure the Multia gets properly ventilated or else you can end up with corrosion on the power connector, which will cause the Multia to not boot up. Unfortunately the Multia uses AT style connectors, but the voltages are different from an AT power supply. Some or all of the 5V pins on an AT power supply were replaced be 3.3V pins on the Multia.

    * The CPU is sometimes soldered onto the motherboard
    There are two or three different versions of the multia. The low end 166MHz model's processor is soldered on. However higher end models have a ZIF socket. Of course I'm almost certain that you can't put a 533MHz processor in.
  • Don't get me wrong, I very much like Alpha's. I have three Multias, one each running Linux RH5.2, OpenBSD and Compaq Tru64 (hobbyist).

    But take a look at the 21264 datasheet: 107W [sic] maximum power/heat at 633 MHz (2.35V & 46Amps). This is HOT, but it's also 0.35 um process. If they shrink it to the current 0.18 um, DEC could bring it down to ~20W @ 1.4V. Very competitive.

    But will Compaq make this investment?

  • I'd like to see a DS10 versus an Athlon system measured performing tasks that I actually need to do quickly. These include:

    • Compiling C and C++ programs, with complex Makefiles.
    • Playing whatever the latest first-person shooter or massive multiplayer RPG is.
    • Running interpreted Python scripts with lots of natively compiled modules.
    • Traversing a file system with a complex "find" and "grep" query.
    • Serving static web pages.
    • Doing RSA, triple-DES and SHA1.
    • Running complex Emacs LISP scripts in XEmacs.
    • Serving as an active server -- running Python scripts in response to (possibly many simultaneous) requests over TCP/IP.
    • Playing audio and video files.
    • Running the occasional Java program (mainly just hushmail.com).

    There might be some things that I've missed. Why don't you folks chime in with what kind of tasks really matter to you. Basically any time you end up _waiting_ on the computer, or anything that would be of higher value to you if it could use more CPU cycles.

    If I did scientific computing and simulation then of course I would have included that here. Hm -- in fact, I might be doing something like that on my next computer, so include something like:

    • Trying to categorize RNA or DNA or protein sequences using suffix arrays with dynamic programming, including sometimes large datasets.
    • Some sophisticated algorithm to infer categories (functional or historical) of sequences.
    • Randomly generating, recombining, and mutating programs in a sophisticated dynamic programming language like Haskell and then running the resulting programs; doing this a zillion times.

    If some of the people who have access to both Alphas and AMDs could do some benchmarks which helped me predict the performance of these kinds of tasks I would much appreciate it. (Hint: hard disk latency might be a big issue.)

    Zooko

  • The most important two things that we get from our Linux Alpha box are:

    1. It keeps you honest. Alpha Linux by default core-dumps on divide-by-zero error, where Intel Linux and SGI Irix silently return infinity. Also, Alpha seems much more likely to dump core on memory access error. Running the heterogenous environment that we do helps us find those errors more quickly. All of you that write perfect code the first time need not consider this reason.

    2. Alpha is the first of the 64 bit systems. When trying out new ideas for animation tools (which is what we build at my company) it is useful to try them on the 64-bit machine, to see what performance gains there are due to the 64 bit-ness. All of our workstations will be 64-bit machines someday; it's useful to have a prototype today.

    One interesting note is that (at least with gcc) a 'long' on an Alpha is 64 bits; while an 'int' is still 32 bits. This revealed not a few bugs (or, perhaps, future anachronisms) in our code.

    So, long live heterogenaity!

    thad

  • Cool! Where do I find the -$3500 deal? ;-)

    If you go here [esis.com.au] and do a search in the bottom frame, for "alpha", you'll see some of the $3000 models listed as being available!

    I still want to actually build my own Alpha... as the Penguin-Ferrari article points out, we need diversity, and anyway I like messing with different stuff. Anybody know of more recent articles, or have personal experiences to offer?

    I bought a refurbished Alpha Digital Alpha Workstation 433au (21164 version, either EV5 or EV56) from egghead.com for US$1,199 a couple of weeks ago. &nbsp It just came in last week and it's up and running Red Hat 6.2 for alpha, kernel 2.2.14-6.0. &nbsp The box (without monitor and keyboard) came with a 433Mhz alpha CPU, 128MB RAM, 2MB cache, an Adaptec 2940 SCSI controller, 2.1GB Seagate SCSI drive, 12X CDROM, Intergraph video card (and something with Cirrus Logic video too), built-in EtherWorks ethernet, and built-in SB-compatible sound. &nbsp It also has 2 64-bit slots and 2, shared ISA/PCI slots. &nbsp It's one of the older boxes (circa '96-'97), but it's processing seti@home work units (non-optimized) at around the same speed as a PIII 450. &nbsp One of the main reasons why I bought it was because it is upgradeable to a 600Mhz 21164 processor (although I am finding that the CPU alone is exorbitant and I would do better buying a mobo w/CPU and memory!).

    During my difficult search for any alpha info, including parts (I am realizing how little is written about alphas as compared to x86s), I stumbled upon this site [eli.com] that sells DEC Alpha hardware (and others like Sun), and they have a link here [eli.com] called "Build Your Own Alpha System", where they list the various parts you can purchase separately to put a system together. &nbsp There are a few other places out there too, but I don't have the links handy.

    I would say go for it - I've seen prices for 500-600MHz alpha mobos, w/64MB RAM for like $1500.

    All in all, this is my first non-intel (or non-intel compatible) machine and it's been a really decent learning experience dealing with the Penguin on the 'ole Ferrari!

    ;-)
  • A recent camparison of many platforms (including various Alpha and two K7's) with benchmarks related to computational chemistry is here. http://www.dl.ac.uk/CFS/benchmarks /compchem.html [dl.ac.uk] The K7's hold up well in a broader range of benchmarks.
  • I recently bought a 600MHz 21164 Alpha and am running it with RedHat 6.1... The machine runs VERY good usually, but I have LOGS of "unaligned traps" (mostly from Postgres and Emacs) showing up in my /var/log/messages. X also freezes at least once a day, and I think it could be a related problem. How can I recompile those apps with the proper bit alignment?

    Thanks!

  • The Alpha is not limited to 2-way SMP. The Alpha can also be put into much larger SMP configurations. The AlphaServer ES40 is 4-way, the GS60 is 8-way, and the GS140 can have up to 14 Alpha processors in it. There is even a new SC that can have 64-512 processors! These are SMP machines, not Beowulf clusters.

    Correction: The Compaq SC systems are not SMP boxes. They are ES40s clustered via high-speed, low-latency Quadrics interconnects. Nice systems, but a few quirks.

    Compaq is working on very large SMP systems. Ask your sales rep about details.

  • What I would like to see is a 64-bit integer arithmetic benchmark. If you are comparing two types of machines with different register widths, then you should do benchmarks, that examine each of their capabilities.

    You could compare a 32-bit machine (say a 386) running in 16-bit mode against a machine that only has 16-bit registers? Sure, the 386 will run DOS a faster than the 8086, but that ignores all the capability that the 386 gave over the 8086, or even the 286.

    What the 32-bit machine gives above the 16-bit machine is enormous. A 16-bit machine is severly limited when trying to do 32-bit arithmetic. The difference in address size alone, allows applications that simply could not run on a 16-bit machine.

    A 32-bit machine is also severly limited when doing 64-bit operations. The same is true about a 64-bit machine compared with a 32-bit machine.

    It is a bit less than 16 years since the 386 (a 32-bit machine) came out. Although it is slightly more than 16 years since DEC's VAX was first released.

    If we assume that high-end application memory requirements double each year, then it will take another 32 years before we need to jump register sizes again.

    Peter

  • Building an Alpha (Score:4)

    by Skald ( 140034 ) on Saturday May 06, 2000 @10:07AM (#1088141)
    Don't let the "budget" label fool you, though, as the system still carries a hefty price tag of +/- $3500.

    Cool! Where do I find the -$3500 deal? ;-)

    Let me point out an article I liked from Linux Gazette [linuxgazette.com] on how to build your own inexpensive Alpha [linuxgazette.com]. It's almost 2 years old, so some of the details aren't as relevant, but I think some of the stuff explaining the gotchas of the Alpha platform for someone used to x86 (like me) are still informative.

    I still want to actually build my own Alpha... as the Penguin-Ferrari article points out, we need diversity, and anyway I like messing with different stuff. Anybody know of more recent articles, or have personal experiences to offer?

  • Did you buy the cpu/mobo from ebay or was this a complete system?
  • Interesting... They compare the lowest end 21264 to the highest end K7. Looks real fair to me! NOT! ;-)

    They're also comparing the 21264, which they're saying is around $3500 (though their system had extra stuff and was more like $5500), to the Athelon system which they seem to be saying is around $2000.

    You're right. It's fair... NOT! :-)

  • Slightly misleading ... (Score:5)

    by LL ( 20038 ) on Saturday May 06, 2000 @10:35AM (#1088144)
    ... in the comment that HP is betting solely on EPIC and MIPS has stopped processor development altogether. Given the slippage in Intel's original roadmap for Merced, HP were smart to keep a hand in the PA-RISC and while MIPS may not be revving their processors like the others, they earn a very tidy income from licensing their IP and creating variants of their processors for the embedded market while continuing to source high-end designs from SGI (who still have a 5 year roadmap for their R14K, R16K).

    From a personal perspective, it is rather disturbing from an architecture point of view that so much attention is focused on the branding and MHz rating. If we use the analogy of cars, the peak revolutions per minute has got nothing to do with the actual real-world engine performance. A lot of factors depend on the I/O subsystem (gas tank + injection system), cache design (suspension), and more recently stlying (bodywork). Just because it runs hot, doesn't mean that it runs well, in fact from a thermodynamic perspective excess heat is an indication of inefficiency. Just as in real-life, there is a distinction between buses (good ol' shared memory Suns), industrial trucks (IBM workhorses), SUVs (SGI drool-designs) or motorcycle packs (Beowulfs). Pretending a souped up scooter with over-granished rusty frame can do the job of everything is a serious indication of cluelessness or delusion.

    Just as in real life, the limitation is the overall transport system (network) which is still an information back-lane despite the heavy hype. Sure a speed-demon Porsche (Alpha) can outrun anything in a speed race but most people settle for a Ford (Intel) or Chevvy (AMD) to commute to work. Some may prefer a flashier Saab (Apple) or stick with the boring but solid Volvo (IBM) or even go upmarket with a BMW (SGI) but they all serve a basic purpose (mobility) and dominate specific niches. You get paid for maintaining a professional non-bias and correctly matching your company's needs to the available choices.

    The quasi-technical mainstream press really has to
    get their act together if reading the IT section in any general newspaper is any indication ... more like unpaid advertising sometimes. I suspect that past 1 GHz, the CPU performance is of only marginal benefit (outside niche areas) as the speed limit is the bandwidth limitations anyway.

    LL
  • Yeah, well, but they didn't use the right compiler for Athlon either. Hint: egcs-1.1.2 is not particularry well optimized for Athlon either. Pgcc is a lot better. See the the thread here [aceshardware.com] for some numbers.
  • IMHO the current drive towards having one power-hungry-but-fast CPU in most commercial systems is not the best approach for delivering high-power and low-cost systems. A large gain in computing power could be made by changing the machine design to one based around several low power CPUs. A CPU design such as the ARM architecture, if modified to support multi-processing (the StrongARM does not support it as-is, I do not know how the situation is with other ARM chips) would be good for this as ARM CPUs are sufficiently fast and are power-friendly. (see http://www.arm.com [arm.com] for more details).

    It would then be possible, once the chipset and motherboard have been designed and manufactured (not a small task by any means) to have a system with a lot of processing power but not costing a great deal more than a typical high-end system today. Also the power requirements for such a system would be on a par with a current system as well, if not less (depending on the number of CPUs employed and the system with which it is being compared).

  • Got the motherboard (with CPU) from a guy on the NetBSD lists. It woulda been $100 cheaper, but word got out and people started tracking down his source and the source raised the price... Memory from the chip merchant. Case from Antec - they botched it pretty bad, actually, but I soldered the relevant contacts to a switch. ;-)

    After the recent spam problems and the amazing amount of lying they've done, I'm not buying from eBay, ever, under any circumstances.

  • Re:Interesting... (Score:1)

    by Anonymous Coward

    Well, this test shows basically just one fact: Alpha is dead !

    Some years ago, Alpha CPUs were much faster than anything else and they also had a higher clock frequency. This has changed dramatically, Alphas are now far behind and they are falling back more and more. The fierce competition between AMD and Intel and the very high revenue brought in by these products enabled a dramatic improvement process which is about to break the neck of all competing RISC-CPUs (maybe with the exception of some embedded chips).

    Come on guys, we have a system costing only 2.000 $ outperforming a different system for 3.500 $. The choice is clear ! Even the optimizing Compaq-Compiler (remember, gcc is also far away from being a perfect x86 optimizer) results in inferior FPU-values (e. g. NeuroNet) of the Alpha.

    Nobody cares which CPU is running inside a system, it is all a matter of performance, price and availability of software.

    Please remember that Athlon and Pentium are already "true" RISC-Chips in a CISC-disguise with lots of internal (invisible) registers and huge pipelines. As soon as there will be a decent L2-Cache for the Athlon and an appropriate SMP-chipset, these little babies will eat all other low-production-volume-RISC-dwarves.

    Alphas will soon be destinct, and MIPS/SPARC/PowerPC will follow ... Rest in Peace

    Jürgen
  • You want to play games on Alpha ?
    I guess, you could play Xsokoban on it but then ... this will run on p100 just as well...
  • Windoze is not a dirty hack. It is an OS that is not UNIX ,ok ?? It is way better at many things including desktop applications,games and just about anything that is not related to high performance services.

    I you insist on name calling ... Linux is a ripoff off 20 years old operating system. Nothing, nothing there is original and this really applies to the whole OpenSource community.
    Show me one product that is not another version of something that was done commercially before.
    And these people talk about innovations ..... ironic.
  • Well you can get one from eBay right now [ebay.com]!
  • Call Compaq about some of the discounts they have available. Through a reseller and with software developer discounts you can get the price REAL close to $2500.
  • Who, praytell, gives a hoot about integer performance? Almost any technical app you can think of that really needs massive CPU performance (web serving really is more about I/O and bandwidth and cache, contrary to what intel would have you belive) that doesn't rely mostly on integer ops?
  • Aheem, correction, I mean to say "doesn't rely mostly on floating point ops."
  • That's true. The Multia does have some limitations:

    • It has one PCI slot, used by a SCSI controller
    • The sound and video are integrated on the motherboard
    • There is room for only a 2.5" laptop hard drive
    • The power supply is insanely under-powered
    • Some systems come with no floppy, hard drive, or memory
    • The CPU is sometimes soldered onto the motherboard


    It is possible to put a 3.5" hard drive in some Multias, but that would be a bad idea. The power supply is only 75W, if I remember correctly. Anything more powerful than a laptop hard drive would suck up too much power. Out of curiosity, I tried hooking up a Seagate Barracuda (7200 RPM, 50 pin Ultra Narrow) to a Multia's power supply, but the system couldn't handle it, of course.

    I agree, buying a "real" motherboard and CPU would be a wise choice, but consider the benefits of the Multia:

    • Everything is integrated. You don't need to know anything about the Alpha architecture.
    • It's plug 'n' play, as long as you have memory, a hard drive, and floppy. These are easy to add, in case you didn't get them.
    • It uses industry standard parts (72 pin SIMMs, SCSI peripherals, etc). A lot of older Alpha hardware is proprietary.
    • They're available in bulk. It's not easy to find cheap Alpha workstations at your average web store.
    • Lots of people have Multias. If you have a question, someone can answer it.
    • Most of the add-ons that you would want for a Multia are obsolete hardware. They sell on ebay for under $50, sometimes under $20.
    • The Multia has two PCMCIA slots - pretty cool.


    I admit, they're not for everyone, but for someone who doesn't want to spend lots of time learning about hardware, Alphas, and what is supported on the Alpha (ie, you can't put a GeForce on your Alpha), the Multia is a great little computer.

  • I have to take issue with the above. Just because I have to be sensible when (if) I chose a computer for work, it doesn't mean we need that on Ace's Hardware (or Slashdot). You can find the mature business focussed coverage elsewhere.

    To use your analogy, your complaint seems to be that a Porsche is an unrealistic car. Yeah, but it's MUCH more fun to read about than the Neon I might actually be able to afford. It's also a lot more beautiful.

    Like a Porsche Turbo with that amazing four-wheel drive system, the Alpha gets a lot of "traction" in high end boxes - it really gets the power "down onto the road", better even than the Athlon with more MHz (which I suppose is a bit like the McLaren - sky-high power output but not much faster along real roads).

    Stuff like the wide, point-to-point memory link, beefy cache and killer compiler really is more interesting than raw MHz and that's why this is a good read and a worthwile (informative) review.

  • That wouldn't make sense to anyone but an electrical engineer or hardware enthusiast.

    Use jargon if you want, but it doesn't help explain complex subjects to people who are genuinely interested in learning more.
  • and we will all be the worse off if these architectures die. the fact is that these architectures do things the *right* way..and sometimes the right way looses out to a quick and dirty hack (or mass production). besides competition is good - since when have you, as a /. reader, been against competition ? thats the only reason linux exists comapred to the dirty hack that is windoze.

  • Question (Score:1)

    by PD ( 9577 )
    Why did they blur out the Compaq logo on the photograph of their test system?
  • SMP Alphas need huge power supplies. We're talking 600W to 800W.

    However, I've got all this running on a 250W power supply:
    • Two 450 MHz Pentium III CPUs
    • Quantum Atlas 10K (10K RPM Ultra160 SCSI drive)
    • Conner 2GB/4GB SCSI tape drive
    • Ricoh SCSI 2x2x6 CD-RW
    • Generic 24X SCSI CDROM
    • Maxtor 8.4GB EIDE hard drive
    • Western Digital 1.0GB EIDE hard drive
    • Tekram Ultra2 Wide SCSI controller
    • Matrox G400 AGP video card
    • SoundBlaster Live! Value
    • LinkSys EtherFast (10/100Mbs PCI Fast Ethernet)
    • Generic ISA NE2000 card

    Now, that's a lot of junk in one computer, especially for a 250W power supply. I've never, ever had a single problem with this system in the 7 or 8 months that I've had it. It stays up for a month straight, at which time I lose electricity because of an electrical storm. However, I just bought a UPS for the computer, so it should stay up forever now. :)

    My point is that the Pentium III doesn't require much power at all. Go ahead and stick two, four, or eight Pentium Pros or Xeons in your system. I doubt you'll need much power.

Slashdot Top Deals

"What man has done, man can aspire to do." -- Jerry Pournelle, about space flight

Close