Catch up on stories from the past week (and beyond) at the Slashdot story archive

 


Forgot your password?
Close
typodupeerror
×
Hardware

Building The Fastest Desktop Possible 228

Posted by Hemos from the cramming-money-into-it dept.
hero_or_what writes "Tom's Hardware has built one the fastest PCs on the planet. Its basically an overclocked Athlon running at 1600MHZ!! The beast is described here. I wonder how long this monster would take to do a "make world"."
This discussion has been archived. No new comments can be posted.

Building The Fastest Desktop Possible More

Building The Fastest Desktop Possible

Comments Filter:

  • Not redundant (Score:1)

    by Anonymous Coward
    This is not redundant. It looks like someone's viewing in newest first, since they modded a later post as funny and this as redundant.

    Redundant is so stupid anyway. I've been modded as redundant for comments posted at the same time (same time stamp) as someone else, but getting just later. It's silly. Fine hide the duplicate comments, but don't take karma away as a result.

  • It's just a PC (Score:1)

    by Anonymous Coward
    Step back into time with me to 1997, when I last had the opportunity to play with SGI Origin 2000 systems.

    You know how us geeks dream about playing on a Cray or otherwise super fast box? I pretty much got that out of my system playing with those boxes (and the 256 spindles I had available).

    Way back then, an Origin 2000 with 8 lowly 195 Mhz R10000 CPU's did a full compile, load and dump of GNU emacs 19.34 in 9.5 seconds. Bandwidth baby, Bandwidth.

    It is something you've really got to see..

    Being able to build larger systems using the CrayLink interconnects was also super cool.

    Back in my day (hah, I'm 33) we built emacs and X11R2 on Sun 3/60's.. Talk about long lunches..

    When considering your latest pre-IPO startup opportunity, SGI is a great example to remember.

    I've worked with everything in the UNIX server and workstation space, and for the most part their hardware, software and support blew everyone else away. Yet they basically failed in the marketplace..

    BK

  • Re:Time to make world (Score:1)

    by Anonymous Coward
    Actually, Moore's "Law" is more 1.5y => 2x speed

    Maybe you should recalculate.

  • Athlon 1,6GHz on HARDiNFO (Score:1)

    by Anonymous Coward
    A Danish hardware page was first. They have a Kryotech 1,6 Athlon. And they postet the article 17/2. http://www.hardinfo.dk/artikler/default-a.asp?ID=1 96

  • Re:Is this really necessary? (Score:1)

    by Anonymous Coward
    Just imagine what a beowolf cluster of these would do.

  • Re:It'll Never Happen (Score:1)

    by Anonymous Coward
    Babes in bikinis dripping over *anything* will sell *anything*. Hell, they have these geeky little IT shows at the Cleveland IX center and I remember a certain company hiring chiKs just for the day to dress up in spandex pants and tight t-shirts to sell t1s. I remember people from our company thinking that there was no way these girls worked for nowonline and we finally got one to admit she worked at Ace Hardware and nowonline just paid her like $50 for the day. Sex sells anything and everything.
  • It's not the creation of Earth, but rather the command to compile XFree86... ;>
  • I have that much! Why not a gig? Seriously. Is there some limit I don't know about beyond which more ram is unused?
  • Perhaps you should use an interpreted language, if that's how you debug your code. If you want to avoid the recompile in these situations then use your brain for those extra 2 to 5 minutes and try to solve the problem properly, rather than experimenting.

    Also, reading /. and k5 whilst "waiting" for it to compile is a waste of your own resources. Again, why not spend it looking at code, letting it sink in, reading documentation? You seem to be spending too much time context-switching here, perhaps even specifically so that you can bum around when you should be thinking instead.

    This isn't at all intended as a flame. I obviously read news sites and the like "in between" patches of work as well, but I admit to myself the damage it does to my work-rate and level of concentration. I don't blame it on the slow speed of my CPU, just my laziness!
  • It's old IBM-speak. Speed of processing and feeding the card readers.

  • Now pardon me if I'm somewhat blase' on this. But it's a refrigerated CPU. Anybody with the cash could have done this well before now. So I don't see what the big deal is.

    Is it entertaining? Sure.

    But is it as entertaining or informative as if someone had introduced one that was simply water-cooled? Or better yet, air-cooled? HECK NO.

    There's absoloutely NO skill involved in buying a Kryotech or Vapochill and dropping it in your system. It's not an interesting system. It's just "another Kryotech/Vapochill system".

    Heck, I'd have been more interested if someone had hand-built the refrigeration system. I look in on these things to see cool hardware hacks. Not expensive, store-bought bells and whistles.

    I'm very sorry if this sounds elitist. But I really don't find commercial soloutions like this all that entertaining. I like a person's system to have more "character".

    But that's just me, the hardware freak.


    Chas - The one, the only.
    THANK GOD!!!

  • If you're an 31337 overclocking ultrageek, why would you deign to run windows for some snarky boss?? Oh, wait....well, I guess there's always WINE... ;o)
  • did you read the article? did you even look? productive work for their bosses means getting people to read their articles. it's been /.ed already, i'd say people are reading it.
    they clearly state on the front page that it's beyond the reach of most readers. just like a porsche, etc.
    This was done for the same reason that most hackers do stuff... to see if it can be done. It's all about fun. The work machine should just chill and do stuff i want it to... agreed. for them this means, crank out more MIPS than any other PC.

    even with that aside, processors are designed around margins based on reliability, and presumably lifetime, while balancing performance.
    just like the guy who puts an eaton blower in his 72 rabbit to get 400 horsepower, some people are more interested in performance than reliability.

    that said...
    ATHLONS suck! ;-)

  • With a 351 Cleveland, tricarbs, v gate, rock cruncher, ported polished, 3/4 race cam. I threw a con rod on the NYS Thruway around new Baltimore going 140+ mph.

    I had a 1965 Porche Carrera 2 - 2000cc Fuhrman tinker toy DOHC, 8 plugs, replaced the Webbers with a Hilborn mechanical fuel injector. Damn close to a single digit weight to power ratio. Holy hell to tune up.

    I redlined an Audi A4 on the A23 in France on the flats with 3 other people in the car. It couldn't push any more air out of the way beyond 190 kph.

    I can't for the life of me imagine why anyone would want one of these things.
  • You have no creativity.
  • I have used both Evolution and KMail with complete success on mailandnews.com. Try one of these out.
  • The router machine on my desk has 16MB RAM, and no swap. The filesystem is on /dev/ramfs. It has several MB free. The PC under my desk has 256MB. With WYSIWYG word processing, web browsing, email reading, text editing, and web servging applications running, it still has 150MB free. 256MB is *not* too "little" for "most applications", just like cars without V10s are not always "underpowered". But then, my personal car and my personal computer certainly don't reflect that minimalist approach. ;)
  • I remember when I was in school, programming my Commodore 64, and the joy was in just how primitive it all was. Just typing in those BASIC commands, waiting while that tape drive chugged away - that was half the fun of it.
    It's all fast food, fast cars, fast living, and it's not good for us

    There were fast cars long before anyone threw their piece of crap Commodore64 in the trash because of its infuriatingly slow tape drive. Overclockers did not invent hot rodding, although the reverse could potentially be true... :) I know that *I* was trying to make my lawnmower go faster before I tried to make my computer faster...

  • Why do you have a "router machine" on your desk?I

    Home network + not large home = router on desk instead of in kitchen or behind toilet

    At work, it's under my desk so I can work with it without having to get up and walk down to the wiring closet or open an ssh session. :)

  • Because waiting the tape drive was causing me stress! Much more than those fast 5,1/4 disk!!

    Why? It wasn't for the time it took, it was the POOR RELIABILITY of the things.

    Half the time, it didn't work..

    IMHO you suffer from "rosy glass nostalgia", when we were young everything was SO much better!
    Yeah right!

  • I remember that, My TRS-80 COCO's tape drive was as fast as the Commodore64 disk drive, and the coco's disk drive was blazingly fast compared to the commodores... But then the C64 was limited to basic... you couldn't run os/9 like the coco, and you couldn't get the graphics the coco offered. The cocoII came out finally, but the toy computer craze had passed... the IBM was here with expandability, and something bizzare in rthe computer world.... a standard hardware interface designation. (GasP!)

    Ahhh those lovely days of the 300bps handset modem.......

  • >For the average user using Win98, spending money on a single processor system is the way to go.

    Uh, since Win9x doesn't do SMP *at all*, this kind of goes without saying, doesn't it?

    (btw, nice sig, "Last Resort" is about my favorite Eagles song...)

    -LjM
  • Correction. Moore's Law (paraphrased) states that processor speed will double every 18 months. So...
    518400/2^(6000/1.5) = 3.932 * 10^1199 seconds.

    Pretty substantial difference....sort of. :)
  • Wake up! There are Axp 21264 processors out there you know.

    Do you have shares in them ? :-)

    Yes, but do they run Diablo ?
  • Like come on, who REALLY needs an Athalon running that fast?

    I remember someone saying "Who needs more than 640K?"

    Whilst I have my doubts about how long the CPU will hold up when overclocked, no matter how fast you make a machine go, someone will want a faster one.

  • Well...only a program that can take advantage of the dual CPU's is going to run any faster than it would on a single cpu 933 mhz machine though. The athlon was one cpu at 1600 Mhz. It gets alot done per Mhz, and it has ALOT of them. Anything will run like hell on there.

    Gotta love the smooth multitasking on a dual though.
  • LMAO. That's another /. hall-of-famer, for sure.

    --
  • Nah, just have a railgun and a bot. Lamers with bots are really starting to clog the servers these days. And they're nowhere near as subtle as they were in the old days, the simply sit on the railgun and shoot in all directions without turning or anything.

    That's why id keeps releasing new binaries, to try to keep a handle on this. They keep coming back faster every time.
  • Actually, a thousand 8086s would be a total of 4.7 gigahertz.....
  • I have to wonder why they have not gone to bigger L2 caches. Even my old sparcserver has a 1MB L2 cache. (It actually keeps up fairly well for a box that came out in 95)

    Is this due to cost factors? It seems to me that by having a larger L2 cache you can get a nice performance boost without needing any high-tech advancements. But then again I'm not a real hardware kinda guy, so maybe I'm being a bit simple here.

  • OOPS (Score:1)

    by BeNude ( 28969 )
    Too bad the Tom's Hardware web site isn't running on it... looks like it's Slashdotted...:)
  • Unix is _unreasonably_ stable.
  • doctorfaustus wrote:
    The Tao of Pooh sufferes from a logical fallacy, overgeneralization.

    Ah-ha, but you forgot the corrolary, Grasshopper...
    Elsewhere, you're too busy working to pay for machines to save you time so you won't have to work so hard.
    All these time-saving devices cost money... how many hours did you have to work to afford that washing machine, that telephone, that computer? Not to mention the cost of utilities and maintennance...
    ----
    Dave
    MicrosoftME®? No, Microsoft YOU, buddy! - my boss
  • <sarcasm>
    What website are you posting to?

    It may not occur to you, but some people have PCs of their own. The keep them at home; you know, not work?

    Besides, who wants to run Windows?
    </sarcasm>
  • But how can you run Winstone2000 on a mac? (-=

    You Like Science?
  • >she worked at Ace Hardware

    Yup, good looking women in overalls could sell me just about anything :)
    --
  • Well, (1600/1400)*5Mkeys/s = ~5.5Mkeys/s

    With the same proc, the scores scale very linearly with MHz (since the client fits in the cache).

    --
  • Dual proc boards are great if you're running apps that are multithreaded. Otherwise, you're only going to be able to take advantage of one of those processors. You also have to take into account what kind of SMP support the OS has. Some OS's scale very well with multiple procs (Solaris, AIX) and some just do ok (Linux, NT).

    Abe
  • For example, the Fibannaci (sp?) series alone requires over 3000 billion calculations, using the recursive method, to find Fib(100). And that's one of the easier functions.

    Algorithms whose runtime grows exponentially aren't really the best justification for faster hardware, as even huge jumps in processor performance yield only marginal improvements in what is practical to compute. Consider an algorithm involving some computation that takes 1 s to do an iteration. An algorithm of time complexity O(2^n) takes over 6.5 months to complete a calculation for n=44. If you reduce your computing time per iteration from 1 s to 1 ns (a 3-order-of-magnitude reduction), you only get to increase n to 54 before you're talking about half a year again to do a computation.

    (Besides, if you had some application that needed to use the Fibonacci sequence, you would find it iteratively, not recursively. Instead of O(2^n), your runtime is O(n). Not only is it much faster, but future increases in hardware performance translate to bigger, more worthwhile gains in the performance of your product.)

  • I don't know for games, but because I handle a very large image database (no, not porn !) I need ACDSee to be able to draw thumbnail fast. And even with a 1,2 Ghz CPU it could still go faster, especially when there's 50 thumbnail to create on the fly of pictures that are 3 megapixel each :)
  • 1) true but the Pentium 200 shipped very quickly afterward, and was favorably compared to the Pentium Pro both in performance and price

    2) the Pentium Pro was faster on 32 bit code, unfortunately the only widely available OS at that time was mostly 16 bit code (Windows), and even 32 bit coded video-games were still faster on the Pentium. NT was still confidential, and well, unix benchmark was worthless to 99.9% of potential buyers. And few people knew what Linux was.

    3) nope the PPro stayed at 0.35 micron die, which was really not small enough given that it had a big L2 cache on die. Intel simply dropped the Pentium Pro altogether and used the core on the Pentium II (easier and cheaper to manufacture since the L2 cache was not on die anymore). The Pentium II core also had some optimisation to run 16 bit code a little better...
  • Windows crashes all the time when you're not overclocking so what have you got to lose? :)
  • And if the NT kernel fails, I'm sure gnome and KDE are up to the task

    I'm pretty sure that neither gnome nor kde are up to the task of replacing the kernel of any operating system.
  • >Somehow I don't think that babes in bikinis
    >dripping over the latest PC system will sell
    >magazines(I could be wrong)

    You're new around here, aren't you?
  • It does work.
    While they're not dripping over the computers, they're still missing some clothes.

    Tomorrow Magazin [tomorrow.de]
  • drop the bikinis, and it may..

    //rdj
  • Dual proc boards are great if you're running apps that are multithreaded.

    Or if you run multiple apps concurrently, like compiling the kernel while applying some heavy filters in the Gimp, or even when using make -j.

  • The thing I can't understand about all this is the fact that I saw, several weeks ago now, a dual Athlon pushing 1.5 Ghz+. It topped 8BIPS, which is easily twice what good 'ole Tom's "powerbox" can do. In fact, the benchmark may have actually been on Tom's site. Can't remember, but anyhow this certainly isn't a dream box. And cryogenic-like cooling costs don't exactly appeal to me.(especially with rising electric bills) Tom is going downhill.
  • Heh, I'd just like to see what its Bovine RC5 [distributed.net] keyrate [distributed.net] is. I mean, an Athlon 1440 can get about 5 Mkeys/sec [distributed.net], so I can only imagine what this beast gets..

    Alex Bischoff
    ---
  • Well you are off by a few years there, its 5761 years not 6000 years. And I'm not sure that this exactly follows more's law.
  • The rather ironic thing is that by Jewish thought Hashem created the world, with a word, the first programing language.
  • The fast computer isn't nesseccarily the problem.

    It is something of a lifestyle thing. People seem to think that getting things done faster mean that they'll have more free time. That's not really true. They'll just find more things to do in the same time frame. The 'faster' thing seems to get into our psychology and we start demanding things come sooner and sooner.

    I know people like to do or handle multiple things at once (call on the phone while driving while shaving / doing makup) but I wonder how much this affects long term emotional stability.
  • Do you?

    That's the point.
  • I think you might be somewhat underestimating the improvement in CPU performance - the 32-bit registers and floating point hardware make a massive difference on many tasks (have a look at a 32-bit multiplication routine written in assembly for a 8-bit processor one day, let alone floating-point operations).

    Comparisons like this are very hard to make, of course, because any benchmark that runs on the older machine can run entirely in L1 or L2 cache on the newer one, representing an entirely unrealistic workload for the new machine.

  • I never claimed that 2 800's are as fast as one 1.6GHz. I understand the limitations of SMP, but for MY main uses (which I outlined in my original posting) the dual processors are more appreciated.

    However, for your run-of-the-mill Win9x gamer, the 1.6GHz is certainly a better option. This isn't my needs..
  • Maybe 20% less time than it takes most desktops?
  • I could do with one too.

    I write code, for fun and profit. Compiling code takes time; not huge amounts of it (I currently do all my coding in Java), but enough. (Previously, in C++, our in-house libraries could take an hour or more...)

    With my current (work) machine, it can take anything between 2 and 5 minutes between making a change to the source, and getting so see whether or not that fixed the problem. That's time spent compiling the code (a few seconds), starting the server (a minute or so in debug mode), parsing and compiling the page(s) (a minute or so for an average jsp), etc. (I currently do jsp/servlet work, for web sites deployed on Linux boxes running Apache and Resin)

    Much past 5:30pm, that's just too long (especially on a Saturday or Sunday...).

    With a faster machine, the whole process would take less time (well, duh), and I'd be happier. I'd also be less inclined to read /. and k5 whilst waiting for stuff to compile, thus wasting more time than is necessary (as it takes time to notice that my stuff is ready and waiting for me, and time to "just finish the bit I'm reading now"). I also wouldn't be sat there waiting for it to be ready, thinking "if only I had a faster machine, this wouldn't take so long" (I have a P3 450)

    No, the average user doesn't need that much speed; gamers, coders, 3d modellers/artists, people running number-crunching simulations, etc, do.

    Cheers,

    Tim
  • "Do we really need this kind of processing ability? Like come on, who REALLY needs an Athalon running that fast? "

    You REALLY haven't tried Visual Studio DotNet Beta1 yet ? Have you?

  • NT reboot times are not really dependent on processor speed. They'll always take a while due to disk access times, services which wait on other services to start, intentional wait periods thrown in by developers to keep their bad software from killing itself, etc.

  • Who would continue using a tape drive on the 64 once the 1541 (5-1/4") drive came out? The Commodore PET was the last computer with which anyone should have been forced to live with a tape drive. Heck, once the 1571 was built into the C=128 - now that was loading in speed and style.

    --
  • Well... let's see here:
    The MSI K7-Master has 4 133MHz DDR slots (or 266 if you want to say that)... and the memory used was Micron PC-2100 CL 2.5 DDR... but of course, that was mentioned on page three and four of the article, which you can't be bothered to read.
    --
  • Yeah, enlarging the L2 increases can increase the die size and decrease yield, both of which help lead to increased pricing. Since AMD likes to be able to battle on cost, this probably won't happen soon. What would be interesting is once the SMP chipset comes out if they would offer several options for cache size, trying to hit the entry server market... I would guess not, but who knows.
    --
  • Its all very well having a 1600GHz Athlon on your desk, but what use it it if Windows crashes every five minutes because you are overheating ? Is it just me or does anyone else agree that we should really confine ourselves to running our CPUs at the speed they were
    designed for rather than some arbitary speed we choose ?

    Did you happen to read the article? Tom used a Vapochill to keep the system at temperatures way below normal. Heat is definitely not a problem for this system.

  • L1 and L2 cache are "small" on the current P4 because the P4 core was designed for 0,13 micron tech. Intel usually does its core with the next die-size in mind, which means version 1.0 of every new Intel CPU is running on an inaddequate die-size. The Pentium Pro also had this problem, when it came out people found it barely faster than a 200 Mhz Pentium. As we have seen with the Pentium III E 1 Ghz, the core can do much much better than a Pentium. It just needs to get a an adequate die-size.

    When 0,13 micron fabs are ready Intel will put out a new version of the Pentium 4 with optimized core, large caches, higher speeds and it will also be much cheaper to manufacture. The format will also change (different socket and motherboards).
  • Why the 760 motherboard? And with a 133MHz bus? Why not the new Asus K7M 761, with its DDR 266MHz bus? I have a feeling this project was started in about November.
  • I don't think the point is that everybody and their grandma should be running a 1.6 GHz machine. Think of motorsports: companies invest a significant amount of money to make their car go around the track as fast as possible. They don't do this because they want to put an 800 horsepower turbo V8 in my Civic. But that racing technology certainly carries over to even practical cars.

    I say overclock the hell out of those CPUs. And then give me one that runs faster and is more stable.
  • I don't want this to be a flame, but it feels like one in my head already.

    Why anything then? Why has man ever accomplished anything? Because accomplishment feels good, it helps us to advance. Go west, go farther, climb higher, have knowledge of the distant stars, see the smallest things imaginable........and yes, make your computer go faster. It's a pretty basic drive.

    I'm sure you pine for the days of BASIC and your C-64, but the rest of us are trying to accomplish wonderful things and fulfill grand visions. It certainly was simpler and easier back then, and I would be the first to agree that the complexity of life, society, and technology may not be doing us a lot of good, but don't ask why.

  • I do not agree with this. I thought exactly the same thing a year and a half ago and built a Dual PII 400mhz system on a Gigabit 6bxd motherboard. When running Linux the system *rocks* compared to similarly priced single processor system at the time.

    However, it sucks as a game machine. Unless you run Win2000, then all windows games are just on a single processor. While 400mhz is fairly fast for my needs on a gaming system, having it dual processor enabled would make flight simulator, Quake or Reader Rabbit's Kindergarten way faster. :-)

    For the average user using Win98, spending money on a single processor system is the way to go. For the average geek running Win2000 or Linux, then I agree that a dual setup is ideal. I really like mine, but it is not for everyone.

    p.s. I have been spending more time in Windows since I can't find an email client that supports SMTP Auth with mailandnews.com. Otherwise Linux would be almost exclusive.
  • fun to wait eh?

    *yoink* no more broadband for you

    instead we've decided to replace it with a trusty good ole 300bps modem which should provide you with years if not decades of fun on the net :)
  • I still think that the Intel P4 is incomplete as it was released. My guess is that the marketing crowd wanted something to respond to the AMD devices and forced engineering to release a P4 more or less as it was.

    It allowed Intel to release a high-clock CPU, and 'show off' the P4. Unfortunately, it kinda backfired, because people who run clocks that fast tend to care (and know) about actual performance VS numbers. The end resuult was a slow machine that gave the P4 a bad name.

    I think that the P4 architecture actually has some hope --- if/when they ever release a full implementation.
    --

  • I expect we'll all be able to buy systems at least this by next year.

  • http://www.pcfactory.co.kr/1600overclock.htm [pcfactory.co.kr]

    These guys took a 1.2G chip to 1.6G and, frankly, the box alone makes the project cool beyond anything you're going to buy. Sure a Peltier (156Watt) isn't exactly standard kit in an OEM PC but 12x133MHz has gotta fizz.

  • Who needs this for games? Why games?!?

    You're going to need this to run Windows XP + Office XP

    that and a T3. Fear .NET
  • Excuse me, where does it say in Tom's writeup that this system is unstable? Where does it say it overheats, and that Windows crashes every 5 minutes? On the contrary:

    "the Power Box is a real system, running quietly and reliably without making any headaches" - Tom's Hardware [tomshardware.com]

    If Tom found that the box was not running stable, he would have lowered the clock speed until he found a speed that worked perfectly. That's what he does, and that's what he's trying to do here.. push the available technology to its limits and see just how fast fast can get.

    "the real reason we have a computer on our desktops - to perform productive work for our bosses"

    First of all, speak for yourself, not everyone else.. there are plenty of other uses for computers besides producing for an employer. Secondly, how can you claim that a faster computer will not aid in productivity? The world is FULL of applications just waiting for faster computers to become available. Real time video processing, ray tracing, language interpretation, gene analysis.. etc.

  • Not too long ago the Linux community was about getting the most out of hardware. In the days of the Pentium 233, folks were able to make 486/66 systems jump sing and dance with Linux. By the time the Pentium II rolled around, a Pentium 200 was more than enough for an awesome desktop. Even today, most linux distros will work great on a 486/100, but it appears that is changing. The Linux world has slowly been turning into yet-another-windows forced churn, just like Apple in recent years. Code optimization has given way to throwing more hardware at the problem. People that once poked fun at the latest, fastest wiz-bang hardware for being overkill are now chomping at the bit to get that 1.5 GHz P4 or 1.6 GHz Athlon. Some are waiting a few more months for 2 GHz systems before they upgrade. Regardless of the hardware, it's the mentality that's changing. Gone are the days of pride, tight code, and making things work. Today the community just waits for faster hardware to solve their problems.
  • we are losing sight of the real reason we have a computer on our desktops - to perform productive work for our bosses.

    I think you are losing sight of why most people got into computers... I don't overclock my machine so my boss can benifit from it. I do it cause I want that little bit of extra power for me... My games, my apps. If my boss want better proformance out of my machine at work he can buy me somethign better than the P2 400 I'm running.
    MG

  • to perform productive work for our bosses

    You need to re-evaluate your priorities. My "Boss" can go fuck himself. Life is short - I exchange 8 hours of my day for money to eat, clothe and shelter myself. Nothing more. My Heart, Soul and Conscience do not belong to my 'Boss'.

    There's no such thing as a free lunch dudes.

    Bullshit - come to my home; Ill serve you a wonderfull free meal (your all invited). Ive got more than enough for everyone. No strings attached. We can talk about Linux, Life, Art and the Universe. Why so cynical?
  • Do we really need this kind of processing ability? Like come on, who REALLY needs an Athalon running that fast?

    I do. One of my hobbies is 3-D graphics. Now, admittedly, I use one of the lower end software packages, mostly because it does what I want, but even it has taken 5 days to render a 1000 x 1000 scene.

    Of course, most of that was due to my (over)use of objects with their material set to diamond... But the multiple (10+) light sources refracted and reflected nicely!

    (Note to self, stop imbedding light sources inside diamond objects.)
  • ...with the big fuzzy dice hanging down in front of your "chop-top" monitor.

    Imagine a 21" monitor that's 6"x20"...a must for a Lowrider PC amigo.

    Ruger
  • Too bad ol' Tommy boy there didn't spend as much time building his server as he spent on his little 1600mhz gameboy! Seems he's been /.'d
  • Just some thoughts:

    The "dark ages of personal computing:"

    IIRC, around mid-80's ... maybe earlier

    8 mhz 8 bit CPU max.
    640K RAM
    300 bps modem.
    10MB Harddisk (if you were priveleged to have a harddrive)
    video: CGA (Color Graphics Adaptor) Max. Resolution 640x200@2 colors, or, color resolution of 320x200@4 colors. 3D acceleration? Forget it.
    Cost of a system like this? 3-5K

    Today's "Average modern PC"

    In 2001

    1GHz (1000 mhz) CPU (32 bit)
    40 GB harddrive (40,000 MB)
    256MB RAM
    Modem: 1MBIT DSL connection (Who still uses a modem??)
    Video: 1600x1200@32,000 colors, 3D acceleration in hardware. Lots of polygons, really fast.
    Price? About 2K

    Comparison?

    We've increased CPU speed 100-fold,
    Memory capacity more than 400 fold.
    Disk capacity 4000 fold.
    And Graphic capabilities are astounding compared to the systems of yesteryear.

    When you look at it this way, Just remember: CPU power has been the slowest to increase.

    So we get more space, faster and cheaper. With numbers like this, you still have to wait ... (Just look at the Windows 2000 service pack (100MB!! even on DSL that gives you enough time to "smell the roses!")


    ---
  • I remember when I was in school, programming my Commodore 64, and the joy was in just how primitive it all was. Just typing in those BASIC commands, waiting while that tape drive chugged away - that was half the fun of it.

    Nowadays, everything's instantaneous, and people don't realize the fun of waiting. This is a problem with our culture I think. Everything has to be so fast.

    It's all fast food, fast cars, fast living, and it's not good for us.

    It is no coincidence that in countries where they take things slowly that they have lower rates of heart disease, and lower incidence of stress-related industry.

    Sure it's nice to have fast things every now and then, but I worry that people will forget the experience of waiting - the thrill of anticipation as that new game installs, the pause while the computer boots up, etc. It would be great to just go to a nice restaurant with nice slow service, and then to come back and use that Commodore 64 again.

    All this speed means that people don't appreciate what they've got - they don't appreciate the joys of living - the call of birdsong, the flowers coming up in the spring - because they're too busy. And busy doing what? Busy doing things too damn quickly. Of course I'm not saying that progress is bad, but just that this is symptomatic of the ever faster pace of life; the way we don't speak to each other, the fact we take minutes for meals, and seconds for just talking. We should take the time out to enjoy life every now and then.
  • Or you could stick with that old comp and keep watching This Thing [rmitz.org] over and over.

  • Time to make world (Score:3)

    by Zachary Kessin ( 1372 ) <zkessin@gmail.com> on Tuesday February 20, 2001 @04:11AM (#418242) Homepage Journal
    6 days, plus 1 day of rest. :)

  • overclocking, big deal (Score:3)

    by scrytch ( 9198 ) <chuck@myrealbox.com> on Tuesday February 20, 2001 @06:42AM (#418243)
    I wonder how long this monster would take to do a "make world"."

    Unless he got ultra-fast hard drives and boatloads of RAM with it, probably not a great deal faster than an 800 MHz box. Goes double for make world because it has sooooo many files to compile.

    Besides, if I'm not doing games, I'd rather have two boxes that were running within tolerance than one with a voided warranty on the verge of melting.
    --
  • I cannot help feeling that with all the 31337 overclocking going on here, we are losing sight of the real reason we have a computer on our desktops - to perform productive work for our bosses.

    Its all very well having a 1600GHz Athlon on your desk, but what use it it if Windows crashes every five minutes because you are overheating ? Is it just me or does anyone else agree that we should really confine ourselves to running our CPUs at the speed they were designed for rather than some arbitary speed we choose ?

    There's no such thing as a free lunch dudes.

  • Intel should be scared at this news. (Score:3)

    by Heidi Wall ( 317302 ) on Tuesday February 20, 2001 @04:09AM (#418245)
    It is well known that the P4 is a huge disappointment. It has been shown to be outperformed by an Athlon running at 1.2 GHz, and is mainly hampered by its small cache size, which is relatively hampered. Its design was led my marketers, rather than engineers, and so its development was stymied for cost reasons.

    AMD, however, still stick to the tradition of engineer led design. The Athlon, simply the most powerful processor on the market, should be much more scalable to higher clock speeds than the PIII, and will continiue to outperform the P4 until Intel get their act together and release it with the large cache it was supposed to have.

    I forsee AMD greatly increasing its share of the processor market this year.

    However, AMD's future still depends on the Sledgehammer. That processor might just give it a long term edge over Intel, for the first time.
    --
    Clarity does not require the absence of impurities,

  • Re:Time to make world (Score:4)

    by radja ( 58949 ) on Tuesday February 20, 2001 @04:43AM (#418246) Homepage
    well.. factoring in Moore's law (+1 year = x2 speed)
    assuming a creationist view of the world (created 6000 years ago)

    1 day = 24 * 60 * 60 sec = 86400 seconds
    6 days = 518400 seconds
    518400 / 2^6000 = 3.425 * 10^-1801 seconds.

    Yup. That's pretty fast

    //rdj

  • Re:Fast is good, but stable is better... (Score:5)

    by volsung ( 378 ) <stan@mtrr.org> on Tuesday February 20, 2001 @04:35AM (#418247)
    Computers are designed to compute. That's it. You can use them comercially or personally. Some people want to tinker, and some want stability. These guys built a 1600MHz Athlon not to slap on their desk at work, but to play with and see what the current technology can do. I don't think anyone is going to tell their boss that making a server out of one of these things is a good idea.

    I say run the processors at whatever speed you like. Just understand the potential consequences.

  • Re:Time to make world (Score:5)

    by smartin ( 942 ) on Tuesday February 20, 2001 @04:23AM (#418248)
    That was a long time ago, things have improved considably since then.

  • I'll take quantity over speed, thanks. (Score:5)

    by benmhall ( 9092 ) on Tuesday February 20, 2001 @04:40AM (#418249) Homepage Journal
    Seriously, 1.6GHz sounds VERY impressive, but I'll take more CPU's and better components over a faster CPU any day.

    I've had my Dual 466 Celeron for over a year and a half now, and it's absolutely fantastic, and rock-solid stable. Sure, I've upgraded the RAM over that time from 128MB to 512, but through it all I've felt no need to upgrade the processor(s)

    The motherboard recently went south on me and I had to replace it. I got looking around and noticed that Asus now has a dual PIII board for ~$230CDN. I ended up just RMAing this board, but I know when I do eventually need to upgrade there's no WAY I'll be going back to a single processor board.

    If you're running Linux, FreeBSD or Win2k (or even BeOS) an SMP system makes a world of difference under heavy load. Recompiling? Encoding MP3's? Running VMWare? These operations are sped up very noticeably.

    For people looking for a new machine: Save your precious dollars on the fastest processor. Fill up on RAM, get a good video card, and get an SMP board. I'd rather have 2 800MHz chips than a 1.6GHz any day of the week.

    AMD: I'd rather get an SMP chipset out of you than Yet Another "Fastest" Processor. I'd much rather own a Duron or Athalon than a crappy Celeron or PIII, but I'd take an SMP Celeron over a single Duron..

  • Lowrider (Score:5)

    by Pope Slackman ( 13727 ) on Tuesday February 20, 2001 @04:33AM (#418250) Homepage Journal
    I wanna see a Lowrider computer magazine.
    I mean, chromed RAID arrays, hydraulic monitor positioning,
    overstuffed, ergonomic workstations, boxes covered with shaggy purple fur, golden G4 cubes buffed to a mirror-like finish...

    THAT's where it's at.

    --K

  • My 486 (Score:5)

    by sharkey ( 16670 ) on Tuesday February 20, 2001 @04:31AM (#418251)
    $make world.c
    Hmmm.... immeaurable by the naked eye. Let's see if it runs.
    $./world
    "Hello, World."
    Yippee! Don't need no 1600MHz Athlon to make my world!

    --

  • Re:Why must everything be so fast? (Score:5)

    by oops ( 41598 ) on Tuesday February 20, 2001 @04:42AM (#418252) Homepage
    I didn't finish this. Who has time to read 6 paragraphs ?

  • "make world"... (Score:5)

    by Snard ( 61584 ) <mike.shawalukNO@SPAMgmail.com> on Tuesday February 20, 2001 @04:07AM (#418253) Homepage
    I wonder how long this monster would take to do a "make world".

    Six days?

  • Non x86 processors? (Score:5)

    by sl3xd ( 111641 ) on Tuesday February 20, 2001 @04:55AM (#418254) Journal
    How 'bout some comparisons between some non-x86 processors?

    I've never seen anything about how fast a fired-up Alpha can go.

    Or how fast the 1.6 GHz Athlon compares to the 733 MHz G4 (Except from Apple, of course)

    I use an X86 processor too... but there's better stuff out there.

  • Re:Fast is good, but stable is better... (Score:5)

    by plover ( 150551 ) on Tuesday February 20, 2001 @05:12AM (#418255) Homepage Journal
    Its all very well having a 1600GHz Athlon on your desk, but what use it it if Windows crashes every five minutes because you are overheating?

    News flash: processor heat is probably not to blame for Windows crashing.

    The reason for the overclocking is really to get the Windows boxes to reboot faster after a crash. Have you timed an NT reboot lately? 1600MHz might make it endurable.

    Is it just me or does anyone else agree that we should really confine ourselves to running our CPUs at the speed they were designed for rather than some arbitary speed we choose?

    Seriously, I think it's just you. Remember, CPUs are designed to run as fast as possible. The limitations being sidestepped by the overclocking crowd are physical world limits: heat will cause failures in the CMOS semiconductor junctions. You sound like you are saying we should remove our fans and heat sinks and run our 500MHz boxes at 33MHz, because that's what God intended.

    Don't confuse the engineering limits with the marketeering limits set by the Intel folks, either. They don't want people overclocking their cheap chips (and so avoid paying the premium for "faster" chips), so they put in circuitry designed to detect and prevent overclocking. They claim it's to "protect their name", but it's strictly financial.

    John

  • Re:Why must everything be so fast? (Score:5)

    by seanmeister ( 156224 ) on Tuesday February 20, 2001 @04:40AM (#418256)
    ....waiting while that tape drive chugged away - that was half the fun of it.Nowadays, everything's instantaneous, and people don't realize the fun of waiting.....We should take the time out to enjoy life every now and then.

    ...Which is exactly what you SHOULD be doing instead of watching your tape drive chug away. While I agree with you in that the "ever faster pace of life" can be kinda hard on the ol' personal well-being, I don't think that (topic)faster computers(/topic) are a part of the problem. I certainly have better, more fulfilling things to do than sit around while my kernel compiles. The quicker my system boots and connects to the net, the quicker I can check the local events, weather, whatever, and get the hell out into the sunlight.


    Sean

  • It'll Never Happen (Score:5)

    by TheNecromancer ( 179644 ) on Tuesday February 20, 2001 @04:11AM (#418257)
    Tom's Hardware Guide is unfortunately not a car magazine (yet...),

    Somehow I don't think that babes in bikinis dripping over the latest PC system will sell magazines(I could be wrong).

Slashdot Top Deals

"What if" is a trademark of Hewlett Packard, so stop using it in your sentences without permission, or risk being sued.

Close