Follow Slashdot blog updates by subscribing to our blog RSS feed

 


Forgot your password?
Close
typodupeerror
×
AMD

AMD Athlon 600 Preview 128

Posted by CmdrTaco from the stuff-to-read dept.
An anonymous reader sent us a Firingsquad article on the AMD K7, which provides some benchmarks comparing the K7 and the PIII.
This discussion has been archived. No new comments can be posted.

AMD Athlon 600 Preview More

AMD Athlon 600 Preview

Comments Filter:

  • HUH? (Score:1)

    by Anonymous Coward
    What's so great about alpha EV6? Is www.freebsd.org offereing jobs? Do millions of people have $8 for a movie? I'm just going to drink pepsi and stay away from that coke. yuk.

    ^--- makes as much sense as what you wrote :P

  • # of Mhz is purely a psychological weapon (Score:1)

    by Anonymous Coward
    checkout www.amd.com for comparisons btw. athlon 600 mhz, athlon 550 mhz and pIII 550 mhz.
    the difference between an a550 and an a600 is very small.

  • VA won't do AMD (Score:2)

    by Anonymous Coward
    Look at who has invested money into VA, and it might get a little more clear to you why they don't sell AMD products.
  • The damned thing runs Linux. Now who feels stupid?

    - A.P.
    --


    "One World, One Web, One Program" - Microsoft Promotional Ad

  • The damned thing runs Linux. Now who feels stupid?



    - A.P.
    --


    "One World, One Web, One Program" - Microsoft Promotional Ad

  • You're comparing it to the wrong chip. Everything I've seen and heard about the K7 indicates that it'll beat the pants off not just the Celeron but the Xeon at the same clock speed... and it clocks faster, too. If you compare a K7's price tag (and we're talking $400 or so for the 500s) to a (slower) Xeon's, AMD comes out ahead by a serious margin...
  • Posted by MaverickPl:

    The thing was that they used to run with an MS-Access database and ASP but that couldn't keep up so they moved to a linux box. They have never really changed the extensions because I guess that there html editors work better with .asp (just speculation). You can check netcraft and it shows that they are using apache though...
  • Equal Mhz. Who cares. It's equal $ that matters. Anyone got prices and benchmarks for these chips?

  • When buying a CPU Mhz doesn't matter, it is price vs performance that matters. CPUs of equal price should be compared in these kinds of tests.

    It's a pity that they don't have a price vs performance graph etc.
  • What's so great about the AMD K7? Is www.firingsquad.com offering jobs? Do millions of people have $950 to spend on Athlons? I'm going to stick to the old overclocked Celeron 466 for $144.
  • I'm sure you're familiar with integers. Whole numbers only, no fractions. Maximum number determined solely by the number of bits available and is a power of two.

    Floating point is different. We're familiar with decimal fractions because we're taught arithmetic in decimal. Binary floating point - which can be made very fast in silicon - is a strange beast. Binary floating point is good when the numeric results stay meaningful only for the PC. Such as 3D rendering. This is because binary floating point has different error characteristics than decimal floating point. You can't represent 0.1 accurately in binary.

    But integer calculations are simpler and faster to do in silicon and many applications don't need binary floating point because they can either synthesized decimal floating point (with it's familiar error characteristics) or they simply don't need fractions.

    Wade.

  • Just to add more fuel to the fire here, the huamn eye *can* precieve more than 30fps, but there is a bandwidth limitation on the visual cortex. What you end up getting is aprox 30fps, but with what would be the equivelant of very very good motion bluring. The reason people claim that we can see the difference between 60fps and 30fps is that your eye is bluring the 60fps giving you the impression of smoother motion. However, if you played back the same sequence at 30fps with good motion bluring, you should get the same effect. Notice that you *can* precieve 60+ fps, if you look out of the corner of your eye at a white part on your monitor you maybe able to see the flicker. This however, doesn't mean that you will be able to notice this when trying to maximize detail and focus on moving scenes. One of the most important things is to keep a consistent framerate. a constant framerate will emmerse the viewer in what he's watching. It's as if the brain likes a steady pattern. It's almost a hypnotizing effect. When you have a variable or uneven framerate, you lose that effect. another interesting thing to note is that around the edges of your vision, the relative resolution is much lower. It'd be interesting if we could utilize that with virtual reality hardware.
  • This was a decent review. The author, Kenn Hwang knew his stuff. It is worth the reading.

    About the chip... I can't wait. I tried to see if VA Research was going to work on optimization for it. I got the salesmans response of we have other machines, wouldn't you rather buy that?
  • Hi.

    When reading those benchmarks I always wonder whether the integer results and the floating point results are in the same scale.

    If they are on the same scale, would not it be much better to do all the calculations in float? (If you've got huge numbers, you loose some accuracy, and you cannot use floating point in the kernel, but what about the rest?)

    Thanks,

    Liran.
  • So far, I've read the human eye thinks "smooth" is 15 fps all the way up to 55 fps. Of all places, FiringSquad has this issue covered. [firingsquad.com]

    I agree with the sentiment; 70 fps can cause motion sickness. :o> Reason why some people refuse to play Quake and EverQuest; it's not because they don't like playing games, they just get physically sick of watching those polygons.

    18 fps was the quality of most .mov videos when Quicktime 1.5 was the big, big deal. I remember those looking choppy as hell. Maybe different eyes are more or less sensitive than "normal" eyes. I dunno. All I know is that you can't tell geeks they can't use technology if it presents itself, so by next year we all could be convulsing in epilepsy.

    J.

  • Where did you get these numbers??? They're all wrong.

    The human eye percieves smooth motion at about 20 frames per second.

    No. It varies depending on contrast and ambient lighting, but 60 frames per second is often cited for purposes of generalization.

    TV (in the US at least) is broadcast at 24 fps

    No, U.S. TV is 30 full frames per second (interleaved from 60 half frames per second to reduce flicker). European TV is 25 frames per second, due to 50 half frames per second interleaved. (In both areas, there's a historical and RF noise connection to the 60/50 hertz power lines.)

    movies are usually at 30 fps

    That doesn't ring a bell; I think it's actually 24 frames per second double-shuttered to give an effective 48 fps. I may be misremembering the precise numbers there.

    At low resolutions used for games (640x480, 800x600), many graphics cards can supply the monitor with a vertical refresh rate of 85 Hz, but at high resolutions, all but the most expensive cards (things like cards designed for CAD, such as the FireGL cards) drop off in maximum vertical refresh rate.

    Yes and no; what you said is very misleading, since recent higher end cards from Matrox, and the Nvidia TNT2, retain high frame rates even at the highest of resolutions (approaching almost 2000x2000 these days). Now it's true these are "high end cards", but that's in a consumer sense of very roughly $200, they are *not* high end CAD market cards, which can cost thousands of dollars.

    I found 1024x768 to be a much more desirable resolution for playing quake. So, what use is it if the Athlon can push out frames faster than your monitor can display them???

    That's largely a matter of taste. It's true that not everyone has high quality monitors, but on the other hand, a fair number of people (including gamers) do in fact have high bandwidth monitors. Your taste (and monitor) isn't everyone's.

    Granted, there are other areas where fast 3D performance is a big plus (rendering movies, etc.), but for games, 70fps is absolutely pointless

    Now that goes way too far. Actually, a monitor refresh rate of a minimum of 70 hertz is highly recommended to avoid perceptual flicker under adverse lighting conditions -- this varies from person to person, but many of us see flicker, at least in peripheral vision, almost always with 60 hertz displays.

    If you had said "85 fps monitor refresh rate is pointless", you'd be closer to the truth, although that too is debatable for more complex reasons that we're getting into here.

    (I usually use 1600x1200 32 bit color 80 hertz for non-game (2d) purposes, and for games, usually have to drop it down, since I don't yet have a TNT2 nor voodoo3. The highest res of my voodoo2 is less than that of my 2d card.)

  • with a TNT2U of course, what's even strange is that the matrox G400 is better than the voodoo3, i think 3dfx lost the battle finaly...
    also if they can test the athlon with a Neon250 [powervr.com] the result should be great also, certainly more than 6000.
    --
    http://www.beroute.tzo.com
  • AMD announced second quarter loss of $162 million, hoping it'll not affect their future, it's hard to fight with bigger (Intel), like BeOS/MacOS/Linux fighting against M$...
    --
    http://www.beroute.tzo.com
  • Take a look at the demo they used - crusher.dm2, the traditional "put a lot of polys and a lot of lightmaps on the screen so the CPU is the bottleneck not the video card" demo. With most game demos, you max out around 90fps because of the fill rate of the video card, because game designers keep their poly counts low to accomodate low end CPUs.

    Off-the-shelf, AGP 3D cards now are competitive in fill rate even with sweet SGI workstations, but for the geometry portion of rendering PCs get blown away, and seriously blown away with high polygon scenes. Until we get PC dedicated hardware to do geometry acceleration and not just texture filling (doesn't the G200/G400 do some geometry on the card?), we need heavy CPU FPUs to do it for us.

    Sure, spitting out a bunch of 60 poly models at 120 fps might be overkill, but displaying 400 poly models at 60 fps won't be. In the near future I can imagine game designers giving "low poly -> high poly" options in the video setup just as they give "320x200 -> 1600x1200" options today to balance quality vs. performance. (doesn't half life do this to some extent already?) We're seeing the first 3D engines that render bezier curves by breaking them up into polys (does Q3 do this when the level is compiled or when it's played?) and the more polys you use the smoother it looks. Hell, the more polys you use the more cool stuff you can do with background effects (remember how the Unreal flies blew away the Q2 buzzing dots?) and realistically moving human models. And man, some of us want to play games with that kind of image quality.

    Oh, yeah, and it'll be a big boon for scientific computing, video editing, professional rendering, etc. when they're reasonable on cheap x86 hardware too.
  • They are currently selling K6-2 and K6-3 systems.
  • This must be an imposter.
    1. You cannot buy one yet.
    2. Motherboards aren't going to be readily available till August or September at the EARLIEST
    3. PRE-ORDERS for the K7 still resemble:
      • $490: K7-500
      • $710: K7-550
      • $940: L7-600
    4. While retail price for an IN STOCK P3 resemble:
      • $680: P3 550
      • $440: P3 500
    5. Intel is due for another BIG price-drop this month.
    6. K7 performs admirably when just stressing the CPU. But when Real-World performance is taken into account, you see a difference of approximately 20% between it, and a chip 10% slower than it in Mhz ratings. I don't know about you, but the extra marginal performance isn't worth the premium pricetag.They noted problems remeniscent of K6+SS7 (AGP cards).
    7. All this overclocking that people keep talking about in conjunction with K7 isn't going to happen. No. The chip isn't hard-locked like P3 and Celeron. But it requires a proprietary programmable connector which AMD is NOT selling. If you're an overclocker, AMD is not a choice for you anymore.

    Fastest? Yippie. I'll wait till the platform is proven stable.


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

  • 6: You have to remember. Overall performance equates out to 10-20% between a P3 550 and the K7 600. That's marginal.

    7: Again. Supposedly you're not going to be able to change ANY of the processor settings without this proprietary and non-available connector. From AMD's site:
    High-performance cache architecture featuring an integrated 128KB L1 cache and a programmable, high-speed backside L2 cache interface.

    Guess what this means. If you cannot get the RAM to function above it's operating specs (which is what PROGRAMMABLE implies) then all you can do is:

    1. Burn it out running above spec while the L2 runs at it's preprogrammed speed.
    2. Disable the L2 (which should hurt performance).

    Don't get me wrong. I know if Intel were in the same position they'd be charging a testicle-tightening amount for their chips. I just don't think an overall performance bump of approximately 10% clock for clock is worth such a high markup. Which is also why you won't find me buying Xeon any time in the near future.


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

  • "Winbond" is a Taiwanese company that makes temperature monitoring chips. All modern Pentium II/III motherboards have one, and Linux has supported them for a while (since at least 2.1.93 or so, and possibly sooner). There's even GNOME/KDE/WindowMaker/Afterstep graphical temperature readout apps - search freshmeat.

  • Oops... errors. (Score:3)

    by juuri ( 7678 ) on Friday July 16, 1999 @12:11PM (#1798582) Homepage
    TV is 30 frames (60 fields per second).

    Movies run at 24 FPS.

    It is a well stated and known fact that 24 FPS and even 30 FPS give away errors in things that are filmed (ie spinning objects like wheels). To fool the human eye into thinking it is seeing real motion it needs not only 60 FPS but motion blur as well. The newest generation of game consoles can do this... Some people can easily tell the difference between 70 and 60 FPS. If you can't then count yourself as lucky because you will be satisfied with less.

    ---
    Openstep/NeXTSTEP/Solaris/FreeBSD/Linux/ultrix/OSF /...
  • The problem is that on many cards, when you're getting "70 fps", you're really getting 50 to 90. And when you're getting "45 fps", you're really getting 30 to 60. And when you're getting "30", you're really getting 15 to 45 (or something like this). Where it gets choppy is when you drop to the lowest point.

    All the "timedemos" that people hear about are really averages. So "70 fps" might be the minimum for one game for playability if it means that it's really dropping down low at some point.

  • I don't think you will be able to find a smp capable motherboard for the k7 till Q4 at least. Motherboards for the k7 are going to be hard enough to come by. I'd love to get a k7-500 and a dual MB and save for the 2nd k7. I don't think thats gonna happen. I read somewhere that you won;t be seeing dual k7 MBs till late November/December. Sucks.

  • But... (Score:2)

    by Axe ( 11122 )
    Which is also why you won't find me buying Xeon any time in the near future

    But isn't Xeon the biggest money maker for
    Intel? It looks like people ARE paying, and
    that's what AMD bets on.

    It's like NT/Linux network benchmarking. Half that would saturate you bandwidth anyway - but people do pay attention.

    Being the best and being just good enough is a huge difference in marketing world.


  • This debate is getting tiresome. Remember, 70 fps is the AVERAGE. While you're right on with the refresh thing, there will be sections of the game where a 70 FPS average will include, say, 130+ FPS staring at a wall (good) or 30 FPS in a heated fight (bad).

    Going from 70 FPS down to 30 FPS is quite a visible jolt (to me anyways). The way around this is to have enough CPU power available to keep the frame rate as high as possible in complex scenes.

    One way is to boost the CPU power dramatically. Another (similar) way is to use SMP, which is what the yet-to-be-released new version of Q3Test has.
  • I'm excited by the results, but I would like to see PIII / Athlon benchmarks for chips operating at the same Mhz.

    In these benchmarks, it appears that their will be only marginal differences, although the Athlon is currently at the low end of a architecture (motherboards, etc), while the PIII has almost maxed out its design.

    -OT
  • finally are the benchmarks including the Xeon, which is what the Athlon is really supposed to be a competitor for. While it handily beats the PIII it is still planned for use in servers/workstations and not in home PC's, yet.
  • some semi-pointless info
    WinBond is now owned by American Megatrends, makers of AMIBIOS among other things.
    As someone else pointed out, they make a LOT of other things besides monitoring chips. Just about any type of important chip inside a pc has a version of it made by winbond I think. From IDE controllers to temperature monitors, winbond makes it, I think.
  • OK, consider these numbers:
    The human eye percieves smooth motion at about 20 frames per second. TV (in the US at least) is broadcast at 24 fps, and movies are usually at 30 fps. At low resolutions used for games (640x480, 800x600), many graphics cards can supply the monitor with a vertical refresh rate of 85 Hz, but at high resolutions, all but the most expensive cards (things like cards designed for CAD, such as the FireGL cards) drop off in maximum vertical refresh rate. I don't know about you, but I found 1024x768 to be a much more desirable resolution for playing quake. So, what use is it if the Athlon can push out frames faster than your monitor can display them??? Granted, there are other areas where fast 3D performance is a big plus (rendering movies, etc.), but for games, 70fps is absolutely pointless.
  • Forget using your hand to see the flicker in your screen. Stretch a rubber band out and hold it up in front of the monitor. Then pluck it like a guitar string. You should be able to see the rubber band pulse up and down with sine waves if you get the tension right.



    Also, the reason that a monitor is partially black isn't really an aliasing thing. It's just that any CRT is really only partially lit at any given time. The entire image never exists on the screen at once, it only exists in video memory and your brain. Simply taking a still photograph of a TV will demonstrate this.



    On the other hand, when you see a monitor on TV the black band will be moving either up or down the screen at some frequency. The rate at which the band moves is an aliasing artifact whose rate is determined by the refresh rate of the monitor and the rate at which it is being "sampled" by the camera.
  • I won't try to speculate as to motherboard costs, but the other information is fairly easy to locate in the various articles that have appeared on the net. Because AMD is attempting to minimize the cost to third-party developers, they are using a design that is physically fairly close to the Pentium II spec. This leads me to believe that it will be an ATX design. Based on the information in this article, the best power supply to use seems pretty obvious: the best one you can afford. Or in other words, any name-brand quality PS should be fine, although I would look for 300w to be on the safe side. And the RAM question is simply a question of what is supported. Since the first chipsets have been stated to support PC100 SDRAM, (and I would guess PC133) then you want to get RAM with a good CAS rating (look for 3) and a quick speed (IE 6-7 ns). As the newer chipsets and MB designs become avaiable you will want to look into RDRAM, but as the yields are still very low I don't think this will be a large concern for quite some time to come. Good PC100 SDRAM looks to be of use for quite a while.
  • I can't find it now (archive's not working), but on amdzone they had a report on the power consumption of the Athlons (not sure who the source was). 60W!!!! Sheesh. Not that I don't want one, though!
  • Try plotting the performance curve of a CPU in relation to performance and speed. You will most likely notice that it is some type of a curve. It is not a linear increase in performance. A 560mhz processor is not 93.3% as fast as a 600mhz processor because of that. A more rational approach would be to calculate the equation of the exponential growth of the P3 line and extrapolate the performance of a 600mhz chip. Your results might turn up quite a bit different than your crude and elementary approach. If you send me some text data on P3 performance I would be happy to do the math for you!

    Noehre
  • Most OEMs are saying ETA for this chip is 8/16. AMD's page doesn't give an exact date, just Q3 99. Large Manufactures should already have them according to the Athlon(bleh) press release. Prices are kinda neat. AMD is shipping them in 1000 quantiy units, with the Athlon 600 at 699 per chip(in the 1000 quanity). Only 324 for the 500 MHZ, 479 for the 550.
  • Most gamers use Win 9x so SMP will not effect them. But when Q3A comes out then you'll see a lot of gamers look around for a OS that can scale! I always liked AMD but lately I have warmed up to the performance/price of Celeron! Of course the PIII tracking system will force me to go AMD all the way now.

  • It was a smaller loss than was projected for them, and much of it can be written off as investing in the fabrication process for the K7 line (sorry, I think "Athalon" is just too cheesy...) -- their Q3 and Q4 numbers should be very interesting.
  • I plan on building my own K7 box when possible. These benchmarks were nice, but I already wanted one anyways. So, what I REALLY want to know is..


    How much will the motherboards cost?

    What type of case is needed? ATX?

    Apparently Athlon needs a lot of power, so what kind of power supply is needed?

    Is PC100 SDRAM the best RAM to use for it?


    That's it for now.. anyone care to fill me in?
  • I was hoping someone would try to estimate the costs of the motherboard, as I'm trying to determine just how much money I'll need. But thanks for the information. I still haven't decided whether I should wait for the K7 to go copper or not. I get the feeling I'll be waiting forever if I'm worried about something better coming out.

    And concerning your sig,
    Why does everyone spend so much time wearing digital watches?
  • maybe I'll start caring when I can get one by itself with a MB. until then it's just a big tease
  • plus, not to mention the fact that once you're used to the power, you can almost never go back...

    Consider quake1/2 on a plain vanilla voodoo1 card... not too shabby.... I upgraded to a voodoo3 and was heartily impressed.. I'd never go back... especially since q3test runs okay but slow on a voodoo1...

    Another good point was made about cars... sure, I could drive around a chevrolet metro, but why, when I'm used to the power of my tweaked Camaro Z28???

    Not going to happen. I'll keep the speed, and any increases (even if not "necessary" by any standards) are more than welcome.....

    VooDoo3 and V8 speed for everyone!!! :)
  • Ok, 70 fps is useless, agreed. But think about the future.. more grahpically boated games. I have bought a Voodoo2 card a while ago, Quake2 was
    FAST at 800x600, I was more than happy. Now I got Q3Test. It is still playable but the frame rate is lower than with Q2 and I play in lower resolutions now..
  • Why is the loss so big anyway? Is x86-development really that expensive? Or is their other chip production not making money anymore?
  • Next time you watch the news look for a computer monitor in the image.. Does it look partially black? If so it's the same general problem you're having the frame rate of the television camera is aliasing the image on the monitor.. Think of what would happen if you turned off all the lights in the room and turned on that old strobe light stuffed in the corner of your closet.. I'll bet that you'll immediately see the same problem you see on the newscast...

    that's not right ether, beacuse the monitor generates it's own light, so you see it weather or not the lights in the room are on.
    if you want an example of how frame rates are effected, wave your hand around infront of the monitor. you'll see several crisp images of your hand's silluete, instaid of a single blurry image
    _
    "Subtle mind control? Why do all these HTML buttons say 'Submit' ?"
  • actualy Multiplys on pX chips use the CPUs floating point unit, as far as I know.

    the thing is, int math is also used for calculating things like RAM location and stuff.
    _
    "Subtle mind control? Why do all these HTML buttons say 'Submit' ?"
  • actualy Multiplys on pX chips use the CPUs floating point unit, as far as I know.

    the thing is, int math is also used for calculating things like RAM locations and stuff. so any errors could cause a complete crash of the system
    _
    "Subtle mind control? Why do all these HTML buttons say 'Submit' ?"
  • Any recommendations on a dual processor MB? I'm building a new machine to use to run my Quake3 server, going to use AMD cpu - I'm interested in the idea of going SMP.

  • His MySql database is starting to timeout. I am not sure what platform the dbase is running on, but I am sure that is the problem.
  • Nice to hear for the Intel employee
  • A new solution to this problem was presented in the 4/1 issue of Computer Graphics. It's called 'temporally bidirectional tesselation' and it uses the CPU idle time during simple scenes to prerender more complex scenes. Rumor has it that Carmack has implemented this for the new version of Q3ATest.

    ( =^) for the humor-impaired)
    --

  • Cool... I was going ot buy a new system anyway.... so should I buy K7 600 in a few months when prices go down a bit or save up for 750MHZ or better yet 1GHZ.... Yeah Baby!!!!

    Ahhh... the difficult choices of a computer geek :)

  • Re:Who needs 70 frames per sec? (Score:3)

    by aressa ( 35705 ) on Friday July 16, 1999 @12:02PM (#1798612) Homepage
    I can see the debate now, it has been on /. many times:

    24/30fps for Movie/TV takes into account a 24/30'th of a second of "motion", on games it is a crisp static image... the more frames you can show, the more it "blurs" like reality.

    It really does make a difference.

    A
  • 550 to 600, that is an increase of under 10% (damn my math rules!), and in the important numbers the Athlon was above even that.

    Look at the Quake2 Crusher benchmark at low res (the best way to test FP performance bar none) and the Athlon did over 70 fps, against the P3 not even managing 50. That is what I call impressive (Are you impressed? :-) ).
  • Without a doubt, the best option today is to get a double slot1 motherboard and get two of the "slockets" that allow you to put a socket 370 Celerons in SMP.

    Which Celeron is still a bit of an issue, but the 300 OCed to 450 (100 fsb) is probably still the best price to performance. I understand the 466 Celerons do 525 (with 75 fsb), so that might be better, but they are at least double the price.

    Or maybe one should wait for the Athlon, but the above MB + Processors still costs less than the cited prices for the Athlons...
  • It is my understanding that it is hard, but still possible, to get ahold of them. I live in Europe though, maybe they are selling out faster over there...
  • It does not matter if your monitor has a refresh rate of 85. If you feel that your FPS is capped by your monitor's refresh rate you can disable VSync. You can do it either through drivers or even in-game. Minimal artifacting might be present, but that is the tradeoff for getting the maximum speed possible. I won't even get into the 24/30/60 FPS debate now.
  • Does anyone know what that means? I get nervous
    when pieces of hardware are named Win-something.
  • It is finally good to see some data from a neutral source.
    It is also very nice to see comments describing waiting for Intel to catch up.
    As for the comment somewhere on the page saying that the K7 eats 60 W, ouch! No wonder the thing has a running temperature of 30 C and greater!

  • What its like (Score:3)

    by oldzoot ( 60984 ) <morton.james@nOSpam.comcast.net> on Friday July 16, 1999 @11:14AM (#1798619)
    I think the most important aspect of this new chip is the continued pressure of competition on Intel.
    This should increase the motivation for the continuation of gains in performance and cost / performance ratio. In terms of what Joe Doakes average hobbiest gets for the money, I think the optimum design right now is a dual processor system using a not-quite bleeding edge CPU - say a PII 450 or therabouts. The cost of adding a second CPU to a system seems to be less than the cost of bumping up the clock another hundred or so MHZ. This does not account for improvements in internal efficiency, but you do get a lot more bang for the buck with 2 CPU's. The more people who have multiprocessor systems, wether SMP, clusters or even shared-memory systems the faster we will see the evolution of new software paradigms that will truly advance the state of the art.
    Z
  • Hey if you think about it for game a K7 will give what almost double the performance? Some people have the extra 400 bucks and want the performance. Especially me, waiting anothe month of so to save up the extra money is worth it in a lot of cases espeacially if you are a gamer, or 3D designer. Very few people otherwise need the power. (I mean if you run a webserver or do other compute intensive stuff, then you can run dual, but most games are only single threaded.)
  • Actually since games have lower detail levels than a movie in each frame, it needs a higher frame rate to give the illusion of smooth motion.
  • Who in compiles and player mp3 while playing a game? I barley get enough fps as it is (PII 400 , TNT)
  • I don't see the hight markup. The 550 version barely cost 40$ more than the PIII.
  • The ping times are good:

    64 bytes from 216.200.57.12: icmp_seq=0 ttl=248 time=316.5 ms
    64 bytes from 216.200.57.12: icmp_seq=1 ttl=248 time=365.4 ms
    64 bytes from 216.200.57.12: icmp_seq=2 ttl=248 time=94.1 ms
    64 bytes from 216.200.57.12: icmp_seq=3 ttl=248 time=335.0 ms

    But the response time for the page is way slow.
    http://www.firingsquad.com/hardware/athlon600pre view/page3.asp
    Is .asp the problem?
  • Who uses Access for high-traffic web sites?. It isn't suited for multi-threading.. so the problem isn't ASP but Access
  • I should have probably used 'was' instead of 'is'.
  • They are commiting a crime. Maybe not a 'hate crime' but crime.
  • The name Winbond has absolutely nothing to do with any particular OS, which I don't need to name. They were making IO boards about 10 years ago. I think at some point they got bought out by AMI.
  • This is called "linear interpolation". It has absolutely no relevance to what the scores would actually be. Some of the benchmarks are more relative of the AGP, the RAM, etc. When the PIII-600 actually comes out, you can watch it lose.
  • Wow, I hope you realize how monumentously stupid and bigoted that proved you are.

    Man, I didn't know that hating specific operating systems made you a bigot.

    Does that mean that people who try to crack micro$oft.com [microsoft.com] are commiting a hate crime?
  • That would be the Nyquist theorem, not the 'Shannon sampling theorem.'

    Ray
    --
    Every truth has a context.
  • Imagine using glasses like Crystal Eyes (lcd shutter glasses) if your card can pump out 70 fps that goes to 35 fps per eye which in turn equals very smooth True 3D graphics :-)
  • I'm not sure if Win2k will solve this problem or not but soundcard companies are notorious for producing NT drivers that are SMP unfriendly. Soundblaster Live had major problems running under SMP until everyone complained and they fixed it. I've tried the Guillemut's HomestudioPro64 and it chokes under SMP. I have a Diamond MX300 now and it's not happy under NT either (even using latest Aureal latest RC drivers). Whenever I load a game in NT like Starcraft or Quake3test, I'd get about 2 seconds of sound before it quit working or static. I've had the wav device choking while running Realaudio or Winamp also. As soon as I pull the second PII out everything works. So I'd recommend you wait till Win2k comes out and see if there are any improvements before going for the dual processor gaming.

    As far as dual motherboards goes, check out Asus, Tyan, or SuperMicro.
  • Ugh. Well, there's no way trying to fool tech folks I guess :(

    As most of you probably imagined, our server was on NT before we made the move to Linux. We kept the .asp extensions to make the transition smoother and avoid breaking external links. I believe the actual machines are 2 load-balanced dual p3-xeons with 1GB ECC each...still a bit flaky though.
  • Winbond makes a LOT more than just temperature monitoring chips. THey produce a lot of different I/O chips that used to be staples of the motherboard business before completely integrated solutions ("motherboard chipsets") became available...
  • When is it planned for release?
    --
    Alan L. * Webmaster of www.UnixPower.org

Slashdot Top Deals

It seems that more and more mathematicians are using a new, high level language named "research student".

Close