Forgot your password?
typodupeerror
AMD

K7 Info 35

Posted by CmdrTaco
from the stuff-to-read dept.
TBM writes "This is an article on the new K7 chipset from AMD. Pretty much all information that has been released to date and even rumors are included. Information on what mainboard manufacturers will be supporting the new slot-A and slot-B and which tier-1 OEMs will be selling K7 systems. A very useful read for anyone interested in a new system... "
This discussion has been archived. No new comments can be posted.

K7 Info

Comments Filter:
  • BFS can handle files larger than 18 terabytes. I'm not going to put down a number, because then I would be the third person that gave a wrong number in this thread :) If you really must know what BFS can handle, go to Barnes and Noble and thumb through Dominic G's filesystem book.

    As for Genki, chill out... It's so close you can almost taste it. Barring nuclear war you should have it installed RSN.
  • I'll take the risk of being wrong :-)
    18 petabytes is the current maximum for BFS.

    Roll on Genki...
  • Windows 2000 supports 64gigabytes of memory too.
  • I think that back in the 80's Bill G said something along the lines of ``nobody will _ever_ need more than 256 bytes".

    Wasn't that the old DOS memory '640K will be enough for anybody' line? .. still, I fail to see why you'd need to be able to support 18TB files *JUST* yet considering noone even makes a 1TB HD or other storage device yet. Okay, you even RAID some HDs (22GB?) together and have yourself 0.1TB of storage .. whooo doesn't look like bouncing off the limits yet :)

    Let's not worry about 'still' having 32-bitness yet... it's good to get the most out of things before moving on to new architectures, can you say 'backward compatible'? :)

  • I won't buy until I get the _full_ specs

    You won't be able to buy one until after you got the full specs. Intel aren't stupid, they are funding gcc and by extension Linux development on IA64, and they will release all the docs before the hardware is on sale.

  • Isn't it time for x86 to die? I mean theres no use for chipmakers to make thease internal risc hardwareemulating x86cisc CPU's enymore is there?
    This must cost lots of money? Why not just go risc all the way? I mean more speed less cost.
    Cars don't look like horsecariges anymore no do they?
  • Right, I checked it out a couple of times and in fact, BFS does support files up to 18 PETAbytes.

    So incredibly huge, the mind boggles. In 1999 there is probably NO practical use for such a file, but hey, it's cool to be able to brag about it.

    Got Genki?
  • it's June already, and I don't have a K7 under my desk. I ask you AMD, why?!

    Ah well, it's not like these K6's will break down anytime soon.
  • 1) Announce chip that will blow away Intel's top of line
    2) Allow delivery date to slip
    3) Finally deliver chip that doesn't meet expectations
    4) Repeat
  • Those "strange practices" are the best AMD can do. They have some excellent designs, but they just don't have the kind of production facilities intel does. Intel has something like 13 fabs. AMD has one. Plus, with this one fab they are making K6-2s, K6-IIIs, mobile K6-2s, mobile K6-IIIs, and now K7s (I believe the K7 chipsets are being made by a Taiwanese company). I'd like to see you do any better. Things will definately improve when AMD's massive new fab in Dresden, Germany is complete.
  • Over in his Computex recap. [tomshardware.com]

    In it, he says that so far the K7 is looking like a "clear winner" and that for once, its floating point performance beats Intel's. He does, however, also mention that AMD has strange practices when it comes to production, and says they need to step it up it they want the K7's performance to be timely.
  • More of the facts than you probably think are true on that page. I tried to place question marks next to the questionable bits or otherwise mark them as rumour.

    The bit about the PLL is 100% true, for example -- it's just that for some stuff like this, I cannot reveal my sources.

    I can't *believe* that I forgot to mention anything regarding SMP. I mean, that was just *wrong* of me! I'll have to update the page in full force, though it might take a bit. I do recall that there is a company making very high SMP (for x86) chipsets for K7, but I keep thinking it's Profusion (the folks who are doing similar stuff on Intel's platform).

    The company making for EV-6 could conceivably make specialized 8-way K7 a possibility for 1999, though it all depends on market demand, of course.

    Oh...I think their name might have been Poseidon.

    Anyway, that page is far from finished. It just covered a bunch of stuff that I think were important at the time, such as the concern over cost, exactly why the floating point unit is so fantastic, who might be supporting the platform. I even ignored MHz, since I'd basically assumed since November that the K7 would be coming out at 600MHz. I mean, wasn't it obvious?

    -JC
    PC News'n'Links
    http://www.jc-news.com/pc
  • by warmi (13527)
    We are very fortunate that there is competition for Intel. All my systems are either Cyrix or lately AMD. It is simply cheaper ...
  • Wasn't there a time when 64 bit processors were going to reign over the world by mid 1999 and Linux was going to take over the world as the only 64 bit operating system? Instead we have yet another generation of 32 bit processors coming out. We're still stuck with the 1 gig maximum file size.
  • Probably the most accurate collection of information concerning the K7 on the net, but still largely based on hardcopy and heresay. Biostar, MSI and Gigabyte will be offering Irongate boards. Asus will as well, but will sit on the idea for a few months (typical Asus. Case and point: the P5A).

    I'm not sure about the 500 MHz. I was pretty sure the first one would be a 550.

    Also, concerning the "MHz" gap between the K6 and K7s: There WILL be a K6 550, so those of you with Super Socket 7 boards will not have to build up all new PCs (yet) to have an AMD with 550 Mhz speeds.

    Another important factoid is that the K7 can be run in tandem....or more. At AMD they have a test machine with 4 K7s running NT. They only use 32MB of RAM in it to prove the lack of any sort of bottleneck in such a system. The x86 has come of age! Booo yeah.
  • Fast forward five years. You have a multi Gbit/sec link into your house, a gigabit ethernet SOHO LAN, and can mount the Internet (with vastly upgraded backbone capacity) as a virtual file system.
  • Microsoft has a similar plan....
  • Uhh, if memory serves, Digital Unix is a 64bit operating.
  • Posted by Lord Kano-The Gangster Of Love:

    >>Well, they don't so you just have to live with it and I can sympathize with the point of view of the programmers. It is a lot of time and energy (and in bussiness those mean money) to add support for 3dnow to your game.

    It can be simpler than that. I've not done too much x86 development, my area is with the PPC it should be as simple as clicking a checkbox in the IDE. Support 3dnow = yes.

    It's the developer of the IDE who does the majority of the work in that area. The compiler and the linker can dynamicly adjust for the target CPU/environment. For example, metrowerks Codewarrior can be configured to optimize for a specific processor within the PowerPC family (601, 603, 604, and I'm sure G3 will be next).

    All in all AMD's biggest shortcoming is and has been for the past few years with FPU performance, it's been abysmal. If (and that's a big IF) they can improve that with the K7 they just might be able to live up to their own hype.

    LK
  • How does that song go? Promises, promises... Well, AMD never made any promises in the past about FPUs being any better than Intel's, but I have always enjoyed AMDs for giving me more processing power OVERALL than any Intel processor for the money. Floating point shortcomings were always present, but I easily could justify buying a processor with twice the processing speed for half the money. Admittingly, in the past I have said that "if I win the lottery I'd buy a dual Xeon system". Well, no promises from AMD on this processor (not even a release date. The June release was a rumor). However, one thing is a FACT and that is the three full FPUs. This will undoubtedly make up for any floating point issues. Unfortunately (for AMD) with recent Intel price drops (and more to come, I'm sure) it will be hard to see how AMD can compete with Intel on price (and make a profit). Even if the processor IS faster in every aspect (which I'm not going to say it will be without knowing the FACTS), if the price is the same as the Intel (I doubt that AMD would price it higher!) than people would continue to buy the Intel, simply because it's an Intel. *sigh*
  • Posted by Lord Kano-The Gangster Of Love:

    My only problem with AMD has been with the naming convention used with the K5 series. The "K5 PR-133" ran at 100 mhz, but they named it in such a way that people who didn't know better thought that it ran at 133mhz. They remedied this with the introduction of the K6 though. They K6 family is great, for half the price you can outperform an Intel chip. However the FPU performance is horribly lacking. My P2-300 seemed faster at FP math than my K62-450, the fast chips don't seem as overclockable as well. I tried taking my 450 up to 500 mhz, a markup of a mere 11 percent and it would cause windows protection errors. I took my P100 up to 133. The chip was stable running at 133% of it's intended speed.

    All it all I think I made a good move by getting the K62 450.

    For the record, if I won the lottery I'd probably get a quad PPro 200 system. "Can you say 15 second kernel compiles?"

    LK
  • AMD never gave a release date. June is a rumored month of release. If you want my "opinion" based on "promises" made to me by AMD, I'll be lucky if I see an eval unit by the end of June. If YOU expected to see K7s in or by June, YOU were lied to.
  • Only the PR133 ran under spec. If you want to rag on a company for not putting the right speed on a chip, look at Cyrix!!! At any rate, the K5 PR133 did suck, but the K5 166 (which actually RAN at 166) kicked ASS. True, floating point has always been an issue, but this is why the three full floating point units in the K7. AMD was always a good "follower" trying to make as best a chip as possible for an "existing interface" and then falling short. I think it'll take AMD biting the bullet and creating their own interface (the Slot A) and redefining the x86 architechture for them to really shine. Lead, follow or get out of the way. I think they're tired of "follow". Hopefully they won't end up "getting out of the way".
  • One big diferance between MS and AMD.

    MS makes money.


    --------------------------------
    ( my music [mp3.com])
  • Posted by Lord Kano-The Gangster Of Love:

    The K5PR-166 ran at 116 mhz. It was 66mhz bus with a 1.75 multiplier. Even though you set the jumpers on the MB to a 2.5 multiplyer the chip only ran at a 116 mhz core speed. Just as a 1.5 multiplier is the same as a 3.5, the 2.5 only does 1.75 on the K5PR-166.

    Don't get me wrong, I like AMD. I just bought a K62-450 last week and I think it's sweet, but they better do a bang up job on the K7.

    LK

Little known fact about Middle Earth: The Hobbits had a very sophisticated computer network! It was a Tolkien Ring...

Working...