nforce2 smp

Archived from groups: comp.sys.ibm.pc.hardware.chips (More info?)

How hard would it be to modify the north bridge for dual processing ?
25 answers Last reply
More about nforce2
  1. Archived from groups: comp.sys.ibm.pc.hardware.chips (More info?)

    flekso <ivan_svaljek@email.hinet.hr> wrote:
    > How hard would it be to modify the north bridge for dual processing ?

    Pretty difficult, considering that only AMD makes dual-processing chipsets
    for the Athlon MP family.

    Yousuf Khan
  2. Archived from groups: comp.sys.ibm.pc.hardware.chips (More info?)

    "Yousuf Khan" <news.tally.bbbl67@spamgourmet.com> wrote in message
    news:Fa6oc.34117$pp.30957@news04.bloor.is.net.cable.rogers.com...
    > flekso <ivan_svaljek@email.hinet.hr> wrote:
    > > How hard would it be to modify the north bridge for dual processing ?
    >
    > Pretty difficult, considering that only AMD makes dual-processing chipsets
    > for the Athlon MP family.
    >
    > Yousuf Khan
    >
    >

    I don't understand, does AMD have a license on SMP ?
    I was wondering more about the kind of electronic mischief they'd run into ?
    Wasn't it true in the early PC days, that there was only one bus line that
    was used for SMP, like #LOCK or something ?

    How hard ($) would it be to only double the FSB lines, and use the dual ram
    controller from there ?
    It would be a real nice transitional platform for workstation people (how
    many threads are you running now), coupled with two mobile bartons at
    2.6GHz.
    http://www.tweakpc.de/tweaking/athlon_xpm_oc/athlon_oc.htm
  3. Archived from groups: comp.sys.ibm.pc.hardware.chips (More info?)

    flekso wrote:

    > How hard ($) would it be to only double the FSB lines, and use the dual ram
    > controller from there ?
    > It would be a real nice transitional platform for workstation people (how
    > many threads are you running now), coupled with two mobile bartons at
    > 2.6GHz.

    Get a nice fine-tipped soldering iron, some aluminum foil, and let us
    know how it goes! :)


    -WD


    -----= Posted via Newsfeeds.Com, Uncensored Usenet News =-----
    http://www.newsfeeds.com - The #1 Newsgroup Service in the World!
    -----== Over 100,000 Newsgroups - 19 Different Servers! =-----
  4. Archived from groups: comp.sys.ibm.pc.hardware.chips (More info?)

    flekso <flekso@vip.hr> wrote:
    >> Pretty difficult, considering that only AMD makes dual-processing
    >> chipsets for the Athlon MP family.
    >>
    >> Yousuf Khan
    >
    > I don't understand, does AMD have a license on SMP ?
    > I was wondering more about the kind of electronic mischief they'd run
    > into ? Wasn't it true in the early PC days, that there was only one
    > bus line that was used for SMP, like #LOCK or something ?

    No-no, there just wasn't any interest by anybody else to make a chipset for
    the Athlon MPs. AMD wouldn't have minded a bit if somebody else had made a
    dual-processor chipset for the MPs, but most of the chipset houses decided
    the money was in making chipsets for desktops and laptops running Athlon
    XPs.

    > How hard ($) would it be to only double the FSB lines, and use the
    > dual ram controller from there ?

    Probably not that hard, in fact it would've been a perfect use for the
    Athlon MPs, but Nvidia just never bothered.

    > It would be a real nice transitional platform for workstation people
    > (how many threads are you running now), coupled with two mobile
    > bartons at
    > 2.6GHz.

    There's no chance it's going to happen now. AMD is concentrating on Opterons
    from now on, they're not going to bother with Athlon MPs anymore, and
    neither is anybody else.

    Yousuf Khan
  5. Archived from groups: comp.sys.ibm.pc.hardware.chips (More info?)

    "Yousuf Khan" <news.tally.bbbl67@spamgourmet.com> wrote in message
    news:_lcoc.39131$n7P1.30797@twister01.bloor.is.net.cable.rogers.com...
    > flekso <flekso@vip.hr> wrote:
    > >> Pretty difficult, considering that only AMD makes dual-processing
    > >> chipsets for the Athlon MP family.
    > >>
    > >> Yousuf Khan
    > >
    > > I don't understand, does AMD have a license on SMP ?
    > > I was wondering more about the kind of electronic mischief they'd run
    > > into ? Wasn't it true in the early PC days, that there was only one
    > > bus line that was used for SMP, like #LOCK or something ?
    >
    > No-no, there just wasn't any interest by anybody else to make a chipset
    for
    > the Athlon MPs. AMD wouldn't have minded a bit if somebody else had made a
    > dual-processor chipset for the MPs, but most of the chipset houses decided
    > the money was in making chipsets for desktops and laptops running Athlon
    > XPs.
    >
    > > How hard ($) would it be to only double the FSB lines, and use the
    > > dual ram controller from there ?
    >
    > Probably not that hard, in fact it would've been a perfect use for the
    > Athlon MPs, but Nvidia just never bothered.
    >
    > > It would be a real nice transitional platform for workstation people
    > > (how many threads are you running now), coupled with two mobile
    > > bartons at
    > > 2.6GHz.
    >
    > There's no chance it's going to happen now. AMD is concentrating on
    Opterons
    > from now on, they're not going to bother with Athlon MPs anymore, and
    > neither is anybody else.
    >
    > Yousuf Khan
    >
    >

    m-kay then :\
  6. Archived from groups: comp.sys.ibm.pc.hardware.chips (More info?)

    On Tue, 11 May 2004 18:13:52 +0200, "flekso" <flekso@vip.hr> wrote:

    >"Yousuf Khan" <news.tally.bbbl67@spamgourmet.com> wrote in message
    >news:Fa6oc.34117$pp.30957@news04.bloor.is.net.cable.rogers.com...
    >> flekso <ivan_svaljek@email.hinet.hr> wrote:
    >> > How hard would it be to modify the north bridge for dual processing ?
    >>
    >> Pretty difficult, considering that only AMD makes dual-processing chipsets
    >> for the Athlon MP family.
    >>
    >> Yousuf Khan
    >>
    >>
    >
    >I don't understand, does AMD have a license on SMP ?
    >I was wondering more about the kind of electronic mischief they'd run into ?
    >Wasn't it true in the early PC days, that there was only one bus line that
    >was used for SMP, like #LOCK or something ?
    >
    >How hard ($) would it be to only double the FSB lines, and use the dual ram
    >controller from there ?
    >It would be a real nice transitional platform for workstation people (how
    >many threads are you running now), coupled with two mobile bartons at
    >2.6GHz.
    >http://www.tweakpc.de/tweaking/athlon_xpm_oc/athlon_oc.htm
    >
    Why bother? K7 is obsolete ever since K8 got out the door. MSI dual
    board (the cheapest) is around $200, Opteron 242 is getting close to
    that - and Barton will never run Win64. If you have money to burn,
    get yourself Tyan K8W Thunder and a couple of 248 - and this combo
    will beat any Barton (or Xeon or anything for that matter) in most any
    benchmark. By the way, MSI board is VIA based. Have not heard about
    dual Nvidia board though.
  7. Archived from groups: comp.sys.ibm.pc.hardware.chips (More info?)

    nobody@nowhere.net <MyGarbage2000@hotmail.com> wrote:
    > that - and Barton will never run Win64. If you have money to burn,
    > get yourself Tyan K8W Thunder and a couple of 248 - and this combo
    > will beat any Barton (or Xeon or anything for that matter) in most any
    > benchmark.

    At least anything halfway affordable. If nothing else, you can always go
    with the Tyan 4-way server board and some 848s.

    --
    Nate Edel http://www.nkedel.com/

    "Elder Party 2004: Cthulhu for President -- this time WE'RE the lesser
    evil."
  8. Archived from groups: comp.sys.ibm.pc.hardware.chips (More info?)

    "nobody@nowhere.net" <MyGarbage2000@hotmail.com> wrote in message
    news:p5g5a0lh2okrhq00ddt1gfdrkh32kafdu7@4ax.com...
    > On Tue, 11 May 2004 18:13:52 +0200, "flekso" <flekso@vip.hr> wrote:
    >
    > >"Yousuf Khan" <news.tally.bbbl67@spamgourmet.com> wrote in message
    > >news:Fa6oc.34117$pp.30957@news04.bloor.is.net.cable.rogers.com...
    > >> flekso <ivan_svaljek@email.hinet.hr> wrote:
    > >> > How hard would it be to modify the north bridge for dual processing ?
    > >>
    > >> Pretty difficult, considering that only AMD makes dual-processing
    chipsets
    > >> for the Athlon MP family.
    > >>
    > >> Yousuf Khan
    > >>
    > >>
    > >
    > >I don't understand, does AMD have a license on SMP ?
    > >I was wondering more about the kind of electronic mischief they'd run
    into ?
    > >Wasn't it true in the early PC days, that there was only one bus line
    that
    > >was used for SMP, like #LOCK or something ?
    > >
    > >How hard ($) would it be to only double the FSB lines, and use the dual
    ram
    > >controller from there ?
    > >It would be a real nice transitional platform for workstation people (how
    > >many threads are you running now), coupled with two mobile bartons at
    > >2.6GHz.
    > >http://www.tweakpc.de/tweaking/athlon_xpm_oc/athlon_oc.htm
    > >
    > Why bother? K7 is obsolete ever since K8 got out the door. MSI dual
    > board (the cheapest) is around $200, Opteron 242 is getting close to
    > that - and Barton will never run Win64. If you have money to burn,
    > get yourself Tyan K8W Thunder and a couple of 248 - and this combo
    > will beat any Barton (or Xeon or anything for that matter) in most any
    > benchmark. By the way, MSI board is VIA based. Have not heard about
    > dual Nvidia board though.

    There's also the n*? issue. Opteron 240 is 314$. I'm reffering to Opteron
    prices in Croatia; on the other hand dual Durons that can easily be modifed
    to dual XPs(and then MPs), now that is affordable.

    My single Duron @ 2330 can beat P4 3.0 in some scenarios, and that's 45 vs
    277$ in a country where average sallary weighs in at about 2/3 K$.

    We people don't have bukco$ enough to go the Opteron route just yet but we
    sure like some other shiny chips...
  9. Archived from groups: comp.sys.ibm.pc.hardware.chips (More info?)

    flekso <flekso@vip.hr> wrote:
    > There's also the n*? issue. Opteron 240 is 314$. I'm reffering to
    > Opteron prices in Croatia; on the other hand dual Durons that can
    > easily be modifed to dual XPs(and then MPs), now that is affordable.
    >
    > My single Duron @ 2330 can beat P4 3.0 in some scenarios, and that's
    > 45 vs 277$ in a country where average sallary weighs in at about 2/3
    > K$.
    >
    > We people don't have bukco$ enough to go the Opteron route just yet
    > but we sure like some other shiny chips...

    Maybe, but modifying Durons to become Athlon MPs is only a solution for now.
    Eventually the Opteron prices will plummet. AMD has already discontinued the
    Opteron x40 series, their lowest performing Opteron family. Maybe if a
    dual-CPU market starts flourishing in places like Croatia, AMD can be
    convinced to start up a special line of older Opterons for sale only within
    these countries like they did with the Durons last year? There has to be big
    enough demand for dual-processing for it to be worth their time.

    I don't think you can get cheap dual-processing Xeons either as it now
    stands.

    Yousuf Khan
  10. Archived from groups: comp.sys.ibm.pc.hardware.chips (More info?)

    Yousuf Khan <news.tally.bbbl67@spamgourmet.com> wrote:
    > I don't think you can get cheap dual-processing Xeons either as it now
    > stands.

    Nope. Cheap older P3s on eBay is as close as you come... in the retail
    market for dual-processor Intels, it's either Xeon(P4-based) or the 512K
    cache Tualatins, which while still available, are highway robbery (as
    expensive as the P4-based Xeons last I checked... and while Mhz-for-Mhz
    they're faster, not enough so to make up for twice the megahertz until power
    comes in as a factor...)

    --
    Nate Edel http://www.nkedel.com/

    "Elder Party 2004: Cthulhu for President -- this time WE'RE the lesser
    evil."
  11. Archived from groups: comp.sys.ibm.pc.hardware.chips (More info?)

    In article <40a147f9_7@corp.newsgroups.com>,
    wdormann@yahoo.com.invalid says...
    > flekso wrote:
    >
    > > How hard ($) would it be to only double the FSB lines, and use the dual ram
    > > controller from there ?
    > > It would be a real nice transitional platform for workstation people (how
    > > many threads are you running now), coupled with two mobile bartons at
    > > 2.6GHz.
    >
    > Get a nice fine-tipped soldering iron, some aluminum foil, and let us
    > know how it goes! :)

    You can't solder to aluminum foil, silly! ;-)


    --
    Keith
  12. Archived from groups: comp.sys.ibm.pc.hardware.chips (More info?)

    KR Williams wrote:

    > In article <40a147f9_7@corp.newsgroups.com>,
    > wdormann@yahoo.com.invalid says...
    >
    >>flekso wrote:
    >>
    >>
    >>>How hard ($) would it be to only double the FSB lines, and use the dual ram
    >>>controller from there ?
    >>>It would be a real nice transitional platform for workstation people (how
    >>>many threads are you running now), coupled with two mobile bartons at
    >>>2.6GHz.
    >>
    >>Get a nice fine-tipped soldering iron, some aluminum foil, and let us
    >>know how it goes! :)
    >
    >
    > You can't solder to aluminum foil, silly! ;-)
    >
    >

    No, but if you lay the foil on a surface you can use
    the soldering iron to deposit traces on that surface.
    I've seen it done with gold foil, but not aluminum.
    Should work OK for wrecking motherboards, etc.
  13. Archived from groups: comp.sys.ibm.pc.hardware.chips (More info?)

    In article <10afdeva7tlqnfe@corp.supernews.com>,
    rob.stow@sasktel.net says...
    > KR Williams wrote:
    >
    > > In article <40a147f9_7@corp.newsgroups.com>,
    > > wdormann@yahoo.com.invalid says...
    > >
    > >>flekso wrote:
    > >>
    > >>
    > >>>How hard ($) would it be to only double the FSB lines, and use the dual ram
    > >>>controller from there ?
    > >>>It would be a real nice transitional platform for workstation people (how
    > >>>many threads are you running now), coupled with two mobile bartons at
    > >>>2.6GHz.
    > >>
    > >>Get a nice fine-tipped soldering iron, some aluminum foil, and let us
    > >>know how it goes! :)
    > >
    > >
    > > You can't solder to aluminum foil, silly! ;-)
    > >
    > >
    >
    > No, but if you lay the foil on a surface you can use
    > the soldering iron to deposit traces on that surface.
    > I've seen it done with gold foil, but not aluminum.

    One *can* solder to gold. I've done it, more often by accident
    than on purpose though. :-( Aluminum is another kettle of stinky
    fish though.

    > Should work OK for wrecking motherboards, etc.

    Sure. I noticed that that was the plan. ;-)

    --
    Keith
  14. Archived from groups: comp.sys.ibm.pc.hardware.chips (More info?)

    KR Williams wrote:
    > In article <10afdeva7tlqnfe@corp.supernews.com>,
    > rob.stow@sasktel.net says...
    >
    >>KR Williams wrote:
    >>
    >>
    >>>In article <40a147f9_7@corp.newsgroups.com>,
    >>>wdormann@yahoo.com.invalid says...
    >>>
    >>>
    >>>>flekso wrote:
    >>>>
    >>>>
    >>>>
    >>>>>How hard ($) would it be to only double the FSB lines, and use the dual ram
    >>>>>controller from there ?
    >>>>>It would be a real nice transitional platform for workstation people (how
    >>>>>many threads are you running now), coupled with two mobile bartons at
    >>>>>2.6GHz.
    >>>>
    >>>>Get a nice fine-tipped soldering iron, some aluminum foil, and let us
    >>>>know how it goes! :)
    >>>
    >>>
    >>>You can't solder to aluminum foil, silly! ;-)
    >>>
    >>>
    >>
    >>No, but if you lay the foil on a surface you can use
    >>the soldering iron to deposit traces on that surface.
    >>I've seen it done with gold foil, but not aluminum.
    >
    >
    > One *can* solder to gold. I've done it, more often by accident
    > than on purpose though. :-( Aluminum is another kettle of stinky
    > fish though.
    >
    >
    >>Should work OK for wrecking motherboards, etc.
    >
    >
    > Sure. I noticed that that was the plan. ;-)
    >

    The reason I saw it being done with gold was because I
    had wrecked the very first motherboard I had bought for
    myself about 17 or 18 years ago and was watching a friend
    repair it. I had been using the tip of a screwdriver to
    force something into a socket or pry loose a locking clip
    - can't remember exactly anymore - and the tip slipped off
    of whatever I was pushing/prying and scraped a trench across
    some traces. I wonder if the circuitry on modern motherboards
    is robust enough to tolerate that kind of repair job ?
  15. Archived from groups: comp.sys.ibm.pc.hardware.chips (More info?)

    flekso wrote:
    > How hard would it be to modify the north bridge for dual processing ?

    It would be completely impossible unless you're nVidia and can
    dramatically redesign the inner chip workings.

    The Athlon fronstside bus (EV6 bus) isn't a "bus," it's a point-to-point
    connection. One "point" is the CPU, and the other "point" is the
    Northbridge. It's the Northbridge chip's job to provide interconnect
    with any other busses (AGP, memory, PCI/V-link/HT/etc).

    The AMD 760MP SMP Northbridge provides *two* EV6 endpoints, and routes
    traffic between those two and other provided busses.

    The 760MP and 760MPX are the only two SMP-supporting Athlon chip sets
    that exist - or ever will exist for K7s, since newer AMD chips have
    moved on to Hypertransport and embedded memory controllers.
  16. Archived from groups: comp.sys.ibm.pc.hardware.chips (More info?)

    Richard Krehbiel <krehbiel@comcast.net> wrote:
    > The AMD 760MP SMP Northbridge provides *two* EV6 endpoints, and routes
    > traffic between those two and other provided busses.
    >
    > The 760MP and 760MPX are the only two SMP-supporting Athlon chip sets
    > that exist - or ever will exist for K7s, since newer AMD chips have
    > moved on to Hypertransport and embedded memory controllers.

    Wasn't there supposed to be some kind of sharing of resources between Alpha
    and Athlon because of their common bus standard? What were the
    multiprocessor solutions available for Alpha?

    Yousuf Khan
  17. Archived from groups: comp.sys.ibm.pc.hardware.chips (More info?)

    In article <10ag8a64o0ikdec@corp.supernews.com>,
    rob.stow@sasktel.net says...
    > KR Williams wrote:
    > > In article <10afdeva7tlqnfe@corp.supernews.com>,
    > > rob.stow@sasktel.net says...
    > >
    > >>KR Williams wrote:
    > >>
    > >>
    > >>>In article <40a147f9_7@corp.newsgroups.com>,
    > >>>wdormann@yahoo.com.invalid says...
    > >>>
    > >>>
    > >>>>flekso wrote:
    > >>>>
    > >>>>
    > >>>>
    > >>>>>How hard ($) would it be to only double the FSB lines, and use the dual ram
    > >>>>>controller from there ?
    > >>>>>It would be a real nice transitional platform for workstation people (how
    > >>>>>many threads are you running now), coupled with two mobile bartons at
    > >>>>>2.6GHz.
    > >>>>
    > >>>>Get a nice fine-tipped soldering iron, some aluminum foil, and let us
    > >>>>know how it goes! :)
    > >>>
    > >>>
    > >>>You can't solder to aluminum foil, silly! ;-)
    > >>>
    > >>>
    > >>
    > >>No, but if you lay the foil on a surface you can use
    > >>the soldering iron to deposit traces on that surface.
    > >>I've seen it done with gold foil, but not aluminum.
    > >
    > >
    > > One *can* solder to gold. I've done it, more often by accident
    > > than on purpose though. :-( Aluminum is another kettle of stinky
    > > fish though.
    > >
    > >
    > >>Should work OK for wrecking motherboards, etc.
    > >
    > >
    > > Sure. I noticed that that was the plan. ;-)
    > >
    >
    > The reason I saw it being done with gold was because I
    > had wrecked the very first motherboard I had bought for
    > myself about 17 or 18 years ago and was watching a friend
    > repair it. I had been using the tip of a screwdriver to
    > force something into a socket or pry loose a locking clip
    > - can't remember exactly anymore - and the tip slipped off
    > of whatever I was pushing/prying and scraped a trench across
    > some traces.

    Oops! I *HATE* it when that happens.

    > I wonder if the circuitry on modern motherboards
    > is robust enough to tolerate that kind of repair job ?

    No. The technology hasn't changed much, if not gotten worse with
    multi-layer boards. The only thing inbetween the screwdriver and
    the traces is the solder-mask. ...little more than paint. One
    learns not to force things (or at least direct the force in a
    direction not compromising several $hundred).

    --
    Keith
  18. Archived from groups: comp.sys.ibm.pc.hardware.chips (More info?)

    In article <rwYpc.43618$0qd.34205
    @twister01.bloor.is.net.cable.rogers.com>, news.tally.bbbl67
    @spamgourmet.com says...
    > Richard Krehbiel <krehbiel@comcast.net> wrote:
    > > The AMD 760MP SMP Northbridge provides *two* EV6 endpoints, and routes
    > > traffic between those two and other provided busses.
    > >
    > > The 760MP and 760MPX are the only two SMP-supporting Athlon chip sets
    > > that exist - or ever will exist for K7s, since newer AMD chips have
    > > moved on to Hypertransport and embedded memory controllers.
    >
    > Wasn't there supposed to be some kind of sharing of resources between Alpha
    > and Athlon because of their common bus standard? What were the
    > multiprocessor solutions available for Alpha?

    ISTR that that was all about the electrical design, rather than
    physical. Alpha had different requirements, so I never expected
    much cross-pollination.

    --
    Keith
  19. Archived from groups: comp.sys.ibm.pc.hardware.chips (More info?)

    KR Williams wrote:

    > In article <10ag8a64o0ikdec@corp.supernews.com>,
    > rob.stow@sasktel.net says...
    >
    >>KR Williams wrote:
    >>
    >>>In article <10afdeva7tlqnfe@corp.supernews.com>,
    >>>rob.stow@sasktel.net says...
    >>>
    >>>
    >>>>KR Williams wrote:
    >>>>
    >>>>
    >>>>
    >>>>>In article <40a147f9_7@corp.newsgroups.com>,
    >>>>>wdormann@yahoo.com.invalid says...
    >>>>>
    >>>>>
    >>>>>
    >>>>>>flekso wrote:
    >>>>>>
    >>>>>>
    >>>>>>
    >>>>>>
    >>>>>>>How hard ($) would it be to only double the FSB lines, and use the dual ram
    >>>>>>>controller from there ?
    >>>>>>>It would be a real nice transitional platform for workstation people (how
    >>>>>>>many threads are you running now), coupled with two mobile bartons at
    >>>>>>>2.6GHz.
    >>>>>>
    >>>>>>Get a nice fine-tipped soldering iron, some aluminum foil, and let us
    >>>>>>know how it goes! :)
    >>>>>
    >>>>>
    >>>>>You can't solder to aluminum foil, silly! ;-)
    >>>>>
    >>>>>
    >>>>
    >>>>No, but if you lay the foil on a surface you can use
    >>>>the soldering iron to deposit traces on that surface.
    >>>>I've seen it done with gold foil, but not aluminum.
    >>>
    >>>
    >>>One *can* solder to gold. I've done it, more often by accident
    >>>than on purpose though. :-( Aluminum is another kettle of stinky
    >>>fish though.
    >>>
    >>>
    >>>
    >>>>Should work OK for wrecking motherboards, etc.
    >>>
    >>>
    >>>Sure. I noticed that that was the plan. ;-)
    >>>
    >>
    >>The reason I saw it being done with gold was because I
    >>had wrecked the very first motherboard I had bought for
    >>myself about 17 or 18 years ago and was watching a friend
    >>repair it. I had been using the tip of a screwdriver to
    >>force something into a socket or pry loose a locking clip
    >>- can't remember exactly anymore - and the tip slipped off
    >>of whatever I was pushing/prying and scraped a trench across
    >>some traces.
    >
    >
    > Oops! I *HATE* it when that happens.

    Yup. Kinda like getting it on with your girlfriend
    and then her father's voice comes on through the
    answering machine. "Honey, you might as well answer
    the damned phone ... "

    >
    >
    >>I wonder if the circuitry on modern motherboards
    >>is robust enough to tolerate that kind of repair job ?
    >
    >
    > No. The technology hasn't changed much, if not gotten worse with
    > multi-layer boards. The only thing inbetween the screwdriver and
    > the traces is the solder-mask. ...little more than paint. One
    > learns not to force things (or at least direct the force in a
    > direction not compromising several $hundred).
    >

    Well, like I said it was my first mobo and I had yet to learn better.
    As well, it is less of an issue with current motherboards because
    things seem to pop in and out of their sockets much better than they
    did way back when. Either that or building a systems regularly for
    more than a decade eventually gave me a knack for managing dimm socket
    clips.
  20. Archived from groups: comp.sys.ibm.pc.hardware.chips (More info?)

    On Mon, 17 May 2004 23:33:08 -0400, KR Williams <krw@att.biz> wrote:
    >In article <rwYpc.43618$0qd.34205
    >@twister01.bloor.is.net.cable.rogers.com>, news.tally.bbbl67
    >@spamgourmet.com says...
    >> Wasn't there supposed to be some kind of sharing of resources between Alpha
    >> and Athlon because of their common bus standard? What were the
    >> multiprocessor solutions available for Alpha?
    >
    >ISTR that that was all about the electrical design, rather than
    >physical. Alpha had different requirements, so I never expected
    >much cross-pollination.

    There was some talk and even some prototype demos that had both AMD K7
    and DEC/Compaq Alpha chips being interchangeable on motherboards with
    only a BIOS flash, but I don't think any such products ever made it to
    market. They were supposed to use a common "Slot B" form factor
    (basically "Slot B" was to Intel's old Xeon "Slot 2" what "Slot A" was
    to Intel's PII/PIII "Slot 1"). I don't think either AMD or DEC ever
    sold any Slot B processors though.

    The closet the two came to cross-pollinating were some
    single-processor Alpha motherboards that used AMD chipsets. Physical
    sockets weren't compatible with any AMD chips, but the chipsets could
    be used. These things were sold into very different markets though,
    so the whole idea was kind of doomed right from the get-go.

    -------------
    Tony Hill
    hilla <underscore> 20 <at> yahoo <dot> ca
  21. Archived from groups: comp.sys.ibm.pc.hardware.chips (More info?)

    In article <10ajaug4ni82faf@corp.supernews.com>,
    rob.stow@sasktel.net says...
    > KR Williams wrote:
    >
    > > In article <10ag8a64o0ikdec@corp.supernews.com>,
    > > rob.stow@sasktel.net says...
    > >
    > >>KR Williams wrote:
    > >>
    > >>>In article <10afdeva7tlqnfe@corp.supernews.com>,
    > >>>rob.stow@sasktel.net says...
    > >>>
    > >>>
    > >>>>KR Williams wrote:
    > >>>>
    > >>>>
    > >>>>
    > >>>>>In article <40a147f9_7@corp.newsgroups.com>,
    > >>>>>wdormann@yahoo.com.invalid says...
    > >>>>>
    > >>>>>
    > >>>>>
    > >>>>>>flekso wrote:
    > >>>>>>
    > >>>>>>
    > >>>>>>
    > >>>>>>
    > >>>>>>>How hard ($) would it be to only double the FSB lines, and use the dual ram
    > >>>>>>>controller from there ?
    > >>>>>>>It would be a real nice transitional platform for workstation people (how
    > >>>>>>>many threads are you running now), coupled with two mobile bartons at
    > >>>>>>>2.6GHz.
    > >>>>>>
    > >>>>>>Get a nice fine-tipped soldering iron, some aluminum foil, and let us
    > >>>>>>know how it goes! :)
    > >>>>>
    > >>>>>
    > >>>>>You can't solder to aluminum foil, silly! ;-)
    > >>>>>
    > >>>>>
    > >>>>
    > >>>>No, but if you lay the foil on a surface you can use
    > >>>>the soldering iron to deposit traces on that surface.
    > >>>>I've seen it done with gold foil, but not aluminum.
    > >>>
    > >>>
    > >>>One *can* solder to gold. I've done it, more often by accident
    > >>>than on purpose though. :-( Aluminum is another kettle of stinky
    > >>>fish though.
    > >>>
    > >>>
    > >>>
    > >>>>Should work OK for wrecking motherboards, etc.
    > >>>
    > >>>
    > >>>Sure. I noticed that that was the plan. ;-)
    > >>>
    > >>
    > >>The reason I saw it being done with gold was because I
    > >>had wrecked the very first motherboard I had bought for
    > >>myself about 17 or 18 years ago and was watching a friend
    > >>repair it. I had been using the tip of a screwdriver to
    > >>force something into a socket or pry loose a locking clip
    > >>- can't remember exactly anymore - and the tip slipped off
    > >>of whatever I was pushing/prying and scraped a trench across
    > >>some traces.
    > >
    > >
    > > Oops! I *HATE* it when that happens.
    >
    > Yup. Kinda like getting it on with your girlfriend
    > and then her father's voice comes on through the
    > answering machine. "Honey, you might as well answer
    > the damned phone ... "

    Kinda like when you look out the window and see her parents
    walking up the drive? ...but that was 34 years ago. ;-)

    > >>I wonder if the circuitry on modern motherboards
    > >>is robust enough to tolerate that kind of repair job ?
    > >
    > >
    > > No. The technology hasn't changed much, if not gotten worse with
    > > multi-layer boards. The only thing inbetween the screwdriver and
    > > the traces is the solder-mask. ...little more than paint. One
    > > learns not to force things (or at least direct the force in a
    > > direction not compromising several $hundred).
    > >
    >
    > Well, like I said it was my first mobo and I had yet to learn better.

    Trust me. We've all been there. I remember blowing up a scope
    plugin in my first week at work (I was a technician in school).
    The damned calibration tool *SHOULD have* worked (right
    mainframe, wrong plugin). What I do is imagine where my tool is
    going to go if something breaks and then make sure the force-
    vector isn't in a line to hit any thing critical. Surely
    something's going to surprise and the force pointing off into
    space is better than one going into the guts. Sorta like
    carving/cutting wood. Think where the sharp end is going to go
    if you slip, and make sure there aren't any important (Bobit?)
    parts in that path.

    > As well, it is less of an issue with current motherboards because
    > things seem to pop in and out of their sockets much better than they
    > did way back when. Either that or building a systems regularly for
    > more than a decade eventually gave me a knack for managing dimm socket
    > clips.

    We'll see how far things have some. The shroud for the K8 looks
    a tad better thought-out than the previous stuff. There is
    several $hundred in there too, so I'll go on the cautious side.

    As TJW said "THINK". ...always good to remember, though not
    guaranteed to work. We are human and don't always.

    --
    Keith
  22. Archived from groups: comp.sys.ibm.pc.hardware.chips (More info?)

    KR Williams wrote:
    >
    > We'll see how far things have some. The shroud for the K8 looks
    > a tad better thought-out than the previous stuff. There is
    > several $hundred in there too, so I'll go on the cautious side.
    >

    The Opty/FX/Athlon64's are different enough that I appreciated
    having the opportunity to watch someone else install a couple
    (his first two, as a matter of fact) before I did it myself.

    It is not really any harder or easier than dealing with an
    Athlon XP or a P4 - just different enough that you want to
    treat it as learning a new little skill.
  23. Archived from groups: comp.sys.ibm.pc.hardware.chips (More info?)

    In article <10alnsi4j1qsg1f@corp.supernews.com>,
    rob.stow@sasktel.net says...
    > KR Williams wrote:
    > >
    > > We'll see how far things have some. The shroud for the K8 looks
    > > a tad better thought-out than the previous stuff. There is
    > > several $hundred in there too, so I'll go on the cautious side.
    > >
    >
    > The Opty/FX/Athlon64's are different enough that I appreciated
    > having the opportunity to watch someone else install a couple
    > (his first two, as a matter of fact) before I did it myself.

    > It is not really any harder or easier than dealing with an
    > Athlon XP or a P4 - just different enough that you want to
    > treat it as learning a new little skill.

    Iv'e been out of x86 land so long that I've not done any of the
    above. (are you trying to scare me? ;-) Perhaps I should have
    bought the CFO even more flowers so she would have gone for a
    G5? ;-)) ...not likely to have worked either.

    I think I can figure it all out, though it's all still in the
    *retail* (no other reasonable choice that I can see) packaging.

    --
    Keith
  24. Archived from groups: comp.sys.ibm.pc.hardware.chips (More info?)

    On Wed, 19 May 2004 22:17:50 -0400, KR Williams <krw@att.biz> wrote:

    >In article <10alnsi4j1qsg1f@corp.supernews.com>,
    >rob.stow@sasktel.net says...
    >> KR Williams wrote:
    >> >
    >> > We'll see how far things have some. The shroud for the K8 looks
    >> > a tad better thought-out than the previous stuff. There is
    >> > several $hundred in there too, so I'll go on the cautious side.
    >> >
    >>
    >> The Opty/FX/Athlon64's are different enough that I appreciated
    >> having the opportunity to watch someone else install a couple
    >> (his first two, as a matter of fact) before I did it myself.
    >
    >> It is not really any harder or easier than dealing with an
    >> Athlon XP or a P4 - just different enough that you want to
    >> treat it as learning a new little skill.
    >
    >Iv'e been out of x86 land so long that I've not done any of the
    >above. (are you trying to scare me? ;-) Perhaps I should have
    >bought the CFO even more flowers so she would have gone for a
    >G5? ;-)) ...not likely to have worked either.

    I just did a retail Athlon64 the other day and found it much easier than
    the Athlons I'd done previously. With the XPs, in fact any heatsink which
    clips to the ZIF socket, there's always been the question of whether I
    needed to tilt the whole heatsink at an angle to get the first latch
    engaged... and whether I can then straighten and seat it on the CPU without
    munging the TIM on the bottom of the heatsink.

    With the Athlon64, I just plopped the heatsink straight on to the CPU, in
    the middle of the retention frame of course, and the spring clip was loose
    enough that I could latch the free side easily (make sure it's fully
    engaged on the latch) and the other (lever) side didn't need much finger
    pressure at all to latch it. Then the lever pushes over and clamps the
    heatsink down - you can feel there's a lot of pressure being applied from
    the lever but since it's a lever.....

    >I think I can figure it all out, though it's all still in the
    >*retail* (no other reasonable choice that I can see) packaging.

    Check your case carefully. The Antec (SX-635) I just bought had a broken
    lower front panel, which I found out about after full assembly: when I
    grabbed the front bottom to pick the system up, the cracked panel mounting
    pegs just broke away. There was no visible damage to the shipping carton
    to bother Fedex with, so I called Antec up (no wait) and they're sending me
    a new front panel - the original is taped on till it arrives.

    Hurry up and get that painting done so we can all hear about this beast!

    Rgds, George Macdonald

    "Just because they're paranoid doesn't mean you're not psychotic" - Who, me??
  25. Archived from groups: comp.sys.ibm.pc.hardware.chips (More info?)

    Rob Stow <rob.stow@sasktel.net> wrote:
    > The Opty/FX/Athlon64's are different enough that I appreciated
    > having the opportunity to watch someone else install a couple
    > (his first two, as a matter of fact) before I did it myself.
    >
    > It is not really any harder or easier than dealing with an
    > Athlon XP or a P4 - just different enough that you want to
    > treat it as learning a new little skill.

    I actually will beg to differ there; heat-spreaders make for a much more
    durable processor and a much easier and less nerve-wracking heat-sink
    installation than the old exposed-core AMD chips did.

    --
    Nate Edel http://www.nkedel.com/

    "Elder Party 2004: Cthulhu for President -- this time WE'RE the lesser
    evil."
Ask a new question

Read More

CPUs Hardware IBM