KR> i dont think that i have a problem getting it to quadruple, there is one 
KR> link on the mb which put in one position causes the 5x86 to give identical 
KR> performance to a dx4-100 (its doing it now to produce this).

 BG> What brand of board, type of chipset, and BIOS ?

the only identifier on the board is the "made in china" label. if
you peel that off, there is a label that says "made in taiwan"
(:. the chipset is sis, the bios award.

KR> if i move it to the other position (it's function is not explained in the 
KR> mb manual) the whole thing collapses, and locks on boot. using a dx4-100 
KR> the same positions make it run at 66mhz and 100mhz respectively. the 
KR> problem is just that the rest of the board cant take the strain of the cpu 
KR> running at 133mhz.

 BG> I find that hard to believe (no offence), as the CPU is only running
 BG> at 4X internally.  As far as the rest of the machine is concerned,
 BG> it's the same 33MHz that would be used for a DX or DX4 chip.

KR> ps. i tried another motherboard of identical model, jumpered as far as i
KR> could tell exactly the same. it still hung in exactly the same way with
KR> the chip set as a quad, but with it set as a tripler, it gave a landmark
KR> of 270 instead of 475 with my motherboard ??????

 BG> Did you try setting it up as both the Enhanced AMD DX2/66 or DX4/100 ?
 BG>  With the former, it should run as a quad if you pull pin 17 low, and
 BG> with the latter, a normal tripler (assuming that the board doesn't
 BG> actually support the 5x86, which it obviously doesn't).

i have now got it working. i removed the cpu, and metered pin a17 to one of
the setup jumper pins. the manual calls out that pin for some cyrix cpus,
but not any amd models. i did this with a supposedly identical board at
work, and got a different answer. god knows what i was tying to vss (:

i have now got it running at 160mhz, even though the bios only reckons it
at 150 (just quadded and not overclocked, it reckons 120) but landmark v6
comes up with 163.949mhz. these are the landmark results:-
    
    dx4-100 .............. 457
    5x86 (33mhz clock) ... 563   (1.23x)
    5x86 (40mhz clock) ... 821   (1.79x)

the only bugbear now is that the cpu series regulator transistor gets quite
hot. it got too hot to touch after about 2 min. i have now put a heatsink
on it and have added a cpu fan blowing on it. that makes 4 of these fans in
the box (one for the cpu, one for each hard disk, and now one for the
regulator.

 BG> FWIW, I've had my fun with the 5x86, and have just bought a Triton II
 BG> 430VX motherboard and a Cyrix 6x86/100 CPU (which runs a treat at
 BG> 120MHz) for a total outlay of just $300 inc tax.  That's only $100
 BG> more than a 5x86 board and CPU, yet the difference in speed is
 BG> staggering.  QDI have just released a similar board, which has inbuilt
 BG> SCSI, so perhaps something like that might be the way to go?  Model
 BG> number is P51430VX-250, and the Award PCI flash BIOS includes NCR 810
 BG> and Adaptec 7850 SCSI BIOS.  Sounds very elegant to me.

think that i'd better give up fixing mainframes and take up vcrs (:
the p5143vx-250 sounds interesting, does it come with aspi drivers? whats
the price and from where?

what do you think of the 686? i hear that it's integer performance is
excellent but that the fpu is a lot slower than the pentium, i also hear
that they dissipate noticably more heat that an equivalent pentium.

                 keith
@EOT:

--- WinPoint 0.2.07 Alpha