Aside from the rumors of the socket 5 motherboard running k6’s on interposers: Dual AMD K6 computer , I’m talking about the fastest dual socket 7 board it’s self like, was there ever one that could reach 75, 83, 100 fsb?
Sphere478 wrote on 2021-02-11, 16:25:Aside from the rumors of the socket 5 motherboard running k6’s on interposers: Dual AMD K6 computer , I’m talking about the fastest dual socket 7 board it’s self like, was there ever one that could reach 75, 83, 100 fsb?
None of the 'super' So7 chipsets supported SMP. You're basically limited to i430NX and i430HX, and i430HX is by far the fastest of the two, both in terms of performance per clock and in attainable clock speed.
i430HX will run at 83MHz if you're lucky, boards like late rev P55T2P4 were famous for running K6Plus CPUs at 6x83=500MHz. Unfortunately the P55T2P4D doesn't feature FSB options beyond 66MHz. That is likely to be the biggest bottleneck.
Looking at boards in TH99 (not exhaustive), most i430HX dual boards only have two jumpers for 50/55/60/66MHz FSB. Only possible exception is the Tyan S1564D - there is a jumper J21 in the speed block that's alway set to open in official documents. There's a small but nonzero chance that this board has a PLL with three-bit input, and that setting this to closed would allow 68/75/83/? MHz. If so, this will be the fastest board, if the CPUs can keep up. P55C CPUs are massively overengineered, so I'd be confident that they won't be the bottleneck.
However this is speculation unless somebody can positively identify the PLL on that board and confirm what J21 does.
Interesting, never noticed j21
I happen to not only have that board but wrote the addendum with the extra jumper settings listed on uh19.
I never thought to look over at the multiplier block for fsb settings
Can the tillamook 4x multiplier mod be done on a 233 mmx?
Sphere478 wrote on 2021-02-11, 17:55:Can the tillamook 4x multiplier mod be done on a 233 mmx?
Don't think so, at least never heard of it (BF2) being possible there.
Bad news: your photograph is good enough to be able to ID the PLL, an IC Works W48C60-402G. Now, I can't find any datasheet for the -402G (*lots* of dead links), but there's one for the -201, and I doubt the pinout of the -402G would be different:
https://www.datasheetarchive.com/pdf/download … pe=O&term=48c60
Just two selector lines, SEL0 and SEL1, giving you 50, 60, 66 and 'test mode'. So whatever J21 does, it's not going to give you higher FSB.
I tried the J21 it doesn’t seem to do anything. It may be a cyrix jumper?
dionb wrote on 2021-02-11, 19:08:Bad news: your photograph is good enough to be able to ID the PLL, an IC Works W48C60-402G. Now, I can't find any datasheet for the -402G (*lots* of dead links), but there's one for the -201, and I doubt the pinout of the -402G would be different:
https://www.datasheetarchive.com/pdf/download … pe=O&term=48c60Just two selector lines, SEL0 and SEL1, giving you 50, 60, 66 and 'test mode'. So whatever J21 does, it's not going to give you higher FSB.
I also tried the hard soldered j7 and it also had no effect. (Before seeing your reply) so. My only hope is higher ratios on cpu. But I also tried temporarily jumping it with tweezers and it stayed at 233 I tried it again with2x and got 133
Ugh.
Looks like my fastest options are 2x 233mmx, single cyrix and single k6 III+ 400
Btw, got my new heatsinks! Had to do some custom trimming.
Ps, to anyone reading this thread in the future. 768mb of ram is possible with 128x6 in slots 3-8
For a while I have toyed with the idea of building a replacement supplemental PLL circuit board.
The idea being, use the onboard PLL for clocking the other buses: floppy, KB, USB, PCI etc.
Then I would find a PLL that has multiple variable CPU clock outputs either configurable via jumpers or maybe even ‘soft’ controlled via a microcontroller for 1MHz adjustments.
Disconnect the onboard PLL CPU clock connection and feed the CPU clock from the secondary board. No idea if it would work or what buffering or synchronization might be needed to make it work.
gbeirn wrote on 2021-02-11, 19:38:For a while I have toyed with the idea of building a replacement supplemental PLL circuit board. […]
For a while I have toyed with the idea of building a replacement supplemental PLL circuit board.
The idea being, use the onboard PLL for clocking the other buses: floppy, KB, USB, PCI etc.
Then I would find a PLL that has multiple variable CPU clock outputs either configurable via jumpers or maybe even ‘soft’ controlled via a microcontroller for 1MHz adjustments.
Disconnect the onboard PLL CPU clock connection and feed the CPU clock from the secondary board. No idea if it would work or what buffering or synchronization might be needed to make it work.
I would love a project like that. It’s so annoying how both of my boards only have stable settings 🤣 I want to find the limits myself 🤣!
