[Info-vax] KZPAC / RAID Array 230 questions

[John Reagan]

On Sunday, December 20, 2020 at 2:58:14 PM UTC-5, Stephen Hoffman wrote:
> On 2020-12-18 01:36:14 +0000, Alan Frisbie said: 
> 
> > I recently acquired a DS20 with a KZPAC (RAID Array 230/Plus) SCSI 
> > controller. The firmware "show config" command also reports it as a 
> > Mylex DAC960.
> The old Mylex DAC RAID controllers are not among my favorite RAID 
> controllers, and yes, those old addressing limits are permanent. 
> Wouldn't ever be my choice for controllers either, both due to the 
> limits, and due to the "fun" I always had with that series. 
> 
> As for the Alpha 21264 processors and processor passes... IIRC, Alpha 
> EV6 (specifically EV6 and not EV67, etc) 500 MHz pass 2.3 didn't play 
> well with other 500 MHz passes, with a timing bug in the probe 
> instruction implementation that could cause weirdness when used in 
> mixed-pass same-MHz configurations, and mixed EV6 passes are thus best 
> avoided with multi-processor configurations. Mixed-MHz wasn't something 
> you'd want to do. IIRC, EV68 is all pass 2.6, and all works. And IIRC, 
> there was an FCO for this. 
> 
> Some very old reading material... 
> 
> > Variants 
> > 
> > Intermixing the KN310-AA (DS20) CPU and the KN310-BA (DS20E) CPU. 
> > 
> > The -AA or -BA variants of the KN310 cannot be used to determine 
> > processor compatibility. The actual part number must be used. 
> > 
> > DS20 processors with numbers are 54-24758-03 & -05, will work together; 
> > they have the same EV chips in them (No longer orderable). 
> > 
> > DS20 and DS20E processors 54-24758-33 & -35 will work together, as they 
> > have the same EV chips in them. These cannot be mixed with -03 and -05 
> > (DS20) or -43 and -45 (DS20 or DS20E). 
> > 
> > DS20 and DS20E processors 54-24758-43 & -45 will work together; they 
> > have the same EV chips in them. These cannot be mixed with -03 and -05 
> > (DS20) or -33 and -35 (DS20 or DS20E). 
> > 
> > DS20E processors 54-30060 -01, -02, -03, -04 can be mixed together. 
> > 
> > 
> > System Model Number CPU Description 
> > 
> > DS20E 54-24758-33 KN310-BA 500 MHz CPU, EV6 4MB Motorola/IBM Pass 2.5 
> > 54-24758-35 
> > 
> > DS20E 54-24758-43 KN310-CA 500 MHz CPU, EV67 4MB IBM Pass 2.5 54-24758-45 
> > 
> > DS20E 54-30060-01 KN311-BA,BB,BC,BD 667MHz CPU, EV67 8MB IBM/Sony Cache 
> > Pass 2.2.3 54-30060-02 
> > 
> > DS20E 54-30060-03 KN311-CA,CB,CC,CD 667MHz CPU, EV67 8MB IBM/Sony Cache 
> > Pass 2.6 54-30060-04 
> > 
> > DS20E 54-30482-02 KN312-BB,BC,BD 833MHz CPU, EV68 8MB IBM/Sony Cache 
> 
> 
> More (old) reading: 
> 
> same issue on AlphaServer ES40 as with AlphaServer DS20 and DS20E processors: 
> 
> https://manx-docs.org/collections/antonio/dec/EK-KN610-CL.A01.pdf 
> 
> And for the Alpha EV6 21264 chip internals and passes and ilk: 
> 
> https://www.star.bnl.gov/public/daq/HARDWARE/21264_data_sheet.pdf 
> 
> And if you don't already have it: 
> 
> https://manx-docs.org/collections/antonio/dec/MDS-2000-01/cd1/ALPHA/AS140SVA.PDF 
> 
> 
> 
> 
> -- 
> Pure Personal Opinion | HoffmanLabs LLC
Ah yes, I remember Andy Moskal and I looking at various ES40 CPU boards when we were cutting back on the ZKO3 labs.  I had several ES40s in the compiler clusters that had boards he was looking for.  We also had magnifying glasses looking at memory DIMMs too looking for particular revs.