karl@ddsw1.MCS.COM (Karl Denninger) (12/21/88)
A while back there was some statement made here that one of the reasons for poor I/O performance on SCO Xenix had to do with SCO's driver resetting the controller after each sector transfer.... Well, if that's the case, then I certainly can't find the problem with the practice. We just got in a load of WD1006-V/SR2 controllers, hot off the assembly line (serial #s under 200!). These, for those of you who aren't familiar, are RLL 2,7 track buffered Hard/Floppy controllers, nearly 100% surface mount construction, and will fit in a XT286 machine (short & only about 1/2 length). They're brand new; Western Digital had some serious production problems with the original design. It now appears the troubles are fixed, although there is a very limited supply at present (I grabbed everything I could get my hands on, and at a reasonable price too. Email, call voice or see my "biz" group posting for more info if you want one). These things work beautifully. I'm getting over 300Kbyte/second through the Xenix drivers now, about three times the previous "best". Raw I/O rates to/from the drive are in the 650Kbyte/second range, about as good as one can ask out of a RLL ST-412 unit (theoretical max is 750Kbyte/sec). This compares with nearly-identical raw transfer rates for the Adaptec ACB2372, but only 120Kbytes/second through the Xenix drivers (boo hiss!). The difference really has to be seen to be appreciated. I'd estimate that from a user's viewpoint, you've got 80%+ of the throughput of an ESDI subsystem at substantially less than half the cost. Drives we've checked out so far that work with these things include the Seagate 251's and ST4096's (if these work, nearly anything ought to be ok) as well as a Miniscribe 20M half-height XT drive. No problems at all. Boots are near-instant -- your first reaction upon seeing the system boot will be "nawwww, it couldn't be up that fast". The board also contains a ROM BIOS, and on-board formatting, surface analysis and defect management software (none of which is needed for Xenix applications, but nice for DOS!) Well recommended -- there's one in the machine this is being posted from. Wherever that nasty rumor about SCO's drivers came from, it's hogwash. We'll be testing these smokers with Microport and 386/ix within the next couple of weeks; email if you'd like results. If sufficient interest is shown, I'll post a followup. -- Karl Denninger (karl@ddsw1.MCS.COM, ddsw1!karl) Data: [+1 312 566-8912], Voice: [+1 312 566-8910] Macro Computer Solutions, Inc. "Quality solutions at a fair price"