l...@garlic.com (Anne & Lynn Wheeler) writes: > 3090 added vector processing as part of playing in the supercomputer > market ... however that required that they also be able to support > 100mbyte/sec (and/or 1gbit/sec) I/O. 3090 was barely able to get up to > 4.5mbyte/sec transfers ... so what to do?
re: http://www.garlic.com/~lynn/2017d.html#4 GREAT presentation on the history of the mainframe as mentioned mainframe bus&tag half-duplex had increasing throughput problems, handling 3mbyte/sec disks and then 4.5mbyte/sec disks ... but unable to handle the engineering&scientific 40mbyte/sec disk arrays. I've also mentioned that ESCON didn't finally get announced until it was already obsolete (1990 w/ES9000) ... at 17mbytes/sec, it couldn't also handle the 40mbyte/sec disk arrays. This also basically precluded a project that the father of 801/risc roped me into ... doing a "wide" disk head that did 18 tracks in parallel. This is somewhat analogous to the 60s 2301 drum, essentially the same as the 2303 drum but read/wrote four tracks in parallel at a time, four times the transfer rate with 1/4th the number of tracks that were four times larger. Original 3380 had 20track width spacing between every data track. This was later reduced to 10track width spacings for double density 3380 with twice the number of cylinders ... and the spacing was reduced again for triple density 3380. Tne 3380 "wide-head" disk would have 16 adjacent data tracks plus a 17th servo track. The "wide-head" would span 18tracks, two servo tracks on either side of the 16 data tracks. It would read/write at 16*3mbytes/sec=48mbytes/sec. The problem for the project was that none of IBM mainframe I/O could handle that data rate. some old email about the wide-head r/w 16 data tracks. http://www.garlic.com/~lynn/2006s.html#email871122 http://www.garlic.com/~lynn/2006s.html#email871230 following year (1988), I get asked to help LLNL standardize some serial stuff they are playing with ... which quickly evolves into fibre channel standard (initially @1gbit/sec full-duplex, 2gbit/sec aggregate) ... but mainframe heavy-duty protocol running over fibre channel standard isn't available until much, much later http://www.garlic.com/~lynn/submisc.html#ficon past posts getting to play disk engineer in bldgs 14&15 in the late 70s and much of the 80s http://www.garlic.com/~lynn/subtopic.html#disk past 16+2 disk head posts: http://www.garlic.com/~lynn/2006s.html#30 Why magnetic drums was/are worse than disks ? http://www.garlic.com/~lynn/2007m.html#23 Bulkiest removable storage media? http://www.garlic.com/~lynn/2007o.html#64 Toshiba Boosts Hard Drive Density By 50% http://www.garlic.com/~lynn/2009e.html#41 "A foolish consistancy" or "3390 cyl/track architecture" http://www.garlic.com/~lynn/2009i.html#66 Was there ever a 10in floppy? http://www.garlic.com/~lynn/2009k.html#75 Disksize history question http://www.garlic.com/~lynn/2009p.html#12 Secret Service plans IT reboot http://www.garlic.com/~lynn/2010m.html#52 Basic question about CPU instructions http://www.garlic.com/~lynn/2011g.html#70 History of byte addressing http://www.garlic.com/~lynn/2012e.html#39 A bit of IBM System 360 nostalgia http://www.garlic.com/~lynn/2012e.html#103 Hard Disk Drive Construction http://www.garlic.com/~lynn/2013e.html#3 The Big, Bad Bit Stuffers of IBM http://www.garlic.com/~lynn/2013g.html#41 A History Of Mainframe Computing http://www.garlic.com/~lynn/2013o.html#92 Cylinder buffer http://www.garlic.com/~lynn/2014b.html#17 Quixotically on-topic post, still on topic http://www.garlic.com/~lynn/2014l.html#78 Could this be the wrongest prediction of all time? http://www.garlic.com/~lynn/2016f.html#2 IBM DASD RAS discussion -- virtualization experience starting Jan1968, online at home since Mar1970 ---------------------------------------------------------------------- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN