return to Short Bio

A start of an autobiography for Randy


> P.S.  I'm CCing my "youngest" - now in his 40s
>    who got into computing about age 16, after
>    punched paper tape was no longer used in computing.
A bit younger, in fact. I wrote my first program in 6th grade. I suspect I was writing them with pencil and paper on green engineering graph paper under the sheets with a flash light the year before; but I didn't have a computer to run them on. (Dad and/or my brother Edward would take me to the Commodore Computer store in San Jose where I'd code them in the keyboard. Then my dad bought one.)

At any rate, a few years before I turned 10 my dad would bring me and my brothers in to Measurex where we played games on TTY terminals connected to main frames (Star Trek was a popular game we played). That terminal had paper tape. And I believe we had a paper tape version of the game. But, it was more for show than use, if I remember correctly. Which I often don't. :-)

Here's a screen shot of what would have been printed on paper then:


COMMAND> srscan


    1 2 3 4 5 6 7 8 9 10
 1  . . . . . . . . . .  Stardate      3600.0
 2  . . . .   . * . . .  Condition     GREEN
 3  . * . . . . * . . .  Position      2 - 1, 8 - 3
 4  . . . . . . . . . .  Life Support  ACTIVE
 5  . . . . . . . . . .  Warp Factor   5.0
 6  . . . . . . . . * .  Energy        5000.00
 7  . . . . * . . * . .  Torpedoes     10
 8  . . E . . . . . . .  Shields       DOWN, 100% 2500.0 units
 9  . . . . . .   . . .  Klingons Left 4
10  . . . * . . . . . .  Time Left     7.00
Ah. Those were the days. Wait, these are the days!

-- Randy


some correspondence triggered by Robert Garner's "Tales of CISC and RISC from Xerox PARC and Sun" - 12/07/11 - scroll down the contents to the bottom.
Hey Robert --

That was great! I really appreciated the historical perspective.

It was fun seeing Steve Kleiman in your photo of those at Sun on the Hardware + Software Team (he was the guy on the far right, I believe, 50min 40sec into the video). I saw Steve and his wife Helen at Clarke's Burgers in Mountain View the other day. They're both fine.

Processor architecture remains a fascination of mine. For example, one of my favorite books is: Readings in Computer Architecture

I'm over at Apple working on mobile silicon helping hardware and software engineers debug DRAM, CPU and SOC issues. Best job I've had in my life. Second was NetApp debugging off-the-shelf chipsets and processors. (NetApp didn't want to spend a great deal of money building custom hardware. For the most part they limited themselves to the relatively-simple-to-implement file system accelerator: a PCI NVRAM card.)

My favorite (and shortest) NetApp story takes place on a Friday evening. It was about 4PM and I was heading out of the office, and my friend Steve Strange and his wife were coming over for dinner. Steve Strange's office was in the same building as mine and I walked over to let him know that we should leave to get home for dinner. He said, "Sure, but I've got this core dump here that I can't make heads or tails of. What do you think?" I looked at it for a few minutes and then asked him to let me sit in his seat. I poked around the core for a while and then I said, "Hmm. The processor (a Pentium 4) wrote the wrong value onto the stack for the return address and when the called function 'returned' it went off to some other piece of code and thus we took this invalid address exception that caused the core dump to be generated. The correct return address would have had bit 19 cleared. See?" Steve looked at me and said, "Wow. I never would have presumed it was a processor error." I said, "Neither would I, but that's what the evidence indicates. Come one, let's file the bug report and get home; Gina's going to be mad if we're late." In the end, we weren't late for dinner. And, the fix was new microcode from Intel to change the drive strength of bit 19 during the write cycle of the return address. Who knew they could fix things like that?

I'm looking forward to watching some of the other presentations in Ron Mak's History of Computing line-up.

-- Randy


Randy --> Yes, that is Steve Kleiman in the pic. He was a key contributor the SPARC architecture and member of the OS group, and, of course, NFS co-designer. (Note that Xeroids had previously designed the equivalent at PARC. ;-) Steve and I worked closely together in bringing up SunOS on the first SPARC, the Sun-4/200. I've got a fun story about debugging a problem with its virtual address cache.

Interesting story that you guys stumbled over an Pentium bit-flip bug and that Intel was able to address it via a microcode patch. They must have been sufficiently pre-worried to have designed in a firmware controlled bus driver. Sun also had a harry bit-flip problem inside our SuperSPARC chip, which wasn't under firmware control. :-(.

It's good to hear that Apple has a functioning and "happy" hardware group these days! I never would have imagined that Apple would become the world's mostly highly valued company. And $100b of cash / cache! ;-)

...

Best Wishes,

- Robert

IBM Almaden Research, San Jose, CA
Office: 408-927-1739
Mobile: 408-679-0976
robgarn@us.ibm.com