Monday, October 16, 2017

Memory management

One of the platters from the 80 Mbyte drive I used in grad 
school.  Note the magnetic materials has been scraped down
to the bare metal. Always back up.  I lost six weeks of work
that had to be retyped in.
It's the start of fall break here and fall weather seems to have arrived with it. (Though not quite to stay, the temperatures for next weekend exceed the 90th percentile for high's at this point in October.)  I've been looking forward to some time to write and some time to do some of the low priority things I'd pushing to the back.  Even as I write this, I keep remembering things I wanted to accomplish!

A couple of weeks ago I spoke at St. Edward's University in Austin for their annual Lucian symposium.  Four speakers, a dinner, a student poster session. My talk was organized around the tools that chemists have for exploring structure, including some we might not think of as "modern" such as paper models, making the case for broadening the base of tools chemists have access to along with modernizing them.

Along the way I showed some of the computational tools I had used, including our "big" drive.  80 Mbytes and the size of a washing machine.  I can now keep 128 Gbytes on my key chain and forget I have it.  One feature of the coding I did in those days was the need to intentionally allocate memory for lists and arrays, including deciding could be moved into memory as a piece. For these computationally demanding tasks, having a carefully designed overlap was key.

My to-do list feels like that old-fashioned overlay.  I have the teaching list, the administrative list, the writing list, the personal lists. Load them in, wiping what is currently in memory.  Repeat.  Can I keep things from accidentally writing over what I need to keep in working memory (the load of wash in while I write?).

No comments:

Post a Comment