Persistent Memory At LinuxCon 2016

Seems like there is a lot of interest currently in “persistent memory” architectures (e.g. flash storage that doesn’t sit on the other side of a slow SATA interface). Quite a few different programming models and libraries are being developed to try to find the best approach to making efficient use of this resource <http://phoronix.com/scan.php?page=news_item&px=Persistent-Memory-Linux-2016>. I suppose this is like the early days of disks (and drums, and tapes), when the first primitive filesystems were being developed. Anybody remember core memories? But they were never used this way...

Anybody remember core memories? But they were never used this way...
In the 1980's I used to maintain a computer in Te Puke that had core memory and no disk drives. As an example, you didn't specifically run a shutdown routine. You just turned off the power. In the milli-seconds before the DC rails collapsed, the computer would interrupt through its power-fail address and pick up the address in core memory to execute its power-fail routine. This routine ended with a "halt" instruction and at some point not long after that the DC rails collapsed. At power on, once the DC rails were stable, the computer would interrupt through its start-up routine address, and execute any startup code for self-testing before jumping to execute the application. The only output from the computer was to print on the console. If the memory was every replaced, then the applications were reloaded into memory using a paper tape reader. At the time, somewhat jokingly, we used to say, "Real computers don't have disk drives". I've been waiting quite a long time, but I envisage it won't be too long before everyone is saying that, "Real computers don't have disk drives". One day we will look back on the history of computer technology and say, "remember that 50 year period when computers had disk drives? ...wasn't that a totally insane concept!". cheers, Ian.
participants (2)
-
Ian Stewart
-
Lawrence D'Oliveiro