4 ms·
The language survey class I took as part of my CSci degree, had us use Lisp on a MicroVAX. The resources that Lisp required versus the machine we were running
by protomyth 8y ago
The language survey class I took as part of my CSci degree, had us use Lisp on a MicroVAX. The resources that Lisp required versus the machine we were running it on made it very much like programing a batch system. The terminals were slow when Lisp was running. It was so bad we typed our programs out on a PC and ftp'd it up to the VAX and then loaded it, check if we got the syntax right, run it, and repeat. The whole interactive lesson was lost on us since it would take forever to do anything. The only good that came out of it was winning enough money at poker played in the lab to pay for pizza that weekend.
- nils-m-holm 8y agoThe VAX was really a beautiful architecture and a good fit for systems that ran lots of small processes subsequently. However, it was also notoriously bad at running LISP because of its huge page table overhead. A page was 512 bytes in size, giving an address space of 2^23 pages. So when a process grew larger, the machine started to swap out parts of the page table and everything went down to a crawl. This is probably what you observed. 512-byte pages worked fine on a 18-bit or 22-bit architectures, like the various models of the PDP-10 and the PDP-11, but were simply too small for the VAX. So the VAX did solve the memory shortage of its predecessors, but in the case of LISP at the price of abysmal performance when running memory-intensive applications, which many LISP programs unfortunately were.