(Ln(x))3

The everyday blog of Richard Bartle.

RSS feeds: v0.91; v1.0 (RDF); v2.0; Atom.

Previous entry. Next entry.


10:56am on Thursday, 14th January, 2010:

Lacking Drive

Outburst

I've come to the conclusion that my PC is too slow not because of its CPU but because of its hard drive. Whenever I have to wait 30 seconds for a program to start up or exit, it's because the hard drive is thrashing away.

OK, it may be because I'm using Windows XP, but even so I didn't used to get this long a wait when I was using a DECsystem-10 under timesharing (well, not often anyway). Defragmenting it makes no difference, and it isn't even half full: it's just slow.

We have more memory and faster CPUs, but hard drives don't seem to have caught up in terms of speed. It's useless having a top of the range graphics card and a multi-core set of processors if they're sitting around twiddling their thumbs while Windows does some stupid act of housekeeping that entails blocking all activity until the hard drive has stopped thundering. Caches make a difference, in that if they didn't exist the drives would be even slower, but in absolute terms my hard drive is just too slow.

In the old days, when disc packs were the size of washing machines, the fastest ones were fixed head. What this meant was that they had a read/write head for each track. They didn't have to scoot across the disc like a floating head system does, they just picked up the data on the next revolution. They could also write multiple pieces of data simultaneously. Why can't we have fixed head drives now? Why do we have to have floating heads that are slow slow slow?

Also, the old drives were multi-platter. Instead of having just one spinning surface, they had several. This, too, meant that you could write several pieces of data at the same time, even if each head was floating rather than fixed. Why can't we have this? I don't mean a RAID, I mean a single-spindle hard drive with multiple platters. That would speed things up, too, especially if files were split between platters so they could be written out a physical chunk at a time.

Maybe I should just wait until we have reliable solid-state permanent storage instead of waiting for old technology to be readopted...


Latest entries.

Archived entries.

About this blog.

Copyright © 2010 Richard Bartle (richard@mud.co.uk).