Unixy system benchmarks

Ivan Voras has finished his benchmark of a system running FreeBSD, NetBSD, DragonFly and Linux. DragonFly does quite well, coming unsurprisingly close to the leader – FreeBSD 4.x. The difference would probably be more pronounced on a multiprocessor system, which wasn’t used in these benchmarks.

5 Replies to “Unixy system benchmarks”

  1. Mm, benchmarks!

    I’m still reading, but it brings to mind a ‘usability’ annoyance in the BSD versioning system; thankfully, “April 7th, 2004” is specified at the bottom, but elsewise, where Linux gets defined odd-numbered releases, at least 3 out of 4 BSDs get “-CURRENT.” Not a big deal… until you’re searching around and land on a page of indeterminate date.

    Nothing wrong with the development model, but anyone have ideas for a notation better or more compact than “-CURRENT as of DD/MM/YY,” which we all know we should write, but seldom do? (Of course, that’s date-format chauvinist, but as long as there’d be a “standard”…)

  2. FreeBSD-5-040407 sound reasonable, doesn’t it? (ISO date, YYMMDD) That’s the way most people note their CVS versions in e.g. Debian packages not based on released tarballs.

  3. Works for me, and/or if there’s a way to extract a CVS version off something that always increments with each ‘touch’ to the repository, that’d make for even better forensics.

  4. You mean something like a named/bind serial number? YYMMDD01, YYMMDD02 etc.

  5. Yes, though that’s probably the worst of both worlds as far as compactness is concerned. (CVS doesn’t quite let you roll around like that, so I suppose it doesn’t really matter, does it? Unless someone wants to hack on an automatic changelog facility, where the changelog conveniently resides in CVS and rolls over per-tagging — or one file just records the ‘last change’ touch and gets versioned? Probably not much fun for diminishing returns…)

Comments are closed.