Michael L. VanLoon writes:
> 
> 1) It doesn't look like you're using tagged command queuing.  That
> will improve performance more than anything you can do.  Rebuild a
> kernel with "options AHC_TAGENABLE" in it, and boot that.
> 
> 3) Try running these tests on a ccd striped over two or more identical
> drives.

AHC_TAGENABLE appears to make a *huge* difference in filesystem
performance. I repeat my previous query - Why isn't this part of the
standard kernel config?

Thanks for the pointer to the config option, Michael!

Single ST15150W w/o AHC_TAGENABLE:

  Disk /dev/rsd0a bandwidth          6677 KB/sec
  File /usr/tmp/XXX write bandwidth: 1466 KB/sec

Single ST15150W with AHC_TAGENABLE:

  Disk /dev/rsd0a bandwidth          6624 KB/sec
  File /usr/tmp/XXX write bandwidth: 3898 KB/sec

Striped (alternating tracks) dual ST15150W's w/o AHC_TAGENABLE

  Disk /dev/rccd0a bandwidth         11388 KB/sec
  File /n/carp/1/XXX write bandwidth: 2368 KB/sec

Striped (alternating tracks) dual ST15150W's with AHC_TAGENABLE

  Disk /dev/rccd1a bandwidth         12437 KB/sec
  File /n/carp/1/XXX write bandwidth: 5663 KB/sec

(note: the former is on NetBSD 1.2-release. The latter is on a snapshot
of OpenBSD from last June)