Unfortunately, there's no meaningful difference between 1.5 with A/UX
interrupts and without. But then again, these numbers can't be right for
two otherwise identical machines:

              -------Sequential Output-------- ---Sequential Input-- --Random--
              -Per Char- --Block--- -Rewrite-- -Per Char- --Block--- --Seeks---
Machine    MB K/sec %CPU K/sec %CPU K/sec %CPU K/sec %CPU K/sec %CPU  /sec %CPU
1.4 Q800  100   502 99.4   385 31.0   392 52.9  1064 94.8  3928 92.5  89.5 24.7
1.5 Q800  100   169 97.6   192 96.4    99 96.4   274 99.4   333 95.5  19.1 12.2
Amiga 3k  100   267 99.5  1526 99.0   876 98.3   271 99.7  1917 96.8  60.9 38.1
Amiga 4k  100  2821 95.8  7131 96.9  3909 96.9  2265 96.6  8519 98.2 149.6 25.2

1.4 Q800 & 1.5 Q800: Quadra 800, 40 mhz 68040, 72 megs of ram, 4 gig Seagate
Amiga 3k: Amiga 3000, 25 mhz 68030/882, 16 megs of ram, 4 gig Seagate
Amiga 4k: Amiga 4000, 66 mhz 68060, 144 megs of ram, 18 gig IBM UW SCSI

Now, I'd expect the Q800 to come out faster than the Amiga 3k, and of
course I'd expect the Amiga 4k to whip the Quadras.

But the 1.5 Quadra is simply pitiful. It reminds me of the old 16 bit
ST-506 controllers where when the interleave was set improperly, the
system would crawl. There is something wrong here.

The A/UX interrupts isn't it. Does anyone have any other ideas?

Thanks,
John Klos