andrew@frip.WV.TEK.COM (Andrew Klossner) (02/23/90)
[]
"You cannot get more than 600KB per second out of the
filesystem in the best of circumstances, and even that is only
achieved, as far I know, by the MIPS UNIX. Others top out at
around 300KB per second."
This doesn't sound right. I just tried sequential disk reads from a
large file on a Tektronix XD88 (88k-based) workstation and got
650KB/sec. Even our wimpy 68020-based systems get 450KB/sec. This
through a slow SCSI channel, and the file system is Berkeley FFS.
-=- Andrew Klossner (uunet!tektronix!frip.WV.TEK!andrew) [UUCP]
(andrew%frip.wv.tek.com@relay.cs.net) [ARPA]ehrlich@cs.psu.edu (Daniel Ehrlich) (02/23/90)
In article <6238@orca.wv.tek.com> andrew@frip.WV.TEK.COM (Andrew Klossner) writes:
Andrew> []
Andrew> "You cannot get more than 600KB per second out of the
Andrew> filesystem in the best of circumstances, and even that is only
Andrew> achieved, as far I know, by the MIPS UNIX. Others top out at
Andrew> around 300KB per second."
Andrew> This doesn't sound right. I just tried sequential disk reads from a
Andrew> large file on a Tektronix XD88 (88k-based) workstation and got
Andrew> 650KB/sec. Even our wimpy 68020-based systems get 450KB/sec. This
Andrew> through a slow SCSI channel, and the file system is Berkeley FFS.
I can not speak to the speed of the file system on AIX, but it is a `new and
improved' file system. :-) I still have not found all of the details in the
documentation. It does have an `atomic transaction log' (my term, not
IBM's) that allows fsck to quickly rebuild a damaged file system. As I wade
through the docs and find more info I will post again.
--
Dan Ehrlich <ehrlich@cs.psu.edu>
Voice: +1 814 863 1142 FAX: +1 814 865 3176prc@erbe.se (Robert Claeson) (02/26/90)
In article <6238@orca.wv.tek.com>, andrew@frip.WV.TEK.COM (Andrew Klossner) writes: > Others top out at around 300KB per second." > This doesn't sound right. I just tried sequential disk reads from a > large file on a Tektronix XD88 (88k-based) workstation and got > 650KB/sec. Even our wimpy 68020-based systems get 450KB/sec. This > through a slow SCSI channel, and the file system is Berkeley FFS. I just tried reading a 10 MB large file sequentially on our Encore Multimax. The disk is a NEC something (655 MB formatted) and it is using the *old* I/O system from Encore (based on single-ended, async SCSI). I got well over 900 KB/sec throughput thru the BSD FFS on a non-striped partition under System V. -- Robert Claeson E-mail: rclaeson@erbe.se ERBE DATA AB
madd@world.std.com (jim frost) (03/09/90)
prc@erbe.se (Robert Claeson) writes: >In article <6238@orca.wv.tek.com>, andrew@frip.WV.TEK.COM (Andrew Klossner) writes: >> Others top out at around 300KB per second." >I just tried reading a 10 MB large file sequentially on our Encore >Multimax. [...] >I got well over 900 KB/sec throughput thru the BSD FFS on a >non-striped partition under System V. Current versions of UMAX (Encore's BSD) do a bit better than that. But for good throughput, I'd like to see what Amdahl's UNIX does. Remember that UNIX describes the interface, not the implementation or the hardware. You got good I/O, you got good throughput. jim frost saber software jimf@saber.com