[comp.sys.amiga] FFS speed. Bytes per second or Bits per second?

dillon@CORY.BERKELEY.EDU (Matt Dillon) (02/26/88)

>cmcmanis@sun.UUCP (Chuck McManis) wrote:
>>Yes it stinks, yes it is fixed by the Fast File system
>
>I've seen an A2000 with Connor 100M disk on a CBM 2090 controller
>running FFS; they're getting around 600 Bytes/sec.  Yum.

	Don't you mean:  600Kbps .. 600K BITS/sec

	Considering the transfer rate for a hard disk, I don't think
FFS can go at 600K Bytes/sec.   600Kbps-1Mbps ==  75KBps-125KBps.
(That is, with normal hard disks).

					-Matt

steveb@cbmvax.UUCP (Steve Beats) (02/29/88)

In article <8802251923.AA23029@cory.Berkeley.EDU> dillon@CORY.BERKELEY.EDU (Matt Dillon) writes:
>	Considering the transfer rate for a hard disk, I don't think
>FFS can go at 600K Bytes/sec.   600Kbps-1Mbps ==  75KBps-125KBps.
>(That is, with normal hard disks).
>
>					-Matt

Imagine me stood on your desk with an indignant look and a belligerant stance..
got it ...... good, I shall begin.

FFS CAN GO 600K BYTES PER SECOND ON READS FROM A CONNER 100MB SCSI DRIVE!!!!!
There, now I feel better :-)  Really, it does go that fast, honest, would I
lie to you?  This figure wasn't just plucked out of thin air, I really timed
it over many test runs with large files.  Admittedly, the file was not fragged
up but the disk was not empty either so header and extension block fetches
caused some extraneous seeks.

	Steve Beats

PS.  You can carry on now, I just got off your desk :-)

dillon@CORY.BERKELEY.EDU (Matt Dillon) (03/01/88)

:FFS CAN GO 600K BYTES PER SECOND ON READS FROM A CONNER 100MB SCSI DRIVE!!!!!
:There, now I feel better :-)  Really, it does go that fast, honest, would I
:lie to you?  This figure wasn't just plucked out of thin air, I really timed
:it over many test runs with large files.  Admittedly, the file was not fragged
:up but the disk was not empty either so header and extension block fetches
:caused some extraneous seeks.
:
:	Steve Beats
:
:PS.  You can carry on now, I just got off your desk :-)
:

	Yah, that's what I get for typing and sleeping at the same
	time.

	I WANT IT!  (Damn it, just when I thought my money was safe,
	yet another reason appears to get an A2000!)

					-Matt

cmcmanis%pepper@Sun.COM (Chuck McManis) (03/01/88)

In article <8802251923.AA23029@cory.Berkeley.EDU> (Matt Dillon) writes:
>	Considering the transfer rate for a hard disk, I don't think
> FFS can go at 600K Bytes/sec.   600Kbps-1Mbps ==  75KBps-125KBps.

In article <3385@cbmvax.UUCP> (Steve Beats) writes:
> FFS CAN GO 600K BYTES PER SECOND ON READS FROM A CONNER 100MB SCSI DRIVE!!!!!

What neither Matt nor Steve realize is that we've substituted an
instant coffee for their regular brew. Seriously, what we have is some
assumptions on the part of the claimants that disagree :-). Basically,
if you have a Hard disk and it has an ST-506 interface to the physical
drive mechanism you get around 650K Bytes per second out of the read
head tops. This is due to the fact that the ST-506 interface spec puts
the Data Clock at 5 Mhz, and 5000000/8 = 625K Bytes. In reality, the
better way to compute the transfer speed is to take the number of
sectors on a track (16) multiply that by the size (512 bytes) and
multiply that by the speed of rotation in seconds (3600 RPM = 60 RPS)
to get 512*16*60 = 491,520 Bytes/Second MAX.  But all these wonderful
formulas go out the window when the drive has an imbedded SCSI
controller because the physical drive interface may no longer be
ST-506. And in fact it usually isn't since ST-506 is out dated anyway.
Steve wasn't using an ST-506 based drive, he was using the Connors
Peripherals drive, which has a SCSI interface on it, and the actual
drive interface is something like RLL or GCR and generates bits of data
at substantially higher rates. And in truth the limiting factor becomes
the SCSI interface, which in its Asynchronous mode is limited to 1.25
Megabytes per second. So they are both right, FFS can't do 600K bytes
per second on an ST-506 drive and FFS can do 600K bytes per second on 
some SCSI drives.


--Chuck McManis
uucp: {anywhere}!sun!cmcmanis   BIX: cmcmanis  ARPAnet: cmcmanis@sun.com
These opinions are my own and no one elses, but you knew that didn't you.