clemon@lemsys.UUCP (Craig Lemon) (11/28/90)
This has been something that's been bugging me for a long time
now.... In the near past I have heard people with low-cost type HD
controllers (even for the 500) giving performances of around 500K/sec
(KBytes), even when used with lower cost drives (ie. Seagate). Some
people with better controllers on stock 2000s have given speeds of
900K/sec. (ie. HardFrame/2000 and Quantum Drive). I have a REV 4.2 B2000
with a HardFrame/2000 and Q105S and since the day I installed it I've
gotten the following performance :
--
DiskPreformance - V3.0 - 03/21/89
Testing drive dh0: with big files (harddisk mode)
File create/delete: create 12 files/sec, delete 40 files/sec
Directory scan: 102 entries/sec
Seek/read test: 114 seek/reads per second
r/w speed: buf 1024 bytes, rd 104509 byte/sec, wr 47880 byte/sec
r/w speed: buf 8192 bytes, rd 204268 byte/sec, wr 122401 byte/sec
r/w speed: buf 32768 bytes, rd 227951 byte/sec, wr 140434 byte/sec
--
What is wrong with this picture? Just now (as I write this article),
the performance with small files (floppy mode) has dropped to 65 K/sec read
and 17 K/sec writes. I got no documentation with my drive and I am using
whatever defaults were set by RDPrep. I have a
capacity of 102MB so I lost some space somewhere, but I'm not worried
about that right now. Could my 64KB read ahead cache be disabled? (I
also heard people mentioning that the Amiga never uses the cache. Is this
true?) What about SCSI termination? Hardframe/2000 ROMS? If needed my
full sys.config is :
B2000 REV 4.2
KS 1.2
WB 1.3.2
A2088 Bridgecard (ancient model)
2x2052 memory (4MB total)
68010
Hardframe/2000 (obviously)
Quantum Q105S
Miniscribe 30MB (BRIDGECARD) with some controller
I thought that it was possible that the Bridgecard was having some
effect (ie. eating cycles for the bus or something) but I've run the test
without the Bridgecard installed and it still gives these numbers. I
believe I even tried it without the memory. This is what RDPrep says the
settings are for my Rigid Block setup :
Existing Partitions on this disk.
block 27, Name "DH0"
TableSize : 16
SizeBlock : 128
SecOrg : 0
Surfaces : 1
SectorPerBlock : 1
BlocksPerTrack : 201
Reserved : 2
PreAloc : 0
Interleave : 0
LowCyl : 1
HighCyl : 1018
NumBufers : 5
BufMemType : 5
MaxTransfer : 131072
Mask : 0x0fe
BotPri : 3
DosType : "DOS^A"
What about my MaxTransfer setting? Could this be the culprit?
I've left it at default as well. I haven't tried playing with it because
I haven't got a backup of this drive and I don't want to accidently do
something or encounter a power out at the worst possible time or something
like that. (BTW, I use addbuffers dh0: 30 in my startup). One thing I do
notice is that when I access a very large file (ie. copying Maker from
DvideoIII, about 500K, and copying it to RAM: or something, disk access
consists of about 3-5 very short bursts about 0.75-1.0 seconds apart. If I
could line these bursts up into one big burst, then I'd be happy.
What about the 68010? I can't see any effect that it would have.
I don't autoboot from my drive (KS 1.2 until 2.01 is released) so that
explains my bootpri.
While I have you, one more Hardframe guru question. If I set my
bootpri as high as it will go to the hardframe, will I still be able to
boot from floppy if one is inserted. This may sound stupid but I have
heard of controllers that won't let you unless you disable it.
--
Craig Lemon - Kitchener, Ontario. Amiga B2000/10--2400 bps--AmigaUUCP 1.03D
clemon@lemsys.UUCP or lemsys!clemon@xenitec.on.ca
....!{uunet}!watmath!xenitec!lemsys!clemonlphillips@lpami.wimsey.bc.ca (Larry Phillips) (11/29/90)
In <3989.659757130@lemsys.UUCP>, clemon@lemsys.UUCP (Craig Lemon) writes: >KS 1.2 >WB 1.3.2 This is not a good idea, in general. Mixing releases can lead to all sorts of subtle problems. >This is what RDPrep says the >settings are for my Rigid Block setup : > >Existing Partitions on this disk. >block 27, Name "DH0" > TableSize : 16 > SizeBlock : 128 > SecOrg : 0 > Surfaces : 1 > SectorPerBlock : 1 > BlocksPerTrack : 201 > Reserved : 2 > PreAloc : 0 > Interleave : 0 > LowCyl : 1 > HighCyl : 1018 > NumBufers : 5 Good idea here would be to increase the number of buffers, at least a little. > BufMemType : 5 > MaxTransfer : 131072 Unless Microbotics has changed their hardware, this value is just a little high. The most you should have here is 130560 (128K-512 bytes). This limitation is because of the limitations in the SCSI chip and the DMA controller. > Mask : 0x0fe This is most likely to be the biggest factor in your lack of speed. Tha mask value is used to determine where host adapter is allowed to use DMA, and where it has to use programmed IO. With the value set as you show it here, the HardFrame will only DMA into even addreses between $00 and $fe. _ALL_ accesses to or from FAST memory, or to CHIP memory above $fe, will be done using programmed IO. The proper value here is 0xfffffe for a 2000, which says that the host adapter may DMA to any even address in the natural addressing range. > BotPri : 3 > DosType : "DOS^A" > > What about my MaxTransfer setting? Could this be the culprit? >I've left it at default as well. The rdprep program has gone through several revisions. One of the revisions may have changed this value to 130560. At any rate, it should be reduced to this value. > I haven't tried playing with it because >I haven't got a backup of this drive and I don't want to accidently do >something or encounter a power out at the worst possible time or something >like that. Do yourself a favour. Make a backup, then play with the figures. Changing MaxTransfer and Mask will not destroy your data, but any time you are playing around with drive stuff like this, a small mistake could cause problems. -larry -- The only things to survive a nuclear war will be cockroaches and IBM PCs. +-----------------------------------------------------------------------+ | // Larry Phillips | | \X/ lphillips@lpami.wimsey.bc.ca -or- uunet!van-bc!lpami!lphillips | | COMPUSERVE: 76703,4322 -or- 76703.4322@compuserve.com | +-----------------------------------------------------------------------+
dzenc@gnu.ai.mit.edu (Dan Zenchelsky) (11/30/90)
In article <2279@lpami.wimsey.bc.ca> lphillips@lpami.wimsey.bc.ca (Larry Phillips) writes: >In <3989.659757130@lemsys.UUCP>, clemon@lemsys.UUCP (Craig Lemon) writes: > >>KS 1.2 >>WB 1.3.2 > >This is not a good idea, in general. Mixing releases can lead to all sorts of >subtle problems. > >-larry > >| // Larry Phillips | While I agree with you that, in general, mixing versions is a "Bad Thing," it is quite safe to mix KS 1.2, and WB 1.3.x. In fact, the changes between Kickstart 1.2 and 1.3 are so small, that Commodore refers to Kickstart 1.3 as Kickstart 1.2.1 in its technical documentation. To quote from the _Native Developers Update 1.3_: VERSION: Kickstart version number is bumped to 34.4 . FEATURES: V1.2.1 is an incremental kickstart release. It is designed to maintain V1.2 compatiblity throughout, while adding two important features to the operating system. These features are 1) autoboot from ROM-based expansion devices and 2) a larger preferences structure for future expansion. -Dan -- ___________________________________________________________________________ | _______ |________________________________________| | || |o| Dan Zenchelsky | | | ||____| | | Any sufficiently advanced bug is | | | ___ | dzenc@gnu.ai.mit.edu | indistinguishable from a feature. | | |_|___|_| |______________-- Rich Kulawiec__________| |__________________________________|________________________________________|
peterk@cbmger.UUCP (Peter Kittel GERMANY) (11/30/90)
In article <2279@lpami.wimsey.bc.ca> lphillips@lpami.wimsey.bc.ca (Larry Phillips) writes: >In <3989.659757130@lemsys.UUCP>, clemon@lemsys.UUCP (Craig Lemon) writes: > >>KS 1.2 >>WB 1.3.2 > >This is not a good idea, in general. Mixing releases can lead to all sorts of >subtle problems. Sorry Larry, you're normally right, but not in this point. It was officially stated by Commodore that you may mix KS 1.2/1.3 with WB 1.2/1.3 in every possible combination! The only disadvantage with KS 1.2 is the loss of autoboot ability with a HD. This was though an exception. It will NOT apply to KS/WB 2.0 mixing with its predecessors, that's impossible. -- Best regards, Dr. Peter Kittel // E-Mail to \\ Only my personal opinions... Commodore Frankfurt, Germany \X/ {uunet|pyramid|rutgers}!cbmvax!cbmger!peterk
jayward@eecs.cs.pdx.edu (Jay Ward) (12/01/90)
In article <600@cbmger.UUCP> peterk@cbmger.UUCP (Peter Kittel GERMANY) writes: >In article <2279@lpami.wimsey.bc.ca> lphillips@lpami.wimsey.bc.ca (Larry Phillips) writes: > >Sorry Larry, you're normally right, but not in this point. It was >officially stated by Commodore that you may mix KS 1.2/1.3 with WB 1.2/1.3 >in every possible combination! The only disadvantage with KS 1.2 is the >loss of autoboot ability with a HD. > I believe that there is a problem in mixing the 1.3 printer.device with the 1.2 printer drivers. I remember hearing (quite a long time ago) that this particular mix-and-match combination would NOT work. It seems that at least one commercial program (DeluxePrint II) was released with this combo before anyone even realized the problem. > >-- >Best regards, Dr. Peter Kittel // E-Mail to \\ Only my personal opinions... >Commodore Frankfurt, Germany \X/ {uunet|pyramid|rutgers}!cbmvax!cbmger!peterk ---------------------------------------------------------------------------- Jay Ward --> jayward@eecs.cs.pdx.edu | if TrailBlazers > Opponents then | TrailBlazerWins++ ----------------------------------------------------------------------------
andy@cbmvax.commodore.com (Andy Finkel) (12/01/90)
In article <806@pdxgate.UUCP> jayward@eecs.UUCP (Jay Ward) writes: >I believe that there is a problem in mixing the 1.3 printer.device with the 1.2 >printer drivers. I remember hearing (quite a long time ago) that this >particular mix-and-match combination would NOT work. It seems that at least >one commercial program (DeluxePrint II) was released with this combo before >anyone even realized the problem. In general, you could mix 1.2 and 1.3 printer device with printer drivers. However, EA did a custom version of the 1.2 printer.device/drivers which was not compatible with the directions we went in 1.3. That was what was on the Deluxe Print disk, and they later switched to standard 1.3 printer drivers, or so I've been told. andy -- andy finkel {uunet|rutgers|amiga}!cbmvax!andy Commodore-Amiga, Inc. "It is much easier to suggest solutions when you know nothing about the problem." Any expressed opinions are mine; but feel free to share. I disclaim all responsibilities, all shapes, all sizes, all colors.
lphillips@lpami.wimsey.bc.ca (Larry Phillips) (12/02/90)
In <600@cbmger.UUCP>, peterk@cbmger.UUCP (Peter Kittel GERMANY) writes: >In article <2279@lpami.wimsey.bc.ca> lphillips@lpami.wimsey.bc.ca (Larry Phillips) writes: >>In <3989.659757130@lemsys.UUCP>, clemon@lemsys.UUCP (Craig Lemon) writes: >> >>>KS 1.2 >>>WB 1.3.2 >> >>This is not a good idea, in general. Mixing releases can lead to all sorts of >>subtle problems. > >Sorry Larry, you're normally right, but not in this point. It was >officially stated by Commodore that you may mix KS 1.2/1.3 with WB 1.2/1.3 >in every possible combination! The only disadvantage with KS 1.2 is the >loss of autoboot ability with a HD. > >This was though an exception. It will NOT apply to KS/WB 2.0 mixing with >its predecessors, that's impossible. I stand corrected, though if I were to run 1.2 KS and 1.3 WB, I would be holding my breath most of the time, regardless of the assurances. :-) -larry -- The only things to survive a nuclear war will be cockroaches and IBM PCs. +-----------------------------------------------------------------------+ | // Larry Phillips | | \X/ lphillips@lpami.wimsey.bc.ca -or- uunet!van-bc!lpami!lphillips | | COMPUSERVE: 76703,4322 -or- 76703.4322@compuserve.com | +-----------------------------------------------------------------------+
bleys@tronsbox.xei.com (Bill Cavanaugh) (12/03/90)
>I stand corrected, though if I were to run 1.2 KS and 1.3 WB, I would be >holding my breath most of the time, regardless of the assurances. :-) > >-larry > >-- >The only things to survive a nuclear war will be cockroaches and IBM PCs. >+-----------------------------------------------------------------------+ >| // Larry Phillips | >| \X/ lphillips@lpami.wimsey.bc.ca -or- uunet!van-bc!lpami!lphillips | >| COMPUSERVE: 76703,4322 -or- 76703.4322@compuserve.com | >+-----------------------------------------------------------------------+ I've been running a combination of KickStart 1.2 (it's on ROM in the A500, which I bought back in March '87), WorkBench 1.3 (which I bought at the first Philly WOC show, in November '88), and ARP 1.3 (which I downloaded sometime in the summer of '88). I've never had a problem, and I've never heard of anyone who did. In March of '89, when I bought my GVP Impact A500, I thought about getting the KS 1.3 chip so I could autoboot, but promises of KS/WB 1.4 RSN held me back.<grin> /**************************************************************** * All of the above copyright by the below. * * Bill Cavanaugh uunet!tronsbox!bleys * * "A language is a dialect with an army and a navy." * ****************************************************************/