11 Oct
2001
11 Oct
'01
3:51 p.m.
Glen Goodwin skrev:
>Hmm -- my personal experience is that, unless the machine is already
>junked-up with a random assortment of TV and radio tuner cards, video
>accelerators, SoundBlasters and DVD decoders, SCSI is easily added to a
>motherboard with onboard IDE ports. I've personally built a few dozen
of
them, using
both IDE and PCI SCSI controllers.
>What sorts of problems are you encountering? You're welcome to try to sort out our pile of SCSI
cards which so far
haven't
worked in our OpenBSD machine.
One machine, or many? If you have a pile of cards, one machine, and none
of the cards work in that specific box, I'd suggest that there may be a
fault in the box, or another device in there which conflicts with your SCSI
cards.
We don't use PCI cards, though (we sold that
one, since we wouldn't use the SCSI for anything else than tapestreamers
and
so forth). Besides, SCSI integration into PC systems
is really clumsy.
Wow - these days the BIOS setup utility on the cards allows a lot of
flexibility and also provides information about the installed SCSI devices,
making it easy to verify IDs, etc.
A few years ago I was building 80MHz 486 systems using Rancho Technologies
RT1000 8-bit SCSI cards which I got for about $8 each. The cards were
shipped labeled "not Win95 compatible," but, not only did they work with
Win95, I never saw a faulty card, or a system in which the RT1000 wouldn't
work. (Okay, so they were *slow*)
So, my SCSI experiences have been good.
and they don't behave like the IDE hard drives.
Well, that's the point, right? ;>)
They
have their own little BIOSes and things which I'm not used to from other
systems.
Those little BIOSes (the ones with a setup program) are a *big* advantage.
Just today I was cursing the fact that the BIOS on the ATA-66 controller I
was installing didn't have a setup program. It took me two hours to get
all six IDE drives working properly. With a decent SCSI card it would have
been 15 minutes, tops (barring any bad drives or cables).
In fact, non-PC systems tend to see IDE as a kind of
bastard SCSI
instead.
They may be onto something there . . .
Glen
0/0
12 Oct
12 Oct
1:12 a.m.
Glen Goodwin skrev:
Glen Goodwin
skrev:
>Hmm -- my personal experience is that, unless the machine is already
>junked-up with a random assortment of TV and radio tuner cards, video
>accelerators, SoundBlasters and DVD decoders, SCSI is easily added to a
>motherboard with onboard IDE ports. I've personally built a few dozen
of
> >them, using both IDE and PCI SCSI controllers.
>
> >What sorts of problems are you encountering? Iggy Drougge replied:
You're welcome to try to sort out our pile of
SCSI cards which so far
haven't
> worked in our OpenBSD machine. One machine, or many? If you have a pile of cards, one
machine, and none
of the cards work in that specific box, I'd suggest that there may be a
fault in the box, or another device in there which conflicts with your SCSI
cards.
Who cares what the cause is? The point is that it won't work.
We don't
use PCI cards, though (we sold that
one, since we wouldn't use the SCSI for anything else than tapestreamers
and
> so forth). Besides, SCSI integration into PC systems is really clumsy. Wow - these days the BIOS setup utility on the cards
allows a lot of
flexibility and also provides information about the installed SCSI devices,
making it easy to verify IDs, etc.
I see that BIOS setup utility on the cards as a sympthom of the low level of
integration. The cards behave as an alien entity in the computer.
A few years ago I was building 80MHz 486 systems using
Rancho Technologies
RT1000 8-bit SCSI cards which I got for about $8 each. The cards were
shipped labeled "not Win95 compatible," but, not only did they work with
Win95, I never saw a faulty card, or a system in which the RT1000 wouldn't
work. (Okay, so they were *slow*)
So, my SCSI experiences have been good.
So have mine, only not with PCs.
> and they don't behave like the IDE hard
drives.
Well, that's the point, right? ;>)
Not really. IMO a drive is a drive is a drive is a drive.
> They
> have their own little BIOSes and things which I'm not used to from other
> systems.
Those little BIOSes (the ones with a setup program) are
a *big* advantage.
Just today I was cursing the fact that the BIOS on the ATA-66 controller I
was installing didn't have a setup program. It took me two hours to get
all six IDE drives working properly. With a decent SCSI card it would have
been 15 minutes, tops (barring any bad drives or cables).
Why? I've never had any need for a SCSI BIOS on my SCSI computers.
> In fact, non-PC systems tend to see IDE as a kind
of bastard SCSI
> instead.
They may be onto something there . . .
For example, the onboard IDE controller in the Amiga 4000 and 1200 appear to
the system as scsi.device. And NetBSD/amiga only recently switched from "IDE
on SCSI" to the machine-independent WDC device. =)
--
En ligne avec Thor 2.6a.
Every Horse has an Infinite Number of Legs (proof by intimidation):
Horses have an even number of legs. Behind they have two legs, and in front
they have fore-legs. This makes six legs, which is certainly an odd number of
legs for a horse. But the only number that is both even and odd is infinity.
Therefore, horses have an infinite number of legs. Now to show this for the
general case, suppose that somewhere, there is a horse that has a finite
number of legs. But that is a horse of another color, and by the [above]
lemma ["All horses are the same color"], that does not exist.
12:18 a.m.
At 02:12 PM 10/12/01 +0100, you wrote:
I see that BIOS setup utility on the cards as a
sympthom of the low level of
integration. The cards behave as an alien entity in the computer.
Which is exactly the point :-) . SCSI cards, unlike IDE subsystems,
are supposed to work by themselves and ease the burden
on the processor. This is considered a "good thing".
Like many others, I've not had problems with the Adaptec 2940 series.
I also like a CMD card that I have somewhere; it has a 72 pin SIMM
slot to increase the cache; at one point I had a 64MB SIMM there and
it helped the web server a lot.
But the nicest one I've seen is the ibm ultra-160 raid controller
in our smp netfinity Linux server. It absolutely rocks.
Do note that I tend to use HP, Sun, DEC or IBM scsi cables.
Mac cables have given me trouble before.
carlos.
--------------------------------------------------------------
Carlos E. Murillo-Sanchez carlos_murillo(a)nospammers.ieee.org
7:22 a.m.
Carlos Murillo skrev:
At 02:12 PM 10/12/01 +0100, you wrote:
>I see that BIOS setup utility on the cards as a sympthom of the low level of
>integration. The cards behave as an alien entity in the computer.
Which is exactly the point :-) . SCSI cards, unlike
IDE subsystems,
are supposed to work by themselves and ease the burden
on the processor. This is considered a "good thing".
Something which doesn't necessarily have to shine through in the user/OS
interface. In this case, it's not due to the inner workings of SCSI, but
rather the PC philosophy.
Like many others, I've not had problems with the
Adaptec 2940 series.
We sold ours. It worked, but was wasted on our SCSI setup after our last SCSI
drive died.
I also like a CMD card that I have somewhere; it has a
72 pin SIMM
slot to increase the cache; at one point I had a 64MB SIMM there and
it helped the web server a lot.
IBM had two 30-pin slots on their MCA SCSI cards.
I like my Fastlane Z3 card. It's about the fastest Zorro3 SCSI card one might
get, and it's got sixteen 30-pin SIMM slots for system memory expansion as
well. =)
Do note that I tend to use HP, Sun, DEC or IBM scsi
cables.
Mac cables have given me trouble before.
I've not had much problem with cables. If a cable doesn't work, it's quite
obvious, and I replace it.
My DECstation runs with an HP cable now, and it works fine. A good testament
to the marriage between the two companies. =)
--
En ligne avec Thor 2.6a.
Programming today is a race between software engineers striving to build
bigger and better idiot-proof programs, and the universe trying to produce
bigger and better idiots. So far, the universe is winning.
-- Rick Cook, Mission Manager, NASA Mars Pathfinder Project
1:58 a.m.
Rumor has it that Iggy Drougge may have mentioned these words:
You're certainly entitled to your opinion, but I cannot subscribe to that
point of view -- especially when I want my drive to do something a normal
drive isn't supposed to do.
And don't get me started on IDE 2.1 Orb Drives... Love the storage, *Hate*
the setup... (and XP doesn't support it yet...) SCSI "just plain works..."
And you won't until you need to boot from any drive that's not ID 0 on the
board (like booting from CD, etc...)
;-)
Glen Goodwin skrev:
If the cause is the motherboard, you can't (justly) accuse the SCSI card of
being the problem...
One machine, or many? If you have a pile of
cards, one machine, and none
of the cards work in that specific box, I'd suggest that there may be a
fault in the box, or another device in there which conflicts with your SCSI
cards.
Who cares what the cause is? The point is that it won't work. I see that BIOS setup utility on the cards as a
sympthom of the low level of
integration. The cards behave as an alien entity in the computer.
With my job, one realizes that *every* card is an alien entity -- I've had
*much* less heartburn getting add-on SCSI cards to work compared to
PCI-based add-on IDE cards (which, BTW, have their own BIOS too - and a
much more useless one at that...) and you apparently have never gotten a
Sound Blaster Live MP3/5.1 card working under Windows 2000... It took me
over 4 hours to research that one, whereas my 4-year-old --ahem -- no
longer supported -- PCI Diamond MultiMedia Fireport 40 Ultra SCSI card
installed without a hitch...
> and they
don't behave like the IDE hard drives.
Well, that's the point, right? ;>)
Not really. IMO a drive is a drive is a drive is a drive. > They
> have their own little BIOSes and things which I'm not used to from other
> systems.
Those little BIOSes (the ones with a setup
program) are a *big* advantage.
Just today I was cursing the fact that the BIOS on the ATA-66 controller I
was installing didn't have a setup program. It took me two hours to get
all six IDE drives working properly. With a decent SCSI card it would have
been 15 minutes, tops (barring any bad drives or cables).
Why? I've never had any need for a SCSI BIOS on my SCSI computers. > In fact,
non-PC systems tend to see IDE as a kind of bastard SCSI
> instead.
>They may be onto something there . . . For example, the onboard IDE controller in the Amiga
4000 and 1200 appear to
the system as scsi.device. And NetBSD/amiga only recently switched from "IDE
on SCSI" to the machine-independent WDC device. =)
And the only way I can get my IDE 16X Plextor CD-RW to work under Linux is
with SCSI emulation - without it I end up with digital frizbees...
Thankfully, SCSI works for me! ;-)
Just my $0.000002 (Canadian - I'm going to Canada today...)
Roger "Merch" Merchberger
7:34 a.m.
Roger Merchberger skrev:
Rumor has it that Iggy Drougge may have mentioned these
words:
>Glen Goodwin skrev:
>
> >One machine, or many? If you have a pile of cards, one machine, and none
> >of the cards work in that specific box, I'd suggest that there may be a
> >fault in the box, or another device in there which conflicts with your
> >SCSI cards.
>
>Who cares what the cause is? The point is that it won't work.
If the cause is the motherboard, you can't (justly)
accuse the SCSI card of
being the problem...
That's most probably not the case.
>I see that BIOS setup utility on the cards as a
sympthom of the low level of
>integration. The cards behave as an alien entity in the computer.
With my job, one realizes that *every* card is an alien
entity -- I've had
*much* less heartburn getting add-on SCSI cards to work compared to
PCI-based add-on IDE cards (which, BTW, have their own BIOS too - and a
much more useless one at that...) and you apparently have never gotten a
Sound Blaster Live MP3/5.1 card working under Windows 2000... It took me
over 4 hours to research that one, whereas my 4-year-old --ahem -- no
longer supported -- PCI Diamond MultiMedia Fireport 40 Ultra SCSI card
installed without a hitch...
Should I envy you??;-)
> >> and they don't behave like the IDE
hard drives.
>
> >Well, that's the point, right? ;>)
>
>Not really. IMO a drive is a drive is a drive is a drive.
You're certainly entitled to your opinion, but I
cannot subscribe to that
point of view -- especially when I want my drive to do something a normal
drive isn't supposed to do.
And don't get me started on IDE 2.1 Orb Drives...
Love the storage, *Hate*
the setup... (and XP doesn't support it yet...) SCSI "just plain works..."
I'd never stand up for IDE. I only use it out of necessity, like everyone
else. But I'm not so sure that my next drive will be an IDE. Not if I'm not
getting a new IDE controller at the same time.
> >> They
> >> have their own little BIOSes and things which I'm not used to from
other
> >> systems.
>
> >Those little BIOSes (the ones with a setup program) are a *big* advantage.
> >Just today I was cursing the fact that the BIOS on the ATA-66 controller I
> >was installing didn't have a setup program. It took me two hours to get
> >all six IDE drives working properly. With a decent SCSI card it would
> >have been 15 minutes, tops (barring any bad drives or cables).
>
>Why? I've never had any need for a SCSI BIOS on my SCSI computers.
And you won't until you need to boot from any drive
that's not ID 0 on the
board (like booting from CD, etc...)
That's just stupid.
On my DECstation, I issue a boot command like:
boot 6/rz2/netbsd
to boot NetBSD from SCSI ID 2. If I'd like to boot Ultrix from ID 4 instead,
I'd type
boot 6/rz4/vmunix
or something like that.
On the Amiga, what matters is the boot priority of any particular drive or
partition. That is independent of the SCSI or IDE ID. If I don't want to boot
the device with the highest priority, I select another boot device in the
Early Startup menu.
And then we have PS/2s, which start booting at device 6 and work their way
down.
AFAIK, none of my SCSI machines have any SCSI device at unit 0, except for the
PS/2s, where the controller is ID 0.
Finding your boot device based on SCSI ID just sounds stupid.
>For example, the onboard IDE controller in the
Amiga 4000 and 1200 appear to
>the system as scsi.device. And NetBSD/amiga only recently switched from "IDE
>on SCSI" to the machine-independent WDC device. =)
And the only way I can get my IDE 16X Plextor CD-RW to
work under Linux is
with SCSI emulation - without it I end up with digital frizbees...
Thankfully, SCSI works for me! ;-)
It should work for everyone except the wallet. =(
And it just seems absolutely stupid that IDE has survived to this day. By now,
IDE has evolved from the set of buffers on an ISA bus that it originally was
into a beats which is almost as complex as SCSI, while still not nearly as
flexible. At the same time, even though drives for both interfaces share more
and more parts, SCSI has only become more and more expensive, and now it's not
even an alternative for a lot of semi-professional applications. It's really
absurd, and a testament to how the mass-market mechanisms are a very dangerous
factor to development.
--
En ligne avec Thor 2.6a.
Life begins at '030. Fun begins at '040. Impotence begins at '86.
9:56 a.m.
Iggy Drougge wrote:
Why? I've never had any need for a SCSI BIOS on
my SCSI computers.
The onboard bios' diagnostics, low level formatting, and media
verification, termination control, among other features, can be
rather convenient. And you don't have to worry about overwriting
the engineering track(s?) that some(all?) IDE has.
You don't have the master/slave issues or incompatibilities that
some IDE drives had
with SCSI, you can do more devices per controller port in
case, if for no other reason than, you want to.
I'd never stand up for IDE. I only use it out of
necessity, like
everyone else. But I'm not so sure that my next drive will be an
IDE. Not if I'm not getting a new IDE controller at the same time.
At the same time, even though drives for both
interfaces share more
and more parts, SCSI has only become more and more expensive, and
now it's not even an alternative for a lot of semi-professional
applications.
If you don't like SCSI? And you don't like IDE? And you don't like
expensive components? What are you going to use? only MFM and RLL?
If you must have the latest SCSI interface technology, yes, you
must "pay the man", but if you're willing to sacrifice a bit on
performance, and go one or two steps back on SCSI hardware technology,
there is nothing expensive about the parts. Plus, as it's
been up till now, the best, newest, most expensive IDE performance
level was only about par with that of whatever SCSI level was one
generation back.
11:16 a.m.
Jan Koller skrev:
Iggy Drougge wrote:
> Why? I've never had any need for a SCSI BIOS on my SCSI computers.
The onboard bios' diagnostics, low level
formatting, and media
verification, termination control, among other features, can be
rather convenient. And you don't have to worry about overwriting
the engineering track(s?) that some(all?) IDE has.
Well, low-level formatting IDE drives is a stupid idea, I concur. =)
You don't have the master/slave issues or
incompatibilities that
some IDE drives had
...
with SCSI, you can do more devices per controller port
in
case, if for no other reason than, you want to.
Why are you telling me this?
> I'd never stand up for IDE. I only use it out
of necessity, like
> everyone else. But I'm not so sure that my next drive will be an
> IDE. Not if I'm not getting a new IDE controller at the same time.
> At the same time, even though drives for both
interfaces share more
> and more parts, SCSI has only become more and more expensive, and
> now it's not even an alternative for a lot of semi-professional
> applications.
If you don't like SCSI? And you don't like IDE?
And you don't like
expensive components? What are you going to use? only MFM and RLL?
What the hell are you talking about? Please point out where I have expressed
my dislike for SCSI.
If you must have the latest SCSI interface technology,
yes, you
must "pay the man", but if you're willing to sacrifice a bit on
performance, and go one or two steps back on SCSI hardware technology,
there is nothing expensive about the parts. Plus, as it's
been up till now, the best, newest, most expensive IDE performance
level was only about par with that of whatever SCSI level was one
generation back.
You're preaching to the already converted. =?
--
En ligne avec Thor 2.6a.
G? med i SUGA, Swedish Usergroup of Amiga!
WWW: http://swedish.usergroup.amiga.tm/
BBS: 08-6582572, telnet://sua.ath.cx:42512
2:52 p.m.
On Fri, 12 Oct 2001, Jan Koller wrote:
Iggy Drougge wrote:
I think a better comparison would substitute ST-506 and ESDI for MFM
and RLL.
- don
Why? I've never had any need for a SCSI BIOS
on my SCSI computers.
The onboard bios' diagnostics, low level formatting, and media
verification, termination control, among other features, can be
rather convenient. And you don't have to worry about overwriting
the engineering track(s?) that some(all?) IDE has.
You don't have the master/slave issues or incompatibilities that
some IDE drives had
with SCSI, you can do more devices per controller port in
case, if for no other reason than, you want to.
I'd never stand up for IDE. I only use it out
of necessity, like
everyone else. But I'm not so sure that my next drive will be an
IDE. Not if I'm not getting a new IDE controller at the same time.
At the same time, even though drives for both
interfaces share more
and more parts, SCSI has only become more and more expensive, and
now it's not even an alternative for a lot of semi-professional
applications.
If you don't like SCSI? And you don't like IDE? And you don't like
expensive components? What are you going to use? only MFM and RLL? If you must have the latest SCSI interface technology,
yes, you
must "pay the man", but if you're willing to sacrifice a bit on
performance, and go one or two steps back on SCSI hardware technology,
there is nothing expensive about the parts. Plus, as it's
been up till now, the best, newest, most expensive IDE performance
level was only about par with that of whatever SCSI level was one
generation back.
6:39 a.m.
A non-text attachment was scrubbed...
Name: not available
Type: text
Size: 437 bytes
Desc: not available
Url :
http://www.classiccmp.org/pipermail/cctalk/attachments/20011012/35ee3613/at…
9:26 a.m.
Tony Duell skrev:
> Who cares what the cause is? The point is that it
won't work.
This comment worries me a lot (mind you, it's
typical of the mentality of
PC-goons, but I thought this list was mostly hackers)...
Whenever something doesn't work, my first reaction
is to fix it. And the
first stage in fixing a problem is to identify what that problem is (at
whatever level of detail is necessary to be able to decide on the right
repair).
I have limited patience with some kind of hardware. Mending PCs is not a hobby
of mine, and if it is the most common platform at the moment, I'll just assume
that that's because it's such a nice and easy to use platform. There is so
much to blame everything on in the PC world, but that's not my problem. It
should at least work as fluently as any other platform.
In short, I don't think some systems are deserving of my undivided attention.
I'll rather go soldering on my Atari ethernet adapter or my C64 AV cable than
fool around with so-called modern hardware which misbehaves.
--
En ligne avec Thor 2.6a.
- Lupin! Vart ska du?
- Tar v?gen ?ver berget f?ljer gr?nsen s? kommer vi till havet.
- Det ?r ju mer ?n 100 km.
- Det ?r v?l ingenting. Jag tillh?r v?rldseliten jag.
Lupin III den otrolige (Lupin III vs. fukusei ningen), TMS 1978
8255
days inactive
8256
days old
10 comments
7 participants
participants (7)
-
acme_ent＠bellsouth.net -
ard＠p850ug1.demon.co.uk -
cmurillo＠emtelsa.multi.net.co -
donm＠cts.com -
optimus＠canit.se -
vze2mnvr＠verizon.net -
zmerch＠30below.com