Probs with Adaptec AVA-1505
Horst Schambeck
I have the problem with the AVA-1505 from Adaptec that the computer does not
recognize it. Means, that ASPI2DOS.SYS can not find the adapter even in DOS
mode.
But some times ago the card worked on an Intel-486-based system very well.
I want to operate it on an Asustek P/I-P55TP4 board with an Intel
Pentium(1)-133. I already disabled the Interrupts for the PCI-Bus which are
possible with the 1505 ("used by ISA").
Please, can anyone help me and give me fast answer?
Thanks a lot!
Horst
Ervier
always problems installing the scsi-card. I have installed this card on
win95 and win98 several times succesfully, but now it just won't. I've
tried to change irqs, memory allocations, old and new installations,
dos-only, but nothing helps. I've taken out the soundcard, and all
other cards that i can miss, so the system is as simple as possible.
Does anyone have any tips?
Ervier
In article <mMY46.240170$Bq.10...@news.chello.at>, Horst Schambeck
Ron Reaugh
Ervier wrote in message <070120010151474897%r4_NO...@mailmij.nl>...
>Same problem here. I have an IBM Aptiva 133, with windows98, and as
>always problems installing the scsi-card. I have installed this card on
>win95 and win98 several times succesfully, but now it just won't. I've
>tried to change irqs, memory allocations, old and new installations,
>dos-only, but nothing helps. I've taken out the soundcard, and all
>other cards that i can miss, so the system is as simple as possible.
Get a PCI SCSI card like a 2906 or 2930U.
Alain-Pierre CHERTIER
AVA-1505 is a simple but very good SCSI-2 controller :-)
You have to choose the IRQ and I/O address on it.
I/O address can be 0x340 or 0x140 only.
IRQ can be one of the allowed jumper positions.
The choosed IRQ must be free, not conflicting with other cards...
The ASPI2DOS.SYS needs a specific parameter to let it speek with
the AVA-1505 at the correct I/O address...
See Adaptec's manuals to find the correct "/x" parameter...
If I remember, ASPI2DOS.SYS uses the 0x340 address by default (?)
In Win95 or 98, you don't need the ASPI2DOS.SYS device driver
if you work only under the 32bits graphical environment...
But you'll have to force W95/98 to recognize the card at the
correct I/O address and IRQ... You must force config in the usual
way in the Configuration Panel...
I hope this will help you ?
Let me know...
Alain-Pierre CHERTIER <ap.ch...@free.fr>
Sent via Deja.com
http://www.deja.com/
Ron Reaugh
Alain-Pierre CHERTIER wrote in message <93cef8$eq9$1...@nnrp1.deja.com>...
>Hello !
>
>AVA-1505 is a simple but very good SCSI-2 controller :-)
No, it's ISA.
Eric Gisin
Alain-Pierre provided a helpful answer. Something you are incapable of.
"Ron Reaugh" <Ron-R...@worldnet.att.net> wrote in message
news:hqj66.6166$fj6.5...@bgtnsc05-news.ops.worldnet.att.net...
Chris Pitzel
to date touch the computer while you're burning.
I'm totally with Ron on this one.. 1505's suitable for use with 386s
and 486s, but not with any of the Pentium-era equipment.
Ron Reaugh
Eric Gisin wrote in message ...
>No Ron, wrong again. It is fine for CDs, and most SCSI-2 stuff.
Nonsense. not on current configurations.
Eric Gisin
even with a SCSI disk. You are as bad as Ron if you don't believe the people
who have first hand experience.
"Chris Pitzel" <chris....@nospam.usask.ca> wrote in message
news:3A5A34E6...@nospam.usask.ca...
Eric Gisin
even tested an ISA SCSI card, yet pretend to be an expert.
"Ron Reaugh" <Ron-R...@worldnet.att.net> wrote in message
news:d8r66.422$Cy.1...@bgtnsc06-news.ops.worldnet.att.net...
Ron Reaugh
Eric Gisin wrote in message ...
>even tested an ISA SCSI card, yet pretend to be an expert.
You are a bald faced liar. I have several ISA SCSI cards.
Ron Reaugh
Eric Gisin wrote in message ...
>even with a SCSI disk. You are as bad as Ron if you don't believe the
people
>who have first hand experience.
Just NO!
lun...@bit-net.com
an aha-1522- using adaptec ez-scsi's cd-copier, get a speed test
value of 4x. That means it was tranfering 4x * 2. I've also seen
the same pc- using scsi explorer, rate 1.8 mb/sec asynchronous xfer
speed from a fujitsu hdd. It may not be up to modern high performance
standards, but it'll do for a lot of things.
An aha-1522 has the same controller chip as an ava-1505. The ava-1505
is just extremely simplfied. No bios, etc.
The only way I'd agree with this is if the thread were about 1542's.
Isa bus-mastering dma does rather poorly on (Intel) pci chipsets.
Ron Reaugh
lun...@bit-net.com wrote in message <3a5a66f3...@news.bit-net.com>...
>I have personally witnessed a p166 w/ (what I now know to be)
>an aha-1522- using adaptec ez-scsi's cd-copier, get a speed test
>value of 4x. That means it was tranfering 4x * 2. I've also seen
>the same pc- using scsi explorer, rate 1.8 mb/sec asynchronous xfer
>speed from a fujitsu hdd. It may not be up to modern high performance
>standards, but it'll do for a lot of things.
While doing that the overall multitasking performance of current PCI chipset
machines are negatively impacted. Use PCI busmastering SCSI cards.
lun...@bit-net.com
cpu usage while this is happening. But it's an Nt machine,
and nt always has a lot of needed processes running in the
backround, so I don't think the scsi card is impacting the
cpu as much as you think.
Ron Reaugh
lun...@bit-net.com wrote in message <3a5a737...@news.bit-net.com>...
>The next time I'm over there, I'll see if I can check the
>cpu usage while this is happening. But it's an Nt machine,
>and nt always has a lot of needed processes running in the
>backround, so I don't think the scsi card is impacting the
>cpu as much as you think.
At a 6X-8X burn it likely is having an impact and I'm not sure if a CPU
usage tools can measure it as what CPU usage is can't even be well defined
and this is just such a case that tests the definition.
Eric Gisin
dozen just like it.
Reading at 600KB while writing at 600KB. Makes Ron's 150/300KB claim look
clueless. Why would any sane person agree with Ron on this one?
<lun...@bit-net.com> wrote in message
news:3a5a66f3...@news.bit-net.com...
Eric Gisin
I know the results for a 154X, post the results for a 1505/1520! Around
2MB/s, right? More than enough for CD-R at 8X, right?
"Ron Reaugh" <Ron-R...@worldnet.att.net> wrote in message
news:Qtt66.22135$7f3.1...@bgtnsc07-news.ops.worldnet.att.net...
Ron Reaugh
Eric Gisin wrote in message ...
>Hey Chris, don't forget to save this article. I've seen at least half a
>dozen just like it.
>
>Reading at 600KB while writing at 600KB. Makes Ron's 150/300KB claim look
>clueless.
You are a bald faced liar. I never made any "150/300KB claim".
Chris Pitzel
> even with a SCSI disk. You are as bad as Ron if you don't believe the people
> who have first hand experience.
First of all, you seem to not be able to give up the intifadah that
you're carrying on against Ron when he's clearly saying stuff that's
accurate.
Secondly, I've personally had the misfortune of dealing with a few
systems with 1505's in them, and believe me, that controller on a
Pentium machine is a piece of garbage. Never could get a 2X CD-R to
work on one well a few years back -- ended up giving my friend a SC-200
instead. And I ran into another one which brought the system to a crawl
whenever even the simplest of disk i/o was performed (mp3 playback).
The last one I've run into was when a friend tried to get an 4X CD-R to
work on one. Worked sometimes, but the slightest move of the mouse
would cause CD-R's to coaster..
I wouldn't wish one of these cards even on my enemies -- I don't know
how you get off saying that these things are perfectly fine. If you
think the 5 1505's that I have sitting on my junk pile are actually
functional SCSI cards in Pentium machines, feel free to make me an
offer. But as it stands right now, I can't even give them away without
having a guilty conscience.
Eric Gisin
who haven't. Every time this thread comes up, someone reports their's works
fine. Don't you beleive other people experience?
You point about a HD on a 1505 is off topic, the same thing happens with PIO
EIDE.
"Chris Pitzel" <chris....@nospam.usask.ca> wrote in message
news:3A5AAE5E...@nospam.usask.ca...
Chris Pitzel
> who haven't. Every time this thread comes up, someone reports their's works
> fine. Don't you beleive other people experience?
In this case, no. I believe either they are tolerant of not being able
to touch their systems while burning, or they are running 486 class
machines that can more acceptably accmodate such cards.
> You point about a HD on a 1505 is off topic, the same thing happens with PIO
> EIDE.
There was no comment about HDs specifically.. That MP3 could have been
played back from a CD-ROM for all I said in the post. Yes, it was
played back from a HDD, but that doesn't make any difference. The card
isn't intrinsically any faster with CDs than it is with hard drives.
Alain-Pierre CHERTIER
A few ones who are pretending to have knowledge about SCSI
should shut up...
I'm burning CDs on a Panasonic CW7502 cd burner (4X write / 8X read)
and reading CDs on a Plextor PX32TS throuh my venerable
SoundBlaster16-SCSI2 ISA card... It has an onboard Adaptec 6360
SCSI chip (SCSI-2) and acts as would a 1505/1510 and even 1520 ctrl...
It's running fine, MSDOS, Win3.x, Win95, Win98, NT3.51, NT4.0 and
various Linux distribs...
The burning tool is Nero 3.5 or 4.0, EasyCD Creator 3.5 or CDRWin...
All this is working very well !
My mbs are ISA/PCI mbs QDI Titanium 1B (with Cyrix-200) or 1B+ (with K6-2
450)
or even old ISA/VLB QDI V4P895 mb...
Hey, Ron & Chris, you still don't want to admit the evidence ?!
AVA-1502/1505/1510/1515/1520 (and all 6260 or 6360 equipped boards)
are E-X-C-E-L-L-E-N-T in ALL systems I ever tried !
Best regards. ;-)
Alain-Pierre CHERTIER <ap.ch...@free.fr>
Ron Reaugh
Eric Gisin wrote in message ...
>who haven't. Every time this thread comes up, someone reports their's works
>fine. Don't you beleive other people experience?
Nonsense. We KNOW that the 1505 is operable. We also KNOW that it is an
overall performance abomination. Now do you get it.
>You point about a HD on a 1505 is off topic,
No you are simply clueless.
> the same thing happens with PIO
>EIDE.
Duh, PIO EIDE is to be avoided like the plague. So now you have simply
confirmed all our assertions.
Ron Reaugh
Alain-Pierre CHERTIER wrote in message ...
>Hey !!!
>
>A few ones who are pretending to have knowledge about SCSI
>should shut up...
>
>I'm burning CDs on a Panasonic CW7502 cd burner (4X write / 8X read)
>and reading CDs on a Plextor PX32TS throuh my venerable
>SoundBlaster16-SCSI2 ISA card... It has an onboard Adaptec 6360
>SCSI chip (SCSI-2) and acts as would a 1505/1510 and even 1520 ctrl...
And you have crippling overall multitasking performance while doing so.
>It's running fine, MSDOS, Win3.x, Win95, Win98, NT3.51, NT4.0 and
>various Linux distribs...
>The burning tool is Nero 3.5 or 4.0, EasyCD Creator 3.5 or CDRWin...
>All this is working very well !
>My mbs are ISA/PCI mbs QDI Titanium 1B (with Cyrix-200) or 1B+ (with K6-2
>450)
>or even old ISA/VLB QDI V4P895 mb...
>
>Hey, Ron & Chris, you still don't want to admit the evidence ?!
>AVA-1502/1505/1510/1515/1520 (and all 6260 or 6360 equipped boards)
>are E-X-C-E-L-L-E-N-T in ALL systems I ever tried !
Nonsense.
lun...@bit-net.com
I called up my friend who has the p166 w/ the adaptec 1522, and
we tried to put some numbers to all of this. Neither of us are Nt
gurus, so hopefully these are the right numbers.
Nt's performance monitor was started up, and told to graph the
% of time in priviledged mode.
Then scsi bench 32 was started and the 1.8 mb/sec fujitsu hdd
was tested, and performance monitor said it was using 95%. This
was a bit odd, because it wasn't effecting the rest of the system
too much except for the mouse pointer.
So then we tried the 4x cd-rom. Same thing. 95%. This is
impossible. If a 4x cd-rom can use 95%, the fujitsu could not
run at 1.8 mb/sec. It would shut down the system first.
At this point, I theorize that "scsi bench 32" is itself a cpu
hog, and will try to use up as much cpu time as it can find, no
matter what it's actually testing.
So then we tried some real-world tasks instead- by comparing
the cpu usage of the scsi hdd vs a 1.7 gb ide hdd. Both these
drives seem to run at about the same speed. The ide drive was
on a pci ide controller built-in to the motherboard. The task
was to xcopy about 200 mb of data on each of the drives.
The fujitsu hdd on the 1522 used about half the cpu time that
the ide drive did. That's just the average, it varied by +- 10%.
Given that, I'd say that non- mastering isa scsi isn't too
bad. Not everyone is using scsi for high performance. A lot of
things don't really need more than 2 mb/sec. Considering the
number of problems people are having w/ Adaptec's newest pci
adaptors, by comparison the isa models are incredibly reliable
and trouble-free
(if you have them configured right in the first place).
On Mon, 8 Jan 2001 19:12:29 -0800, "Eric Gisin" <er...@techie.com>
wrote:
Ron Reaugh
lun...@bit-net.com wrote in message <3a5bd98...@news.bit-net.com>...
>Seeing that this has started turning into a umm ... a discussion,
>I called up my friend who has the p166 w/ the adaptec 1522, and
>we tried to put some numbers to all of this. Neither of us are Nt
>gurus, so hopefully these are the right numbers.
> Nt's performance monitor was started up, and told to graph the
>% of time in priviledged mode.
> Then scsi bench 32 was started and the 1.8 mb/sec fujitsu hdd
>was tested, and performance monitor said it was using 95%. This
>was a bit odd, because it wasn't effecting the rest of the system
>too much except for the mouse pointer.
"95%. ....except for the mouse pointer" There's your proof!
> So then we tried the 4x cd-rom. Same thing. 95%. This is
>impossible. If a 4x cd-rom can use 95%, the fujitsu could not
>run at 1.8 mb/sec.
Not necessarily.
> It would shut down the system first.
> At this point, I theorize that "scsi bench 32" is itself a cpu
>hog, and will try to use up as much cpu time as it can find, no
>matter what it's actually testing.
>
> So then we tried some real-world tasks instead- by comparing
>the cpu usage of the scsi hdd vs a 1.7 gb ide hdd. Both these
>drives seem to run at about the same speed. The ide drive was
>on a pci ide controller built-in to the motherboard. The task
>was to xcopy about 200 mb of data on each of the drives.
> The fujitsu hdd on the 1522 used about half the cpu time that
>the ide drive did. That's just the average, it varied by +- 10%.
Was DMA mode enabled on the EIDE HD? If not then the test wasn't valid.
> Given that, I'd say that non- mastering isa scsi isn't too
>bad. Not everyone is using scsi for high performance. A lot of
>things don't really need more than 2 mb/sec. Considering the
>number of problems people are having w/ Adaptec's newest pci
>adaptors,
They are having very few problems with Adaptec PCI SCSI cards.
Eric Gisin
problems. Enough said.
We now have two people reporting 4X CD-R with Adaptec ISA works fine. Why
haven't you replied to them, to tell them their lying? I'm sure they don't
have 486s.
Face it Chris, you're not a PC tech. You can't solve simple problems like a
broken video driver that causes other problems. Or MP3 players skip if they
don't have enough buffering.
"Chris Pitzel" <chris....@nospam.usask.ca> wrote in message
news:3A5B57B3...@nospam.usask.ca...
Ron Reaugh
newbies should be ignored.
Eric Gisin wrote in message ...
Chris Pitzel
This whole intifidah thing of yours has gone far enough.. No, I'm not a
PC tech -- such a job wouldn't pay enough for me to live on. But I
don't see how a 'tech' would do any better than I would in configuring a
CD-R on a 1505.
Once again, if you think the 1505's are so great, then you're entitled
to my stash of them for not much more than the price of postage.. And
I'm not convinced that a few people have gotten them to work in anything
but extremely well-controlled scenarios.
And I do not understand your obscure reference to broken video drivers
or MP3 players at all. Where did you pull that one out of your ass?
lun...@bit-net.com
I don't happen to have at my fingertips at the moment. So
assume that it wasn't.
These attempts at a cpu benchmark had nothing to do with
"ide vs. scsi" or "busmastering vs. not".
It was only about whether/ pio mode scsi uses so much cpu
time that it cripples the system, and makes it impossible to
burn a 4x cd while moving the mouse- this is the claim that
you and Chris have been making.
If dma-mode wasn't enabled, then it was a comparison between
(pio mode ide) and (pio mode scsi). The pio mode scsi only used
about half the cpu time.
BUT, (pio mode ide) was the normal state of affairs on ide
pc's, until about 3 or 4 years ago. And nobody's system was
crippled by this. Mouses could moved while disk activity was
taking place. Ide cd-roms could be read at 8x even, and not
bring all activity to a halt.
(Pio mode scsi) used only half the cpu time of (pio mode ide)
So how will it cripple a system when the ide couldn't?
This test is also extra evidence that the 95% cpu usage
while scsibench 32 was running, is inaccurate. If the scsi
was really using 95%, then the ide sub-system would be using
190% of the cpu's max. Impossible.
"Ron Reaugh" <Ron-R...@worldnet.att.net> wrote:
|lun...@bit-net.com wrote in message <3a5bd98...@news.bit-net.com>...
| >Seeing that this has started turning into a umm ... a discussion,
| >I called up my friend who has the p166 w/ the adaptec 1522, and
| >we tried to put some numbers to all of this. Neither of us are Nt
| >gurus, so hopefully these are the right numbers.
| > Nt's performance monitor was started up, and told to graph the
| >% of time in priviledged mode.
| > Then scsi bench 32 was started and the 1.8 mb/sec fujitsu hdd
| >was tested, and performance monitor said it was using 95%. This
| >was a bit odd, because it wasn't effecting the rest of the system
| >too much except for the mouse pointer.
|
|
| "95%. ....except for the mouse pointer" There's your proof!
|
|
| > So then we tried the 4x cd-rom. Same thing. 95%. This is
| >impossible. If a 4x cd-rom can use 95%, the fujitsu could not
| >run at 1.8 mb/sec.
|
|
| Not necessarily.
|
| > It would shut down the system first.
| > At this point, I theorize that "scsi bench 32" is itself a cpu
| >hog, and will try to use up as much cpu time as it can find, no
| >matter what it's actually testing.
| >
| > So then we tried some real-world tasks instead- by comparing
| >the cpu usage of the scsi hdd vs a 1.7 gb ide hdd. Both these
| >drives seem to run at about the same speed. The ide drive was
| >on a pci ide controller built-in to the motherboard. The task
| >was to xcopy about 200 mb of data on each of the drives.
| > The fujitsu hdd on the 1522 used about half the cpu time that
| >the ide drive did. That's just the average, it varied by +- 10%.
|
|
| Was DMA mode enabled on the EIDE HD? If not then the test wasn't valid.
|
| > Given that, I'd say that non- mastering isa scsi isn't too
| >bad. Not everyone is using scsi for high performance. A lot of
| >things don't really need more than 2 mb/sec. Considering the
| >number of problems people are having w/ Adaptec's newest pci
| >adaptors,
|
|
| They are having very few problems with Adaptec PCI SCSI cards.
|
Ron Reaugh
lun...@bit-net.com wrote in message <3a5d309a...@news.bit-net.com>...
>Whether or not the Ide dma-mode was enabled, is info that
>I don't happen to have at my fingertips at the moment. So
>assume that it wasn't.
>
> These attempts at a cpu benchmark had nothing to do with
>"ide vs. scsi" or "busmastering vs. not".
> It was only about whether/ pio mode scsi uses so much cpu
>time that it cripples the system, and makes it impossible to
>burn a 4x cd while moving the mouse- this is the claim that
>you and Chris have been making.
Nope, the claim is that the overall performance is heavily impacted when
ISA I/O of 900KB/sec. or faster is underway.
> If dma-mode wasn't enabled, then it was a comparison between
>(pio mode ide) and (pio mode scsi). The pio mode scsi only used
>about half the cpu time.
No.
> BUT, (pio mode ide) was the normal state of affairs on ide
>pc's, until about 3 or 4 years ago. And nobody's system was
>crippled by this.
Wrong.
> Mouses could moved while disk activity was
>taking place. Ide cd-roms could be read at 8x even, and not
>bring all activity to a halt.
Wrong.
> (Pio mode scsi) used only half the cpu time of (pio mode ide)
Wrong.
>So how will it cripple a system when the ide couldn't?
Obviously an off question.
> This test is also extra evidence that the 95% cpu usage
>while scsibench 32 was running, is inaccurate. If the scsi
>was really using 95%, then the ide sub-system would be using
>190% of the cpu's max. Impossible.
Clueless.
lun...@bit-net.com
there was content. Apparently you are suffering from road
rage on the information super-highway. Have a nice day,
Mr Murr. Try not to be involved with too many vehicular
homicides. It's bad for your karma.
On Thu, 11 Jan 2001 06:15:52 GMT, "Ron Reaugh"
<Ron-R...@worldnet.att.net> wrote:
<snip>
Chris Pitzel
> >burn a 4x cd while moving the mouse- this is the claim that
> >you and Chris have been making.
>
> Nope, the claim is that the overall performance is heavily impacted when
> ISA I/O of 900KB/sec. or faster is underway.
900kb/sec? Try more like 300kb/sec. And 600kb/sec just isn't reliably
achievable with the 1505's that I've tried. Maybe I'm just doing
something totally wrong, maybe the drivers that were shipped with Win9x
or downloaded from Adaptec are defective, but 4X burning on those
controllers is pretty much a no-go on Pentium-era boards.
> > BUT, (pio mode ide) was the normal state of affairs on ide
> >pc's, until about 3 or 4 years ago. And nobody's system was
> >crippled by this.
> Wrong.
Exactly. 3-4 years ago, people didn't run CD burners, and those who did
had busmastering solutions available to them. And anyone who ran SCSI
HDD's 3-4 years ago had PCI equipment, or used 1505 (or similar) in
486-type systems.
> > Mouses could moved while disk activity was
> >taking place. Ide cd-roms could be read at 8x even, and not
> >bring all activity to a halt.
> Wrong.
I agree with you totally.. Start any sort of operation involving a 1505
on such board, and the whole thing becomes totally dead. Mouse
movements are jerky, and moving the mouse significantly affected data
rates.
> > (Pio mode scsi) used only half the cpu time of (pio mode ide)
>
> Wrong.
>
> >So how will it cripple a system when the ide couldn't?
>
> Obviously an off question.
The whole line of logic seems to be flawed.. There has been no evidence
that SCSI PIO used less CPU time than PIO IDE, so the other statements
have zero meaning...
Ron Reaugh
Chris Pitzel wrote in message <3A5DEB6B...@nospam.usask.ca>...
I wish you'd stop snipping the respondents name from the top of your posts.
Erik Mooney
<chris....@nospam.usask.ca> wrote:
>> > BUT, (pio mode ide) was the normal state of affairs on ide
>> >pc's, until about 3 or 4 years ago. And nobody's system was
>> >crippled by this.
>> Wrong.
>
>Exactly. 3-4 years ago, people didn't run CD burners, and those who did
>had busmastering solutions available to them. And anyone who ran SCSI
>HDD's 3-4 years ago had PCI equipment, or used 1505 (or similar) in
>486-type systems.
Nope. Almost 3 years ago, in summer 1998, I was running a 2x burner on an
ISA AHA-1505, on a Pentium 166 Win 95 machine. The system also had ISA
sound (44 KHz 16-bit stereo = consistent 176KB/sec ISA bus bandwidth
consumption) and network (on a broadband pipe where 50KB/sec downloading
was common). I could burn CDs while playing MP3s and doing moderately
heavy network (web and nntp) activity. And the IDE HD was PIO. I don't
think I ever got a buffer underrun when burning from HD (did at times when
copying CD-to-CD.)
ISA is capable of quite a lot. Remember that ISA is 16-bit, so it tops
out at 1600KB/sec, not 800.
Ron Reaugh
Erik Mooney wrote in message
<02mv5to7ko41ppof3...@news.alt.net>...
On older mobos but on FX-BX mobos ISA can only do about 1000KB/sec.
lun...@bit-net.com
tested a scsi hdd at 1.8 mb/sec. And motherboards with Via Apollo
chipsets have reported getting almost 2.5 mb/sec through their
Isa slots.
If YOUR motherboard can't do more than 1 mb/sec through an Isa slot,
then you either have it set up wrong- which I wouldn't doubt in your
case, or you have a motherboard that is poorly designed.
On Sat, 13 Jan 2001 08:27:13 GMT, ("Ron Reaugh"/"Michael A Murr, Jr")
<Ron-R...@worldnet.att.net> wrote:
>
>On older mobos but on FX-BX mobos ISA can only do about 1000KB/sec.
>
Ron Reaugh
busmastering SCSI cards.
Eric Gisin
(amoung other things).
The 154X is the one with a 1MB/s limit on Intel chipsets. The 1505/1520 can
do 2MB/s on any chipset. Hey Ron, do I have to kick you in the head for this
to sink in?
<lun...@bit-net.com> wrote in message
news:3a60605...@news.bit-net.com...
lun...@bit-net.com
problem. Perhaps that's what he wants. He already flames himself,
under the guise of MM.
I was just re-reading the linux scsi-howto, and found this section
which applies to aic-6x60 scsi cards. I guess they're better than
I've seen them to be. The howto appears to be saying that an Isa
1520 can do 4 mb/sec w/ 75% cpu, but the system still remained
usable.
<blockquote>
FIFO Polled
Boards using FIFO polled I/O put a small (typically 8K) buffer
between the CPU and the SCSI bus, and often implement some
amount of intelligence. The net effect is that the CPU is only
tied up when it is transferring data at top speed to the FIFO
and when it's handling the rest of the interrupt processing for
FIFO empty conditions, disconnect/reconnect, etc.
Peak transfer rates should be sufficient to handle most SCSI
devices, and have been measured at up to 4M/sec using raw SCSI
commands to read 64K blocks on a fast Seagate Baracuda with an
Adaptec 1520.
CPU utilization is dependent on the transfer rates of the
devices, with faster devices generating more interrupts per unit
time which require more CPU processing time. Although CPU
usage may be high (perhaps 75%) with fast devices, the system
usually remains usable. These boards will provide excellent
interactive performance with broken devices which don't
disconnect/reconnect (typically cheap CDROM drives)
Recommended for persons on a budget.
</blockquote>
On Sat, 13 Jan 2001 16:29:56 -0800, "Eric Gisin" <er...@techie.com>
wrote:
>This has been explained to Ron many times. He is incapable of learning
Chris Pitzel
> which applies to aic-6x60 scsi cards. I guess they're better than
> I've seen them to be. The howto appears to be saying that an Isa
> 1520 can do 4 mb/sec w/ 75% cpu, but the system still remained
> usable.
>
> <blockquote>
> FIFO Polled
> Boards using FIFO polled I/O put a small (typically 8K) buffer
> between the CPU and the SCSI bus, and often implement some
> disconnect/reconnect (typically cheap CDROM drives)
>
> Recommended for persons on a budget.
> </blockquote>
Your quote definitely has truth on i386 and i486 class machines, but not
Pentium machines with Intel chipsets. Keep in mind that the driver and
that howto was written in the days when 486's were the state-of-the-art,
and ISA still was well supported.
Ron Reaugh
lun...@bit-net.com wrote in message <3a61179...@news.bit-net.com>...
The above information is absolute best case on non-Intel PCI chipset mobos.
On Intel PCI chipset mobos the situation is much worse and when running the
card in saturated PIO at 1.5MB/sec. the CPU usage approaches 100& and the
system is not usable.
lun...@bit-net.com
On 1/14 you say an Intel pci mb will hit 100% cpu usage
transferring 1.5 mb/sec over an Isa slot.
On 1/13 you said that no Isa slot could ever xfer more
than 1.0 mb/sec
On 1/11 you said that a system would be crippled after
900 kb/sec.
Previous to that, you've stated that 4x cd-burning at
600 kb/sec would be impossible.
If I kept talking long enough, perhaps I could have you
admitting the truth. Or would it be like the task of
Sisyphus? A week later you might be back to the same
old thing.
I arranged a test of the equipment in question, and you
refuse to believe the results. You refuse to believe
everyone else too. What would you believe? Does someone
have to come to your house, set up a Pc, strap you to
a chair, and make you watch a benchmark?
Or perhaps you don't believe the things you say even while
you say them? Perhaps you only state these things in order
to attract flames?
At least Chris Pitzel appears to be honest and consistant
in views. When he said that he couldn't get a 1505 to go
over 2x, I get the impression that he at least tried. I
also think that his problem may have been the Os & drivers
he was using, but least that would be something which could
be talked about.
In terms of Isa vs pci scsi, I find it interesting that on a
different thread, someone with a 2940uw pro had 100% cpu
usage from windows and normal performance from linux. Does
that mean I can say pci cards are bad? No. But I can say
that the drivers can make a big difference.
I'm quitting this thread. It's been chatted to death, to the
point of boredom. You sound like Mojo Jojo- I don't want to
end up that way too.
On Mon, 08 Jan 2001 21:41:40 GMT, Chris Pitzel
<chris....@nospam.usask.ca> wrote:
>Yeah, if you have a 1X or 2X burner, and have enough patience not even
>to date touch the computer while you're burning.
>
>I'm totally with Ron on this one.. 1505's suitable for use with 386s
>and 486s, but not with any of the Pentium-era equipment.
>
>Eric Gisin wrote:
>>
>> No Ron, wrong again. It is fine for CDs, and most SCSI-2 stuff.
>>
>> Alain-Pierre provided a helpful answer. Something you are incapable of.
>>
>> "Ron Reaugh" <Ron-R...@worldnet.att.net> wrote in message
>> news:hqj66.6166$fj6.5...@bgtnsc05-news.ops.worldnet.att.net...
>> >
>> > Alain-Pierre CHERTIER wrote in message <93cef8$eq9$1...@nnrp1.deja.com>...
>> > >Hello !
>> > >
>> > >AVA-1505 is a simple but very good SCSI-2 controller :-)
>> >
>> >
>> > No, it's ISA.
>> >
Ron Reaugh
lun...@bit-net.com wrote in message <3a61e169...@news.bit-net.com>...
>To Ron Reaugh/(MA Murr) ->
>
>On 1/14 you say an Intel pci mb will hit 100% cpu usage
> transferring 1.5 mb/sec over an Isa slot.
>On 1/13 you said that no Isa slot could ever xfer more
> than 1.0 mb/sec
>On 1/11 you said that a system would be crippled after
> 900 kb/sec.
>Previous to that, you've stated that 4x cd-burning at
> 600 kb/sec would be impossible.
You're a bald faced LIAR.
Eric Gisin
reduces this to 2MB/s (not 1MB/s).
You can downgrade the bridge to PCI 2.0 by disabling Delayed Transaction in
BIOS/Chipset, which gives you better ISA but degrades PCI bus mastering.
<lun...@bit-net.com> wrote in message
news:3a61179...@news.bit-net.com...
Eric Gisin
replies. That would make you a LIAR, right?
The only person who actually measured throughput in this thread reports
1.8MB/s. Two or three reported 4X CD burning worked fine, one not. If
you dig through the archives, several more have reported success, almost no
one else reporting failure. I have seen 2MB/s reported a few times.
I've completely forget what the original post was. Reposting may help solve
the problem.
"Ron Reaugh" <Ron-R...@worldnet.att.net> wrote in message
news:w0n86.1294$4y6.1...@bgtnsc07-news.ops.worldnet.att.net...
Folkert Rienstra
"Chris Pitzel" <chris....@nospam.usask.ca> wrote in message news:3A5DEB6B...@nospam.usask.ca...
: > >time that it cripples the system, and makes it impossible to
: > >burn a 4x cd while moving the mouse- this is the claim that
: > >you and Chris have been making.
: >
: > Nope, the claim is that the overall performance is heavily impacted
: > when ISA I/O of 900KB/sec. or faster is underway.
:
: 900kb/sec? Try more like 300kb/sec. And 600kb/sec just isn't
: reliably achievable with the 1505's that I've tried.
Maybe because you are not a PC-Tech?
: Maybe I'm just doing something totally wrong, maybe the drivers that
: were shipped with Win9x or downloaded from Adaptec are defective,
: but 4X burning on those controllers is pretty much a no-go on Pentium-
: era boards.
Now you are dismissing each and every (PCI) chipset.
I have no idea about Sandra's effectiveness on deriving at the speeds
from my 430TX ISA bus but it says 9 MHz for busspeed and 5 MHz
for DMA speed.
I just installed a TRIPACE TC-1550 ISA SCSI controler card I had
laying around in my junkbox. Installed a DCHS-02U (~10 MB/s) on it
and benchmarked it with Adaptecs SCSIBench.
It gave an STR of 2.9 MB/s at 100% CPU (W9x System Monitor)
IIRC it was ~1 MB/s on my ASUS 430HX based board tested with a DFHS (~7MB/s).
:
: > > BUT, (pio mode ide) was the normal state of affairs on ide
: > >pc's, until about 3 or 4 years ago. And nobody's system was
: > >crippled by this.
: >
: > Wrong.
:
: Exactly. 3-4 years ago, people didn't run CD burners, and those who did
: had busmastering solutions available to them. And anyone who ran SCSI
: HDD's 3-4 years ago had PCI equipment, or used 1505 (or similar) in
: 486-type systems.
:
: > > Mouses could moved while disk activity was taking place. Ide cd-roms
: > > could be read at 8x even, and not bring all activity to a halt.
: >
: > Wrong.
:
: I agree with you totally. Start any sort of operation involving a 1505
: on such board, and the whole thing becomes totally dead.
: Mouse movements are jerky, and moving the mouse significantly affected
: data rates.
I'm not completely sure what system type you are describing but on the TX mouse movement is direct. Screen(re)draws are very slow
though.
Still, a second harddrive still manages ~4MB/s on the SYMBIOS PCI controller at the same time.
:
:
: > > (Pio mode scsi) used only half the cpu time of (pio mode ide)
Ron Reaugh
Folkert Rienstra wrote in message ...
>
>"Chris Pitzel" <chris....@nospam.usask.ca> wrote in message
news:3A5DEB6B...@nospam.usask.ca...
>: > >time that it cripples the system, and makes it impossible to
>: > >burn a 4x cd while moving the mouse- this is the claim that
>: > >you and Chris have been making.
>: >
>: > Nope, the claim is that the overall performance is heavily impacted
>: > when ISA I/O of 900KB/sec. or faster is underway.
>:
>: 900kb/sec? Try more like 300kb/sec. And 600kb/sec just isn't
>: reliably achievable with the 1505's that I've tried.
>
>Maybe because you are not a PC-Tech?
Clueless.
>: Maybe I'm just doing something totally wrong, maybe the drivers that
>: were shipped with Win9x or downloaded from Adaptec are defective,
>: but 4X burning on those controllers is pretty much a no-go on Pentium-
>: era boards.
>
>Now you are dismissing each and every (PCI) chipset.
No, just dismissing PIO mode SCSI on any current platform.
>I have no idea about Sandra's effectiveness
That pretty much outlines your competence.
> on deriving at the speeds
>from my 430TX ISA bus but it says 9 MHz for busspeed and 5 MHz
>for DMA speed.
Not on a 1505 it doesn't.
>I just installed a TRIPACE TC-1550 ISA SCSI controler card I had
>laying around in my junkbox. Installed a DCHS-02U (~10 MB/s) on it
>and benchmarked it with Adaptecs SCSIBench.
>It gave an STR of 2.9 MB/s at 100% CPU (W9x System Monitor)
>
>IIRC it was ~1 MB/s on my ASUS 430HX based board tested with a DFHS
(~7MB/s).
Again at 100% CPU so you test clearly demonstrates that ISA SCSI is
completely useless in current multitasking OSs.:
>: > > BUT, (pio mode ide) was the normal state of affairs on ide
>: > >pc's, until about 3 or 4 years ago. And nobody's system was
>: > >crippled by this.
>: >
>: > Wrong.
>:
>: Exactly. 3-4 years ago, people didn't run CD burners, and those who did
>: had busmastering solutions available to them. And anyone who ran SCSI
>: HDD's 3-4 years ago had PCI equipment, or used 1505 (or similar) in
>: 486-type systems.
>:
>: > > Mouses could moved while disk activity was taking place. Ide cd-roms
>: > > could be read at 8x even, and not bring all activity to a halt.
>: >
>: > Wrong.
>:
>: I agree with you totally. Start any sort of operation involving a 1505
>: on such board, and the whole thing becomes totally dead.
>
>: Mouse movements are jerky, and moving the mouse significantly affected
>: data rates.
>
>I'm not completely sure what system type you are describing but on the TX
mouse movement is direct. Screen(re)draws are very slow
>though.
>Still, a second harddrive still manages ~4MB/s on the SYMBIOS PCI
controller at the same time.
Just NO. That's DMA mode and doesn't use the CPU. Try and run something
that needs the CPU when using a 100% CPU usage PIO ISA SCSI card
actively....you can't because it's frozen.
Folkert Rienstra
"Ron Reaugh" <Ron-R...@worldnet.att.net> wrote in message news:2Gx96.804$cN.5...@bgtnsc07-news.ops.worldnet.att.net...
:
: Folkert Rienstra wrote in message ...
: >
: >"Chris Pitzel" <chris....@nospam.usask.ca> wrote in message
: news:3A5DEB6B...@nospam.usask.ca...
: >: > >time that it cripples the system, and makes it impossible to
: >: > >burn a 4x cd while moving the mouse- this is the claim that
: >: > >you and Chris have been making.
: >: >
: >: > Nope, the claim is that the overall performance is heavily impacted
: >: > when ISA I/O of 900KB/sec. or faster is underway.
: >:
: >: 900kb/sec? Try more like 300kb/sec. And 600kb/sec just isn't
: >: reliably achievable with the 1505's that I've tried.
: >
: >Maybe because you are not a PC-Tech?
:
:
: Clueless.
Huh, why are you answering for Chris?
:
: >: Maybe I'm just doing something totally wrong, maybe the drivers that
: >: were shipped with Win9x or downloaded from Adaptec are defective,
: >: but 4X burning on those controllers is pretty much a no-go on
: >: Pentium-era boards.
: >
: >Now you are dismissing each and every (PCI) chipset.
:
:
: No, just dismissing PIO mode SCSI on any current platform.
You are answering for Chris again. Sleep deprivation perhaps?
Or afraid Chris might have an decent answer and you wanting to preempt that?
And yes, current platform is another description for ISA going through
the PCI to ISA bridge aka the chipset.
What you made it out to be is, PIO SCSI is bad, on every platform: Period.
:
: >I have no idea about Sandra's effectiveness
:
:
: That pretty much outlines your competence.
So what is YOUR competence in matters SANDRA?
:
: > on deriving at the speeds from my 430TX ISA bus but it says 9 MHz
: > for busspeed and 5 MHz for DMA speed.
I will reinstall my ASUS HX board sometime soon and see if it gives
different values for that one. Perhaps it is just telling the ISA specifica-
tion and not measuring anything.
:
:
: Not on a 1505 it doesn't.
Huh?
This is ISA busspeed, nothing to do with the 1505.
So much for YOUR competence.
:
: >I just installed a TRIPACE TC-1550 ISA SCSI controler card I had
: >laying around in my junkbox. Installed a DCHS-02U (~10 MB/s) on it
: >and benchmarked it with Adaptecs SCSIBench.
: >It gave an STR of 2.9 MB/s at 100% CPU (W9x System Monitor)
: >
: >IIRC it was ~1 MB/s on my ASUS 430HX based board tested with a
: >DFHS (~7MB/s).
:
:
: Again at 100% CPU so you test clearly demonstrates that ISA SCSI is
: completely useless in current multitasking OSs.
:
Nope, see below.
Meanwhile, a simple Windows drag and drop filecopy between partitions
on the same drive on a PCI controller will net you 100% Kernal cpu time
in System Monitor. No one will assert that Windows is useless because of
that. That's just how Windows works. What matters is how much of that
100% is actual working code instead of a NOOP loop.
:
: >: > > BUT, (pio mode ide) was the normal state of affairs on ide
: >: > >pc's, until about 3 or 4 years ago. And nobody's system was
: >: > >crippled by this.
: >: >
: >: > Wrong.
: >:
: >: Exactly. 3-4 years ago, people didn't run CD burners, and those who
: >: did had busmastering solutions available to them. And anyone who ran
: >: scsi HDD's 3-4 years ago had PCI equipment, or used 1505 (or similar)
: >: in 486-type systems.
: >:
: >: > > Mouses could moved while disk activity was taking place. Ide cd-
: >: > > roms could be read at 8x even, and not bring all activity to a halt.
: >: >
: >: > Wrong.
: >:
: >: I agree with you totally. Start any sort of operation involving a 1505
: >: on such board, and the whole thing becomes totally dead.
: >
: >: Mouse movements are jerky, and moving the mouse significantly af-
: >: fected data rates.
: >
: >I'm not completely sure what system type you are describing but on the
: >TX mouse movement is direct. Screen(re)draws are very slow though.
: >Still, a second harddrive still manages ~4MB/s on the SYMBIOS PCI
: >controller at the same time.
:
:
: Just NO. That's DMA mode and doesn't use the CPU.
Huh? Of course it does, it just does not use much of it. In the SYMBIOS
case just some 10-15% . So the TRIPACE card got 85% and slowed down
a little while the SYMBIOS was active.
This is a question of scheduling time between the processes and since
drivers have a high priority it's the desktop and other applications that suffers most, not the mouse.
: Try and run something
: that needs the CPU when using a 100% CPU usage PIO ISA SCSI card
: actively....you can't because it's frozen.
Nope, not frozen, just not much CPU time allowed for it.
It just crawls on the small amount of cpu time alloted to it.
:
: >: > > (Pio mode scsi) used only half the cpu time of (pio mode ide)
:
:
:
Eric Gisin
>
> Folkert Rienstra wrote in message ...
>
Nice response from a moRon.
> That pretty much outlines your competence.
>
Mo-Ron-ic!
> > on deriving at the speeds
> >from my 430TX ISA bus but it says 9 MHz for busspeed and 5 MHz
> >for DMA speed.
>
Sandra is just picking up the ISA clock from the chipset, which equals the
theoritical rate of ISA.
> >I just installed a TRIPACE TC-1550 ISA SCSI controler card I had
> >laying around in my junkbox. Installed a DCHS-02U (~10 MB/s) on it
> >and benchmarked it with Adaptecs SCSIBench.
> >It gave an STR of 2.9 MB/s at 100% CPU (W9x System Monitor)
> >
> >IIRC it was ~1 MB/s on my ASUS 430HX based board tested with a DFHS
> (~7MB/s).
>
> Again at 100% CPU so you test clearly demonstrates that ISA SCSI is
> completely useless in current multitasking OSs.:
>
Ron, you are such a maRoon.
Did you know millions of people run printers on NT using PIO LPT, and CPU is
close to 100%. The system slows down, it is NOT unusable.
Tim Shoppa
>
> Alain-Pierre CHERTIER wrote in message <93cef8$eq9$1...@nnrp1.deja.com>...
> >Hello !
> >
> >AVA-1505 is a simple but very good SCSI-2 controller :-)
>
> No, it's ISA.
No, it's not even a SCSI-2 controller. It's a *host adapter*.
(As are all the other ISA and PCI boards sold by Adaptec.)
Tim.