Best configuration of drives?

On Tue, 11 May 2004 15:36:20 +0200, "Folkert Rienstra"
wrote:

Actually, it can be problematic when you access the opt. drive
that is on the same cable with the source drive while burning.

Why?

Because it may block the source drive for several seconds, spinning-up
the motor/disk to speed and seeking to the desired data and transfer.

Hmmm... Doesn't sound like a very intelligent way to implement an
optical drive. I would have assumed they could issue a spin-up
instruction and wait for a "ready" interrupt while they do other
things. Holding the IDE channel up while waiting for the disk to spin
up is idiotic.


--

Map Of The Vast Right Wing Conspiracy:
http://www.freewebs.com/vrwc/

"You can all go to hell, and I will go to Texas."
--David Crockett

In article ,
Bob wrote:
On Tue, 11 May 2004 15:36:20 +0200, "Folkert Rienstra"
wrote:

Actually, it can be problematic when you access the opt. drive
that is on the same cable with the source drive while burning.

Why?

Because it may block the source drive for several seconds, spinning-up
the motor/disk to speed and seeking to the desired data and transfer.

Hmmm... Doesn't sound like a very intelligent way to implement an
optical drive. I would have assumed they could issue a spin-up
instruction and wait for a "ready" interrupt while they do other
things. Holding the IDE channel up while waiting for the disk to spin
up is idiotic.



You're dealing with hardware made to be just good enough for the mass
market, and an OS that is lagging behind what the best hardware can
do. 99% of the PCs have one HD and one CD. That's the target for
"good enough".

You can buy a PCI IDE card with 2 channels and put your optical disks
to that.

--
Al Dykes
-----------
adykes at p a n i x . c o m

On 11 May 2004 11:22:14 -0400, (Al Dykes) wrote:

You can buy a PCI IDE card with 2 channels and put your optical disks
to that.

I thought modern computers came with 2 IDE channels on the mainboard.
That's 4 devices.


--

Map Of The Vast Right Wing Conspiracy:
http://www.freewebs.com/vrwc/

"You can all go to hell, and I will go to Texas."
--David Crockett

In article ,
Bob wrote:
On 11 May 2004 11:22:14 -0400, (Al Dykes) wrote:

You can buy a PCI IDE card with 2 channels and put your optical disks
to that.

I thought modern computers came with 2 IDE channels on the mainboard.
That's 4 devices.



yea, but it's not 4 channels, and if you mix a fast and a slow
device on a channel it slows down to the speed of the slower one.

A PCI IDE card is only a few bucks.

Contention between two disks on a channel might be a bottleneck for
your application, but the only way to tell for sure is to use a tool
like perfmon.exe (part of NT) to see what your system is
doing. There's no way to tell, otherwise.
--
Al Dykes
-----------
adykes at p a n i x . c o m

"Bob" wrote in message ...
On Tue, 11 May 2004 15:36:20 +0200, "Folkert Rienstra" wrote:

Actually, it can be problematic when you access the opt. drive
that is on the same cable with the source drive while burning.

Why?

Because it may block the source drive for several seconds, spinning-up
the motor/disk to speed and seeking to the desired data and transfer.

Hmmm... Doesn't sound like a very intelligent way to implement an
optical drive.

Welcome to the wonderful world of IDE.

I would have assumed they could issue a spin-up instruction
and wait for a "ready" interrupt while they do other things.

No such command, I'm afraid (AFAICT).

Holding the IDE channel up while waiting for the disk to spin up is idiotic.

Well, IDE has known overlapped commands for a while now but it is hardly
implemented by any devices and also needs driver support for it to work.

"Al Dykes" wrote in message ...
In article ,
Bob wrote:
On 11 May 2004 11:22:14 -0400, (Al Dykes) wrote:

You can buy a PCI IDE card with 2 channels and put your optical disks
to that.

I thought modern computers came with 2 IDE channels on the mainboard.
That's 4 devices.



yea, but it's not 4 channels, and if you mix a fast and a slow
device on a channel it slows down to the speed of the slower one.

Obviously not.
When the slow drive has the bus the other drive isn't doing anything, obviously.

If and how much slowdown is experienced is decided by the duty cycle of the 2
devices, the bandwidth that they actually need and at what busspeed they run.


A PCI IDE card is only a few bucks.

Contention between two disks on a channel might be a bottleneck for
your application, but the only way to tell for sure is to use a tool
like perfmon.exe (part of NT) to see what your system is doing.

And what will that tell you?

There's no way to tell, otherwise.

On Tue, 11 May 2004 22:42:01 +0200, "Folkert Rienstra"
wrote:

Hmmm... Doesn't sound like a very intelligent way to implement an
optical drive.

Welcome to the wonderful world of IDE.

I would have assumed they could issue a spin-up instruction
and wait for a "ready" interrupt while they do other things.

No such command, I'm afraid (AFAICT).

Holding the IDE channel up while waiting for the disk to spin up is idiotic.

Well, IDE has known overlapped commands for a while now but it is hardly
implemented by any devices and also needs driver support for it to work.

Even Microsoft did a better job of implementing the floppy drive than
that.

What you have just said explains why optical disk makers insist you
use the master channel and nothing on the slave - they take over the
entire channel. Dumb and dumber.


--

Map Of The Vast Right Wing Conspiracy:
http://www.freewebs.com/vrwc/

"You can all go to hell, and I will go to Texas."
--David Crockett

In article ,
Folkert Rienstra wrote:

"Al Dykes" wrote in message ...
In article ,
Bob wrote:
On 11 May 2004 11:22:14 -0400, (Al Dykes) wrote:

You can buy a PCI IDE card with 2 channels and put your optical disks
to that.

I thought modern computers came with 2 IDE channels on the mainboard.
That's 4 devices.



yea, but it's not 4 channels, and if you mix a fast and a slow
device on a channel it slows down to the speed of the slower one.

Obviously not.
When the slow drive has the bus the other drive isn't doing anything, obviously.

If you put a DMA and a non-DMA capable device on an interface the drivers
dumbs down to the lower one, right ? That's what I'm refering to.


If and how much slowdown is experienced is decided by the duty cycle of the 2
devices, the bandwidth that they actually need and at what busspeed they run.


An optimal system/application configuration wuld spread disk activity
across both disks. One way to do this is to set them uup as RAID0
(striping).


A PCI IDE card is only a few bucks.

Contention between two disks on a channel might be a bottleneck for
your application, but the only way to tell for sure is to use a tool
like perfmon.exe (part of NT) to see what your system is doing.

And what will that tell you?


Perfom will allow you to determine where the bottleneck is in your
system when you are running your application and wishing it would go
faster. It may be that you need a faster CPU, more memory, or
faster disk I/O. If you have more than one disk it will show you
which disk is working hardest.




--
Al Dykes
-----------
adykes at p a n i x . c o m

"Folkert Rienstra" wrote in message
...

"Bob" wrote in message
...
On Mon, 10 May 2004 20:59:15 +0200, "Folkert Rienstra"
wrote:

Actually, it can be problematic when you access the opt. drive
that is on the same cable with the source drive while burning.

Why?

Because it may block the source drive for several seconds,

That's not relevant if the burner hasn't spun up yet. What makes you think
anything is blocked during spin-up?

spinning-up
the motor/disk to speed and seeking to the desired data and transfer.

Only after that does the HD have to UDMA burst data to the host's buffers to
maintain an unbroken stream of maybe a paltry 4-8 MB/sec. Nevermind
burn-proof technology.

"Bob" wrote in message
...
On Tue, 11 May 2004 15:36:20 +0200, "Folkert Rienstra"
wrote:

Actually, it can be problematic when you access the opt. drive
that is on the same cable with the source drive while burning.

Why?

Because it may block the source drive for several seconds, spinning-up
the motor/disk to speed and seeking to the desired data and transfer.

Hmmm... Doesn't sound like a very intelligent way to implement an
optical drive. I would have assumed they could issue a spin-up
instruction and wait for a "ready" interrupt while they do other
things. Holding the IDE channel up while waiting for the disk to spin
up is idiotic.

And therefore likely isn't happening.