Hard drive and CD drive on same IDE channel?
Nosmo King
Hello all,
A computer magazine writes that it is not advisable to mix drives of
very different performance on the same IDE channel.
It says that for example, a CD ROM drive will slow down a hard disk
drive (HDD) if placed on the same channel. Is this true?
I thought that it would result in better performance to have the HDD
holding the operating system (C:)on one IDE channel, as master, and
another HDD holding the paging file/swap file, on the second IDE
channel, as master also. I understood that the two HDDs would be able to
'talk' to each other more efficiently when connected this way.
The other IDE accessories, such as CD ROM drive, CDRW drive, would of
necessity, have to be connected as slaves, on one or another of the IDE
channels.
Would it really be better to have the two HDDs configured as master and
slave on the same IDE channel?
Another question which occurs to me is:- Seeing that both the HDD and
CD/CDRW drives would only be working simultaneously when copying files
from one to the other, and that the CD/CDRW drives would be idle for the
majority of the time that the computer is on, would the idle CD/CDRW
drive still slow down the HDD merely by its presence?
Your views would be much appreciated.
N. King.
Mikhail Zhilin
Nosmo, that was in DOS/Win3.1 and the earlier Win95. Even before Win95a
there was an MS patch (update) that eliminated this problem, and since
then you can connect HDD and CD-ROM to the same IDE channel without the
performance degradation.
In theory, if Windows and the swap file are on the HDDs that are on the
different IDE channels -- the performance should be increased (not very
noticeably though), but if DMA is off -- CPU will not be able to serve
the both drives on the highest speed anyway.
As for the different types of CD-ROMs -- usually it is not important on
what of the channel they are, and are they master or slave, -- but
sometimes manufacturer specifies the way how it better to connect it.
--
Mikhail Zhilin (MVP MPS-D)
http://www.aha.ru/~mwz
Sorry, no technical support by e-mail.
Please reply to the newsgroups only.
======
Greg Thiedecke
hard drives.
What you will have issues with in your config is copying from the cd to the
harddrive on the same channel. because only one ide device can talk at a
time...
"Mikhail Zhilin" <m...@aha.ru_NOSPAM> wrote in message
news:7g9aps81jb2l6ec5j...@4ax.com...
Mikhail Zhilin
<soft...@ozemail.com.au> wrote:
>Well actually for best performance split the win2k swap file across both
>hard drives.
Greg, for W2K -- I don't argue. But not applicable for Win9x (the
original message was crossposted to the both OSes groups).
>What you will have issues with in your config is copying from the cd to the
>harddrive on the same channel. because only one ide device can talk at a
>time...
No problems here: Windows uses memory as the temporary buffer. But this
will slow down the system a little, so I prefer to have at the second
channel at least CD-RW.
--
Mikhail Zhilin (MVP MPS-D)
http://www.aha.ru/~mwz
Sorry, no technical support by e-mail.
Please reply to the newsgroups only.
======
>"Mikhail Zhilin" <m...@aha.ru_NOSPAM> wrote in message
>news:7g9aps81jb2l6ec5j...@4ax.com...
>> Nosmo, that was in DOS/Win3.1 and the earlier Win95. Even before Win95a
>> there was an MS patch (update) that eliminated this problem, and since
>> then you can connect HDD and CD-ROM to the same IDE channel without the
>> performance degradation.
>>
>> In theory, if Windows and the swap file are on the HDDs that are on the
>> different IDE channels -- the performance should be increased (not very
>> noticeably though), but if DMA is off -- CPU will not be able to serve
>> the both drives on the highest speed anyway.
>>
>> As for the different types of CD-ROMs -- usually it is not important on
>> what of the channel they are, and are they master or slave, -- but
>> sometimes manufacturer specifies the way how it better to connect it.
<.....>
Greg Thiedecke
two totally different OSes...
FWIW I used to have that configuration but noticed that you never really get
the top speed out of ide cdroms that way, drive transfers a bit of data,
switches to writing it to HDD, cd starts to spin down, spins up again. Much
faster if they are on different controllers.
And given the price of a ultra66 card - I now have hard disc on master 1 ,
ide cd on master 2, cdburner on scsi, and 2x20g's on ultra66 (admittedly I
hacked it to fastrak66) mainly because while nt supports software raid, 9x
doesn't
"Mikhail Zhilin" <m...@aha.ru_NOSPAM> wrote in message
news:81dapssdb8ltkh3nk...@4ax.com...
Mikhail Zhilin
>Well you have to wonder why someone would crosspost a request for help in
>two totally different OSes...
I usually don't trim the crossposting when there are no more than 4
groups, so I simply hadn't mentioned that first :)
>FWIW I used to have that configuration but noticed that you never really get
>the top speed out of ide cdroms that way, drive transfers a bit of data,
>switches to writing it to HDD, cd starts to spin down, spins up again. Much
>faster if they are on different controllers.
Yes, that makes sense.
>And given the price of a ultra66 card - I now have hard disc on master 1 ,
>ide cd on master 2, cdburner on scsi, and 2x20g's on ultra66 (admittedly I
>hacked it to fastrak66) mainly because while nt supports software raid, 9x
>doesn't
BTW, I moved my swap file to the drive that is on the UDMA-66
controller. Seems to work faster.
--
Mikhail Zhilin (MVP MPS-D)
http://www.aha.ru/~mwz
Sorry, no technical support by e-mail.
Please reply to the newsgroups only.
======
>"Mikhail Zhilin" <m...@aha.ru_NOSPAM> wrote in message
>news:81dapssdb8ltkh3nk...@4ax.com...
>> On Sat, 12 Aug 2000 20:47:07 +1000, "Greg Thiedecke"
>> <soft...@ozemail.com.au> wrote:
>>
>> >Well actually for best performance split the win2k swap file across both
>> >hard drives.
>>
>> Greg, for W2K -- I don't argue. But not applicable for Win9x (the
>> original message was crossposted to the both OSes groups).
>>
>> >What you will have issues with in your config is copying from the cd to
>the
>> >harddrive on the same channel. because only one ide device can talk at a
>> >time...
>>
>> No problems here: Windows uses memory as the temporary buffer. But this
>> will slow down the system a little, so I prefer to have at the second
>> channel at least CD-RW.
<......>
Royston Tin
says...
> Subject: Re: Hard drive and CD drive on same IDE channel?
> From: "Greg Thiedecke" <soft...@ozemail.com.au>
> Newsgroups: microsoft.public.win2000.general, microsoft.public.win2000.hardware, microsoft.public.win98.disks.general, microsoft.public.win98.gen_discussion
>
> Date: Sat, 12 Aug 2000 21:46:21 +1000
> Well you have to wonder why someone would crosspost a request for help in
> two totally different OSes...
>
>
>
response.
I cross posted the request for help in two different OSes because I run
a dual boot configuration of WIN98 and WIN2K, on the same computer,
where the IDE channels have one CD ROM and an HDD on one and a CD/RW and
an HDD on the other.
By the way, no one has replied whether the HDD will still slow down when
the CD drives are not working.
N. King
Mikhail Zhilin
<roy...@tin7.fsnet.co.uk> wrote:
<snip>
>By the way, no one has replied whether the HDD will still slow down when
>the CD drives are not working.
>
>N. King
Sorry, I thought I answered that in my first reply, in a half an hour
after the question...
If in Win95b or higher (or in Win-95/95a with the patch installed) --
no, it will not.
Wayne
as the OS the mode for the 2 devices has to be the same and it used to be
that CD Rom drives were behind disk drive as far as modes and UDMA etc..
Only one mode is allowed per channel
Wayne
"Royston Tin" <roy...@tin7.fsnet.co.uk> wrote in message
news:MPG.1402551f...@msnews.microsoft.com...
> In article <OXcgfLFBAHA.242@cppssbbsa05>, soft...@ozemail.com.au
> says...
> > Subject: Re: Hard drive and CD drive on same IDE channel?
> > From: "Greg Thiedecke" <soft...@ozemail.com.au>
> > Newsgroups: microsoft.public.win2000.general,
microsoft.public.win2000.hardware, microsoft.public.win98.disks.general,
microsoft.public.win98.gen_discussion
> >
> > Date: Sat, 12 Aug 2000 21:46:21 +1000
> > Well you have to wonder why someone would crosspost a request for help
in
> > two totally different OSes...
> >
> >
> >
> Greg and other respondents, Hello, and thank you for your informative
> response.
>
> I cross posted the request for help in two different OSes because I run
> a dual boot configuration of WIN98 and WIN2K, on the same computer,
> where the IDE channels have one CD ROM and an HDD on one and a CD/RW and
> an HDD on the other.
>
Mikhail Zhilin
only computer, though mostly with the motherboards P2B and P3B-F) -- I
could set DMA independently, in any combination. Diagnostic programs
(HdTach and several CD-ROM diagnostic programs) confirmed that, too, and
showed the same performance for the HDD whether CD-ROM was connected or
not.
--
Mikhail Zhilin (MVP MPS-D)
http://www.aha.ru/~mwz
Sorry, no technical support by e-mail.
Please reply to the newsgroups only.
======
wrote:
Lee Chapelle
I do believe however (fwiw) that UltraDMA (ATA/66/100) drives will default
to ATA33 if any non-UDMA drive is connected to any channel.
--
..................................................................
Lee Chapelle MVP
Microsoft Platform Support - Desktop
http://www.webdev.net/orca/
http://members.home.com/dts-l/
http://support.microsoft.com/support/
............................................................................
o /' )
/' ( ,
__/' ) .' `;
o _.-~~~~' ``---..__ .' ;
_.--' b) ``--...____.' .'
( _. )). `-._ <
`\|\|\|\|)-.....___.- `-. __...--'-.'.
`---......____...---`.___.'----... .' `.;
`-`
`
"Wayne" <power...@mailcity.com> wrote in message
news:uR$ZCfKCAHA.255@cppssbbsa05...
Ron Sommer
"Lee Chapelle" <l...@NSmvps.org> wrote in message news:elJWSqWCAHA.196@cppssbbsa05...
Walter Clayton
I guess my UDMA-2 and PIO-4 CD-Rom drives don't know that. :-)
The connector is the standard 40 pin connector. You can use the 80 wire
cables on devices slower than UDMA-4 and gain a 5-10% performance increase
due to the improved noise shielding of the 80 wire cables..
--
Walter Clayton - MS MVP(DTS)
Any technology distinguishable from magic is insufficiently advanced.
http://www.members.home.com/dts-l
Want to know where each version of a MS module came from?
Try http://support.microsoft.com/servicedesks/fileversion/default.asp
"Ron Sommer" <rso...@ktis.net> wrote in message
news:uK6r4$YCAHA.251@cppssbbsa05...
Ron Sommer
Thanks for the info. :-)
"Walter Clayton" <w-cla...@SPmailandnewsAM.com> wrote in message news:eq#IGZZCAHA.245@cppssbbsa04...
Richard G. Harper
A "true" UDMA66 cable has a blocked hole that prevents it from plugging into
a standard IDE or DMA33 port. It also has one of the ribbon cables "cut"
near the motherboard plug so it looks like there's a hole in it.
--
Richard G. Harper (MVP MPS-D) rgha...@email.com
* NOTE - Private EMail is generally not replied to. Please post all
* questions and comments in the newsgroup so all can benefit.
* Help US help YOU - http://members.home.com/dts-l/goodpost.htm
Erio Michelangelo Carangella
U-DMA devices starts from old 16 MB/s... also called Ultra ATA(1)... U-ATA2
is U-DMA2 is for 33 MB/s devices (note that the number there is a MULTIPLIER
for the original, basic, 16 MB/s speed, so U-DMA4 later is 4x16), and for
these first 2 implementations the genral rule that the slower device on the
same IDE channel will slow down the faster too it's sadly true.
But with U-ATA66 (U-DMA4) , they changed things in a finally intelligent
way, so U-DMA66 compliant disk drives (only hard disks for now as far as I
know) use an 80 pin cable instead of the standard 40 ones, and this, even if
can be used for other tyoe of U-DMA devices, allows the faster disks to run
at full speed because they use special signals, so they are not slowed down
by slower ones.
You can check it witj utilities like the one inlcuded in the latest Intel's
U-ATA (called Companion) drivers for their chipsets.
Cheer up!!!
Ciao,
E=MC^2 (Erio "U2" Michelangelo C.)
"Lee Chapelle" <l...@NSmvps.org> ha scritto nel messaggio
Erio Michelangelo Carangella
connector, one for an U-DMA66 device and one for a "generic" U-ATA device.
Regards,
E=MC^2 (Erio "U2" Michelangelo C.)
"Richard G. Harper" <rgha...@email.com> ha scritto nel messaggio
news:uHgzw2dCAHA.243@cppssbbsa04...
>
> A "true" UDMA66 cable has a blocked hole that prevents it from plugging
into
> a standard IDE or DMA33 port. It also has one of the ribbon cables "cut"
> near the motherboard plug so it looks like there's a hole in it.
>
> --
> Richard G. Harper (MVP MPS-D) rgha...@email.com
> * NOTE - Private EMail is generally not replied to. Please post all
> * questions and comments in the newsgroup so all can benefit.
> * Help US help YOU - http://members.home.com/dts-l/goodpost.htm
>
>
>
> "Ron Sommer" <rso...@ktis.net> wrote in message
> news:uK6r4$YCAHA.251@cppssbbsa05...
> It is my understanding that UltraDMA uses a blue cable with more pins
that
> will not connect to a ATA33 drive.
>
> "Lee Chapelle" <l...@NSmvps.org> wrote in message
Richard G. Harper
Not according to my reference materials. Yours might, but it's not a
standard cable.
--
Richard G. Harper (MVP MPS-D) rgha...@email.com
* NOTE - Private EMail is generally not replied to. Please post all
* questions and comments in the newsgroup so all can benefit.
* Help US help YOU - http://members.home.com/dts-l/goodpost.htm
"Erio Michelangelo Carangella" <emc...@hotmail.com> wrote in message
news:8nm904$6ki$1...@news.flashnet.it...
Erio Michelangelo Carangella
just buy *ANY* motherboards out there based on a U-DMA66 capable chipset and
you'll see... the standard IDE cables (even U-ATA66) are normally intended
to be used with ONE hard drive and ONE CD-ROM drive. It doesn't really
matter what is the 2nd device in the serial chain, nevertheless they assume
one could attach a CD-ROM drive at leasty in their system... yes, you could
use a second IDE cable, but not many mainbard makers put 2 IDE cables in
their boxes.
Anyway, I've been building U-DMA66 capable PCs since last March now (some 20
systems up to now), and they all are like that. It's pretty enough standard
to my understanding, then... no?
Chuss,
E=MC^2 (Erio "U2" Michelangelo C.)
"Richard G. Harper" <rgha...@email.com> ha scritto nel messaggio
news:eNG79hfCAHA.254@cppssbbsa05...
Richard G. Harper
thing. And yes, I have bought several UDMA66 motherboards recently and no,
I have never seen what you describe, nor read what you claim is "assumed" in
any documentation.
But no matter, thanks for your input anyway.
--
Richard G. Harper (MVP MPS-D) rgha...@email.com
* NOTE - Private EMail is generally not replied to. Please post all
* questions and comments in the newsgroup so all can benefit.
* Help US help YOU - http://members.home.com/dts-l/goodpost.htm
"Erio Michelangelo Carangella" <emc...@hotmail.com> wrote in message
news:8nme4v$8e6$1...@news.flashnet.it...
Erio Michelangelo Carangella
I hope this excerpt (even if not so well written) from the manual found in
Asus's latest 815E based motherboard called CUSL2 will set your mind at
ease... if not, I just hope that other "gurus" like you will convince you
about that...
10) Primary (Blue) / Secondary IDE Connectors (two 40-1 pin IDE)
These connectors (on the motherboards, that is) support the provided IDE
hard disk ribbon cable. Connect the cable's blue connector to the
motherboard's primary (reccomended) or secondary IDE connector, and then
connect the gray connector to your U-DMA/100 slave device (hard disk drive)
and black connector to your U-DMA/100 master device. It is reccomended that
non-U-DMA/100 devices be connected to the secondary IDE connector. If you
install 2 hard disks, you must configure the second drive to Slave mode by
setting its jumper accordingly. (Pin 20 is removed to prevent inserting in
the wrong orientation when using riboon cables with pin 20 plugged).
Note (by me): the IDE cable supplied with this motherboard has 3 40-pin
connectors on it: one blue, one grey and one black. The wiring between the
grey and the blue has 40+40 wires for U-DMA/66-100 signals... making a total
of 80 wires.
It's true that you need 80 cables (and a special connector in the mainboard)
to send signals to/from U-DMA/66-100 devices, but they are integrated in
standard 40-pin cables, which can as well connect standard IDE drives.
And... I didn't mean to be sarchastic in my original reply... sorry if you
got it personally.
E=MC^2 (Erio "U2" Michelangelo C.)
"Richard G. Harper" <rgha...@email.com> ha scritto nel messaggio
news:#OHpNlhCAHA.252@cppssbbsa05...
Walter Clayton
assumptions.
You also contradicted yourself. Up stream you stated
"The U-DMA66 cables have 3 connectors, one for the motherboard interface
connector, one for an U-DMA66 device and one for a "generic" U-ATA device."
then a bit further downstream you stated
"the standard IDE cables (even U-ATA66) are normally intended
to be used with ONE hard drive and ONE CD-ROM drive. It doesn't really
matter what is the 2nd device in the serial chain, nevertheless they assume
one could attach a CD-ROM drive at leasty in their system"
and then you quoted the Asus manual which states
"It is reccomended that non-U-DMA/100 devices be connected to the secondary
IDE connector".
I haven't seen a CD-Rom, CD-R or CD-RW (yet) that operates at anything
faster than PIO-4, which is a 16Mbs protocol.
But we'll go a bit further in the same Asus quote it states:
"If you install 2 hard disks, you must configure the second drive to Slave
mode by setting its jumper accordingly."
Based on the fact that the IDE specs (regardless of protocol) require that
one of the devices on the channel must be the master then the previous
statement must mean that it is OK to hookup 2 HDs up to the same channel.
But notice that nowhere in the manual information quoted, nor anywhere on
the entire document is there any mention of the secondary device having to
be non ATA-100. Is this an oversight on the part of Asus?
Care to explain this?
Since you've only been building systems since March of this year, your
confusion is understandable.
First your assumption that MB vendors assume that the user will be
installing only one HD and one CD are false. The MB vendor makes no such
assumption. At most the only assumption they make is that the customer will
_possibly_ be installing at least one IDE device. Of course if the user
elects to go with SCSI devices, then the IDE controllers are completely
superfluous as is the cable. The included cable is strictly a courtesy.
For a bit further education refer to the following:
http://developer.intel.com/design/chipsets/815e/
ftp://download.intel.com/design/chipsets/datashts/29068701.pdf (WARNING:
This is a 5M download of the chipset specification)
Notice that nowhere in any of these documents is there mention of anything
to validate your presumptions. Quite the opposite in fact.
Now I will explain why most vendors will recommend not mixing devices of
different speeds on the same IDE channel. Part of it is habit since the
original IDE controllers would drop the channel speed to that of the slowest
device. With EIDE however, that constraint no longer exists. The problem
however is with interleaved IO operations. If you have a ATA-100 device on
the same channel as a PIO-4 device, both doing bulk data transfer, then the
faster device will have to wait for the slower device to give up the channel
before it can process another IO request. This will effectively throttle the
faster device, but the device will run full speed when it has control of the
channel.
--
Walter Clayton - MS MVP(DTS)
Any technology distinguishable from magic is insufficiently advanced.
http://www.members.home.com/dts-l
Want to know where each version of a MS module came from?
Try http://support.microsoft.com/servicedesks/fileversion/default.asp
"Erio Michelangelo Carangella" <emc...@hotmail.com> wrote in message
news:8nnaos$h1d$1...@news.flashnet.it...
Nick Goetz
pin. The regular IDE cables that ship with the boards I use all have
that pin hole blocked. I always took it to be an orientation guide.
The UDMA66 cables (with the hole blocked) will fit on them.
Nick Goetz
Richard G. Harper
--
Richard G. Harper (MVP MPS-D) rgha...@email.com
* NOTE - Private EMail is generally not replied to. Please post all
* questions and comments in the newsgroup so all can benefit.
* Help US help YOU - http://members.home.com/dts-l/goodpost.htm
"Nick Goetz" <ngo...@niagara.com> wrote in message
news:399FCCB0...@niagara.com...
Erio Michelangelo Carangella
Thanks for your 2 cents that confirmed what I was trying to explain.
It's good to see some opinions we have are not subjective but objective
things.
Ciao,
E=MC^2 (Erio "U2 ?!" Michelangelo C.)
"Nick Goetz" <ngo...@niagara.com> ha scritto nel messaggio
Walter Clayton
I just checked two cables, both of which came with different IDE devices.
One is an old 40 wire that I'm not sure where came from. None of the
connector pins are blocked.
I have an 80 wire cable that came with my latest UDMA-4 drive purchase.
Since I already had an UDMA-4 drive cabled up, there was no need to use this
cable. None of the pins on it are blocked.
I starting to suspect it depends on who makes the cable as to whether the
middle pin opposite the key is blocked or not to force installation on the
MB instead of the HD.
--
Walter Clayton - MS MVP(DTS)
Any technology distinguishable from magic is insufficiently advanced.
http://www.members.home.com/dts-l
Want to know where each version of a MS module came from?
Try http://support.microsoft.com/servicedesks/fileversion/default.asp
"Nick Goetz" <ngo...@niagara.com> wrote in message
Nick Goetz
Erio, I agree. It could be important for anywone adding drives,
separate IDE cards, buying new cables etc. to know what fits where.
Nick Goetz
Erio Michelangelo Carangella wrote:
>
> I think we helped giving more words about this, isn't it Nick?
>
> Thanks for your 2 cents that confirmed what I was trying to explain.
>
> It's good to see some opinions we have are not subjective but objective
> things.
>
> Ciao,
>
> E=MC^2 (Erio "U2 ?!" Michelangelo C.)
>
> "Nick Goetz" <ngo...@niagara.com> ha scritto nel messaggio
Nick Goetz
Walter, that pin is usually missing on the hard drive as well.
Nick Goetz
Walter Clayton wrote:
>
> I just checked two cables, both of which came with different IDE devices.
> One is an old 40 wire that I'm not sure where came from. None of the
> connector pins are blocked.
> I have an 80 wire cable that came with my latest UDMA-4 drive purchase.
> Since I already had an UDMA-4 drive cabled up, there was no need to use this
> cable. None of the pins on it are blocked.
>
> I starting to suspect it depends on who makes the cable as to whether the
> middle pin opposite the key is blocked or not to force installation on the
> MB instead of the HD.
>
> --
> Walter Clayton - MS MVP(DTS)
> Any technology distinguishable from magic is insufficiently advanced.
> http://www.members.home.com/dts-l
> Want to know where each version of a MS module came from?
> Try http://support.microsoft.com/servicedesks/fileversion/default.asp
>
> "Nick Goetz" <ngo...@niagara.com> wrote in message
Walter Clayton
Curioser and curioser. I've never really looked at the connectors other than
verifying the key before inserting. So far, since this sub-topic has come
about, the only thing I haven't looked at is the male connectors on all the
drives I have (other than making sure there are no bent or recessed pins).
I'll have to pull the cable on one of my UDMA-4 drives and see if the pin is
missing or not.
I can understand the MB vendor shipping a cable with the pin blocked, but I
can not fathom why the MB vendor would ship a cable with the corresponding
pin blocked on the device end. This sort of goes contrary to the implied
purpose of insuring that only the MB end of the cable goes to the MB.
--
Walter Clayton
Microsoft MVP (DTS)
Any technology distinguishable from magic is insufficiently advanced
Want to know where each version of a MS module came from?
Try http://support.microsoft.com/servicedesks/fileversion/default.asp
"Nick Goetz" <ngo...@niagara.com> wrote in message
news:39A11E9C...@niagara.com...
Walter Clayton
it's nothing new.
--
Walter Clayton - MS MVP(DTS)
Any technology distinguishable from magic is insufficiently advanced.
http://www.members.home.com/dts-l
Want to know where each version of a MS module came from?
Try http://support.microsoft.com/servicedesks/fileversion/default.asp
"Nick Goetz" <ngo...@niagara.com> wrote in message
news:39A11E9C...@niagara.com...