D
David Vanderschel
I am unsure about whether or not DMA is working on my
Windows 98 system. I have a 7200RPM Maxtor drive
(Ultra ATA/133), running off of an Iwill KD266
motherboard with a 1.2G Athlon using ALi chip set
(only does ATA/100). ALi has an IDE Cache Utility
which shows the current mode, and it shows "PIO" for
my hard drive, CD-RW, and DVD-ROM. The settings pages
for these devices in the Device Manager no longer
offer a DMA check box. (The DMA boxes used to be
there, but I have not seen them in about a year. I do
not remember what I might have done to cause them to
disappear, but I suspect a motherboard change from a
motherboard using a different (VIA?) chip set.)
The implication would seem to be that I am running in
PIO mode. The ALi utility offers me a choice of "DMA"
for the CD and DVD drives, but it does not 'take'.
For the hard drive, the utility only offers "Auto" and
"PIO". It is set on "Auto". There is even an
(undocumented) "SetDMA.exe" utility from ALi which has
no apparent impact on the mode. (The display it
creates does not look right to me either.)
My BIOS does recognize all three of my IDE devices as
having DMA capability.
To further diagnose my situation, I have run HD Tach.
I get an average read burst rate around 22MB/sec
(surprisingly uniform across the whole disk) with CPU
utilization running in the 4-5% range. These are not
fantastic specs, but I am doubtful that they could be
achieved without DMA. OTOH, my burner has 32x
capability; and, when I burn CDs, I can never burn at
more than about 16x - even coming straight from the
hard drive with, say, an ISO image. It strikes me
that an otherwise unloaded 1.2G Athlon should be able
to keep up with the 32x burner. You can see that the
burner itself _can_ go much faster because it will do
so when its buffer is full. Ie., the limit is imposed
by the CPU's ability to keep the burner's buffer full.
(Copying directly from the DVD-ROM drive to the CD-RW
consistently fails altogether - lots of coasters.)
My experience in this area is very weak, so I am
hoping that someone more knowledgeable can interpret
my numbers and tell me if I am really running in PIO
mode all the time.
Assuming that I am running in DMA mode after all, why
I am getting contrary indications? I hope that this
is not my situation because then there is no hope for
getting full speed out of my CD burner.
Assuming that I am really just in PIO mode, is there
anything that can be done about it? I have done a lot
of searching at Google Groups; and, as far as I can
tell, there are plenty of others who have experienced
the disappearance of the DMA checkboxes in Device
Manager. Yet I have never found comprehensive advice
about how to correct the problem which would be
applicable in my situation. (I do recall a case in
which the 'victim', who had tried a number of things,
finally gave up and reinstalled Windows. I am very
loathe to do that as it would literally take many days
of extremely boring grunt work to reinstall and
configure all the stuff I run at one time or another.
Indeed, if I install a new OS, it will be XP. But
even then, I would still like to be able to dual boot
back to my old Windows 98 system.) I am currently
running the standard MicroSoft driver.
I am familiar with the following thread:
http://www.google.com/groups?hl=en&lr=&ie=UTF-8&oe=UTF-8&th=7de18fdf2f5364c6&rnum=8
My Windows 98 was installed from scratch - no upgrade
from 95. To my knowledge, no non-MS drivers for IDE
controller or hard drives have ever been installed. I
have not seen "no DMA checkbox" under such
circumstances. (However, I cannot swear that
something else was not installed when I was using the
motherboard with the VIA chipset, as I did not
personally do the motherboard changes. (I could have,
but each involved warranty coverage from a local
shop.) I do have the appropriate two lines in my
MSHDC.INF file, though I am not even sure that those
entries even apply to ATA UDMA devices.
I don't see anything that looks relevant in my CMOS
BIOS setup. "PnP OS" is enabled.
Thanks in advance for any advice.
David V.
PS - One more symptom: Several months ago, I made the
mistake of trying to move my Windows 98 installation
from a 6G Fujitsu drive to a 40G Fujitsu drive.
Apparently Windows 98 cannot deal with a drive larger
than 32G. Oddly, Windows would run from the 40G drive
in Safe Mode; but, in regular mode, I got all sorts of
really bizarre symptoms - culminating in the
corruption of the 40G drive (and not just the
partition from which Windows was running). When I
tried the same move from the 6G drive to the Maxtor,
which is a 30G drive, all went immediately as I
expected. The fact that Windows behaved differently
with the 40G drive in Safe Mode (where I am sure that
it uses PIO) suggests that the difference when not in
Safe Mode may have been an attempt to use DMA. But
the DMA check boxes had already disappeared well
before this 'moving' experience.
Windows 98 system. I have a 7200RPM Maxtor drive
(Ultra ATA/133), running off of an Iwill KD266
motherboard with a 1.2G Athlon using ALi chip set
(only does ATA/100). ALi has an IDE Cache Utility
which shows the current mode, and it shows "PIO" for
my hard drive, CD-RW, and DVD-ROM. The settings pages
for these devices in the Device Manager no longer
offer a DMA check box. (The DMA boxes used to be
there, but I have not seen them in about a year. I do
not remember what I might have done to cause them to
disappear, but I suspect a motherboard change from a
motherboard using a different (VIA?) chip set.)
The implication would seem to be that I am running in
PIO mode. The ALi utility offers me a choice of "DMA"
for the CD and DVD drives, but it does not 'take'.
For the hard drive, the utility only offers "Auto" and
"PIO". It is set on "Auto". There is even an
(undocumented) "SetDMA.exe" utility from ALi which has
no apparent impact on the mode. (The display it
creates does not look right to me either.)
My BIOS does recognize all three of my IDE devices as
having DMA capability.
To further diagnose my situation, I have run HD Tach.
I get an average read burst rate around 22MB/sec
(surprisingly uniform across the whole disk) with CPU
utilization running in the 4-5% range. These are not
fantastic specs, but I am doubtful that they could be
achieved without DMA. OTOH, my burner has 32x
capability; and, when I burn CDs, I can never burn at
more than about 16x - even coming straight from the
hard drive with, say, an ISO image. It strikes me
that an otherwise unloaded 1.2G Athlon should be able
to keep up with the 32x burner. You can see that the
burner itself _can_ go much faster because it will do
so when its buffer is full. Ie., the limit is imposed
by the CPU's ability to keep the burner's buffer full.
(Copying directly from the DVD-ROM drive to the CD-RW
consistently fails altogether - lots of coasters.)
My experience in this area is very weak, so I am
hoping that someone more knowledgeable can interpret
my numbers and tell me if I am really running in PIO
mode all the time.
Assuming that I am running in DMA mode after all, why
I am getting contrary indications? I hope that this
is not my situation because then there is no hope for
getting full speed out of my CD burner.
Assuming that I am really just in PIO mode, is there
anything that can be done about it? I have done a lot
of searching at Google Groups; and, as far as I can
tell, there are plenty of others who have experienced
the disappearance of the DMA checkboxes in Device
Manager. Yet I have never found comprehensive advice
about how to correct the problem which would be
applicable in my situation. (I do recall a case in
which the 'victim', who had tried a number of things,
finally gave up and reinstalled Windows. I am very
loathe to do that as it would literally take many days
of extremely boring grunt work to reinstall and
configure all the stuff I run at one time or another.
Indeed, if I install a new OS, it will be XP. But
even then, I would still like to be able to dual boot
back to my old Windows 98 system.) I am currently
running the standard MicroSoft driver.
I am familiar with the following thread:
http://www.google.com/groups?hl=en&lr=&ie=UTF-8&oe=UTF-8&th=7de18fdf2f5364c6&rnum=8
My Windows 98 was installed from scratch - no upgrade
from 95. To my knowledge, no non-MS drivers for IDE
controller or hard drives have ever been installed. I
have not seen "no DMA checkbox" under such
circumstances. (However, I cannot swear that
something else was not installed when I was using the
motherboard with the VIA chipset, as I did not
personally do the motherboard changes. (I could have,
but each involved warranty coverage from a local
shop.) I do have the appropriate two lines in my
MSHDC.INF file, though I am not even sure that those
entries even apply to ATA UDMA devices.
I don't see anything that looks relevant in my CMOS
BIOS setup. "PnP OS" is enabled.
Thanks in advance for any advice.
David V.
PS - One more symptom: Several months ago, I made the
mistake of trying to move my Windows 98 installation
from a 6G Fujitsu drive to a 40G Fujitsu drive.
Apparently Windows 98 cannot deal with a drive larger
than 32G. Oddly, Windows would run from the 40G drive
in Safe Mode; but, in regular mode, I got all sorts of
really bizarre symptoms - culminating in the
corruption of the 40G drive (and not just the
partition from which Windows was running). When I
tried the same move from the 6G drive to the Maxtor,
which is a 30G drive, all went immediately as I
expected. The fact that Windows behaved differently
with the 40G drive in Safe Mode (where I am sure that
it uses PIO) suggests that the difference when not in
Safe Mode may have been an attempt to use DMA. But
the DMA check boxes had already disappeared well
before this 'moving' experience.