Extremely low burn speeds, High CPU usage

[BetaMaster]

I'm using an LG H10N Burner on a P4 3.0GHz with 512 Ram using Nero Express 7. Everything worked fine last week. I took a small break and when I started burning again I started having issues with drive speed. I did have a run in with the Antivirus XP 2008 malware, but it has been removed.

It started with some new Kodak CD media. I burned one at 3 min but it would not close out. The next ones I burned took about 17 min each, and about half of them had errors towards the end. Typically, burning and verification took about 6 minutes before with my Philips CD media.

When I tried to burn a DVD+R (either TDK or Fuji) it was extremely slow. My current test is typical - 15 minutes to burn to the 36% mark. I aborted the burns. I've tried rebooting - no help.

When I look at task manager, the pattern is that Nero jumps to the 200,000K memory mark. Just as it's changing from lead-in to the burn, the CPU usage jumps to the 50% range from the 10% range. It stays that high as the disc burns.

In one case, the memory dropped back down just before the burn started, but when the burn started my Norton Live Update came online and the burn was still slow. That's a common problem on my machine but it's usually solved by aborting the burn, allowing Norton to run it's course, and starting all over again.

I'm thinking I might need a new burner, but would I just be wasting my money if the problem is something else? any help would be appreciated here.

[BetaMaster]

WOOHOO! Solved it!

In the event this happens to anyone else, check your drive's DMA settings. Somehow, mine became disabled. A lot of things can cause this and a lot of things can make it nearly impossible for you to re-enable DMA.

In my case, the Secondary channel IDE controller had permanently set the mode to PIO, meaning it would not enable DNA even if I uninstalled the drive and reinstalled it. The trick was to uninstall the Secondary Channel IDE controller and restart. At that point, both my DVD drives (My ROM drive was suffering the same problem) showed as DMA enabled in Nero.

Nero is still taking a chunk of RAM (about 268,000K), but the overall CPU use is in the 10-20% range and the disk looks like it will burn in about 11 minutes - slow but acceptable and not untypical for this media.

I'm guessing a new burner would not have solved the problem.

Any other comments would be welcome!

[tomlee59]

I was just going to suggest checking this very thing -- and it will likely happen again to you. Windows has a "helpful" behavior -- if there is a recurring problem that requires it to throttle back in drive speed (say, because of a balky disc), Windows will helpfully decide -- without informing you -- that DMA is too much to handle, and will disable it! It's the gift from the folks at Redmond. Now, mind you, Windows will annoyingly pop up with all sorts of useless status information about every other thing ("Updates are now ready to install" "Network connection available" etc.), but it won't tell you "I've done you this favor of permanently prohibiting DMA, so you'll never burn fast again."

[jagabo]

This problem, along with the same fix, is addressed here all the time. There is also a registry patch which will keep Windows from switching to PIO when it's having trouble reading the drive.

[BetaMaster]

Thanks jagabo. I did search the forum first, but couldn't find anything. It was only after I tested my other DVD ROM drive and it exhibited slow reading behaviour that I thought of the DMA settings.

I have had slow burning behaviour before that was from another cause. It was usually my first disc of the day after starting up that was slow. I finally found that in the first 15 minutes after a reboot either Norton, Windows or something else was checking for or, horror of horrors, attempting to install updates. One of the behaviours that often slows me right down on every task is when Norton feels its the "right time" to update its virus definitions. This new problem acted much like that old problem (which still happens of course)

[redwudz]

What I have posted before for DMA/PIO problems:

To check DMA/PIO mode within Windows:

Control Panel>System>Hardware>Device Manager>IDE ATA/ATAPI controllers.

From there, right click on one of the channels and choose 'Properties>Advanced Settings'. All drives should be DMA mode. The 'Current Transfer Mode' for Hard drives is usually DMA 4-6 and DVD burners DMA 2-4, DVD ROMs usually DMA 2. If you see any in PIO mode, that can slow things down.

Changing them back may be easy or complicated. First see if you can change them in that window. If not, I usually uninstall the channel the drive is on and let the OS reinstall it. This will usually take a reboot. This will not damage any files on the computer.

From there, if no luck, get back to us.

[Seeker47]

This problem, along with the same fix, is addressed here all the time. There is also a registry patch which will keep Windows from switching to PIO when it's having trouble reading the drive.

Would you happen to have a link for that Registry patch ? (Hopefully it is applicable to W2K also, rather than just XP.)

[jagabo]

This problem, along with the same fix, is addressed here all the time. There is also a registry patch which will keep Windows from switching to PIO when it's having trouble reading the drive.

Would you happen to have a link for that Registry patch ? (Hopefully it is applicable to W2K also, rather than just XP.)

Maybe these:
http://forum.doom9.org/archive/index.php/t-63560.html
http://winhlp.com/node/10
http://support.microsoft.com/kb/817472/

[Seeker47]

This problem, along with the same fix, is addressed here all the time. There is also a registry patch which will keep Windows from switching to PIO when it's having trouble reading the drive.

Would you happen to have a link for that Registry patch ? (Hopefully it is applicable to W2K also, rather than just XP.)

Maybe these:
http://forum.doom9.org/archive/index.php/t-63560.html
http://winhlp.com/node/10
http://support.microsoft.com/kb/817472/

Thanks, jagabo, I will go over these and see what I can find out. The last of those links I also discovered a short while ago, coincidentally. And I found these:

http://support.microsoft.com/default.aspx?scid=http://support.microsoft.com:80/support...NoWebContent=1
(refers to a utility called DMACHK, which hails from the days of Win NT-4)

http://sniptools.com/vault/getting-back-to-dma-mode-in-windows-xp
(but this may be XP-specific; note the many comments after, with other ideas)

There is also a MS page for a hotfix one may request, though it may only apply to certain particular circumstances.

I'm starting to think my case falls into the more dire examples, and will have to involve some sort of Registry fix.

A Quick Recap:
W2K-SP4 on a P4-based Shuttle. Two separate IDE channels, with only one device on each (HDD on the Primary, DVD burner on the Secondary), so no Master / Slave type concerns. An apparent IDE controller failure forced me to replace the MB, which was supposedly the same Rev. as the original, but maybe not "same" enough. Win somewhat freaked out after this rebuild, at the driver level. The usual arguments with some Win-folks who insist that *everything* must be re-installed from scratch (with me saying "Like Hell I will !"), vs. a few others who report that they've migrated major hardware (much more than in this case), more than once, and re-done all their necessary underlying drivers, without having to reinstall the OS, reinstall all the apps, redo all the user preferences and customization from scratch. If the first camp was correct, Windows sucks even worse than I thought.

Anyway, I think I've recovered a lot of it. Graphics was not a problem, just a temporary inconvenience. I still don't have audio (non-acknowledgment of drivers that are present on the HDD), but this is a much lower priority issue. The real dealbreaker is being stuck in a very stubborn PIO mode. I tried what Redwudz suggested for the Primary, even rebooting after each change, and it did not work. It still shows "DMA Mode if Available" with Auto Detection but Current Transfer Mode = PIO mode for 'Device 0'; Auto Detection, 'DMA if Available', and Not Applicable for Device 1 -- because there is no Device 1, I take it. For the Secondary Controller I uninstalled it, then let Windows re-enable it at the reboot. Doing that made the Advanced tab for the Primary channel disappear entirely . . . and so I now find no report as to whether it is in DMA or PIO mode, but I fear the worst.

I was reluctant to try uninstalling the Primary controller -- not just because of the Advanced Tab going away but because I did not want to risk being able to boot W2K from this HDD.

Maybe I'll find the answer in one of these links, but I think this problem can be one tough bastard, and won't be going away without a Registry hack, a hotfix, or something. It does make this box pretty useless for ripping, burning, etc. Which is why I've been using it for just internet or word-processing for a few months now, and falling back on an older laptop + ext. USB for the rest -- which isn't a much preferable solution. I'd like to get this corrected.

[redwudz]

If the rest of your OS is working, you can delete the controller and it will reinstalled. But I think you have a registry problem. And I don't know if this was mentioned, but a bad IDE cable can cause the same problem. I would replace it. Also make sure your drive has the jumper properly installed, master or slave.

This is a long article, and I don't have a link to the original, but it gives more information about the problem.

DMA reverts to PIO
The stuttering DVD drive
General description

DMA is an abbreviation for Direct Memory Access, an access method for external
devices where the data transfer is not done by the central processor, but by a
small special processor called DMA controller. It uses a procedure called cycle
stealing, where the central processor memory access cycles are delayed for very
short times to intersperse DMA controller memory access cycles. Some newer,
faster DMA modes are called UDMA (Ultra DMA).

The alternative, slow and inefficient data transfer mode is called PIO,
Programmed Input-Output, where the central processor transfers data byte for
byte or word for word. This requires many processor commands for each data word
and therefore causes a high and unwanted processor load.
Possible causes for falling back to PIO mode
The most frequent use why a CD or DVD port falls back to PIO mode is a scratched
or otherwise unreadable CD or DVD.

However, there are a few reasons why a computer may use PIO instead of DMA,
particularly when it's the hard disk port that falls back, not a CD/DVD drive
port. For example, David Duberman reported in 2005 that some Dell computers have
DMA disabled in their BIOS by default for the second hard disk. So it is a good
idea to check the BIOS settings first.

Other reasons can show up in the event log, so check this first and see if you
can find repeated Atapi errors recorded. If so, you likely have a hardware
defect. You can use the procedure described below, but your computer will
probably fall back to PIO mode again and again, until you solve the underlying
problem, which may be located inside the device, on the motherboard, or in the
IDE data cable and its connectors.

The trap
Windows contains a trap in which quite a few computers seem to get caught sooner
or later. The trap was described in a Web article whose link no longer works
(and also in another one mentioned below):

The crucial paragraphs are:
PIO mode is enabled by default in the following situations:
...
For repeated DMA errors. Windows XP will turn off DMA mode for a device after
encountering certain errors during data transfer operations. If more that six
DMA transfer timeouts occur, Windows will turn off DMA and use only PIO mode
on that device.
In this case, the user cannot turn on DMA for this device. The only option for
the user who wants to enable DMA mode is to uninstall and reinstall the
device.

Windows XP downgrades the Ultra DMA transfer mode after receiving more than
six CRC errors. Whenever possible, the operating system will step down one
UDMA mode at a time (from UDMA mode 4 to UDMA mode 3, and so on).
...
Of course, drive firmware being quite complex and certainly containing
programming defects of its own, it is not all that difficult to produce such
errors. In my case a scratched DVD and later also an unreadable (overburned) CD
did the trick, got the drive to choke and Windows to disable DMA for good. Later
my hard disk hiccupped just once and also went back to PIO for good.
I had been using my laptop for DVD viewing for years, until I inserted a
borrowed and heavily scratched DVD.

The player and apparently even the DVD drive
choked on it, and when I finally got the DVD to play, I found that playing was
jerky and processor load was 100%, roughly half of which was system overhead.
This indicated that the drive had reverted from the usual UDMA (Ultra Direct
Memory Access) mode 2 to PIO (Programmed Input Output) mode. No amount of
resetting or changing the relevant registry parameters from 1 (try DMA) to 2
(force DMA) helped. Stubbornly the drive kept using PIO mode, and Windows even
changed these settings back to 0 (use PIO only).

The following text will refer to the secondary IDE port because that is more
often affected, but essentially the same also holds for the primary IDE port, to
which the main hard disk is connected in most computers.
Before you begin to work on the problem, log on as Administrator or as a user
with administrator rights.

Check Your IDE Port Mode
First check what mode your secondary IDE port is currently working in. Go to
Device Manager: right-click on My Computer, select Properties, click on the
Hardware tag, click on the Device Manager button, click on the plus sign to the
left of IDE ATA/ATAPI Controller, double-click on the secondary IDE channel,
click on Extended Settings and check whether it is set to DMA when available.
Directly underneath that setting is a grey field that shows the actual working
mode of your IDE channel. You want the highest possible DMA or Ultra DMA mode
there, and you definitely don't want PIO mode.

Normally you don't have to use the registry editor for this, because the normal
settings are also available through the properties dialog for the IDE port, but
if you want to look at it anyway, the parameter for the secondary IDE port can
be found through regedit.exe at

HKEY_LOCAL_MACHINE\HARDWARE\DEVICEMAP\Scsi\Scsi Port 1

It is named Scsi only for historic reasons. Scsi Port 0 is the primary IDE port,
to which presumably your hard disk is connected.
After trying various remedies—in vain—I found the abovementioned article and
went to work again. I uninstalled the DVD drive in Device Manager and rebooted,
but that did not help either.

So I searched for more and better information, then I went on and did the
following.
Re-enable DMA using the Registry Editor
My thanks go to my fellow MVP Alexander Grigoriev who taught me this method.
Run REGEDIT. Go to the following key:
HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Contro l\Class\{4D36E96A-E325-11CE-BFC1-08002BE10318}
It has subkeys like 0000, 0001, 0002, etc. Normally 0001 is the primary IDE
channel, 0002 the secondary, but other numbers can occur under certain
circumstances. You have to go through these subkeys and check the DriverDesc
value until you find the proper IDE channel.

Delete MasterIdDataChecksum or SlaveIdDataChecksum, depending on whether the
device in question is attached as master or slave, but it can't actually hurt to
delete both. Reboot. The drive DMA capabilities will be redetected.
Open Device Manager again and check whether the device is now actually using DMA
mode. If so, congratulations, you've made it (at least until the next time
Windows disables DMA).

2005-10-24 – Tomáš Souček wrote, if this doesn't work, check also the dword
value MasterDeviceTimingModeAllowed, whose default value is hex 0xFFFFFFFF. If
you have a much smaller value, you can try to set it back to its default and
reboot for a test.

Alternative Method—Uninstalling the Port

1. Uninstall the secondary IDE port
To do that, open Device Manager as follows. Right-click on My Computer, select
Properties, click on the Hardware tag, click on the Device Manager button, click
on the plus sign to the left of IDE ATA/ATAPI Controller, right-click on
Secondary IDE Channel, click on Uninstall. Deactivating is not enough.
Reboot to make the changes active and permanent.
After booting Windows will automatically reinstall the IDE channel and the DVD
(or CD) drive. This Plug-n-Play process can take a little while, so give it a
minute after the boot process finishes.

2. Reactivate DMA
But this is not enough, because unfortunately Windows does not automatically
activate DMA on a DVD or CD drive. You have to tell Windows to try to use DMA
first.
For that, go to Device Manager again. Right-click on My Computer, select
Properties, click on the Hardware tag, click on the Device Manager button, click
on the plus sign to the left of IDE ATA/ATAPI Controller, double-click on the
secondary IDE channel, click on Extended Settings and change the relevant
setting from PIO only to DMA when available.

On Windows NT and 2000 you now have to reboot a second time, but Windows XP
applies the change instantly. Then you can go to the same place in Device
Manager again and check whether the device is now actually using DMA mode. If
so, all is well.

3. Driver is not intended for this platform
If you keep getting the following error message:
There is a problem installing this hardware.
IDE channel
An error occurred during the installation of the device. Driver is not
intended for this platform.

2005-03-30 – Johannes B. wrote: The reason for this error is often that Daemon
Tools or Alcohol 120% are installed. In this case the solution described below
would not work. But when you uninstall these programs and then restart Windows,
it will then install the device drivers without any further problems.
If these programs are not installed, then one possible way out is to rename
C:\WINDOWS\system32\drivers\atapi.sys (or a similar path on your computer) to
something like atapi.old.

If that's not possible, you can try it from the repair console (boot from the
Windows install CD and select the repair console).
If Windows always automatically recreates atapi.sys, you can try renaming it in
safe mode or from a command line window or you can try to rename or remove it in
the driver cache as well.

Desensitize Your Computer's IDE Channels
There's a bit more to it. The following article offers a way to reduce the
incidence of this problem, although it still doesn't solve it altogether.
IDE ATA and ATAPI Disks Use PIO Mode After Multiple Time-Out or CRC Errors Occur
http://support.microsoft.com/kb/817472/

Do read this article because it contains a useful long-term workaround. But you
have to go through the procedure described here to re-enable DMA first. Assuming
you've done that, insert the ResetErrorCountersOnSuccess registry values
mentioned in this article into both the primary and the secondary IDE port
registry keys as described.

Unfortunately this is only a half solution, because when you enter an unreadable
DVD, you will get 6 errors in a row, and the IDE channel will revert to PIO
mode, but at least when you pull out the DVD in time and then insert a good one,
the error counter will be reset and it will at least be a bit more difficult for
Windows to hobble your IDE drive.

[Seeker47]

If the rest of your OS is working, you can delete the controller and it will reinstalled. But I think you have a registry problem. And I don't know if this was mentioned, but a bad IDE cable can cause the same problem. I would replace it. Also make sure your drive has the jumper properly installed, master or slave.

Thanks, redwudz. I believe that article is from one of the links mentioned in the last few posts.

Both the HDD and the burner are still jumpered the way they always were. The computer was first assembled (from the Shuttle "kit") in 2004, and gave me over 4 good years with no occurrence of this problem. The current HDD has been in there about 2 years now, the burner closer to 3. As I stated, the box now has a new replacement MB, of the same type. Since I knew I was going to have to re-do the hardware anyway, and it really wouldn't amount to any more work, I took the opportunity to put in an upgraded PSU (a spare that had been put aside for future need) and a new set of cables identical to the set that came with the original kit. They are the same type of 40-wire IDE cables that served me well all those years, but in proprietary lengths and shapes made for the Shuttle box. I would put 80-wire cables into a more contemporary computer, but these are appropriate for this type of IDE HDD and Pioneer pre-110 (it's a 108) burner.

When you read through a long bunch of comments like those in the Sniptools link, you invariably see a lot of "I tried THIS and it didn't work," "I tried (same) THIS and it did work," but the same problem may well have been possible from more than one cause. You also see a lot of 'maybe this, maybe that' -- everything from device failure to bent pins to _______ (fill in the blank), most of which are conceivable but most of which are also towards the lower end of the likely curve. It would be unwise to rule any of them out, but I wouldn't bet any money on them. I'm thinking your Registry idea is much more likely to be the culprit.

That Advanced Tab going away for the Secondary channel after letting Win reinstall it still puzzles me, but it may involve the difference between the built-in Win drivers and those from the mfr.'s driver CD.

[Seeker47]

UPDATE: Just as I predicted, it probably wasn't anything to do with bad hardware, because I managed what appears to be a fix without doing anything further to devices, cables, etc. Even though I printed out all the links in this thread, plus several other articles on the subject, in the end I did not need to try any of their suggestions -- the Registry Fix ones in particular -- with one possible exception. In one of those articles, I noticed a mention of the "Update Rollup for SP4" (remember, this is W2K, not XP) patch from MS Update. This is a big one -- like 250 megs or so. I always assumed that it was a catch-up catchall, but redundant with the numerous hotfixes I've been applying all along. So, I just ignored that one since it came out, a couple years back. Just recently, I read that this one is a.k.a. the "Unofficial SP-5 for W2K", and elsewhere that it includes a new version of Atapi.Sys. By this point, I've been stuck in PIO mode hell for like 3 months, which is like limping along on half an engine, so I figured there wasn't that much to lose. I backed up the registries in both of the W2K boot partitions on the drive, and applied the Rollup.

On one of the boot partitions, this proved to be the cure in and of itself: I was able to get right back to UltraDMA mode for both Primary and Secondary IDE channels. What a difference ! Everything from the hard drive has snappy response again; DVDs burn again in minutes, rather than hours.

Things did not go quite so smoothly in the other boot partition. It was still an unwilling resident of PIO-ville. I went back to the driver CD for this computer, intending to reinstall the ATA drivers, but before getting that far I noticed an option to install the Intel Application Accelerator -- something that had been mentioned as a fix that worked for one contributor to one of those DMA-fix articles. So I decided to try that, and it seems to have put the other boot partition over the top. That one is working properly again also . . . although that one still has no 'Advanced' Tab where DMA status is reported. That DMACHK utility says there are NO Atapi devices connected ! Well, a rose by any other name: I'll settle for the UltraDMA performance of two ghost devices (HDD and burner), so long as they keep working, and not lose any sleep over it. Maybe there still is something not right about those registries. The key issue will be: does this fix stick ? And it begs the question of why everything worked fine for 3 years, before that Update Rollup came along ?

[redwudz]

I've only had one system refuse to work in DMA mode. I finally gave up and repartitioned, reformatted and reinstalled the OS. I had tried the upper/lower filter stuff and all the rest with no luck.

I suspect in some cases it is a bit the registry, and also the IDE controller or MB drivers and maybe even the cabling. I did have one motherboard, I forget the brand, that didn't have the DMA/PIO settings. They were automatic and you never knew what mode it was in unless the drives were really slow. Reverting to PIO mode can also happen to a hard drive, though I have never seen that.

I did learn early on if you put a DVD/CD in the player and you hear a lot of 'seeking' noise from the laser assembly, you'd best shut it down and get that disc out of there before Windows decides to revert the drive to PIO. I have yet to understand why the OS doesn't just reset it automatically on the next boot. The only OS reasoning for the downgrade that I could find is to let you finish reading the data, albeit at a very slow rate. But rebooting should be able to reset it. Even worse, the OS never tells or warns you that it reverted to PIO. That part seems really dumb. But I assume they have a reason or purpose to this.

To me, it's a major flaw that should have been corrected a long time ago.

[Seeker47]

I suspect in some cases it is a bit the registry, and also the IDE controller or MB drivers and maybe even the cabling. I did have one motherboard, I forget the brand, that didn't have the DMA/PIO settings. They were automatic and you never knew what mode it was in unless the drives were really slow. Reverting to PIO mode can also happen to a hard drive, though I have never seen that.

I had never seen this myself -- for either device -- until a few months ago. In reading those articles, I got the impression that XP is fairly prone to this problem, and sometimes on a recurring basis. Or it may just be that most Win users have been using XP for the last several years. I do hope SATA is going to be more resistant to this issue than IDE, though I guess not that many people have been moving to SATA burners, so it wouldn't really make a difference.

The only OS reasoning for the downgrade that I could find is to let you finish reading the data, albeit at a very slow rate. But rebooting should be able to reset it. Even worse, the OS never tells or warns you that it reverted to PIO. That part seems really dumb. But I assume they have a reason or purpose to this.

To me, it's a major flaw that should have been corrected a long time ago.

AMEN.