Dungeon Siege-legends of Aranna data corruption

[Guest]

Spent weeks trying to troubleshoot the crc / corruption errors. ( MS official position that it is a hardware problem. They are wrong, in my humble opinion).

I verified memory, installed on different known good hard drive, replaced IDE cables, etc, etc. Tried running th game without sound, hardware acceleration, etc. The system is an Abit IS7G mobo, Radeon 9800 pro, Crucial matched DDR and Seagate 120 gig HD. NO TWEAKS! All PCI slots empty. BIOS and all drivers current. Games, drivers re-installed multiple times.... Done it all. Twice.

The interesting data point is that I ran DS with the verifydata=true option, verified all the CRC values on the res files. Next time the game crashed after a save, I rebooted and ran the CRC checks again, and this time the res files had CRC errors.They were now corrupt! So, some nasty app wrote to the res files. Seems likely to me that the app was DS. ( I had disabled all non-essential apps and services using msconfig. This is not a HD issue, because replacing the HD did not change the symptom. )

My guess is that there is a programming error in DS. I assume it is written in c++ ( with some assembler routines?) and that the save file pointer/handle is pointing to the resource file during a write. ( Maybe the io routines are using low level file calls, not standard XP APIs?)

The sheer number of users reporting this problem ought to clue MS onto the fact that it is a real problem. I suspect that it will not ever be fixed in the current games, because the economics dont warrant it. Maybe in DS2?

In any case, I have junked the game. Not wasting any more of my time on it. I just wish MS would 'fess up, acknowledge the problem even if they dont intend to fix it. Telling us that DS may be the only app on the system that fails because it pushes the hardware is not fair to the fans of this otherwise great game.

( As a footnote, DS runs error free on my Windows 98 installation.)

My two cents....

 

[Chris H.]

Have you contacted Gas Powered Games? They're the creators, not Microsoft.
You can find contact information here: http://www.gaspowered.com/team.shtml
Make sure to give them factual details, i.e., "when I'm here in the game and
do this . . ." Be prepared to supply them with any system files and full
information of what you've done to track this problem down. You might want
to send them any error messages from Event Viewer which correspond with the
issues you're seeing, for instance.
--
Chris H.
Microsoft Windows MVP
Associate Expert
Expert Zone -

Spent weeks trying to troubleshoot the crc / corruption errors. ( MS

official position that it is a hardware problem. They are wrong, in my
humble opinion).

I verified memory, installed on different known good hard drive, replaced

IDE cables, etc, etc. Tried running th game without sound, hardware
acceleration, etc. The system is an Abit IS7G mobo, Radeon 9800 pro, Crucial
matched DDR and Seagate 120 gig HD. NO TWEAKS! All PCI slots empty. BIOS and
all drivers current. Games, drivers re-installed multiple times.... Done it
all. Twice.

The interesting data point is that I ran DS with the verifydata=true

option, verified all the CRC values on the res files. Next time the game
crashed after a save, I rebooted and ran the CRC checks again, and this time
the res files had CRC errors.They were now corrupt! So, some nasty app wrote
to the res files. Seems likely to me that the app was DS. ( I had disabled
all non-essential apps and services using msconfig. This is not a HD issue,
because replacing the HD did not change the symptom. )

My guess is that there is a programming error in DS. I assume it is

written in c++ ( with some assembler routines?) and that the save file
pointer/handle is pointing to the resource file during a write. ( Maybe the
io routines are using low level file calls, not standard XP APIs?)

The sheer number of users reporting this problem ought to clue MS onto the

fact that it is a real problem. I suspect that it will not ever be fixed in
the current games, because the economics dont warrant it. Maybe in DS2?

In any case, I have junked the game. Not wasting any more of my time on

it. I just wish MS would 'fess up, acknowledge the problem even if they
dont intend to fix it. Telling us that DS may be the only app on the system
that fails because it pushes the hardware is not fair to the fans of this
otherwise great game.

 

[Paul H]

interesting view - the game corrupts therefore the must be a programming
bug. What about all the thousands of people who play the game happily?
Presumably you think they don't encounter the bug, just you.
I had problems with this games install. I assume due to the copy
protection. I tried all sorts, all the suggestions on the net (in which
there was no consistent answer) Eventually I told the install routine to
Ignore all the errors it found and then manually copied the .res files over
and it worked. Since then the game works fine and both my son and I have
played the single player through to the end, saving regularly with no
problems. I also had it crashing out for a period, along with Diablo 2
which I thought was down to video drivers but after many weeks was found to
be a faulty PC400 stick. Have you tried someone else's memory for a few
hours - I borrowed some to establish the problem.
Anyway that's just my experience. I don't think there are enough people
with crashing corrupt games for there to be a inherent programming error. I
do think the copy protection is trying to be too sophisticated for its own
good and pisses off people like me who buy the program legit (and two copies
so that my son & can play)
ah well

 

[Guest]

Not concluding that programming error is the only possibility, just a likely one. Your anecdotal experience is interesting, but it doesnt address these points

1. I stress tested my memory 8 hours, a stick at a time and then both sticks together. No errors
2. No other applications ( including a lot of DirectX games ) experience the problem
3.The res files ( the ones already installed on the hard drive) were verified as being non-corrupt prior to the game run. These same files were changed during the run. They were corrupt after.The only way thay could have changed would be as a result of a write to the disk.( Or suddenly bad media. But remember, this fails in the same way on a different hard drive ).

As to the number of reports, there were certainly enough reports to compel DS to formulate a position that the problem is hardware or driver based. ( Just type "Dungeon Siege CRC" into Google and look at the results).The problem I have with this position is that all the other games I have use the same hardware and drivers without this problem

I believe, reading between the lines, the DS application is using "optimized" code to improve performance. Optimized can mean writing directly to the hardware (which is very fast), but is more fragile because hardware varies. Optimized code can be very sensitive to system timings and the component mix. Bottom line here, it seems that other games save/restore mechanisms are more robust than the ones in DS.

The argument that thousands of others are using it without problems is really specious. By the same logic, no bugs should ever be reported against any software since thousands of beta testers will have used it successfully. But bugs always are reported because it is impossible to test all permutations of software and hardware combinations that will be encountered. And that is fine. But when large numbers of folks are experiencing a problem, the problem needs to be addressed in a way that brings relief to the user. And in the case of Dungeon Siege, I dont believe that has happened.

It may be that the problem wont be fixed, which I can understand. You dont spend a lot of money patching games if they are not generating much new income. But, as in the case of DS, when thousands of problem reports have been automatically submitted ( which I believe to be the case ) a decent explanation should be provided. What hardware and drivers are believed to be the culprits? Does it fail more often on XP than Windows 98? What are the common factors, if any, in all the reports? Saying bad drivers or bad hardware is a broad assertion that doesnt really help resolve the issue or inform the user

DS is asserting that their is no programming problem. I have been developing software for nearly two decades, and I would never assert an application's innocence unless I had unequivocal proof. And maybe, in fact, the DS folks have that proof. But it would be nice if they could share it with the technical community. If we understood the details of the assertion,then we might be able to actually resolve it on our systems. As it is, we have a fairly generic "not our fault" statement that doesnt go far enough in helping users to fix the problem. Even if a patch cannot be provided, a synopsis of the error reports received and a reasonably detailed explanation of the data flow would be in order.

This is not a flame mail. As I said before, I believe this is a great game. And I believe DS2 will be better still. But I want to be able to play it. I am sure it will be built with the same engine as DS, so the potential for propogating this problem is very high. And I think that if enough people complain (loudly), it will get fixed. Or, worst case, it will be explained in such a way that we can resolve any hardware/driver compatibility issues before the release of DS2.

----- Paul H wrote: ----

interesting view - the game corrupts therefore the must be a programmin
bug. What about all the thousands of people who play the game happily
Presumably you think they don't encounter the bug, just you
I had problems with this games install. I assume due to the copy
protection. I tried all sorts, all the suggestions on the net (in which
there was no consistent answer) Eventually I told the install routine to
Ignore all the errors it found and then manually copied the .res files over
and it worked. Since then the game works fine and both my son and I have
played the single player through to the end, saving regularly with no
problems. I also had it crashing out for a period, along with Diablo 2
which I thought was down to video drivers but after many weeks was found to
be a faulty PC400 stick. Have you tried someone else's memory for a few
hours - I borrowed some to establish the problem.
Anyway that's just my experience. I don't think there are enough people
with crashing corrupt games for there to be a inherent programming error. I
do think the copy protection is trying to be too sophisticated for its own
good and pisses off people like me who buy the program legit (and two copies
so that my son & can play)
ah well

 

[Chris H.]

Have you contacted Gas Powered Games as I suggested?
--
Chris H.
Microsoft Windows MVP
Associate Expert
Expert Zone -

Not concluding that programming error is the only possibility, just a

likely one. Your anecdotal experience is interesting, but it doesnt address
these points:

1. I stress tested my memory 8 hours, a stick at a time and then both sticks together. No errors.
2. No other applications ( including a lot of DirectX games ) experience the problem.
3.The res files ( the ones already installed on the hard drive) were

verified as being non-corrupt prior to the game run. These same files were
changed during the run. They were corrupt after.The only way thay could have
changed would be as a result of a write to the disk.( Or suddenly bad media.
But remember, this fails in the same way on a different hard drive ).

As to the number of reports, there were certainly enough reports to compel

DS to formulate a position that the problem is hardware or driver based. (
Just type "Dungeon Siege CRC" into Google and look at the results).The
problem I have with this position is that all the other games I have use the
same hardware and drivers without this problem.

I believe, reading between the lines, the DS application is using

"optimized" code to improve performance. Optimized can mean writing directly
to the hardware (which is very fast), but is more fragile because hardware
varies. Optimized code can be very sensitive to system timings and the
component mix. Bottom line here, it seems that other games save/restore
mechanisms are more robust than the ones in DS.

The argument that thousands of others are using it without problems is

really specious. By the same logic, no bugs should ever be reported against
any software since thousands of beta testers will have used it successfully.
But bugs always are reported because it is impossible to test all
permutations of software and hardware combinations that will be encountered.
And that is fine. But when large numbers of folks are experiencing a
problem, the problem needs to be addressed in a way that brings relief to
the user. And in the case of Dungeon Siege, I dont believe that has
happened.

It may be that the problem wont be fixed, which I can understand. You dont

spend a lot of money patching games if they are not generating much new
income. But, as in the case of DS, when thousands of problem reports have
been automatically submitted ( which I believe to be the case ) a decent
explanation should be provided. What hardware and drivers are believed to be
the culprits? Does it fail more often on XP than Windows 98? What are the
common factors, if any, in all the reports? Saying bad drivers or bad
hardware is a broad assertion that doesnt really help resolve the issue or
inform the user.

DS is asserting that their is no programming problem. I have been

developing software for nearly two decades, and I would never assert an
application's innocence unless I had unequivocal proof. And maybe, in fact,
the DS folks have that proof. But it would be nice if they could share it
with the technical community. If we understood the details of the
assertion,then we might be able to actually resolve it on our systems. As
it is, we have a fairly generic "not our fault" statement that doesnt go far
enough in helping users to fix the problem. Even if a patch cannot be
provided, a synopsis of the error reports received and a reasonably detailed
explanation of the data flow would be in order.

This is not a flame mail. As I said before, I believe this is a great

game. And I believe DS2 will be better still. But I want to be able to play
it. I am sure it will be built with the same engine as DS, so the potential
for propogating this problem is very high. And I think that if enough people
complain (loudly), it will get fixed. Or, worst case, it will be explained
in such a way that we can resolve any hardware/driver compatibility issues
before the release of DS2.

 

[Paul H]

I don't disagree with any of your points
except one
but I stand humbled on it myself.
My RAM passed all the mem tests and soak tests available. I was CONVINCED
it was okay, so I replaced every other component in the system, even the
motherboard and cpu (I work in an IT firm and have access to parts) Still
the problems persisted, looking like driver issues.
Then I changed the memory - even though I knew mine was fine and it wouldn't
make any difference.
But it did.
So I RMAd the sticks for replacement, had a massive row with the supplier
who said it had passed their tests okay and that they would either return it
(to which I said no, it's of no use to me) or restock it for a 30% fee. I
asked what the value would be in restocking and sending it out to some other
unsuspecting...anyway they eventually, after I complained up the chain sent
me two different sticks, same spec.
and the problem stayed gone away. Go figure. I know with hindsight it was
faulty memory but at the time...
I'm not saying that's definitely your problem, just that in the same way
these games programmers aren't infallible, nor are these testing programmes.
Either way I hope you solve it - it IS a great game and my son & I enjoyed
the expansion pack (we had clocked up 7 hrs each by Christmas evening hehe)

Regards
P

 

[Chris H.]

Have you contacted Gas Powered Games as I suggested? What have they said?
--
Chris H.
Microsoft Windows MVP
Associate Expert
Expert Zone -

Paul,
I really appreciate your thoughtful reply, and the sharing your

(horrific) experience. It got me thinking about the problem. Couldnt
logically explain your experience. It didnt fit the software bug theory.
Chewed on it for a while, and I think I might have had an insight. It is
this : the memory tests used to stress our respective RAM dont stress it
through the AGP port! They likely use only the standard memory bus. So....
The games that suck resources (DS) actually pound the AGP to RAM path. ( I
believe this path even on is on a different clock than the regular memory
bus.) So... all the RAM stress testers could show good results, but the
testing doesnt represent a real world scenario. Real world would be
different memory fetches and stores simultaneously(?) on two different data
memory buses. Marginal memory might work ok with a single app accessing from
a single bus, but what happens when the AGP (8x in my case) starts hammering
as well? Anyway, I am going to run with a bad memory stick as my working
hypothesis. Will update in a few days to let you know, assuming you are
curious. ( What would be really helpful is a system stress tester that runs
under XP. Maybe a multi-threaded thing that hammers the RAM through CPU and
AGP fetch/stores and twiddles the disk drive at the same time. MS, are you
listening?) Anyway, thanks for your updates. If it in fact turns out to be
a RAM problem, I will gladly eat humble pie and extend heartfelt apologies
to MS and the developers at Gas Powered Games. Cheers!

 

[Guest]

I have resolved this problem. It is in fact software, not hardware. However, the DS development team is absolved ( sorry guys). The problem is in the ATI Radeon Catalyst Drivers for my video card. It was introduced sometime after version 3.4 of the drivers. I reproduced failures on driver version 3.9, 3.10 and 4.1. I back-leveled the driver to 3.4. The game performs flawlessly. ( My motherboard supports AGP 8, which may not be real common yet,and may have some bearing.) But for other users who have this problem, they might wan to install an older version of the driver. One additional sad note, in testing the newest Radeon Catalyst 4.1 drivers on my machine, Age of Mythology displays same file corruption problems. For Radeon users, stay away from the new drivers unless you MUST have them for a specific fix.