I'm getting game crashes on windows 10, someone can look at the debugger output (dump)?

I have reproducable crashing on software that has compatibility across all previous versions, but now on the Windows 10 preview it continuously crashes, as for the game's part, it's not caused by the software itself as everything points in the direction of an ATI driver component that causes the crashes, thus game-technically nothing much can be reached or debugged. If anyone here has experience with these sort of crashdumps or atleast the stack messages this dump clearly provides and can give an probable cause of the crash I would be very happy.. I'm not into D3D/graphics development myself nor can I make anything readable out of the debugger output.

It's the game Multi Theft Auto that's crashing on an ATI module, an MP engine for SA that functions across everything but for limited group of users (as for me now) not, or limited to me, no idea.

Debugger output: http://pastebin.com/Bd4NPvKB

If you have a debugger with more (ati) symbols then here's the dumpfile itself: https://mega.co.nz/#!BEESVAqL!ynbyXuyad7vOSMuaxYzsRypa9f__6_9mHbungAUIDes

These are the crash params:

Version = 1.4-release-6949.0.000 Time = Fri Nov 28 17:44:01 2014 Module = C:\WINDOWS\SYSTEM32\atiumdag.dll Code = 0xC0000005 Offset = 0x00027ED5

EAX=00000000 EBX=06D7899C ECX=00000001 EDX=00000000 ESI=06D7899C EDI=0E322B44 EBP=0029F15C ESP=0029EF48 EIP=6CCF7ED5 FLG=00010246 CS=0023 DS=002B SS=002B ES=002B FS=0053 GS=002B

Thanks in advance

There isn't much to be seen here. The faulting instruction was

6ccf7ed5 f7421800040000  test    dword ptr [edx+18h],400h

But edx had a zero in it, so this referred to address 0x18, which of course is a never-defined address (as are all va's from 0 through 0xFFFF).

As to how edx got a zero in it, we can't tell. It may have been loaded from memory by the previous instruction:

6ccf7ecf 8b9624040000    mov     edx,dword ptr [esi+424h]

esi contains a reasonable-looking address; this instruction would have loaded edx from the dword at 0x06d78dc0. But we can't see that memory location in a minidump, and anyway there is no proof that the instruction flow came through the previous instruction; the test instruction might have been JMPd or B'd to from someplace else.

Is the system completely stable under a released version of Windows? If not, suspect hardware (video card or power supply).

Other than that there is probably nothing you can do to fix it except to get an updated driver from ATI. (Or try an older one - sometimes that helps.)

And, keep in mind: These aren't even beta versions of Windows 10; they're not expected to be stable.