Silverfrost Forums

Welcome to our forums

OpenGL on newer ATI graphics cards

14 Jul 2010 9:59 #6623

You need to run it from a command prompt (otherwise the message disappears). It should be run on a machine that has the problem.

14 Jul 2010 10:16 #6624

Is this program supposed to work on a client's PC?

Yes. It is supposed to not crash as it only queries a function and does not execute it.

14 Jul 2010 12:48 #6628

Quoted from Robert You need to run it from a command prompt (otherwise the message disappears). It should be run on a machine that has the problem.

Okay, have tried it and on my PC gives an address. I am assuming that on a computer which has the problem it would give en error?

14 Jul 2010 1:49 #6631

Like I've already said, I tested it on Win7+ATI graphics card and it behaves as expected, i.e. does not crash and prints out the address of the wglCreateContext function.

15 Jul 2010 9:41 #6633

I have tried the OpenGL examples on a Win 7 x64 bit system with an ATI card. However, it was an oldish card and the examples ran okay.

Sebastian, can you post details of the crashes you see: tracebacks preferably from within salflibc.

Additionally, does the issue manifest itself through an RDP connection? If it does, is someone willing to let me have a debug via such a connection?

15 Jul 2010 10:02 #6634

I have tried the OpenGL examples on a Win 7 x64 bit system with an ATI card.

Only newer models are affected (ATI Radeon HD 5770 here, also see lozzer's list), we could not reproduce it with older models either.

Sebastian, can you post details of the crashes you see: tracebacks preferably from within salflibc.

The messages are not too helpful, for our main OpenGL application (Win7 message) it's an Access violation at instruction address 0, reading from 0. The small OpenGL test application (sources above, plain %og) just hangs and triggers some Win7 timeout. Maybe lozzer has a more useful error message on the systems he tested his application.

When running the applications under an external debugger (OllyDbg) it halts at the fpu exception inside the ATI opengl driver. I can try to produce some screenshots of such a debugging session (salflibc's OpenGL window proc up to the wglCreateContext call) if that'd be helpful.

Additionally, does the issue manifest itself through an RDP connection?

I don't think RDP connections from outside are possible here, but I'll check that.

Thank you!

21 Jul 2010 7:43 #6655

Has anyone at Silverfrost done any more about this one?

21 Jul 2010 7:08 #6663

Additionally, does the issue manifest itself through an RDP connection? If it does, is someone willing to let me have a debug via such a connection?

We offered to debug it on a failing machine.

21 Jul 2010 7:20 #6664

We cannot provide remote access to our machines (maybe somebody else can arrange that, but always tricky).

What/how do you intend to debug that? sdbg surely won't get you far. As already mentioned above I can provide e.g. screenshots of a low-level debug session depicting the problem spots. Or you can post details of what information you need/how you want to get them.

21 Jul 2010 9:43 #6665

Please send all the information you can.

23 Jul 2010 11:32 #6677

I've sent it to the ftn95 email account at silverfrost since this forum does not seem to handle attachments/non-external images.

23 Jul 2010 1:43 #6678

Quoted from Robert

Additionally, does the issue manifest itself through an RDP connection? If it does, is someone willing to let me have a debug via such a connection?

We offered to debug it on a failing machine.

I may be able to get a customer to allow us access remotely to their computer. Is the debug program something you can send us to run on their computer, or can it be emailed to the customer to run?

Please login to reply.