Archives Discussions

[dukeleto]

exception generated with new brcc, whereas brook 0.5 parses ok

Hello,

I have a project which compiles and runs fine with the latest brook 0.5 from sourceforge, but for which the new amd brrc generated a windows exception but no error message. Is there a way of finding out what is wrong in my .br?
Many thanks

[michael_chu]

Hi dukeleto,

Can you reduce it down to a small test case which reproduces the problem?

Also, what is the system and GPU configuration you are using?

Also, give v1.00.2-beta a try. It is available for download on the website now:

http://ati.amd.com/technology/...mputing/sdkdwnld.html

Hopefully, it may have fixed whatever you are encountering.

Michael.

[dukeleto]

Hello Michael,
I have just tried with the latest beta, and met the same problem. My system is
winxp 32 bit, a core 2 processor, and an HD3870 card, running catalyst drivers.
I wanted to avoid running through the code kernel by kernel, but I will try, and
see if I can identify the culprit.
Thanks

[michael_chu]

Hi dukeleto,

Let me know how it goes.

Does building it in debug mode help narrow down where the exception is happening?

Michael.

[dukeleto]

Hi there,
after a little break, I've got back to trying out my brook, and have identified the source of the problem;
it was unfortunate/incorrect bracketing which amd_brcc could not pass, but which was (mysteriously!)
correctly interpreted by brcc_0.5.
It might be helpful if the amd_brcc parser were able to inform of the first line that gives a parse problem?
Thanks