Hello

I have an AMD Llano system, and it don't want to start the WU.
I have Bionc Client 7.0.2, Windows 7 x64, 11.11 Driver. I have 8GB RAM.

11.12.2011 19:59:39 | | ATI GPU 0: AMD SUPERSUMO (CAL version 1.4.1607, 512MB, 479MB available, 512 GFLOPS peak)
11.12.2011 19:59:39 | | OpenCL: ATI GPU 0: BeaverCreek (driver version CAL 1.4.1607 (VM), device version OpenCL 1.1 AMD-APP-SDK-v2.5 (793.1), 1024MB)
11.12.2011 19:59:39 | | ATI GPU is OpenCL-capable

The Opencl divice lists that the GPU has 1GB RAM. But the WU won't start.

I've been trying on an AMD Fusion E350 box too. I have the latest drivers; updated to the AMD SDK 2.5 and have 1GB of VRAM. BOINC is working fine for me and sending the GPU work.

However the OpenCL ATI BRP4 application crashes for me each time with this output:

http://albert.phys.uwm.edu/result.php?resultid=61097

[15:42:24][2052][ERROR] Application caught signal 11.

------> Obtained 3 stack frames for this thread.
------> Backtrace:
Frame 3:
Binary file: ../../projects/albert.phys.uwm.edu/einsteinbinary_BRP4_1.19_i686-pc-linux-gnu__atiOpenCL (0x80b6ae2)
Offset info: pthread_mutex_lock+0x64a
Source file: erp_boinc_wrapper.cpp (Function: sighandler / Line: 166)
Frame 2:
Binary file: /opt/AMDAPP/lib/x86/libamdocl32.so (0xf6054199)
Offset info: +0xae199
Frame 1:
Binary file: /opt/AMDAPP/lib/x86/libamdocl32.so (0xf6054199)
Offset info: +0xae199
------> End of backtrace

15:42:24 (2052): called boinc_finish

---

Hi Gaurav!

I've been trying on an AMD Fusion E350 box too. I have the latest drivers; updated to the AMD SDK 2.5 and have 1GB of VRAM. BOINC is working fine for me and sending the GPU work.

However the OpenCL ATI BRP4 application crashes for me each time with this output:

[...]
Binary file: /opt/AMDAPP/lib/x86/libamdocl32.so (0xf6054199)
Offset info: +0xae199
Frame 1:
[...]

I suspect your environment setup is not complete, that lib should have been removed from the resp. directory. See this document:

http://developer.amd.com/sdks/amdappsdk/assets/AMD_APP_SDK_Installation_Notes.pdf

Section 3:

3 Important Note
• For a successful build and correct operation of individually downloaded s when using the Linux
Catalyst 11.11 drivers or later and the Linux AMD APP SDK 2.5, delete the old 2.5 runtime
files from /opt/AMDAPP/ .
To do this for 32-bit Linux OS:

a. Go to /opt/AMDAPP/lib/x86.

b. Delete the libamdocl32.so and the libOpenCL.so.1 files.

c. Create a symlink under /usr/lib using the command
# ln –s libOpenCL.so.1 libOpenCL.so


To do this for 64-bit Linux OS:

a. Go to /opt/AMDAPP/lib/x86.

b. Delete the libamdocl32.so and the libOpenCL.so.1 files.

c. Go to /opt/AMDAPP/lib/x86_64.

d. Delete the libamdocl64.so and the libOpenCL.so.1 files.

e. Create a symlink under /usr/lib using the command
# ln –s libOpenCL.so.1 libOpenCL.so

f. Create a symlink under /usr/lib64 using the command
# ln –s libOpenCL.so.1 libOpenCL.so

CU
HBE

P.S.: It escapes me why this cannot be done by the driver installer (at least it could issue a warning....

---

Thanks, HBE. Unfortunately, this does not appear to be the problem, it seems. I am planning to experiment with slightly older drivers (11.5 -- 11.10) to see if anything by chance just works out ..

Will report back soon.

---

An update: With the the just-released 11.12 catalyst, I actually have things working on the APU! I seem to have a minor problem left to address. After the first data-set processing in a workunit is complete, it is unable to start the next one. For example see below:

http://albert.phys.uwm.edu/result.php?resultid=64181
http://albert.phys.uwm.edu/result.php?resultid=64159
http://albert.phys.uwm.edu/result.php?resultid=64137

They all crash with: Error during OpenCL kernel setup: PS (error: -5) during the processing of the second data-file AFTER completing the processing of the first data file successfully.

Comments?

---

This is a problem Oliver needs to look into, which means it won't be fixed this year.

BM