Re: [Starpu-devel] [LU factorisation: gdb debug output]

[Maxim Abalenkov <maxim.abalenkov@gmail.com>]

Hello Samuel,

I believe I know where the problem comes from. Please see the routines in the attached file. They switch the “view” from blocks (tiles) of the matrix to panels (columns of tiles). I have adopted the lines 25–29 and 73–78 from the StarPU example on multiple views. I think it was a mistake.

—

Best wishes,

Maxim

Attachment: tile2panel.c
Description: Binary data

Maxim Abalenkov \\ maxim.abalenkov@gmail.com
+44 7 486 486 505 \\ http://mabalenk.gitlab.io

On 20 Dec 2017, at 14:46, Maxim Abalenkov <maxim.abalenkov@gmail.com> wrote:

Hello Samuel,

Please find below the gdb output after switching to the crashing thread:

(gdb) thread 3
[Switching to thread 3 (Thread 0x7fffeecde700 (LWP 27673))]
#0  0x00007ffff31ca860 in raise () from /usr/lib/libc.so.6

(gdb) frame 8
#8  0x00007ffff73bd75e in _starpu_fetch_task_input (j=0x555555a5ed40) at datawizard/coherency.c:972
972 ret = fetch_data(handle, node, local_replicate, mode);

(gdb) print node
$1 = 1

(gdb) print local_memory_node
$2 = 0

(gdb) print descrs[index].node
$3 = 1

—

Best wishes,

Maxim

Maxim Abalenkov \\ maxim.abalenkov@gmail.com
+44 7 486 486 505 \\ http://mabalenk.gitlab.io

On 20 Dec 2017, at 14:42, Samuel Thibault <samuel.thibault@inria.fr> wrote:

Maxim Abalenkov, on mer. 20 déc. 2017 14:41:14 +0100, wrote:

Please find attached the output of the gdb commands that you suggested. It seems there is no node.

I meant to adapt the thread number according to the stack, since the
fault will not necessarily happen always in the same thread.

Samuel