Joseph Huber 134401deea

[Offload] Move RPC server handling to a dedicated thread (#112988) ...

Summary:
Handling the RPC server requires running through list of jobs that the
device has requested to be done. Currently this is handled by the thread
that does the waiting for the kernel to finish. However, this is not
sound on NVIDIA architectures and only works for async launches in the
OpenMP model that uses helper threads.

However, we also don't want to have this thread doing work
unnnecessarily. For this reason we track the execution of kernels and
cause the thread to sleep via a condition variable (usually backed by
some kind of futex or other intelligent sleeping mechanism) so that the
thread will be idle while no kernels are running.

2025-01-24 11:36:45 -06:00

..

assert.c

…

fwrite.c

…

global_ctor_dtor.cpp

…

malloc.c

…

puts.c

…

server.c

[Offload] Move RPC server handling to a dedicated thread (#112988)

2025-01-24 11:36:45 -06:00