The version of nvprof in CUDA 9.0 causes a hang when used to profile an OpenACC program. This is because it calls acc_get_device_type from a callback called during device initialization, which then attempts to acquire acc_device_lock while it is already taken, resulting in deadlock. This works around the issue by returning acc_device_none from acc_get_device_type without attempting to acquire the lock when initialization has not completed yet. 2020-07-14 Tom de Vries <tom@codesourcery.com> Cesar Philippidis <cesar@codesourcery.com> Thomas Schwinge <thomas@codesourcery.com> Kwok Cheung Yeung <kcy@codesourcery.com> libgomp/ * oacc-init.c (acc_init_state_lock, acc_init_state, acc_init_thread): New variable. (acc_init_1): Set acc_init_thread to pthread_self (). Set acc_init_state to initializing at the start, and to initialized at the end. (self_initializing_p): New function. (acc_get_device_type): Return acc_device_none if called by thread that is currently executing acc_init_1. * libgomp.texi (acc_get_device_type): Update documentation. (Implementation Status and Implementation-Defined Behavior): Likewise. * testsuite/libgomp.oacc-c-c++-common/acc_prof-init-2.c: New. |
||
---|---|---|
.. | ||
config | ||
lib | ||
libgomp.c | ||
libgomp.c++ | ||
libgomp.c-c++-common | ||
libgomp.fortran | ||
libgomp.graphite | ||
libgomp.hsa.c | ||
libgomp.oacc-c | ||
libgomp.oacc-c++ | ||
libgomp.oacc-c-c++-common | ||
libgomp.oacc-fortran | ||
Makefile.am | ||
Makefile.in | ||
libgomp-site-extra.exp.in | ||
libgomp-test-support.exp.in |