https://gcc.gnu.org/bugzilla/show_bug.cgi?id=105873
Bug ID: 105873
Summary: [amdgcn][OpenMP] task reductions fail with "team
master not responding; slave thread aborting"
Product: gcc
Version: 13.0
Status: UNCONFIRMED
Keywords: openmp, wrong-code
Severity: normal
Priority: P3
Component: target
Assignee: unassigned at gcc dot gnu.org
Reporter: burnus at gcc dot gnu.org
CC: ams at gcc dot gnu.org, jakub at gcc dot gnu.org
Target Milestone: ---
Target: amdgcn-amdhsa
Created attachment 53100
--> https://gcc.gnu.org/bugzilla/attachment.cgi?id=53100&action=edit
C testcase – compile with -fopenmp + AMD GCN offloading
Long example:
https://github.com/SOLLVE/sollve_vv/blob/master/tests/5.0/taskgroup/test_taskgroup_task_reduction_device.c
Reduced example attached (thanks Jakub!). Works with host fallback or on nvptx.
With AMD GCN, it runs into the error:
https://gcc.gnu.org/git/?p=gcc.git;a=blob;f=libgomp/config/gcn/team.c;h=254dd4da23498fc54f196dbe9cbb70236f1d63a2;hb=HEAD#l112
133 const char msg[] = ("team master not responding;"
134 " slave thread aborting");
Namely:
team master not responding; slave thread abortingteam master not responding;
slave thread abortingteam master not responding; slave thread abortingteam
master not responding; slave thread abortingteam master not responding; slave
thread abortingteam master not responding; slave thread abortingteam master not
responding; slave thread abortingteam master not responding; slave thread
abortingGCN Kernel Aborted
GCN Kernel Aborted
GCN Kernel Aborted
GCN Kernel Aborted
GCN Kernel Aborted
GCN Kernel Aborted
GCN Kernel Aborted
GCN Kernel Aborted
Barrier sync failed (another thread died?); aborting.Barrier sync failed
(another thread died?); aborting.Barrier sync failed (another thread died?);
aborting.Barrier sync failed (another thread died?); aborting.Barrier sync
failed (another thread died?); aborting.Barrier sync failed (another thread
died?); aborting.GCN Kernel Aborted
GCN Kernel Aborted
GCN Kernel Aborted
GCN Kernel Aborted
Barrier sync failed (another thread died?); aborting.GCN Kernel Aborted
GCN Kernel Aborted
GCN Kernel Aborted
Barrier sync failed (another thread died?); aborting.GCN Kernel Aborted
Command terminated by signal 6
