forked from OSchip/llvm-project
Fix a problem on Windows machines with >32 processors
llvm-svn: 217024
This commit is contained in:
parent
e499db0184
commit
12ad3b1911
|
@ -754,7 +754,7 @@ __kmp_affinity_bind_thread( int proc )
|
|||
KMP_DEBUG_ASSERT((proc >= 0) && (proc < (__kmp_num_proc_groups
|
||||
* CHAR_BIT * sizeof(DWORD_PTR))));
|
||||
ga.group = proc / (CHAR_BIT * sizeof(DWORD_PTR));
|
||||
ga.mask = 1 << (proc % (CHAR_BIT * sizeof(DWORD_PTR)));
|
||||
ga.mask = (unsigned long long)1 << (proc % (CHAR_BIT * sizeof(DWORD_PTR)));
|
||||
ga.reserved[0] = ga.reserved[1] = ga.reserved[2] = 0;
|
||||
|
||||
KMP_DEBUG_ASSERT(__kmp_SetThreadGroupAffinity != NULL);
|
||||
|
|
Loading…
Reference in New Issue