Stanley Tsang
0b2bfdd6d8
* Re-enabling mp unit tests * Fixing shared memory leak and other bugs related to shared mem for MP unit tests * Revert 43bfbfc97bf9edbae1f386d461439091618ff8ed * Further tightening up unlinks * Moving test check macros to separate header file * Tightening up shared memory unlinking for clique kernels, add munmap for host barrier for MP unit tests * Updating new MP unit test * Fixing mqueue bug * Fixing memory leak in MP unit tests
64 linhas
1.6 KiB
C++
64 linhas
1.6 KiB
C++
#ifndef TESTCHECKS_HPP
|
|
#define TESTCHECKS_HPP
|
|
|
|
#define HIP_CALL(x) ASSERT_EQ(x, hipSuccess)
|
|
#define NCCL_CALL(x) ASSERT_EQ(x, ncclSuccess)
|
|
|
|
#define SYSCHECK_TEST(call, name) do { \
|
|
int retval; \
|
|
SYSCHECKVAL_TEST(call, name, retval); \
|
|
} while (false)
|
|
|
|
#define SYSCHECKVAL_TEST(call, name, retval) do { \
|
|
SYSCHECKSYNC_TEST(call, name, retval); \
|
|
if (retval == -1) { \
|
|
printf("Call to %s failed : %s\n", name, strerror(errno)); \
|
|
fflush(stdout); \
|
|
return ncclSystemError; \
|
|
} \
|
|
} while (false)
|
|
|
|
#define SYSCHECK_GOTO_TEST(call, name, label) do { \
|
|
int retval; \
|
|
SYSCHECKVAL_GOTO_TEST(call, name, retval, label); \
|
|
} while (false)
|
|
|
|
#define SYSCHECKVAL_GOTO_TEST(call, name, retval, label) do { \
|
|
SYSCHECKSYNC_TEST(call, name, retval); \
|
|
if (retval == -1) { \
|
|
printf("Call to %s failed : %s\n", name, strerror(errno)); \
|
|
fflush(stdout); \
|
|
goto label; \
|
|
} \
|
|
} while (false)
|
|
|
|
#define SYSCHECKSYNC_TEST(call, name, retval) do { \
|
|
retval = call; \
|
|
if (retval == -1 && (errno == EINTR || errno == EWOULDBLOCK || errno == EAGAIN)) { \
|
|
} else { \
|
|
break; \
|
|
} \
|
|
} while(true)
|
|
|
|
#define NCCLCHECK_BARRIER_TEST(call, name, rank) do { \
|
|
ncclResult_t retval; \
|
|
retval = call; \
|
|
if (retval != ncclSuccess) { \
|
|
printf("Rank %d call to %s failed : %s\n", rank, name, strerror(errno)); \
|
|
fflush(stdout); \
|
|
return; \
|
|
} \
|
|
} while (false)
|
|
|
|
#define NCCLCHECK_TEST(call, name) do { \
|
|
ncclResult_t retval; \
|
|
retval = call; \
|
|
if (retval != ncclSuccess) { \
|
|
printf("Call to %s failed : %s\n", name, strerror(errno)); \
|
|
fflush(stdout); \
|
|
return retval; \
|
|
} \
|
|
} while (false)
|
|
|
|
#endif
|