Open PetrKryslUCSD opened 2 years ago
Looks like a negative number cast to unsigned? It says "decrease" so I assume the calling code is computing a difference. Can you confirm whether free_memory
itself is returning a bad value?
Yes, the difference is between the reading of the free memory initially and the subsequent reading. So, yes, I suspect the subsequent memory reads return a negative value.
Has this been fixed in 1.8 or 1.9 by any chance?
At the moment tested only on Windows 10 with Julia 1.7.2.
At random intervals,
free_memory
will report totally wrong number of bytes (note the second invocation):In the third invocation of the script,
free_memory
breaks halfway through: