# SystemException[“MemoryAllocationFailure”] with DSolve and using TimeConstrained

Report to WRI CASE:4238154

There is a similar question where SystemException["MemoryAllocationFailure"] is generated by DSolve here memoryallocationfailure-in-dsolve but that due to user using real number and not exact in the ODE.

I am getting SystemException["MemoryAllocationFailure"] where there are no real values in it. And the problem seem to happen when wrapping the call by TimeConstrained to limit the time used, which is very strange.

From a clean kernel

pde = a*D[w[x, y], x] + b*Sinh[lambda*x]*D[w[x, y], y] ==
c*Sinh[mu*x]^m + s*Sinh[beta*y]^k;
sol = TimeConstrained[DSolve[pde, w[x, y], {x, y}], 60*10]


And on my PC, after about only 3 minutes, I get

SystemException["MemoryAllocationFailure"]
Throw::sysexc: Uncaught SystemException returned to top level.
Can be caught with Catch[\[Ellipsis], _SystemException].


And again, from clean kernel, doing this

pde = a*D[w[x, y], x] + b*Sinh[lambda*x]*D[w[x, y], y] ==
c*Sinh[mu*x]^m + s*Sinh[beta*y]^k;
DSolve[pde, w[x, y], {x, y}]


Produces no error, but it hangs. I waited for 15 minutes and nothing happened. No crash and no result. So I terminate the kernel to stop it.

Why would using TimeConstrained cause this? I think this is a bug, but would like to know if others confirm it before I send bug report to WRI.

Using 11.3 on windows 10 pro, with 64 GB RAM.

• Well, it crashed my FE.... – Michael E2 Mar 24 '19 at 0:23