UtilLoadTest_0 fails due to SolverTest-related timeout

**Describe the bug**
The system test times out, with an instance of the SolverTest hogging each of the threads.

With tests chosen at random in that kind of load test, the tests that *don't* overrun are completed quickly, and replaced by another test from the bucket. So, eventually, all of the concurrent threads are held by one of the test types that *does* overrun, which is what's happened here.

**To Reproduce**
[Test_openjdk8_hs_extended.system_s390x_linux #256/UtilLoadTest_0](https://trss.adoptopenjdk.net/output/test?id=5f1a07cc0bec7975aef8ace0)

**Additional context**
```
[2020-07-23T19:44:52.150Z] ULT "load-2" #10 prio=5 os_prio=0 tid=0x000003ffb047ade0 nid=0xa2b runnable [0x000003ff9de52000]
[2020-07-23T19:44:52.150Z] ULT java.lang.Thread.State: RUNNABLE
[2020-07-23T19:44:52.150Z] ULT at net.adoptopenjdk.test.util.solver.SolverTest.processNextCard(SolverTest.java:160)
[2020-07-23T19:44:52.150Z] ULT at net.adoptopenjdk.test.util.solver.SolverTest.processNextCard(SolverTest.java:175)
[2020-07-23T19:44:52.150Z] ULT at net.adoptopenjdk.test.util.solver.SolverTest.processNextCard(SolverTest.java:175)
[2020-07-23T19:44:52.150Z] ULT at net.adoptopenjdk.test.util.solver.SolverTest.processNextCard(SolverTest.java:175)
[2020-07-23T19:44:52.150Z] ULT at net.adoptopenjdk.test.util.solver.SolverTest.processNextCard(SolverTest.java:175)
[2020-07-23T19:44:52.150Z] ULT at net.adoptopenjdk.test.util.solver.SolverTest.processNextCard(SolverTest.java:175)
[2020-07-23T19:44:52.150Z] ULT at net.adoptopenjdk.test.util.solver.SolverTest.processNextCard(SolverTest.java:175)
[2020-07-23T19:44:52.150Z] ULT at net.adoptopenjdk.test.util.solver.SolverTest.processNextCard(SolverTest.java:175)
[2020-07-23T19:44:52.150Z] ULT at net.adoptopenjdk.test.util.solver.SolverTest.processNextCard(SolverTest.java:175)
[2020-07-23T19:44:52.150Z] ULT at net.adoptopenjdk.test.util.solver.SolverTest.processNextCard(SolverTest.java:175)
[2020-07-23T19:44:52.150Z] ULT at net.adoptopenjdk.test.util.solver.SolverTest.processNextCard(SolverTest.java:175)
[2020-07-23T19:44:52.150Z] ULT at net.adoptopenjdk.test.util.solver.SolverTest.processNextCard(SolverTest.java:175)
[2020-07-23T19:44:52.150Z] ULT at net.adoptopenjdk.test.util.solver.SolverTest.processNextCard(SolverTest.java:175)
[2020-07-23T19:44:52.150Z] ULT at net.adoptopenjdk.test.util.solver.SolverTest.testSolution(SolverTest.java:78)
```

Start with SolverTest.java, especially processNextCard at lines 160 and 175 and testSolution at line 78, then reproduce the failure using the linked UtilLoadTest_0 system-test result. Done means the SolverTest-related overrun no longer causes all concurrent load-test threads to remain occupied until timeout.

Written by the indexing model from the issue text.