Disable cmd shrinking for int64 ref Thread test

[jmid]

We are seeing long src/neg_tests/lin_tests_thread_ref.exe causing timeouts and long runs, primarily on Cygwin part 2 (examples below).

Since the experimental Lin Thread mode does not produce test failures as consistently as the Domain mode, this little PR disables cmd argument shrinking for the test, to limit how many shrinking candidates we have to traverse.

I've run focused tests which shows runs limited to at most 771.6s on Cygwin with the fix:

• https://github.com/ocaml-multicore/multicoretests/actions/runs/6066711863/job/16457611765 5.1
• https://github.com/ocaml-multicore/multicoretests/actions/runs/6066711878/job/16457612074 trunk
  Interesting observation unrelated to this PR: the first (positive test) goes from 600-800s on 5.1 to about 1200s on trunk (which causes the former not to complete 20 iterations before the 6h timeout)

Examples of timeouts and long runs:

• https://github.com/ocaml-multicore/multicoretests/actions/runs/6049285277/job/16426696553
  timeout during 9221.8s Lin ref int64 test with Thread (shrinking: 22.0013)
• https://github.com/ocaml-multicore/multicoretests/actions/runs/6042037294/job/16404453112
  long run 5144.7s Lin ref int64 test with Thread
• https://github.com/ocaml-multicore/multicoretests/actions/runs/6026531927/job/16358872865
  timeout during 11039.8s Lin ref int64 test with Thread (shrinking: 50.0002)
• https://github.com/ocaml-multicore/multicoretests/actions/runs/6026072299/job/16356758666
  long run 2381.5s Lin ref int64 test with Thread
• https://github.com/ocaml-multicore/multicoretests/actions/runs/6026072300/job/16358585587
  timeout during 8343.3s Lin ref int64 test with Thread (shrinking: 38.0002)
• https://github.com/ocaml-multicore/multicoretests/actions/runs/6025647317/job/16355293724
  long run 3214.6s Lin ref int64 test with Thread

[jmid]

CI summary

• MingW bytecode 5.1 + trunk fails threadomain timeouts [ocaml5-issue] Windows failures on threadomain #203
• linux-ppc64le-5.2 STM Array and Bytes test SEGVs [ocaml5-issue] Crashes and hangs on ppc64 trunk/5.2 #380
• macos-arm64-5.1 fails to find 5.1.0~rc2 Mac ARM64 5.1.0~rc2 upgrade failure ocurrent/multicoretests-ci#23
• macos-arm64-5.2 unexpectedly found a Sys counterexample

That's 5 failures out of 55 (46+9) with only one being CI noise and no timeouts due to this PR.
I'll go ahead and merge.