Parallelize the exhaustive midpoint test across hardware threads (~75x faster)

[fcostaoliveira]

exhaustive32_midpoint sweeps all 2³² float bit-patterns one at a time, which takes
~30 min on a fast machine. The values are independent, so this splits the range across
std::thread::hardware_concurrency() threads, with an atomic flag for fail-fast. The
per-value work is factored into a check_word() helper; the checks and pass/fail
behavior are unchanged.

On a 96-core machine the runtime drops from ~1900 s to ~25 s (≈75×), identical result:

gcc 13  : PASS in 26s
clang 18: PASS in 22s

Notes:

• The exhaustive tests are gated behind FASTFLOAT_EXHAUSTIVE (off by default and not
  built in CI), so this has no CI impact — it's purely a faster local/dev sweep.
• tests/CMakeLists.txt now links Threads::Threads (with
  THREADS_PREFER_PTHREAD_FLAG).
• C++11, compiles clean under the existing -Werror -Wall -Wextra -Weffc++ -Wconversion -Wsign-conversion -Wshadow set on gcc and clang; clang-format clean.

The same pattern applies to the sibling sweeps (exhaustive32, exhaustive32_64); I
kept this PR to one test for review, happy to extend it if you'd like.