Implement parallel marking (#48600)

Using a work-stealing queue after Chase and Lev, optimized for
weak memory models by Le et al.

Default number of GC threads is half the number of compute threads.

Co-authored-by: Gabriel Baraldi <baraldigabriel@gmail.com>
Co-authored-by: Valentin Churavy <v.churavy@gmail.com>

src/cluster.jl

@@ -1331,7 +1331,10 @@ function process_opts(opts)
     end
 
     # Propagate --threads to workers
-    exeflags = opts.nthreads > 0 ? `--threads=$(opts.nthreads)` : ``
+    threads = opts.nthreads > 0 ? `--threads=$(opts.nthreads)` : ``
+    gcthreads = opts.ngcthreads > 0 ? `--gcthreads=$(opts.ngcthreads)` : ``
+
+    exeflags = `$threads $gcthreads`
 
     # add processors
     if opts.nprocs > 0