[opt] [async] Fix constant folding in async mode

[xumingkuan]

Related issue = #954 #2032

Kernel::Kernel() does not support multi-threads unless we lock the following part with a mutex:

  taichi::lang::context = std::make_unique<FrontendContext>();
  ir = taichi::lang::context->get_root();

  {
    CurrentKernelGuard _(program, this);
    program.start_function_definition(this);
    func();
    program.end_function_definition();
    ir->as<Block>()->kernel = this;
  }

This PR simply locks get_jit_evaluator_kernel() monolithically to prevent two threads from calling Kernel::Kernel() concurrently.

[Click here for the format server]

---

[xumingkuan]

Off-topic change: num_pending_tasks() matches the following Ended up with {} nodes better.

[k-ye]

Great, thanks!

---

I'll merge to see if this fixes problems on my end.

[codecov]

Codecov Report

Merging #2034 (59bdedc) into master (3e2d3e6) will not change coverage.
The diff coverage is n/a.

@@           Coverage Diff           @@
##           master    #2034   +/-   ##
=======================================
  Coverage   43.54%   43.54%           
=======================================
  Files          45       45           
  Lines        6267     6267           
  Branches     1110     1110           
=======================================
  Hits         2729     2729           
  Misses       3366     3366           
  Partials      172      172           

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update 3e2d3e6...59bdedc. Read the comment docs.