Notable changes

• Implements memory recycling on a per-goroutine basis to reduce memory allocations and GC pressure.
• Speed improvement of approx. 18% in nodes per second (NPS) over v0.1.1.
• New UCI option to set number of search threads, defaulting to one search process (goroutine) per logical core.
• Expanded time control options:
  • per-game
  • per-game plus time increment
  • fixed-depth
  • fixed time per move
• Minor bug-fixes and optimizations