Multithreading

Threads

• Process with own instruction/data
• Thread may be a part of a larger program or the whole program
• Each thread has all the state (instructions, data, PC, register state, etc.) necessary to allow it to execute

Types

Coarse-Grained

• Switches thread on costly stalls, such as L2 cache misses
• Advantages
  • Does not require very fast thread switching
  • Doesn’t slow down thread, since instruction from other threads issued only when the thread reaches a costly stall
• Disadvantage
  • Context switching is expensive
    • Is hard to overcome throughput losses from shorter stalls due to pipeline start up calls

Fine-Grained

• Switches between threads on each instruction, causing the execution of multiple threads to be interleaved
• Usually done in a round-robin fashion, skipping any stalled threads
• CPU must be able to switch threads each clock
• Advantage
  • It can hide both short and long stalls
  • Since instructions from other threads executed when one thread stalls
• Disadvantage
  • Slows down the execution of individual threads
• More expensive since it requires specialized hardware

Simultaneous Multithreading

• Requires fine grained ability
• Does not need multiple processors, just parallel execution units