Explain GPU Register Spills and Their Performance Cost

What is a register spill in a GPU kernel, why does it happen, and how would you determine whether it is hurting performance? ### Constraints & Assumptions Discuss compiler-managed per-thread registers and the storage used when values cannot remain in registers. Do not confuse a spill with explicitly declared shared/local scratchpad memory; terminology differs across programming systems. ### Clarifying Questions What is the kernel's register usage and live-variable pressure? Is occupancy limited by registers? Does the compiler report spills? What memory traffic and stalls appear in profiling? ### What a Strong Answer Covers Live ranges, finite register allocation, spill loads/stores, occupancy tradeoffs, and changes justified by measurements. ### Follow-up Questions Can forcing a lower register limit make a kernel slower? How can unrolling or large local arrays affect pressure? When might recomputation be cheaper than preserving a value?