Bob Nystrom
1e842bbb2e
Cache the FNV-1a hash code in the ObjString struct at creation time, replacing the on-the-fly sampling hash in hashObject() and enabling constant-time string equality checks. Refactor string allocation into allocateString() and hashString() helpers, remove the old wrenNewUninitializedString() declaration, and add a CONST_STRING macro. Disable the unimplemented subscript range operator with a runtime error and skip the related UTF-8 tests. Add a string_equals benchmark to measure the speedup.
The benchmarks in here attempt to faithfully implement the exact same algorithm in a few different languages. We're using Lua, Python, and Ruby for comparison here because those are all in Wren's ballpark: dynamically-typed, object-oriented, bytecode-compiled.
A bit about each benchmark:
binary_trees
This benchmark stresses object creation and garbage collection. It builds a few big, deeply nested binaries and then traverses them.
fib
This is just a simple naïve Fibonacci number calculator. It was the first benchmark I wrote when Wren supported little more than function calls and arithmetic. It isn't particularly representative of real-world code, but it does stress function call and arithmetic.
for
This microbenchmark just tests the performance of for loops. Not too useful, but i used it when implementing for in Wren to make sure it wasn't too far off the mark.
method_call
This is the most useful benchmark: it tests dynamic dispatch and polymorphism. You'll note that the main iteration loop is unrolled in all of the implementations. This is to ensure that the loop overhead itself doesn't dwarf the method call time.