Encoding stream benchmarks

TextEncoderStream and TextDecoderStream, under both implementations. Suite rules: src/bench/streams/AGENTS.md.

The shapes are 256 x 16, 64 x 4 Ki and 16 x 64 Ki characters or bytes (SHAPES in helpers.js): a stream of short chunks, like generated tokens, and longer ones. TEXT names the text: ASCII, and two- and three-byte characters in UTF-8.

Module map

Comparing

• Both TextDecoderStreams decode with TextDecoder's native decoder, so the decoder differences are in the streams around it.
• The TextEncoderStreams encode differently. The TypeScript one calls TextEncoder.prototype.encode() on each chunk, which writes UTF-8 through V8, or with simdutf under the workerd-autogate-enable-fast-textencoder autogate (@all-autogates); the C++ one converts each chunk with simdutf itself. The encoder differences therefore include the encoders'.
• The decoder's chunks are consecutive pieces of one encoded text. No chunk size is a multiple of three, so the three-byte chunks end mid-character and the decoder carries the partial character over to the next chunk.