lumbda/examples/rpc-chain-bench.lsp
russell@unturf.com ccf86e3c3f rpc-chain-bench: Python → C relay → asm, timing end-to-end
Adds a transparent S-expression relay (examples/rpc-relay.lsp) plus a
sequential load generator (examples/rpc-chain-bench.lsp) and a bench
script (tests/rpc-chain-bench.sh) that wires them into multi-hop
chains across runtimes.

The relay is pure byte-forwarding: tcp-accept, tcp-recv, tcp-connect
to backend, tcp-send, tcp-recv reply, tcp-send back. Never parses.
Which is the point — S-expressions are the envelope.

Same rpc-relay.lsp runs as relay in any impl; chains are arbitrary
combinations of {Py, C, asm} nodes.

Measured (200 requests, ping, same laptop):

  (A) Py client → asm backend            direct, 1 hop   2061 rps
  (B) Py client → C relay → asm          2 hops          1234 rps
  (C) Py client → Py → C → asm           3 hops           766 rps
  (D) asm client → Py → C → asm          3 hops           796 rps

Per-hop cost ≈ 600-700 µs/request (TCP round-trip + context switch).

Safety: every server spawn used the six-layer pattern from CLAUDE.md
(ulimit -v 512MB + timeout 30 + trap + explicit kill + pgrep verify).
Four benchmark cells × up to 3 servers each = 10+ server spawns.
Zero strays, zero safety-net activations.
2026-04-17 09:18:42 -04:00

39 lines
1.3 KiB
Text

;;; rpc-chain-bench.lsp — time N sequential RPC calls to a given port.
;;;
;;; Used to benchmark multi-hop S-expression relays. The same .lsp runs
;;; in all three impls, so the *client* is also portable — any impl
;;; can drive any chain.
;;;
;;; Override *target-port* / *n-requests* / *request* before load.
;;; Default: hit :9080 with (ping) 200 times.
(define *host* "127.0.0.1")
(define *target-port* 9080)
(define *n-requests* 200)
(define *request* "(ping)")
(define (one-call)
(let ((s (tcp-connect *host* *target-port*)))
(tcp-send s *request*)
(let ((r (tcp-recv s 8192)))
(tcp-close s)
(if (and r (> (string-length r) 0)) 1 0))))
(define (call-loop n ok snap)
(if (= n 0) ok
(let ((got (one-call)))
(heap-restore snap)
(call-loop (- n 1) (+ ok got) snap))))
(define t0 (current-time-ms))
(define ok (call-loop *n-requests* 0 (heap-snapshot)))
(define t1 (current-time-ms))
(define elapsed (- t1 t0))
(display "port : ") (display *target-port*) (newline)
(display "requests : ") (display *n-requests*) (newline)
(display "ok : ") (display ok) (newline)
(display "elapsed : ") (display elapsed) (display " ms") (newline)
(display "rps : ")
(display (if (> elapsed 0) (quotient (* *n-requests* 1000) elapsed) 0))
(newline)