Michael Schimmel
9072cfaa14
This commit updates the benchmark timings and repeat counts for various example files. The `tester.rs` script has been modified to: - Correctly parse and use the `Benchmark-Repeat` directive. - Implement an adaptive sampling mechanism for more accurate median calculation, especially for long-running benchmarks. - Improve the logic for updating and inserting benchmark and repeat lines in example files. - Handle potential compilation and runtime errors during benchmark execution.
18 lines
420 B
Plaintext
18 lines
420 B
Plaintext
;; Takeuchi Function Benchmark
|
|
;; heavily recursive, benefits significantly from specialization of integer arithmetic
|
|
;; and direct function calls (skipping dynamic dispatch).
|
|
;; Output: 5
|
|
|
|
;; Benchmark: 461.4us
|
|
;; Benchmark-Repeat: 6
|
|
|
|
(do
|
|
(def tak (fn [x y z]
|
|
(if (<= x y)
|
|
z
|
|
(tak (tak (- x 1) y z)
|
|
(tak (- y 1) z x)
|
|
(tak (- z 1) x y)))))
|
|
|
|
(tak 12 8 4)
|
|
)
|