parallel pi computations using threads

2025-09-18 10:26:40 +02:00
commit d0586da7c1
1 changed files with 11 additions and 0 deletions
--- a/main.ua
+++ b/main.ua
@@ -0,0 +1,11 @@
+f  ← ˜÷1+1ⁿ2
+ι  ← +⟜(\+↯⤙(˜÷⌵-))        # linspace
+i! ← ÷⊃⧻/+ ^0 ι            # integrate numerically (mean function value)
+t! ← ⊙(&p×1e3⌵-)⊙now^0⊙now # time 1-ary function
+
+0 1 4                       # interval [a, b] with n partitions (threads)
+⊂⟜(\+↯⤙(˜÷⌵-))              # partition into n subintervals
+⊃(t!(⧈₂/(i!f⊙⊙1e8))         # sequential: integrate each area
+| t!(wait⧈₂/pool(i!f⊙⊙1e8)) # threading with n threads: this is faster
+)
+×4÷⊃⧻/+ ◌ # discard one of the benchmarked calculations, collect result