92/uiua: even faster with chunked threading and memoization

2025-10-07 09:03:10 +02:00
parent f4f8a9fb3f
commit 7223098285
1 changed files with 4 additions and 6 deletions
--- a/92/main.ua
+++ b/92/main.ua
@@ -1,7 +1,5 @@
-# next number in chain
-N ← /+⍉°√⊥10
+N ← memo(/+⍉°√⊥10) # memoize, because we get many repeated 1s

-⍜-⇡1 1e7
-⍥N10       # the longest 1-chain is less than 10
-⊙1e7 ⧻⊚=1 # count 1s to subtract from 1e7
-1         # there are 999999 nums, because no 0
+°△ 1e4_1e3    # chunk the input
+wait≡pool⍥₁₀N # thread each chunk of size 1e3
+-⊃(⧻⊚=1|-₁⧻♭)