Talk:CalmoSoft primes
Raku incorrect
I'm getting
For primes up to two hundred and fifty: The following sequence of 49 consecutive primes yields a prime sum: 11 + 13 + 17 + 19 + 23 + 29 +..+ 223 + 227 + 229 + 233 + 239 + 241 = 5,813
which disagrees with the Raku output. For ease of reference, the relevant Raku output is
For primes up to two hundred fifty: Longest sequence of consecutive primes yielding a prime sum: elements: 47 7 + 11 + 13 + 17 + 19 + 23...199 + 211 + 223 + 227 + 229, sum: 5,107 11 + 13 + 17 + 19 + 23 + 29...211 + 223 + 227 + 229 + 233, sum: 5,333
Sadly, I don't think there is an OEIS entry I can check this against (tee hee) --Petelomax (talk) 02:48, 9 April 2023 (UTC)
- I agree with your result for primes < 250 and also disagree with some of the other Raku results. Your 2 second target for the stretch goal may be a bit ambitious for the interpreted languages as it takes Wren 2.6 seconds to sieve for primes up to 50 million before even starting to figure out the longest CalmoSoft prime sequence! --PureFox (talk) 10:29, 9 April 2023 (UTC)
- I also agree your 250 prime result. I also agree that interpreted languages may struggle to do it in 2 seconds. See the notes on the Algol 68 stretch sample. --Tigerofdarkness (talk) 10:53, 9 April 2023 (UTC)
- You are right - I've added a fixed version. - the sum is now 72618848632313, which agrees with the Phix sample.
- Sadly takes somewhat longer than the Wren sample - does the Wren interpreter us a JIT compiler ? --Tigerofdarkness (talk) 13:26, 9 April 2023 (UTC)
- No, there's a single pass compiler which parses the Wren source directly to bytecode (no intermediate AST) which the VM then interprets. A number of devices are used to make Wren faster though I can't see JIT compilation happening as it would make the VM too complicated. --PureFox (talk) 14:03, 9 April 2023 (UTC)