Closed GravisZro closed 1 month ago
Hi @GravisZro,
Yes, this remake is not following the original proposal and does not match the Dlang implementation. It's not ideal and I'll need to review the current implementation.
This current implementation was a response to the initial proposal below. https://github.com/kassane/sieve-cache-cpp/commit/20cdbd1c8758319364db6f7afe4e3f7d0b98e349
The Benchmark previously carried out based on the old implementation, resulting poor performance with std::unordered_map
.
My suggestion is to revert to the poorly performing version that worked correctly and then modify it. Slow is always better than non-functional.
My suggestion is to revert to the poorly performing version that worked correctly and then modify it. Slow is always better than non-functional.
Ok! I do. https://github.com/kassane/sieve-cache-cpp/commit/f9825c9d40d49dd51322225c614eeaeba96a3331
This doesn't implement SIEVE correctly.
Specifically, I see no indication that...