Vectordave - Add vectorization on x86/ARM

[davepl]

This adds vectorization approaches for SSE2, AVX, AVX512, and NEON. It's worth about 30% perf on the Mac and ubdellamd.

It also enables restrict in the hopes it keeps the compiler assured that sieve memory is not being aliased.

• I read the contribution guidelines in CONTRIBUTING.md.
• I placed my solution in the correct solution folder.
• I added a README.md with the right badge(s).
• I added a Dockerfile that builds and runs my solution.
• I selected drag-race as the target branch.
• All code herein is licensed compatible with BSD-3.

[rbergen]

This question applies to all main and nested while loops in this section of code: is there a particular reason for using while instead of for?

I know this is dangerously close to a coding style comment and I'm the first to say we shouldn't spend too much time on those, certainly in this project. However, I think combining the loop condition and the increment expression (which is neatly at the end of every loop) in one line clarifies the "cycle logic" in this particular case.

[davepl]

Good point - the structure required it the way I was doing the preprocessor stuff originally, but now it looks silly. Any cases that have init, condition, AND increment should be a for ! Thanks Dave

…

On Oct 26, 2025, at 6:27 AM, Rutger van Bergen ***@***.***> wrote: @rbergen commented on this pull request. In PrimeCPP/solution_5/PrimeCPP_array.cpp <#1024 (comment)>: > + while (wordIndex + cycleLen <= fullWordCount) + { + size_t idx = 0; + while (idx + 8 <= cycleLen) + { This question applies to all main and nested while loops in this section of code: is there a particular reason for using while instead of for? I know this is dangerously close to a coding style comment and I'm the first to say we shouldn't spend too much time on those, certainly in this project. However, I think combining the loop condition and the increment expression (which is neatly at the end of every loop) in one line clarifies the "cycle logic" in this particular case. — Reply to this email directly, view it on GitHub <#1024 (review)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AA4HCFY357O2RCD4CRVG5JD3ZSVZHAVCNFSM6AAAAACKGMZ53OVHI2DSMVQWIX3LMV43YUDVNRWFEZLROVSXG5CSMV3GSZLXHMZTGOBRGA2DCMBXGA>. You are receiving this because you authored the thread.

[davepl]

So I went through the code and did an AI refactor to for(), and it’s uglier :-). Most of the fors() can’t have the indexer in score because it’s used in multiple loops. You wind up with a lot of this, and I actually prefer the while()! - Dave 

…

On Oct 26, 2025, at 6:27 AM, Rutger van Bergen ***@***.***> wrote: @rbergen commented on this pull request. In PrimeCPP/solution_5/PrimeCPP_array.cpp <#1024 (comment)>: > + while (wordIndex + cycleLen <= fullWordCount) + { + size_t idx = 0; + while (idx + 8 <= cycleLen) + { This question applies to all main and nested while loops in this section of code: is there a particular reason for using while instead of for? I know this is dangerously close to a coding style comment and I'm the first to say we shouldn't spend too much time on those, certainly in this project. However, I think combining the loop condition and the increment expression (which is neatly at the end of every loop) in one line clarifies the "cycle logic" in this particular case. — Reply to this email directly, view it on GitHub <#1024 (review)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AA4HCFY357O2RCD4CRVG5JD3ZSVZHAVCNFSM6AAAAACKGMZ53OVHI2DSMVQWIX3LMV43YUDVNRWFEZLROVSXG5CSMV3GSZLXHMZTGOBRGA2DCMBXGA>. You are receiving this because you authored the thread.

[rbergen]

The screen snippet doesn't come through when you email respond to GitHub comments, but I imagine it's an example of a for loop with an empty first clause (for (; wordIndex + cycleLen <= fullWordCount; wordIdex += cycleLen) or for(; idx + 8 <= cycleLen; idx += 8)).

If that's right then I was aware of this, but I still think the looping logic itself is more clearly expressed in a for context. However, the fact I'm one of the maintainers of this project doesn't change it's your solution, so if you prefer the while, then that's what we'll merge.

[davepl]

I prefer the current while structure! So, LGTM* - Dave * let’s get this merged!

…

On Oct 26, 2025, at 10:49 AM, Rutger van Bergen ***@***.***> wrote: rbergen left a comment (PlummersSoftwareLLC/Primes#1024) <#1024 (comment)> The screen snippet doesn't come through when you email respond to GitHub comments, but I imagine it's an example of a for loop with an empty first clause (for (; wordIndex + cycleLen <= fullWordCount; wordIdex += cycleLen) or for(; idx + 8 <= cycleLen; idx += 8)). If that's right then I was aware of this, but I still think the looping logic itself is more clearly expressed in a for context. However, the fact I'm one of the maintainers of this project doesn't change it's your solution, so if you prefer the while, then that's what we'll merge. — Reply to this email directly, view it on GitHub <#1024 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AA4HCF36UCQX7RKOBCTIFU33ZT3SJAVCNFSM6AAAAACKGMZ53OVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZTINBYG4ZDQOJUGY>. You are receiving this because you authored the thread.