Conversation
- I don't have gcc, so I'm not too sure how well this'll compile
- I used McDougall/Wotherspoon Newton-Raphson Method because I wasn't quite sure how to code the inline assembler equivalent
` _asm {
fild i
fsqrt
}`
- mostly revised for the odd j, and because i wanted to try and code `sqrt(i)`
split McDougall to run once/first, then Raphson thereafter
less opps, same diff.
|
Hi there! I'm working on compiling the program right now, but I'm having some issues 😝 I'll get back to you once I can confirm - but this program does compile. |
|
It works, but it actually seems to be a bit slower than the current version (not sure since I just tested it by hand, I didn't actually time it). Also, maybe you should try squashing your commits so that there aren't 10 1-line changes 😄 (@IanSeyler can also use this) |
|
sorry still a bit new to the community, is it still possible to squash these via the website now that i've done a pull request? or do i need to check them out again to another branch squash them via commandline and redo the request? also note the multi-threaded version is likely to be slower overall, because the original skipped over the processes==3 step. but it seems odd to me that the updated algo runs slower since it reduces the loop count by roughly a half... |
2 participants
I think these updates should be ok, but I'm going to need someone with gcc to try and compile them to confirm.