By: Michael S (already5chosen.delete@this.yahoo.com), April 1, 2021 11:02 am
Room: Moderated Discussions
Hi all
Last week or two I was busy researching (with no relationship to my real work) algorithm options of implementation of standard C library function strtod() on fast 64-bit CPUs.
I was interested both in average conversion time of typical inputs and in various weird worst cases.
The missing part is a broad picture of state of the art.
Unfortunately, right now I don't have a physical access to a lot of platforms and too lazy to try cloud options. So I'll greatly appreciate a help of the readers of this forum.
Test programs are available at github: https://github.com/already5chosen/strtod/tree/master/big
Brief test instruction are in how-to-test.txt
Naturally, I am most intrigued by aarch64, but various x86-64 measurements would be also appreciated.
In fact, right I didn't test anything except WIntel.
Thanks in advance,
Michael
Last week or two I was busy researching (with no relationship to my real work) algorithm options of implementation of standard C library function strtod() on fast 64-bit CPUs.
I was interested both in average conversion time of typical inputs and in various weird worst cases.
The missing part is a broad picture of state of the art.
Unfortunately, right now I don't have a physical access to a lot of platforms and too lazy to try cloud options. So I'll greatly appreciate a help of the readers of this forum.
Test programs are available at github: https://github.com/already5chosen/strtod/tree/master/big
Brief test instruction are in how-to-test.txt
Naturally, I am most intrigued by aarch64, but various x86-64 measurements would be also appreciated.
In fact, right I didn't test anything except WIntel.
Thanks in advance,
Michael