Adding cubic root
Dolorousrtur opened this issue · comments
Hey, thanks to the developers for this incredible library!
In my project, I need to precisely compute the cubic roots of double-precision variables. But it seems that warp does not have it implemented (warp.pow(x, 1/3) is not precise enough).
I reckon it shouldn't be too difficult to add it since it only needs to call cbrt function in CUDA. I tried to find a way to add cbrt to the source code of Warp same as it is done with sqrt, but couldn't do that yet.
Could you direct me to how (in which files) the sqrt support is implemented, s.t. I can add cbrt the same way?
Many thanks in advance!
Hi @Dolorousrtur, thanks for your request. cbrt will be in the next Warp release.
cbrt was included in the v1.0.0-beta.5 release: https://github.com/NVIDIA/warp/releases