touisteur

Thanks a lot to both of you, I have everything I need. Le jeu. 23 juin 2022 à 05:53, Haicheng Wu ***@***.***> a écrit : > We have many f64...

Hi @Ravenwater, very nice of you to answer so quickly and with such details. I'm (sadly) not a subject matter expert and feel I wouldn't do justice to the task....

Heh I did try my hand with sollya for a similar topic (I want to take into account some information I have about the input of the log or log1p,...

Oh thanks! I had no idea rlibm had a posit32 implementation. I need to dive in and check how it fares against float and double logs from libcs... Edit: seems...

Oh yeah I just re-read the code and fell on that double accumulator, errrr.

> Well, not quite: it still uses doubles for compute, and maps back to posit on the final conversion. > > ``` > what is different is the polynomial: just...

I am probably going to end up trying fpm for log/log1p computation. I know it's fixed point, but it allows defining 'where' the 'fixed point' is. So if I know...

Hi, Yes indeed. I'll have to I think. Fpm would be a stepping stone only. Not even sure it does the job! Le lun. 6 sept. 2021 à 13:31, Theodore...

That is very nice of you. Please PLEASE take your time, there's no emergency on my side and it may take me some time afterwards to check the result in...

Hi thanks for all the kind and interesting answers! My use case is : I'm given a (sometimes) huge list of blocks. Each block is a variable-sized ByteBuffer that, once...