Raul Murillo Montero
Raul Murillo Montero
Since the latest standard for posits (https://github.com/posit-standard/Posit-Standard-Community-Feedback) fixes the exponent size to 2 bits, it would be desirable to have fast specialization of posit types with 2 exponent bits (at...
Navigating through the source code, I found 2 different implementations of the Google Brain floating point format. One is a specialization of cfloat (https://github.com/stillwater-sc/universal/blob/e6ef6d76d2012c0261c6b480a37b78b30470d419/include/universal/number/cfloat/cfloat.hpp#L97C7-L97C13), while the other is defined as...
Hello everyone, I am trying to improve the performance of the following kernel: ```c #define N_I 20 #define N_J 20 void kernel_hadamard(int ni, int nj, float alpha, float beta, float...
When setting the **posit environment (8,0)**, there is a small interval (0, (1/(64^2)+1/64)/2) where the posits take a rounding value greater than 0, which is not consistent with the Posit...
According to the Posit notebook, is the posit environment which supports the following fused operations, not the quires. However, some of the fused operations from posits would recall any of...
The names of the variables (`a`, `b`) should match with the ones at the function headers (`p1`, `p2`). https://github.com/mightymercado/PySigmoid/blob/f77b8e2a57c7bb758f689837770a096505280532/PySigmoid/Quire.py#L43-L53