Idioma: Español
Fecha: Subida: 2022-05-06T00:00:00+02:00
Duración: 2h
Lugar: Espinardo - Facultad de Informática - Salón de Actos
Lugar: Directo
Intervalo:
desde 06/05/2022 11:00
hasta 06/05/2022 13:00
Visitas: 540 usuarios

C++ Old Dog, New Tricks

C++ Perro viejo, Nuevos Trucos

Descripción

The C++ language is around for over 40 years. A lot of profound software libraries were written with its help, e.g. the latest platforms TensorFlow and PyTorch for training of the deep learning networks, easily available via the topmost Python framework, in their lower layers are written in C++. However, in the last decade C++ has undergone an enormous refurbishment, not to say a revolution, which brought it into the top technological level among available programming languages. In this talk I’ll present an overview of these newest features of C++20 & 23, such as concepts, modules, ranges, co-routines and many more.

Propietarios

UMtv (Universidad de Murcia)

Publicadores

Jose Tomas Palma Mendez

Comentarios

Nuevo comentario
Eduardo Jose Gomez Hernandez
06/05/2022 13:13:32
Thanks you
Emilio Dominguez Sanchez
06/05/2022 13:11:28
Important advice for our university lecturers... :)
Eduardo Jose Gomez Hernandez
06/05/2022 13:08:36
Thanks for the answers
Emilio Dominguez Sanchez
06/05/2022 13:06:03
nowadays*
Diego Sevilla Ruiz
06/05/2022 13:05:56
Absolutely
Emilio Dominguez Sanchez
06/05/2022 13:05:37
Well, compilers do much more than humans now a day. We are better at optimizing algorithms (what needs to be done), as opposed to simplifying expressions, even more when code needs to be generated.
Eduardo Jose Gomez Hernandez
06/05/2022 13:04:25
But, in the other hand, if it is optimizable even better, you can figure it out and do it by yourself in C. My point here is that I feel that C++ depends even more in the compiler than C.
Emilio Dominguez Sanchez
06/05/2022 13:04:17
The real cost is in compilation time, yes.
Emilio Dominguez Sanchez
06/05/2022 13:02:45
Actually the features some times make it faster, because the compiler can recognize patterns and do things that cannot be done with C code because it would be making assumptions that could be false in some extreme cases.
Eduardo Jose Gomez Hernandez
06/05/2022 12:56:52
First of all, thanks for the presentation. I understand that new features make easier for programmers, but I feel that it also make compilers even more complex and generate more inefficient code. As an example objects and lambda expressions usually requires following more pointers to execute the expected code, while in C it will be a direct call to that code. What do you think?

Relaccionados