Home » Community » PR, media coverage, articles and documentation » Second CodeProject article
|
|
|
|
|
|
|
Re: Second CodeProject article [message #45853 is a reply to message #45851] |
Wed, 13 January 2016 15:10   |
cbpporter
Messages: 1427 Registered: September 2007
|
Ultimate Contributor |
|
|
mirek wrote on Wed, 13 January 2016 14:43cbpporter wrote on Wed, 13 January 2016 11:26Those benchmarks are really impressive and hard to beat. Great job!
Not sure that raw performance will help us convert a lot of new people over to U++ in this day and age, but it is good to have.
PS: I'm interested on how the benchmark looks with long strings (512+ characters) in the vector? Especially in the std one. Did they switch their implementation over to move internal elements?
Seriously, std::sort vs Sort with long strings is not as clear win (if I remember well, about 2x).
Anyway, both GCC and MSC now has moving std::string, but it is till much slower than plain nice memcpy... (I am speaking about e.g. Vector::Insert here).
Oh, BTW, now that you have brought it up: I have totally borrowed the idea of memcpy inside of container implementation from you.
Hope you don't mind .
|
|
|
Re: Second CodeProject article [message #45856 is a reply to message #45853] |
Wed, 13 January 2016 17:03  |
 |
mirek
Messages: 14255 Registered: November 2005
|
Ultimate Member |
|
|
cbpporter wrote on Wed, 13 January 2016 15:10mirek wrote on Wed, 13 January 2016 14:43cbpporter wrote on Wed, 13 January 2016 11:26Those benchmarks are really impressive and hard to beat. Great job!
Not sure that raw performance will help us convert a lot of new people over to U++ in this day and age, but it is good to have.
PS: I'm interested on how the benchmark looks with long strings (512+ characters) in the vector? Especially in the std one. Did they switch their implementation over to move internal elements?
Seriously, std::sort vs Sort with long strings is not as clear win (if I remember well, about 2x).
Anyway, both GCC and MSC now has moving std::string, but it is till much slower than plain nice memcpy... (I am speaking about e.g. Vector::Insert here).
Oh, BTW, now that you have brought it up: I have totally borrowed the idea of memcpy inside of container implementation from you.
Hope you don't mind .
It is open source, after all, is not it... 
So, you have started your own framework?
Mirek
|
|
|
Goto Forum:
Current Time: Sat Apr 26 01:23:50 CEST 2025
Total time taken to generate the page: 0.01992 seconds
|