Don Scott wrote:I guess my thought process was that if the JVM implemented multi-threading already (and is implemented in C++)-what will having C++ with this capability change anything? Will the JVM be more portable across OS since the mutli-threading capabilities will be part of the C++ code vs having to write more native files to deal with certain OS threading differences?
It just seems weird how C++ was not mutli-threaded but used to implement a mutli-threaded language. As such, now that it is itself is multi-threaded-does that impact the implementation of the JVM at all?
If you want to get into speed/performance area, ie how C is faster than Java, etc-which I don't want to get into-I wonder if adding mutli-threading to C++ will improve the speed of these benchmarks even more.
Okay, I think I get where you're coming from.
In the sense of what's possible to implement, nothing much has changed. It's never been that C or C++
as programming languages have supported multi-threading but the operating systems and libraries available for those operating systems. And then we have compilers that may or may not support multi-core systems etc. This is one of the reasons why it's been difficult to implement concurrent programs – everything is different between different operating systems, different threading libraries, different compilers, etc.
With that said, even if the capabilities don't change, bringing multi-threading capabilities into the standard libraries (or even into a programming language itself through new syntax or notation) does tend to lead to better performing applications – somebody else has figured out how to do concurrent stuff correctly so the application programmer is left with less opportunities to shoot himself in the foot, so to speak. (Not that it can't be done, still.)