Size of .NET runtime and Compiler and Runtime speed

[Roger Wright]

I keep reading the term "managed" code - can you point me to an article that explains the concept? I've always thought that I managed my code fairly well... it's all behaved nicely and didn't pick on other processes..

[Nish Nishant]

Roger Wright wrote: I keep reading the term "managed" code - can you point me to an article that explains the concept? I've always thought that I managed my code fairly well... it's all behaved nicely and didn't pick on other processes.. LOL Managed code basically implies auto-garbage-collected code. Nish My most recent CP article :- A newbie's elementary guide to spawning processes www.busterboy.org

[Roger Wright]

Oh. Like Java, only harder to understand...

[Nish Nishant]

http://msdn.microsoft.com/vstudio/techinfo/articles/upgrade/managedext.asp Nish My most recent CP article :- A newbie's elementary guide to spawning processes www.busterboy.org

[Roger Wright]

Thanks!

[Nish Nishant]

Roger Wright wrote: Oh. Like Java, only harder to understand... The idea behind it is to allow you to new stuff without having to delete them. Nish My most recent CP article :- A newbie's elementary guide to spawning processes www.busterboy.org

[Roger Wright]

That sounds like a good thing! I got the .NET SDK, the MDAC2.7 update, the Platform SDK, and all the security fixes for IIS and Win2K this weekend, so I guess I'd better start learning:laugh:

[Nish Nishant]

Roger Wright wrote: so I guess I'd better start learning "Never, too late they say. That's why I never began ;-)" Nish My most recent CP article :- A newbie's elementary guide to spawning processes www.busterboy.org

[Kannan Kalyanaraman]

Nish [BusterBoy] wrote: Around 19 MB I think. Most machines would have it soon enough Nish, the 19mb thing is a installers size. I have a doubt whether after installing it would take say somewhere around 80 mb or so. Any idea on this. regards Kannan

[Nish Nishant]

Kannan Kalyanaraman wrote: Nish, the 19mb thing is a installers size. I have a doubt whether after installing it would take say somewhere around 80 mb or so. Any idea on this Hmmm. Yeah you are right there. In fact I'd guess it would be more like 50-60 rather than 80. Nish My most recent CP article :- A newbie's elementary guide to spawning processes www.busterboy.org

[Jorgen Sigvardsson]

Ewwwwww, the asymmetry! My philosophy: Either you allocate it and deallocate it later, or you don't allocate at all! :D Sonorked as well: 100.13197 jorgen

[Daniel Pratt]

Nish [BusterBoy] wrote: CodeProjectSQ wrote: 2. Does anyone have a comment as to whether the .NET runtime slows down their .exe programs? Managed code runs slower than unmanaged code. This is generally true, but its not a simple as that. Typically a performance comparison between managed and unmanaged code would be dominated by the fact that managed code has more overhead (initialization of variables, bounds-checking, run-time security checks, etc.). Furthermore the unmanaged C++ compiler is probably a lot better at optimizations than the managed/JIT compilers at this early stage of .NET. On the other hand, managed code has a couple advantages over unmanaged code. For example, object memory allocation (heap) will probably be faster in managed code. Because managed code uses a garbage collection scheme the heap is never fragmented. New objects always go on the top of the heap. In the future managed code will be able to take advantage of the fact that it is compiled on-site by making code optimizations specific to the processor/machine that it is deployed on. Additionally, IL (and C#) supports "unsafe" code. Unsafe code is still technically managed code, but it allows the use of pointers and all that implies. For specific tasks this can make a huge difference in performance v.s. "safe" code. Finally, if we accept that managed code will generally run slower than unmanaged code, then the next logical question is how much slower. I'm not a C++ coder by trade, so I'm not qualified to answer that question, but I get the impression that the difference in most cases is not something that end users would notice. The following quote is from an ng post by Ronald Laeremans: "While you may be able to construct a synthetic test that shows performance [of managed C++ code] for a very specific case to be 7x slower [than unmanaged C++ code], that is far from the typical results we have been getting. More typical would be performance differences within the single digit or low double digit percent range." Regards, Dan

[CodeProjectSQ]

Thanks!