Charles said:
kettch said:
*snip*
We will continue to do both. There is not, nor has there ever been, an EitherOr monad, just a Both 
Sometimes, the studio is the best place to shoot demo-heavy pieces, lectures and conversational pieces. Sometimes, the office is the best place (like today's impromptu Archivist interview). Sometimes, Michael comes with us to offices to shoot high quality
OldSchool (like the IE team interviews, Larry's NUI pieces, etc). Since you want more of this, and we asked you what you want, we will deliver.
Plurality is a good thing!
Keep on posting,
C
I would really like to know more about the compilation process of managed code. Specifically how it relates to optimizations. Both for the language compiler (C#) but more importantly the jit compiler.
- How does it determine when to inline
- How does it determine when to unroll loops
- Does it insert mmx/sse code, and if so, when and how. How about newer extensions, AVX etc.? How about a SIMD extension to the language like in mono. Simple stuff, like when I do a div, does it use x87 code or mmx? how is it determined.
I always kinda thought it weird that the c# compiler itself apparently does very little optimization and the bulk of it is left to the jitter, which makes sense for platform specific optimizations, but even so, perhaps the c# compiler should simply provide
more time consuming analysis of the source in form of the metadata to the jitter. What I'm getting at is that the c# compiler seems too damn fast to me. Make it Slower! Also jitter seems fast, but dont make that one slower, hehe. Engage 
Other than that, more Kernel stuff. How about a statement from the recent W8 leak, lol, though I doubt MS would like to comment on that.
Oh, also, get Eric Lippert on camera. I'd love to see a lecture series with him.
Thread Closed
This thread is kinda stale and has been closed but if you'd like to continue the conversation, please create a new thread in our Forums,
or Contact Us and let us know.