On Sun, Feb 23, 2014 at 12:20:02PM -0500, Alan Grimes wrote:
David T. Lewis wrote:
The major issue was that the refactoring of object memory and interpreter into separate class hierarchies (which is a very good thing IMHO) requires the use of accessor methods, and this leads to name conflicts in the generated code if those accessor methods are not fully inlined.
YAY!!!!
I did a half-assed attempt at that ten years ago....
Eliot gets credit for the original refactoring in the Cog VMMaker, although I have since extended his work somewhat in VMM trunk to more clearly separate the "classic" object memory and NewObjectMemory, and the stack and context interpreters.
Dave
My thinking was to run an Interpreter for each physical CPU (I had a SMP machine back then).
Now that we're starting the process of moving away from x86, there are could be specialized interpreters for different processors in the system and, possibly, several Object Memories for things such as the GPU memory area... Then you get Xeon Phi or NUMA type machines...
-- IQ is a measure of how stupid you feel.
Powers are not rights.