Displaying 3 results from an estimated 3 matches for "kong98precise".
2007 Apr 03
0
[LLVMdev] Live Intervals vs. Live Variables
...llocLinearScan.cpp should be the best place to start learning LLVM from.
linearscan gives the best results among LLVM's allocators so it also should
be the first allocator to compete with :)
If you think about better register allocator I'd suggest you to look at
http://citeseer.ist.psu.edu/kong98precise.html and find a paper of
Sid-Ahmed-Ali Touati named "Register Saturation in Instruction Level
Parallelism". These works propose faster and better algorithms than graph
coloring, at least as I understand it.
Best regards,
Anton.
-------------- next part --------------
An HTML attachment w...
2007 Apr 03
2
[LLVMdev] Live Intervals vs. Live Variables
Toward a better register allocator, I'm attempting to understand
the dataflow information available to the allocator.
What's the difference between LiveInterval information and LiveVariable
information? If a LiveInterval is based on a linear ordering of
the machine instructions, isn't it rather conservative in nature?
Let's say I have a typical diamond CFG:
A
2007 Apr 03
3
[LLVMdev] Live Intervals vs. Live Variables
...d be the best place to start learning LLVM from.
> linearscan gives the best results among LLVM's allocators so it also should
> be the first allocator to compete with :)
>
> If you think about better register allocator I'd suggest you to look at
> http://citeseer.ist.psu.edu/kong98precise.html and find a paper of
> Sid-Ahmed-Ali Touati named "Register Saturation in Instruction Level
> Parallelism". These works propose faster and better algorithms than graph
> coloring, at least as I understand it.
>
> Best regards,
> Anton.
>