Use LLVM memset intrinsic for compatibility and performance #795
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
The
runtime.Zeroinit
function isn't compatible withssa.NaiveForm
mode (fully disable SSA optimizations for debugging), because the compiler will generate recursive code.Another reason is that in the disassembly code,
runtime.Zeroinit
isn't inlined; it calls the C function_memset
. These should be replaced with theLLVM memset intrinsic
and be compiled to inlined code. See https://llvm.org/docs/LangRef.html#llvm-memset-intrinsicsc.Memset
maybe also need to be replaced with anllgo:link
instruction to rewrite it to thememset intrinsic
.Test code:
Before:
After:
The instruction
call void @llvm.memset(ptr %0, i8 0, i64 24, i1 false)
will be compiled to 3str
instructions.Calls
memset
on a large more memory range will be compiled like below: