[llvm-commits] CVS: llvm/lib/Target/README.txt

Chris Lattner lattner at cs.uiuc.edu
Thu Feb 2 22:21:55 PST 2006



Changes in directory llvm/lib/Target:

README.txt added (r1.1)
---
Log message:

target independent notes


---
Diffs of the changes:  (+67 -0)

 README.txt |   67 +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
 1 files changed, 67 insertions(+)


Index: llvm/lib/Target/README.txt
diff -c /dev/null llvm/lib/Target/README.txt:1.1
*** /dev/null	Fri Feb  3 00:21:53 2006
--- llvm/lib/Target/README.txt	Fri Feb  3 00:21:43 2006
***************
*** 0 ****
--- 1,67 ----
+ Target Independent Opportunities:
+ 
+ ===-------------------------------------------------------------------------===
+ 
+ FreeBench/mason contains code like this:
+ 
+ static p_type m0u(p_type p) {
+   int m[]={0, 8, 1, 2, 16, 5, 13, 7, 14, 9, 3, 4, 11, 12, 15, 10, 17, 6};
+   p_type pu;
+   pu.a = m[p.a];
+   pu.b = m[p.b];
+   pu.c = m[p.c];
+   return pu;
+ }
+ 
+ We currently compile this into a memcpy from a static array into 'm', then
+ a bunch of loads from m.  It would be better to avoid the memcpy and just do
+ loads from the static array.
+ 
+ ===-------------------------------------------------------------------------===
+ 
+ Get the C front-end to expand hypot(x,y) -> llvm.sqrt(x*x+y*y) when errno and
+ precision don't matter (ffastmath).  Misc/mandel will like this. :)
+ 
+ ===-------------------------------------------------------------------------===
+ 
+ For all targets, not just X86:
+ When llvm.memcpy, llvm.memset, or llvm.memmove are lowered, they should be 
+ optimized to a few store instructions if the source is constant and the length
+ is smallish (< 8). This will greatly help some tests like Shootout/strcat.c
+ and fldry.
+ 
+ //===---------------------------------------------------------------------===//
+ 
+ Solve this DAG isel folding deficiency:
+ 
+ int X, Y;
+ 
+ void fn1(void)
+ {
+   X = X | (Y << 3);
+ }
+ 
+ compiles to
+ 
+ fn1:
+ 	movl Y, %eax
+ 	shll $3, %eax
+ 	orl X, %eax
+ 	movl %eax, X
+ 	ret
+ 
+ The problem is the store's chain operand is not the load X but rather
+ a TokenFactor of the load X and load Y, which prevents the folding.
+ 
+ There are two ways to fix this:
+ 
+ 1. The dag combiner can start using alias analysis to realize that y/x
+    don't alias, making the store to X not dependent on the load from Y.
+ 2. The generated isel could be made smarter in the case it can't
+    disambiguate the pointers.
+ 
+ Number 1 is the preferred solution.
+ 
+ //===---------------------------------------------------------------------===//
+ 
+ 






More information about the llvm-commits mailing list