|
From: <sv...@va...> - 2012-06-13 11:12:56
|
sewardj 2012-06-13 12:12:49 +0100 (Wed, 13 Jun 2012)
New Revision: 12635
Log:
Update with recent notes.
Modified files:
trunk/docs/internals/avx-notes.txt
Modified: trunk/docs/internals/avx-notes.txt (+19 -1)
===================================================================
--- trunk/docs/internals/avx-notes.txt 2012-06-13 12:12:06 +01:00 (rev 12634)
+++ trunk/docs/internals/avx-notes.txt 2012-06-13 12:12:49 +01:00 (rev 12635)
@@ -2,4 +2,22 @@
Cleanups
~~~~~~~~
-(none at present)
+* Important: iropt: Make sure XorV128 and XorV256 of identical
+ args gets folded to zero
+
+* add more iteration in test cases
+
+* math_UNPCKxPS_128: use xIsH ? InterleaveHI32x4 : InterleaveLO32x
+ I think this is safe w.r.t. the backend
+
+* math_UNPCKxPD_128: ditto
+
+* math_UNPCKxPD_256: split into 128 bit chunks and use math_UNPCKxPD_128
+
+
+Known limitations
+~~~~~~~~~~~~~~~~~
+
+* for many (all?) of the vector shift-by-imm cases (pre-existing as
+ well as AVX), out of range shifts are not handled properly and only
+ work I think because the host happens to have the same semantics.
|