[LLVMbugs] [Bug 1210] Constant and Type uniquing should use a DenseMap/FoldingSet

[bugzilla-daemon at llvm.org]

http://llvm.org/bugs/show_bug.cgi?id=1210

Meador Inge <meadori at gmail.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|RESOLVED                    |REOPENED
         Resolution|FIXED                       |

--- Comment #40 from Meador Inge <meadori at gmail.com> 2012-02-21 23:25:57 CST ---
I am reopening this until the GCC 4.4 strict aliasing problem with the Hashing
implementation gets worked out
(http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20120220/137516.html).

As for the benchmarks, they are relative to the baseline.  I was surprised too.
 I think for many cases in practice we don't hit the worst case for key
comparisons (i.e. even with large vector comparisons we find inequality
relatively quickly and we don't end up scanning long chains in the hash
buckets).  The memory improvements were very obvious.  I will drill down into
the runtime aspects a little more to see if anything pops out.

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.