diff options
author | Simon Marlow <marlowsd@gmail.com> | 2012-01-25 14:49:29 +0000 |
---|---|---|
committer | Simon Marlow <marlowsd@gmail.com> | 2012-01-25 14:49:29 +0000 |
commit | 5d5a01483643ea7b06189878975e67801950a8cc (patch) | |
tree | 82107effa506a1a14ad644f8517a692b3c1128e2 /compiler/cmm/cmm-notes | |
parent | 01903ebf23b0041dec721aec9a6d6ff168d2b804 (diff) | |
download | haskell-5d5a01483643ea7b06189878975e67801950a8cc.tar.gz |
delete some no-longer-relevant notes
Diffstat (limited to 'compiler/cmm/cmm-notes')
-rw-r--r-- | compiler/cmm/cmm-notes | 10 |
1 files changed, 0 insertions, 10 deletions
diff --git a/compiler/cmm/cmm-notes b/compiler/cmm/cmm-notes index 81882c8c0e..bbad34cfa5 100644 --- a/compiler/cmm/cmm-notes +++ b/compiler/cmm/cmm-notes @@ -81,10 +81,6 @@ Things to do: Old.Cmm. We should abstract it to work on both representations, it needs only to
convert a CmmInfoTable to [CmmStatic].
- - The MkGraph currenty uses a different semantics for <*> than Hoopl. Maybe
- we could convert codeGen/StgCmm* clients to the Hoopl's semantics?
- It's all deeply unsatisfactory.
-
- Improve performance of Hoopl.
A nofib comparison of -fasm vs -fnewcodegen nofib compilation parameters
@@ -101,18 +97,12 @@ Things to do: When compiling nofib, ghc-head + libraries compiled with -fnew-codegen
is 31.4% slower (http://fox.auryn.cz/msrc/0517_hoopl/32bit.oldghcoldgen.zipghcoldgen.txt).
- So we generate a bit better code, but it takes us longer!
-
EZY: Also importantly, Hoopl uses dramatically more memory than the
old code generator.
- Are all blockToNodeList and blockOfNodeList really needed? Maybe we could
splice blocks instead?
- In the CmmContFlowOpt.blockConcat, using Dataflow seems too clumsy. Still,
- a block catenation function would be probably nicer than blockToNodeList
- / blockOfNodeList combo.
-
- lowerSafeForeignCall seems too lowlevel. Just use Dataflow. After that
delete splitEntrySeq from HooplUtils.
|