[mailto:carter.schonw...@gmail.com]
Sent: 11 December 2012 21:23
To: Simon Peyton-Jones
Cc: Johan Tibell; Greg Fitzgerald; cvs-ghc@haskell.org; GHC Users Mailing List
Subject: Re: Hoopl vs LLVM?
Cool info!
Would love to see that report if you can dig it up :)
-Carter
On Tue, Dec 11, 2012 at 2:16 PM, Simon
| > My question was more: why do we CPS transform? I guess it's because we
| > manage our own stack?
|
| Right. In fact, LLVM does its own CPS transform (but doesn't call it
| that) when the code contains non-tail function calls. We give LLVM code
| with tail-calls only.
|
| The choice about wh
On 12/12/12 17:37, Johan Tibell wrote:
On Wed, Dec 12, 2012 at 4:35 AM, Simon Marlow wrote:
On 11/12/12 21:33, Johan Tibell wrote:
I'd definitely be interesting in understanding why as it, like you
say, makes it harder for LLVM to understand what our code does and
optimize it well.
The exam
On 12/12/12 17:06, Greg Fitzgerald wrote:
On Wed, Dec 12, 2012 at 4:35 AM, Simon Marlow mailto:marlo...@gmail.com>> wrote:
Now, all that LLVM knows is that z was read from Sp[8], it has no
more information about its value.
Are you saying Hoopl can deduce the original form from t
On Wed, Dec 12, 2012 at 4:35 AM, Simon Marlow wrote:
> On 11/12/12 21:33, Johan Tibell wrote:
>> I'd definitely be interesting in understanding why as it, like you
>> say, makes it harder for LLVM to understand what our code does and
>> optimize it well.
>
>
> The example that Simon gave is a good
On 12.12 09:06, Greg Fitzgerald wrote:
> On Wed, Dec 12, 2012 at 4:35 AM, Simon Marlow wrote:
>
> > Now, all that LLVM knows is that z was read from Sp[8], it has no more
> > information about its value.
>
>
> Are you saying Hoopl can deduce the original form from
lustration:
f() {
x = blah
z = blah2
p,q = g(x)
res = z + p - q
return res
}
In this function, for example, a Hoopl pass would be able to derive
something about the value of z from its assignment (blah2), an
On Tue, Dec 11, 2012 at 11:16 AM, Simon Peyton-Jones
wrote:
> Notice that the stack is now *explicit* rather than implicit, and LLVM has no
> hope of moving the assignment to z past the call to g (which is trivial in
> the original). I can explain WHY we do this (there is stuff on the wiki) but
Cool info!
Would love to see that report if you can dig it up :)
-Carter
On Tue, Dec 11, 2012 at 2:16 PM, Simon Peyton-Jones
wrote:
> | In my opinion we should only implement optimizations in Hoopl that
> | LLVM cannot do due to lack high-level information that we might have
> | g
| In my opinion we should only implement optimizations in Hoopl that
| LLVM cannot do due to lack high-level information that we might have
| gotten rid of before we reach the LLVM code generator*. I don't think
Indeed. And I think there is probably quite a lot that is in reach for C--,
agh
Date: Tue Jul 24 21:05:12 2012 +0100
Remove some old temporary warning suppression for hoopl warnings
>---
mk/validate-settings.mk |9 -
1 files changed, 0 insertions(+), 9 deletions(-)
diff --git a/mk/va
low
Date: Thu Jul 5 13:23:21 2012 +0100
Remove "fuel", adapt to Hoopl changes, fix warnings
compiler/cmm/BlockId.hs |2 +-
compiler/cmm/Cmm.hs |8 +-
compiler/cmm/CmmBuildInfoTables.hs| 11 +-
compiler/cmm/CmmCallConv.hs |
New patches in /srv/darcs/git-mirrors/hoopl
commit 1ef68d6f2a8a7e89a5ed4cc78e3fefc89dc8ccab
Author: Norman Ramsey
Date: Wed May 9 16:03:11 2012 -0400
instructinos for developers
commit a520508d8a02c80e0633aab0739a80301d486eef
Author: Norman Ramsey
Date: Wed May 9 15:58:28 2012 -0400
New patches in /srv/darcs/git-mirrors/hoopl
commit e578e938e1ff8dd1233bcf89a7c6e62ed4d892ac
Author: Norman Ramsey
Date: Sat Apr 21 12:52:13 2012 -0400
update version number and remove Stability: field from cabal file
commit 76c3f2566d53124772a5bb54db360a1db08d7e5d
Author: Norman Ramsey
New patches in /srv/darcs/git-mirrors/hoopl
commit ab1439dc866c5b9ca29b8f8209850cef4568082c
Author: Norman Ramsey
Date: Tue Mar 20 18:42:43 2012 -0400
patch from Paolo Capriotti, ignore some GHC build files
(I've changed the .gitignore to operate just in the root dire
age
Simon
| -Original Message-
| From: cvs-ghc-boun...@haskell.org [mailto:cvs-ghc-boun...@haskell.org] On
| Behalf Of Ian Lynagh
| Sent: 17 March 2012 15:56
| To: Simon Marlow
| Cc: cvs-ghc@haskell.org
| Subject: Re: [commit: hoopl] simonmar-hoopl-opt: Remove layers of newtype in
| Unique
On Thu, Mar 15, 2012 at 06:18:39AM -0700, Simon Marlow wrote:
>
> Remove layers of newtype in Unique and Label
>
> Improves performance due to eliminating some unnecessary maps in
> e.g. mapToList, setToList. This was quite a significant effect in
> GHC.
It seems a pity we h
On Mon, Aug 22, 2011 at 07:53:23AM +, Simon Peyton-Jones wrote:
>
> It's not clear that we need a darcs mirror for Hoopl any more. Shall we
> simply nuke it?
>
> Ian, just to check, you aren't using anything Darcs-ish for Hoopl, are you?
> http://hacka
New patches in /srv/darcs/git-mirrors/hoopl
commit b98db91cd0c53ddb2c275c04823f9c379774104b
Author: Simon Peyton Jones
Date: Mon Aug 22 08:47:05 2011 +0100
Add a kind annotation on Graph'
data Graph' block (n :: * -> * -> *) e x where
GNil :: Gr
Norman (and Ian)
I have just pushed a 1-line change to the master hoopl repo
linux.cs.tufts.edu:/r/c--/papers/dfopt.
But I got the error message below. I have no idea what it is saying, but I
suspect the problem is somewhere in the script that pushes the changes to the
(now defunct) darcs
New patches in /srv/darcs/git-mirrors/hoopl
commit 6ffca823d05780f8b1c0de598da814bf248d1c06
Author: David Terei
Date: Tue Aug 16 15:27:04 2011 -0700
Fix safe haskell issue with ghc 7.2.1
We can't rely on the containers package using Safe Haskell
yet, so must have XUt
New patches in /srv/darcs/git-mirrors/hoopl
commit 6ffca823d05780f8b1c0de598da814bf248d1c06
Author: David Terei
Date: Tue Aug 16 15:27:04 2011 -0700
Fix safe haskell issue with ghc 7.2.1
We can't rely on the containers package using Safe Haskell
yet, so must have XUt
New patches in /srv/darcs/git-mirrors/hoopl
commit 6ffca823d05780f8b1c0de598da814bf248d1c06
Author: David Terei
Date: Tue Aug 16 15:27:04 2011 -0700
Fix safe haskell issue with ghc 7.2.1
We can't rely on the containers package using Safe Haskell
yet, so must have XUt
l.org/trac/ghc/wiki/Repositories to
> specify this URL for pushing, and git://code.eecs.tufts.edu/hoopl/hoopl.git
> for pulling. If you want something different, do edit the wiki to say what
> you want
>
> | Attached is a patch for Hoopl for it to use Safe Haskell. Would be great
| > | This URL should set you right:
| > |
| > |linux.cs.tufts.edu:/r/c--/papers/dfopt.git
| >
| > You are sure? That is, "dfopt.git"?
OK I have updated http://hackage.haskell.org/trac/ghc/wiki/Repositories to
specify this URL for pushing, and git://code.eecs.
New patches in /srv/darcs/git-mirrors/hoopl
commit cf8723c8e7d83a12d6f95abffd58ccb78aa0e0cb
Author: Norman Ramsey
Date: Fri Jul 22 15:20:55 2011 -0400
added and exported liftFuel method of FuelMonadT
method runWithFuel remains hidden and should stay hidden
thanks Justin
New patches in /srv/darcs/git-mirrors/hoopl
commit cf8723c8e7d83a12d6f95abffd58ccb78aa0e0cb
Author: Norman Ramsey
Date: Fri Jul 22 15:20:55 2011 -0400
added and exported liftFuel method of FuelMonadT
method runWithFuel remains hidden and should stay hidden
thanks Justin
New patches in /srv/darcs/git-mirrors/hoopl
commit cf8723c8e7d83a12d6f95abffd58ccb78aa0e0cb
Author: Norman Ramsey
Date: Fri Jul 22 15:20:55 2011 -0400
added and exported liftFuel method of FuelMonadT
method runWithFuel remains hidden and should stay hidden
thanks Justin
New patches in /srv/darcs/git-mirrors/hoopl
commit cf8723c8e7d83a12d6f95abffd58ccb78aa0e0cb
Author: Norman Ramsey
Date: Fri Jul 22 15:20:55 2011 -0400
added and exported liftFuel method of FuelMonadT
method runWithFuel remains hidden and should stay hidden
thanks Justin
New patches in /srv/darcs/git-mirrors/hoopl
commit cf8723c8e7d83a12d6f95abffd58ccb78aa0e0cb
Author: Norman Ramsey
Date: Fri Jul 22 15:20:55 2011 -0400
added and exported liftFuel method of FuelMonadT
method runWithFuel remains hidden and should stay hidden
thanks Justin
New patches in /srv/darcs/git-mirrors/hoopl
commit e2a7d30117f865d933a9d1b71a2a8516e32e
Merge: 6a3637d... 0febb7f...
Author: Norman Ramsey
Date: Mon Jul 18 14:33:39 2011 -0400
Merge branch 'master' of linux.cs.tufts.edu:/r/c--/papers/dfopt
the 'validate&
d Z. Yang
Date: Tue Jun 28 08:20:50 2011 -0400
Another variant of hashStr, this one caught a Hoopl bug.
Signed-off-by: Edward Z. Yang
>---
tests/ghc-regress/codeGen/should_compile/all.T |1 +
.../cgrun
from Edward Z. Yang's message of Thu Apr 07 11:37:05 -0400 2011:
> I've been working on new optimization passes for Hoopl to get
> the code back to the quality of the old code generator.
> With the new optimization pass, the binaries in the test suite
> are, on average,
On 07/04/2011 16:37, Edward Z. Yang wrote:
I've been working on new optimization passes for Hoopl to get
the code back to the quality of the old code generator.
With the new optimization pass, the binaries in the test suite
are, on average, 7% larger than the old code generator.
Here ar
I've been working on new optimization passes for Hoopl to get
the code back to the quality of the old code generator.
With the new optimization pass, the binaries in the test suite
are, on average, 7% larger than the old code generator.
Here are the worst offenders, with T3294 clocking in w
se of the register. Here, the register is
+-- assigned multiple times, so it won't actually get inlined...
+
+type AssignmentMap = UniqFM (LocalReg, WithTop CmmExpr)
+
+-- ToDo: Move this into somewhere more general (UniqFM? That will
+-- introduce a Hoopl dependency on that file.)
+j
I've started looking at profiling Hoopl (I figure we should
speed it up before we start adding lots of optimization passes).
I was wondering if there was any useful current state here.
Some initial profiling of concurrent/prog001 indicates the following
(indented sections are shared betwee
low
Date: Thu Feb 3 20:57:05 2011 +
only the GHC repo is in git for now; add hoopl
packages.git | 81 +
1 files changed, 41 insertions(+), 40 deletions(-)
Diff suppressed because of size. To see it, use:
git s
hc-boun...@haskell.org [mailto:cvs-ghc-boun...@haskell.org]
>> *On Behalf Of *Alexander McPhail
>> *Sent:* 01 February 2011 09:10
>> *To:* cvs-ghc@haskell.org
>> *Subject:* Problem with hoopl x86_64 linux ghc 7.1
>>
>>
>>
>> ghc-7.1.20110131
>> lin
Thu Feb 3 12:57:05 PST 2011 Simon Marlow
* only the GHC repo is in git for now; add hoopl
M ./packages.git -40 +41
View patch online:
http://darcs.haskell.org/cgi-bin/darcsweb.cgi?r=ghc;a=darcs_commitdiff;h=20110203205705-12142-b53c042ceca76f461c4ec302416cdf9bdb051b27.gz
-- Forwarded message --
From: Vivian McPhail
Date: 2 February 2011 00:43
Subject: Re: Problem with hoopl x86_64 linux ghc 7.1
To: Simon Peyton-Jones
On 2 February 2011 00:37, Simon Peyton-Jones wrote:
> What version of ghc is in /home/ghc/bin/ghc?
>
>
>
$ ghc -
-boun...@haskell.org] On
Behalf Of Alexander McPhail
Sent: 01 February 2011 09:10
To: cvs-ghc@haskell.org
Subject: Problem with hoopl x86_64 linux ghc 7.1
ghc-7.1.20110131
linux x86_64
$ make dist-clean && perl boot && ./configure && make -j6
[..]
"/home/ghc/bin/ghc
irectory-1.1.0.0 -package filepath-1.2.0.0 -package hoopl-3.8.6.1 -package
hpc-0.5.0.6 -package old-time-1.0.0.6 -package process-1.0.1.4 -package
unix-2.4.1.0 -Wall -fno-warn-name-shadowing -fno-warn-orphans -XHaskell98
-XCPP -XMagicHash -XUnboxedTuples -XPatternGuards -XForeignFunctionInterface
-XEmp
This patch also needs a unyet-committed patch to Hoopl to fix some
warnings, or it will break the validate script (it builds fine normally).
Excerpts from Edward Z. Yang's message of Fri Jan 14 11:52:43 -0500 2011:
> 2 patches for repository http://darcs.haskell.org/ghc:
>
> Thu S
darcs thing?
Simon
| -Original Message-
| From: Edward Z. Yang [mailto:ezy...@mit.edu]
| Sent: 09 December 2010 16:13
| To: n...@cs.tufts.edu; d...@cs.tufts.edu; Simon Peyton-Jones
| Subject: darcs patch: Make Hoopl work with MonoLocalBinds.
|
| 1 patch for repository http://darcs.haskell.org
45 matches
Mail list logo
