Fri Jul 6 13:54:10 PDT 2007 Ian Lynagh <[EMAIL PROTECTED]>
* Remove foldl' from Util; use the Data.List one instead
M ./compiler/utils/FiniteMap.lhs +2
M ./compiler/utils/Util.lhs -11 +1
___
Cvs-ghc mailing list
Cvs-ghc@haskell.org
http://ww
Hi Simon,
> The attached charcount.zip shows the optimised cmm and the assembler.
> As far as I can tell, the optimised cmm for the hotspot function in
> both cases is identical, and the code should be spending all its time
> in the hotspot. The names are Main_zdspreludezu942zull85_entry for 6.7
Neil Mitchell wrote:
Hi Simon,
- 6.6.1 and 6.7 generate exactly the same Cmm
Yes, as far as I can tell.
- but their performance differs significantly,
- with -fasm or -fvia-C?
Both in the default configuration for -O2, then I redid the tests
using -fasm on both.
If -fasm, did y
Build description = HEAD on i386-unknown-linux
(cam-02-unx.europe.corp.microsoft.com)
Build location= /playpen/ghc/nightly/HEAD-cam-02-unx
Build config file = /home/simonmar/nightly/site/msrc/conf-HEAD-cam-02-unx
Nightly build started on cam-02-unx at Fri Jul 6 19:30:01 BST 2007.
checkin
Fri Jul 6 08:21:32 PDT 2007 Simon Marlow <[EMAIL PROTECTED]>
* undo: Get the path right for :list
This is breaking :list for someone. If there's a problem here, we
should fix it another way.
M ./compiler/ghci/InteractiveUI.hs -2 +1
___
Cvs-
On Jul 6, 2007, at 10:14 AM, Simon Marlow wrote:
Peter Tanski wrote:
I'm suggesting that instead of defining them in a .mk file and
plumbing them into GHC, that you just define in GHC somewhere
clOmitFramPointer = "-Oy"
the definitions would still all be in one place (definitely a good
th
Peter Tanski wrote:
On Jul 6, 2007, at 3:57 AM, Simon Marlow wrote:
Personally I'd do away with msvc_tools.mk and put it all into GHC.
You'd still have it all in one place, but there'd be much less
plumbing. Is there a reason you can't do that?
No reason I can't; I just don't want to. W
On Jul 6, 2007, at 3:57 AM, Simon Marlow wrote:
Peter Tanski wrote:
The two (proposed) additions are attached below: build.mk.msvc
does the same job as build.mk.sample: it defines presets. It
would probably be no trouble to combine build.mk.msvc presets with
msvc_tools.mk and conditionall
Fri Jul 6 05:14:50 PDT 2007 Ian Lynagh <[EMAIL PROTECTED]>
* Use subprocess.call rather than os.spawnv in the testsuite driver
os.spawnv is broken on non-cygwin Windows (looks like it goes via a shell).
M ./driver/testlib.py -5 +8
___
Cvs-ghc
Thu Jul 5 06:53:20 PDT 2007 Simon Marlow <[EMAIL PROTECTED]>
* -ddump-hi-diffs shouldn't imply -fforce-recomp, that defeats the purpose
M ./compiler/main/DynFlags.hs -1 +1
___
Cvs-ghc mailing list
Cvs-ghc@haskell.org
http://www.haskell.org/mailm
The HEAD now has a 'validate' script at the top level, which you can use for
testing that patches in your tree pass minimal testing. It won't work on
Windows yet (no args are passed to configure), and parallelism is fixed at -j2.
'make fast' in the testsuite now runs with zero failures on x86-
Hi Simon,
Is this the same as you and Simon have been subsequently discussing (stack,
heap checks), or is it another performance bug?
This is an entirely separate performance bug. The heap check makes one
of the two variants run 10% slower. This bug effects both variants,
and gives a 150% slo
Is this the same as you and Simon have been subsequently discussing (stack,
heap checks), or is it another performance bug?
Simon
| -Original Message-
| From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] On Behalf Of Neil Mitchell
| Sent: 05 July 2007 21:21
| To: cvs-ghc@haskell.org
| Sub
Fri Jul 6 02:53:31 PDT 2007 [EMAIL PROTECTED]
* drv015 should fial
M ./tests/ghc-regress/deriving/should_compile/all.T -1 +1
M ./tests/ghc-regress/deriving/should_compile/drv015.hs +9
A ./tests/ghc-regress/deriving/should_compile/drv015.stderr
_
Fri Jul 6 02:47:31 PDT 2007 [EMAIL PROTECTED]
* tcfail093 should succeed
M ./tests/ghc-regress/typecheck/should_fail/all.T -1 +1
M ./tests/ghc-regress/typecheck/should_fail/tcfail093.hs +3
___
Cvs-ghc mailing list
Cvs-ghc@haskell.org
http://
Fri Jul 6 02:25:48 PDT 2007 Michael D. Adams <[EMAIL PROTECTED]>
* Added tests for Cmm and CPS (they are currently skipped)
Also the test driver doesn't understand 'cmm' files
so something will have to be written before these tests
can even run.
A ./tests/ghc-regress/cps/
A ./tes
Thu Jul 5 10:07:03 PDT 2007 Michael D. Adams <[EMAIL PROTECTED]>
* Added comment to the Cmm parser showing code for use one CPS is enabled
M ./compiler/cmm/CmmParse.y +1
___
Cvs-ghc mailing list
Cvs-ghc@haskell.org
http://www.haskell.org/mailman
Thu Jul 5 10:06:37 PDT 2007 Michael D. Adams <[EMAIL PROTECTED]>
* Eliminated use of 'fromJust' from the CPS pass
M ./compiler/cmm/CmmCPSGen.hs -2 +1
___
Cvs-ghc mailing list
Cvs-ghc@haskell.org
http://www.haskell.org/mailman/listinfo/cvs-ghc
Thu Jul 5 10:05:15 PDT 2007 Michael D. Adams <[EMAIL PROTECTED]>
* Fixed a bug in the CPS pass
M ./compiler/cmm/CmmCPSGen.hs -1 +1
___
Cvs-ghc mailing list
Cvs-ghc@haskell.org
http://www.haskell.org/mailman/listinfo/cvs-ghc
Thu Jul 5 07:50:31 PDT 2007 Michael D. Adams <[EMAIL PROTECTED]>
* Updated the call to the CPS pass for the factoring out of cmmToRawCmm
That call was and is commented out, but it serves as documentation
for when the CPS pass will be turned on.
M ./compiler/main/HscMain.lhs -2 +2
Thu Jul 5 07:48:20 PDT 2007 Michael D. Adams <[EMAIL PROTECTED]>
* Added support for GC block declaration to the Cmm syntax
M ./compiler/cmm/Cmm.hs -1 +1
M ./compiler/cmm/CmmCPS.hs -50 +31
M ./compiler/cmm/CmmCPSGen.hs -19 +8
M ./compiler/cmm/CmmParse.y -9 +16
M ./compiler/
Thu Jul 5 02:47:02 PDT 2007 Michael D. Adams <[EMAIL PROTECTED]>
* Ensure no statements after CmmReturn are emitted
M ./compiler/codeGen/CgMonad.lhs +1
___
Cvs-ghc mailing list
Cvs-ghc@haskell.org
http://www.haskell.org/mailman/listinfo/cvs-ghc
Hi Simon,
- 6.6.1 and 6.7 generate exactly the same Cmm
Yes, as far as I can tell.
- but their performance differs significantly,
- with -fasm or -fvia-C?
Both in the default configuration for -O2, then I redid the tests
using -fasm on both.
If -fasm, did you look at -ddump-opt
Neil Mitchell wrote:
Hi,
I am working on a Core -> Core optimising pass in Yhc Core, which then
spits out GHC Core at the back end. In response to Simon M's comment
that the lack of using GHC's IO Monad meant the results were
unreliable, I have now moved to integrating with GHC's notion of IO.
W
Just to be clear here: you're saying that
- 6.6.1 and 6.7 generate exactly the same Cmm
- but their performance differs significantly,
- with -fasm or -fvia-C?
If -fasm, did you look at -ddump-opt-cmm? (we have a cmm-to-cmm optimiser pass
that runs before the NCG, that flag shows its out
Fri Jul 6 01:11:57 PDT 2007 Simon Marlow <[EMAIL PROTECTED]>
* tcfail140 is expect_broken(451)
M ./tests/ghc-regress/typecheck/should_fail/all.T -1 +1
___
Cvs-ghc mailing list
Cvs-ghc@haskell.org
http://www.haskell.org/mailman/listinfo/cvs-ghc
Fri Jul 6 00:59:31 PDT 2007 Simon Marlow <[EMAIL PROTECTED]>
* FIX BUILD (Windows)
M ./compat/Compat/Directory.hs -1 +1
___
Cvs-ghc mailing list
Cvs-ghc@haskell.org
http://www.haskell.org/mailman/listinfo/cvs-ghc
Peter Tanski wrote:
On Jul 5, 2007, at 9:17 AM, Simon Marlow wrote:
Peter Tanski wrote:
... in mk/config.mk.in...
ifeq "$(TargetOS_CPP)" "windows"
< a whole boatload of extra settings >
endif
That would be better yes. I'm curious to see what the "boatload of
extra settings" are - there m
Build results:
x86-64 Linux head:pass
x86 Windows head: fail (failed stage1)
x86 Windows head fast:pass pass pass fail (failed stage1) fail (failed
stage1) fail (failed stage1)
mnemosyne x86-64 Gentoo head: pass
phil P4 SuSE Linux head: pass
tnaur x86 Linux hea
29 matches
Mail list logo
