sac2c issueshttps://gitlab.sac-home.org/sac-group/sac2c/-/issues2017-11-19T21:47:13Zhttps://gitlab.sac-home.org/sac-group/sac2c/-/issues/1993-check c issue on WL body conformity constraint2017-11-19T21:47:13ZSven-Bodo Scholz-check c issue on WL body conformity constraint| | |
| --- | --- |
| Bugzilla Link | [440](http://bugs.sac-home.org/show_bug.cgi?id=440) |
| Created on | Jun 23, 2008 21:11 |
| Resolution | FIXED |
| Resolved on | Feb 19, 2009 11:33 |
| Version | 1.00beta |
| OS | Linux |
| Archite...| | |
| --- | --- |
| Bugzilla Link | [440](http://bugs.sac-home.org/show_bug.cgi?id=440) |
| Created on | Jun 23, 2008 21:11 |
| Resolution | FIXED |
| Resolved on | Feb 19, 2009 11:33 |
| Version | 1.00beta |
| OS | Linux |
| Architecture | PC |
| Attachments | [aes.sac](/uploads/183e629c41b40572a245631845ea8b15/aes.sac) |
## Extended Description
<pre>Created an attachment (id=480)
source code
as of sac2c rev 15736
compiling the attached source file with the following options( each of which is crucial to produce the error): -maxspec 0 -noINL -check c
leads to
** 14: Introducing explicit memory management instructions ...
**** AUD/SCL distinction ...
**** Making copy operations explicit ...
**** Removing alias results from conformity checks ...
**** Introducing explicit allocation statements ...
**** Removing dead code ...
**** Inferring reuse candidates ...
**** Activating display of alias information ...
**** Interface aliasing analysis ...
**** Applying loop reuse optimization ...
**** Aliasing analysis ...
**** Removing non-local reuse-candidates ...
**** Removing invalid reuse candidates ...
**** Static reuse ...
**** Introducing reuse branches ...
TRAVERSE ERROR: node of type undefined found where N_assign was expected!
OOOPS your program crashed the compiler 8-((
Please send a bug report to bugs@sac-home.org.</pre>Kai TrojahnerKai Trojahnerhttps://gitlab.sac-home.org/sac-group/sac2c/-/issues/1992UFIV traversal breaking CUDA partial folds2017-11-19T21:47:07ZMiguel Sousa DiogoUFIV traversal breaking CUDA partial folds| | |
| --- | --- |
| Bugzilla Link | [1011](http://bugs.sac-home.org/show_bug.cgi?id=1011) |
| Created on | Aug 01, 2012 17:09 |
| Resolution | FIXED |
| Resolved on | Aug 02, 2012 16:24 |
| Version | svn |
| OS | All |
| Architecture...| | |
| --- | --- |
| Bugzilla Link | [1011](http://bugs.sac-home.org/show_bug.cgi?id=1011) |
| Created on | Aug 01, 2012 17:09 |
| Resolution | FIXED |
| Resolved on | Aug 02, 2012 16:24 |
| Version | svn |
| OS | All |
| Architecture | PC |
## Extended Description
<pre>Compiling SaC code with fold with-loop for the CUDA backend with the CUDA partial folding optimization fails as of rev 18118. I got this error in previous revisions, but I was working on the cudahybrid backend so I though the issue was specific to that. It does not seem to be the case however:
------------------------------------------------
$ sac2c relax_fix.sac -target MBpro_cuda -dopfd
...
** 16: Introducing memory management instructions ...
**** Unshare index vectors in WL-folds ...
**** Propagating constants ...
**** CUDA Flatten generators ...
**** AUD/SCL distinction ...
**** Making copy operations explicit ...
**** Introducing explicit allocation statements ...
memory/alloc.c:1010 Assertion "( NODE_TYPE( prf) == N_prf && PRF_PRF( prf) == F_cond_wl_assign)" failed!
Result of partial fold is not defined by F_cond_wl_assign
------------------------------------------------
I briefly looked into this and the result of the partial fold is now a F_unshare prf which was introduced
by the UFIV traversal. I don't know if the assertion is now outdated or if this condition is really necessary. I'm assigning this to Jara, as he committed UFIV according to the svn logs.</pre>Jaroslav SýkoraJaroslav Sýkorahttps://gitlab.sac-home.org/sac-group/sac2c/-/issues/1991bug652.breaks.sac dies in back end, due to unflattened AST node2017-11-19T21:47:01ZRobert Berneckybug652.breaks.sac dies in back end, due to unflattened AST node| | |
| --- | --- |
| Bugzilla Link | [1002](http://bugs.sac-home.org/show_bug.cgi?id=1002) |
| Created on | Jul 10, 2012 20:55 |
| Resolution | FIXED |
| Resolved on | Aug 27, 2012 18:49 |
| Version | svn |
| OS | Linux |
| Architectu...| | |
| --- | --- |
| Bugzilla Link | [1002](http://bugs.sac-home.org/show_bug.cgi?id=1002) |
| Created on | Jul 10, 2012 20:55 |
| Resolution | FIXED |
| Resolved on | Aug 27, 2012 18:49 |
| Version | svn |
| OS | Linux |
| Architecture | PC |
## Extended Description
<pre>sac2c bug652.breaks.sac -v4
...
** 20: Generating Code ...
**** Tag preparation ...
**** Converting to old type representation ...
**** Creating intermediate code macros ...
TRAVERSE ERROR: node of type 46:N_bool found where 31:N_id was expected!
sac@rattler:~/sac/testsuite/optimizations/awlf$ sac2c -V
sac2c v1.00-beta (Haggis And Apple)
developer rev 18058:MODIFIED linux-gnu_x86_64
(Tue Jul 10 10:45:30 EDT 2012 by sac)
We crash around or below codegen/compile.c:2663
A few levels up from the crash, we have this:
call PRTdoPrintNode(arg_node)
-----------------------------------------------
_unshare_( false, SACp_pinl_3510_iv)
--------------------------------------------
The clean approach, I suspect, would be to flatten "false".
FLATGflattenExpression can do this for you.</pre>Jaroslav SýkoraJaroslav Sýkorahttps://gitlab.sac-home.org/sac-group/sac2c/-/issues/1990DLIR crashes ipape.sac in two ways2017-11-19T21:46:55ZRobert BerneckyDLIR crashes ipape.sac in two ways| | |
| --- | --- |
| Bugzilla Link | [952](http://bugs.sac-home.org/show_bug.cgi?id=952) |
| Created on | May 12, 2012 19:31 |
| Resolution | FIXED |
| Resolved on | May 16, 2012 20:20 |
| Version | svn |
| OS | Linux |
| Architecture...| | |
| --- | --- |
| Bugzilla Link | [952](http://bugs.sac-home.org/show_bug.cgi?id=952) |
| Created on | May 12, 2012 19:31 |
| Resolution | FIXED |
| Resolved on | May 16, 2012 20:20 |
| Version | svn |
| OS | Linux |
| Architecture | PC |
| Attachments | [ipape.sac](/uploads/7e7a21fa66ca88f60b74de8e97c4c60c/ipape.sac) |
## Extended Description
<pre>Created an attachment (id=886)
source code to reproduce fault
This failure is recent; it definitely appears in:
sac2c v1.00-beta (Haggis And Apple)
developer rev 17803 linux-gnu_x86_64
(Wed May 9 11:06:42 EDT 2012 by sac)
I see two failure modes, but it is also possible
that the non-failing case is not working properly; I have
not investigated that.
sac2c -doawlf -nowlf ipape.sac -nodlir -v1
WARNING: AWLF is enabled: -ecc enabled.
WARNING: AWLF is enabled: -extrema enabled.
WARNING: AWLF is enabled: -maxoptcyc=20
note: finished parsing.
OOOOOOOPS, your program crashed the compiler 8-((
[The crash here occurs in AWLFI, but if you look into
that code, you'll see a call to DLIR.]
sac2c -doawlf -nowlf ipape.sac -v1
WARNING: AWLF is enabled: -ecc enabled.
WARNING: AWLF is enabled: -extrema enabled.
WARNING: AWLF is enabled: -maxoptcyc=20
note: finished parsing.
tree/traverse_helper.c:440 Assertion "FALSE" failed!
Illegal nodetype found!</pre>Jaroslav SýkoraJaroslav Sýkorahttps://gitlab.sac-home.org/sac-group/sac2c/-/issues/1989Cuda needs architecture information about the card2017-11-19T21:46:48ZDaniel RollsCuda needs architecture information about the card| | |
| --- | --- |
| Bugzilla Link | [770](http://bugs.sac-home.org/show_bug.cgi?id=770) |
| Created on | Nov 04, 2010 12:07 |
| Resolution | FIXED |
| Resolved on | Nov 30, 2010 11:44 |
| Version | svn |
| OS | All |
| Architecture |...| | |
| --- | --- |
| Bugzilla Link | [770](http://bugs.sac-home.org/show_bug.cgi?id=770) |
| Created on | Nov 04, 2010 12:07 |
| Resolution | FIXED |
| Resolved on | Nov 30, 2010 11:44 |
| Version | svn |
| OS | All |
| Architecture | PC |
| Attachments | [add_property_to_sac2crc.patch](/uploads/f3565218f3624b0930b9e7cad6cd5ee8/add_property_to_sac2crc.patch) |
## Extended Description
<pre>Created an attachment (id=771)
patch
According to Jing the cuda backend can now use architecture information about the card to produce more efficient code. Currently there is no way to store this information in the sac2crc file but Jing has promised to fix this.
Maybe we need a property called something like cuda_arch?
I found a patch from Clemens where he added a property to sac2crc and used it so by copying the attached patch and changing the code this shouldn't be too much of a problem, I hope.</pre>Jing GuoJing Guohttps://gitlab.sac-home.org/sac-group/sac2c/-/issues/1988make checkxml complains about CUDA entries2017-11-19T21:46:42ZRobert Berneckymake checkxml complains about CUDA entries| | |
| --- | --- |
| Bugzilla Link | [768](http://bugs.sac-home.org/show_bug.cgi?id=768) |
| Created on | Nov 03, 2010 16:34 |
| Resolution | FIXED |
| Resolved on | Dec 09, 2010 11:26 |
| Version | svn |
| OS | Linux |
| Architecture...| | |
| --- | --- |
| Bugzilla Link | [768](http://bugs.sac-home.org/show_bug.cgi?id=768) |
| Created on | Nov 03, 2010 16:34 |
| Resolution | FIXED |
| Resolved on | Dec 09, 2010 11:26 |
| Version | svn |
| OS | Linux |
| Architecture | PC |
## Extended Description
<pre>$ make checkxml
xsltproc check_integrity.xsl ast.xml
Node Args of traversal CUDA Tag execution mode unknown
This has been around quite a while now.
Definitely fails on:
product rev 17138:MODIFIED linux-gnu_x86_64
(Tue Nov 2 18:23:55 EDT 2010 by sac)</pre>Jing GuoJing Guohttps://gitlab.sac-home.org/sac-group/sac2c/-/issues/1987ast.xml make checkxml fails on CUDA2017-11-19T21:46:37ZRobert Berneckyast.xml make checkxml fails on CUDA| | |
| --- | --- |
| Bugzilla Link | [723](http://bugs.sac-home.org/show_bug.cgi?id=723) |
| Created on | Jun 14, 2010 20:44 |
| Resolution | FIXED |
| Resolved on | Dec 09, 2010 11:26 |
| Version | svn |
| OS | Linux |
| Architecture...| | |
| --- | --- |
| Bugzilla Link | [723](http://bugs.sac-home.org/show_bug.cgi?id=723) |
| Created on | Jun 14, 2010 20:44 |
| Resolution | FIXED |
| Resolved on | Dec 09, 2010 11:26 |
| Version | svn |
| OS | Linux |
| Architecture | PC |
## Extended Description
<pre>make checkxml
xsltproc check_integrity.xsl ast.xml
Node Args of traversal CUDA Tag execution mode unknown
Build #16887</pre>Jing GuoJing Guohttps://gitlab.sac-home.org/sac-group/sac2c/-/issues/1986Unable to build sac2c using cmake-follow on Ubuntu2017-11-19T21:46:32ZRobert BerneckyUnable to build sac2c using cmake-follow on Ubuntu| | |
| --- | --- |
| Bugzilla Link | [1184](http://bugs.sac-home.org/show_bug.cgi?id=1184) |
| Created on | Jan 10, 2017 20:28 |
| Resolution | FIXED |
| Resolved on | Jan 28, 2017 15:30 |
| Version | svn |
| OS | Linux |
| Architectu...| | |
| --- | --- |
| Bugzilla Link | [1184](http://bugs.sac-home.org/show_bug.cgi?id=1184) |
| Created on | Jan 10, 2017 20:28 |
| Resolution | FIXED |
| Resolved on | Jan 28, 2017 15:30 |
| Version | svn |
| OS | Linux |
| Architecture | PC |
| Attachments | [nm_output.txt](/uploads/06d53559b748173cd3df31426642474a/nm_output.txt), [sacprelude_d.sacbugreport](/uploads/4332e962359543bdefd80c1162495a0e/sacprelude_d.sacbugreport) |
## Extended Description
<pre>I tried to build the Ubuntu version of sac2c, using the instructions in
"suggestions for newbie developers". I get this failure:
CMake Error at /home/sac/sac2c/cmake/place-sac2crc-prelude-file.cmake:30 (MESSAGE):
No ~/.sac2crc directory found, please create it. If you have a `.sac2crc'
file, please move this to the ~/.sac2crc directory and rename it
`sac2crc.config'.
I did build an INSTALL script, which does a mkdir ./.sac2crc build_d build_p.
That is apparently not enough.
I am not sure how to proceed from here.</pre>Hans-Nikolai ViessmannHans-Nikolai Viessmannhttps://gitlab.sac-home.org/sac-group/sac2c/-/issues/1985TGTL breaks many AWLF and CF unit tests2017-11-19T21:46:20ZRobert BerneckyTGTL breaks many AWLF and CF unit tests| | |
| --- | --- |
| Bugzilla Link | [1032](http://bugs.sac-home.org/show_bug.cgi?id=1032) |
| Created on | Nov 05, 2012 19:54 |
| Resolution | FIXED |
| Resolved on | Nov 20, 2012 09:30 |
| Version | svn |
| OS | Linux |
| Architectu...| | |
| --- | --- |
| Bugzilla Link | [1032](http://bugs.sac-home.org/show_bug.cgi?id=1032) |
| Created on | Nov 05, 2012 19:54 |
| Resolution | FIXED |
| Resolved on | Nov 20, 2012 09:30 |
| Version | svn |
| OS | Linux |
| Architecture | PC |
## Extended Description
<pre>A new traversal, TGTL, was released on Halloween. It breaks
many of the AWLF and CF unit tests. Please fix any failing
unit tests in these directories, AFTER adding your own
unit tests (see next paragraph) for tgtl, and getting them to
run cleanly.
TGTL does not appear to have any of its own unit tests, so it
is perhaps not surprising that failures in it (or failures
induced by it) are happening.
In addition, there is zero documentation in the traversal
code that gives any rationale for why this traversal was
invented. Please add appropriate documentation to the
relevant C code. [This may have been discussed in sacdev
email, but if so, I missed it.]
sac2c -V
sac2c v1.00-beta (Haggis And Apple)
product rev exported linux-gnu_x86_64
(Mon Nov 5 09:19:30 EST 2012 by sac)
[sac2c -V is not very useful any more, is it?]</pre>Fangyong TangFangyong Tanghttps://gitlab.sac-home.org/sac-group/sac2c/-/issues/1984Different tvd result when switching from vanilla sac2c to the sl_pp backend2017-11-19T21:46:14ZDaniel RollsDifferent tvd result when switching from vanilla sac2c to the sl_pp backend| | |
| --- | --- |
| Bugzilla Link | [678](http://bugs.sac-home.org/show_bug.cgi?id=678) |
| Created on | Feb 07, 2010 23:11 |
| Resolution | FIXED |
| Resolved on | Feb 11, 2010 12:37 |
| Version | svn |
| OS | All |
| Architecture |...| | |
| --- | --- |
| Bugzilla Link | [678](http://bugs.sac-home.org/show_bug.cgi?id=678) |
| Created on | Feb 07, 2010 23:11 |
| Resolution | FIXED |
| Resolved on | Feb 11, 2010 12:37 |
| Version | svn |
| OS | All |
| Architecture | PC |
| Attachments | [caj.sac](/uploads/dca2a60563747f1555c50c3de5c3af28/caj.sac), [tvd2db.sac](/uploads/1f458befa098d6cf7cc5aaaa48de3e56/tvd2db.sac) |
## Extended Description
<pre>The attached code returns NAN for the entire outputted array with -target sl_ppp. When I run with vanilla SaC I get the usual output. I ran this on a private build of Carl's and never recorded the sac2c version.
This is strange since the target option should only change backend code.
I have already tried giving different values (round, unround, zero and negative) to sqrt to compare the microgrid and C sqrt functions but see no difference. I'll try ruling out rounding errors next.</pre>Daniel RollsDaniel Rollshttps://gitlab.sac-home.org/sac-group/sac2c/-/issues/1983product compiler error: request for member ‘attribs’ in something not a struc...2017-11-19T21:46:07ZDaniel Rollsproduct compiler error: request for member ‘attribs’ in something not a structure or union| | |
| --- | --- |
| Bugzilla Link | [576](http://bugs.sac-home.org/show_bug.cgi?id=576) |
| Created on | Oct 23, 2009 19:07 |
| Resolution | FIXED |
| Resolved on | Oct 26, 2009 17:16 |
| Version | 1.00beta |
| OS | All |
| Architect...| | |
| --- | --- |
| Bugzilla Link | [576](http://bugs.sac-home.org/show_bug.cgi?id=576) |
| Created on | Oct 23, 2009 19:07 |
| Resolution | FIXED |
| Resolved on | Oct 26, 2009 17:16 |
| Version | 1.00beta |
| OS | All |
| Architecture | PC |
## Extended Description
<pre>A make prod of 16489 (today's head) causes this compiler failure on my mac:
precompile/lift_with3_bodies.c: In function ‘InitFolds’:
precompile/lift_with3_bodies.c:561: warning: dereferencing ‘void *’ pointer
precompile/lift_with3_bodies.c:561: error: request for member ‘attribs’ in something not a structure or union
The file was changed last in 16486 on Wednesday (21st October).
gcc version 4.0.1 (Apple Inc. build 5490)</pre>Daniel RollsDaniel Rollshttps://gitlab.sac-home.org/sac-group/sac2c/-/issues/1982MTSTFMOD can't deal with recursion2017-11-19T21:46:01ZSven-Bodo ScholzMTSTFMOD can't deal with recursion| | |
| --- | --- |
| Bugzilla Link | [1137](http://bugs.sac-home.org/show_bug.cgi?id=1137) |
| Created on | Sep 20, 2014 23:28 |
| Resolution | FIXED |
| Resolved on | Sep 20, 2014 23:48 |
| Version | svn |
| OS | All |
| Architecture...| | |
| --- | --- |
| Bugzilla Link | [1137](http://bugs.sac-home.org/show_bug.cgi?id=1137) |
| Created on | Sep 20, 2014 23:28 |
| Resolution | FIXED |
| Resolved on | Sep 20, 2014 23:48 |
| Version | svn |
| OS | All |
| Architecture | PC |
| Attachments | [1136.sac](/uploads/55342df86d1c8e258e43525d02455b2a/1136.sac) |
## Extended Description
<pre>Created an attachment (id=1029)
source code
the ST version of evil, in its spmd function calls the ST version again leading to a deadlock.
compile with -mt -minmtsize 10 -bmt | vi -
shows this</pre>Clemens GrelckClemens Grelckhttps://gitlab.sac-home.org/sac-group/sac2c/-/issues/1981wrong sped function abstraction when using states and loops2017-11-19T21:45:55ZSven-Bodo Scholzwrong sped function abstraction when using states and loops| | |
| --- | --- |
| Bugzilla Link | [1136](http://bugs.sac-home.org/show_bug.cgi?id=1136) |
| Created on | Sep 19, 2014 15:35 |
| Resolution | FIXED |
| Resolved on | Mar 26, 2015 21:14 |
| Version | svn |
| OS | All |
| Architecture...| | |
| --- | --- |
| Bugzilla Link | [1136](http://bugs.sac-home.org/show_bug.cgi?id=1136) |
| Created on | Sep 19, 2014 15:35 |
| Resolution | FIXED |
| Resolved on | Mar 26, 2015 21:14 |
| Version | svn |
| OS | All |
| Architecture | PC |
| Attachments | [TotientRange.sac](/uploads/0a0d05cb0ed3b6e70cf13ae673b87e38/TotientRange.sac), [bound.sac](/uploads/d9c6ac57305d8ef2acf038c41c3db784/bound.sac), [1130swap.sac](/uploads/e5738db0a2c9bdb630762fba381a5a0c/1130swap.sac) |
## Extended Description
<pre>Created an attachment (id=1025)
source code
after mtstf ( sac2c-d -mt -o lib -Llib 1130.sac -noprelude -bmt:mtstf) we have:
_flat_0 = with2 (_pinl_420_iv=[_pinl_421__eat_14] (IDXS:_wlidx_581__flat_0)) {....}
where _pinl_1116__flat_59 is locally defined and used in the body...
after mtspmdf (sac2c-d -mt -o lib -Llib 1130.sac -noprelude -bmt:mtspmdf) we have:
cs, scs__SSA0_2 = maxclique:_ST::_mtspmdf_1354_expandFS( _pinl_1116__flat_59, c, scs, _flat_0) ;
this results in broken C code where the variable is used when setting up the SPMD frame without being defined.</pre>Clemens GrelckClemens Grelckhttps://gitlab.sac-home.org/sac-group/sac2c/-/issues/1980run simple applications2017-11-19T21:45:48ZVolkmar Wieserrun simple applications| | |
| --- | --- |
| Bugzilla Link | [1104](http://bugs.sac-home.org/show_bug.cgi?id=1104) |
| Created on | Dec 04, 2013 12:15 |
| Resolution | FIXED |
| Resolved on | Dec 06, 2013 14:20 |
| Version | 1.00beta |
| OS | Linux |
| Archi...| | |
| --- | --- |
| Bugzilla Link | [1104](http://bugs.sac-home.org/show_bug.cgi?id=1104) |
| Created on | Dec 04, 2013 12:15 |
| Resolution | FIXED |
| Resolved on | Dec 06, 2013 14:20 |
| Version | 1.00beta |
| OS | Linux |
| Architecture | PC |
| Attachments | [example.zip](/uploads/f2f26c01145a58bd3726cf9208857db0/example.zip) |
## Extended Description
<pre>Created an attachment (id=996)
Screenshot of output + files
Hi,
During SACDevCamp we have installed an updated SAC compiler version to my laptop. I'm not sure if every feature is working. Currently, I'm able to compile every SAC Program but not to execute them.
In the attached file you can see a screenshot about the cmd output, my SAC program, and the configuration file of the compiler
Many thanks,
Volkmar</pre>Clemens GrelckClemens Grelckhttps://gitlab.sac-home.org/sac-group/sac2c/-/issues/1979development compiler broken in WLTR2017-11-19T21:45:42ZRobert Berneckydevelopment compiler broken in WLTR| | |
| --- | --- |
| Bugzilla Link | [1103](http://bugs.sac-home.org/show_bug.cgi?id=1103) |
| Created on | Dec 01, 2013 22:01 |
| Resolution | FIXED |
| Resolved on | Jan 10, 2014 16:42 |
| Version | svn |
| OS | Linux |
| Architectu...| | |
| --- | --- |
| Bugzilla Link | [1103](http://bugs.sac-home.org/show_bug.cgi?id=1103) |
| Created on | Dec 01, 2013 22:01 |
| Resolution | FIXED |
| Resolved on | Jan 10, 2014 16:42 |
| Version | svn |
| OS | Linux |
| Architecture | PC |
| Attachments | [1103.sac](/uploads/535126f60b4e56f3adbec5842d354649/1103.sac) |
## Extended Description
<pre>sac2c transpose.sac -v1
wltransform/wltransform.c:4575 Assertion "! WLSTRIDE_ISMODIFIED( stride1)" failed!
stride was modified
sac2c -V
sac2c v1.00-beta (Haggis And Apple)
developer rev 18411 linux-gnu_x86_64
(Sun Dec 1 13:45:58 EST 2013 by sac)
This failure appears to have appeared with Clemens' DupTree
copy-all-flags fix. It only affects the develment compiler.
Here is transpose.sac:
use Array: {transpose,genarray,++,*,iota,sum};
int main()
{
a = with {
( [ 0, 0] <= iv=[i,j] < [ 20, 15]) : j;
} : genarray( [40,20], 42);
b = transpose(a);
z = _sub_SxS_( sum(b), 23100);
StdIO::show(b);
StdIO::show(sum(b));
return( z);
}</pre>Clemens GrelckClemens Grelckhttps://gitlab.sac-home.org/sac-group/sac2c/-/issues/1978-pritnfun main -printfunsets d does not print main's LOCALFUNs2017-11-19T21:45:34ZRobert Bernecky-pritnfun main -printfunsets d does not print main's LOCALFUNs| | |
| --- | --- |
| Bugzilla Link | [1079](http://bugs.sac-home.org/show_bug.cgi?id=1079) |
| Created on | May 05, 2013 16:09 |
| Resolution | FIXED |
| Resolved on | Apr 05, 2015 15:14 |
| Version | svn |
| OS | Linux |
| Architectu...| | |
| --- | --- |
| Bugzilla Link | [1079](http://bugs.sac-home.org/show_bug.cgi?id=1079) |
| Created on | May 05, 2013 16:09 |
| Resolution | FIXED |
| Resolved on | Apr 05, 2015 15:14 |
| Version | svn |
| OS | Linux |
| Architecture | PC |
| Attachments | [condfun.sac](/uploads/b496dc577c0e6030efc64b12c7e1089f/condfun.sac) |
## Extended Description
<pre>sac2c -V
sac2c v1.00-beta (Haggis And Apple)
product rev 18133 linux-gnu_x86_64
(Sun May 5 10:06:38 EDT 2013 by sac)
cd ~/sac/testsuite/optimizations/awlf
sac2c buglacsoinline.sac -doawlf -nowlf -bopt -printfun main -printfunsets d >crud
The above printed the wrapper for main, and main itself, but not the
CONDFUN that main() calls.</pre>Clemens GrelckClemens Grelckhttps://gitlab.sac-home.org/sac-group/sac2c/-/issues/1977SAC_ND_RC_TO_NORC" redefined in -mt compilation2017-11-19T21:45:28ZRobert BerneckySAC_ND_RC_TO_NORC" redefined in -mt compilation| | |
| --- | --- |
| Bugzilla Link | [1006](http://bugs.sac-home.org/show_bug.cgi?id=1006) |
| Created on | Jul 18, 2012 16:18 |
| Resolution | FIXED |
| Resolved on | Jul 18, 2012 17:17 |
| Version | svn |
| OS | Linux |
| Architectu...| | |
| --- | --- |
| Bugzilla Link | [1006](http://bugs.sac-home.org/show_bug.cgi?id=1006) |
| Created on | Jul 18, 2012 16:18 |
| Resolution | FIXED |
| Resolved on | Jul 18, 2012 17:17 |
| Version | svn |
| OS | Linux |
| Architecture | PC |
| Attachments | [crud200.sac](/uploads/90a5b81efaa2afe0e2ca4227379583b5/crud200.sac) |
## Extended Description
<pre>Created an attachment (id=918)
source code to reproduce fault
sac2c crud200.sac -doawlf -nowlf -wls_aggressive -O3 -o crud200wlsagMT -v1 -mt
WARNING: AWLF is enabled: -ecc enabled.
WARNING: AWLF is enabled: -extrema enabled.
WARNING: AWLF is enabled: -maxoptcyc=20
In file included from crud200wlsagMT.c:156:
/home/sac/sac2c/include/sac.h:3214: warning: "SAC_ND_RC_TO_NORC" redefined
/home/sac/sac2c/include/sac.h:2340: note: this is the location of the previous definition
/home/sac/sac2c/include/sac.h:3215: warning: "SAC_ND_RC_FROM_NORC" redefined
/home/sac/sac2c/include/sac.h:2345: note: this is the location of the previous definition
sac2c -V
sac2c v1.00-beta (Haggis And Apple)
developer rev 18084 linux-gnu_x86_64
(Tue Jul 17 18:32:29 EDT 2012 by sac)</pre>Clemens GrelckClemens Grelckhttps://gitlab.sac-home.org/sac-group/sac2c/-/issues/1976Lost extrema values cause AWLF unit test prd2AKD.sac to fail2017-11-19T21:45:22ZRobert BerneckyLost extrema values cause AWLF unit test prd2AKD.sac to fail| | |
| --- | --- |
| Bugzilla Link | [984](http://bugs.sac-home.org/show_bug.cgi?id=984) |
| Created on | Jun 19, 2012 21:45 |
| Resolution | FIXED |
| Resolved on | Jul 08, 2012 19:07 |
| Version | svn |
| OS | Linux |
| Architecture...| | |
| --- | --- |
| Bugzilla Link | [984](http://bugs.sac-home.org/show_bug.cgi?id=984) |
| Created on | Jun 19, 2012 21:45 |
| Resolution | FIXED |
| Resolved on | Jul 08, 2012 19:07 |
| Version | svn |
| OS | Linux |
| Architecture | PC |
## Extended Description
<pre>A simple unit test for AWLF now fails, due to lost extrema.
The problem is that, if I compile with this:
cd ~/sac/testsuite/optimizations/awlf
sac2c prd2AKD.sac -doawlf -nowlf -nocyc -bopt:glf >crud.glf
I get this code fragment in crud.glf:
_flat_5 = [ shp ];
_idc_46, _icc_34_pred = _non_neg_val_V_( _flat_5);
...
_flat_0 = [ shp ];
What we want is to have all references to _flat_5 that
follow the _non_neg_val_V_( _flat_5) to become references
to _idc_46. The assignment of _flat_0 is, in a sense,
a violation of this, because _flat_0 does not carry the
AVIS_MIN generated by the guard.
The code that inserts the guards is correct as it stands.
The _flat_0 assign exists by the end of TC (phase 7).
So, we want to get rid of the common expression before
we insert constraints, so we clearly need a CSE/DCR traversal
in CWC, just before ICC. Ah, you mean, like this?
cse : Applying common subexpression elimination
dcr : Removing dead code
icc : Inserting Conformity Checks
Hmmm. Let's look at what we have in iota() after that DCR:
_flat_5 = [ shp ];
_flat_4 = [ shp ];
_flat_3 = 0;
_flat_2 = _mul_SxV_( _flat_3, _flat_4);
_flat_1 = 0;
_flat_0 = [ shp ];
Not exactly what we were hoping for.
CSE never traverses iota(), but it does traverse a lot of wrappers:
grep CSEfundef crud.cwcdcr
CSEfundef: CSE: Begin traversing (wrapper) partitionIntersectMax
CSEfundef: CSE: Done traversing (wrapper) partitionIntersectMax
CSEfundef: CSE: Begin traversing (wrapper) partitionSlicer
CSEfundef: CSE: Done traversing (wrapper) partitionSlicer
CSEfundef: CSE: Begin traversing (wrapper) isPartitionIntersectNull
CSEfundef: CSE: Done traversing (wrapper) isPartitionIntersectNull
CSEfundef: CSE: Begin traversing (wrapper) partitionIntersectMin
CSEfundef: CSE: Done traversing (wrapper) partitionIntersectMin
CSEfundef: CSE: Begin traversing (wrapper) adjustLacFunParams
CSEfundef: CSE: Done traversing (wrapper) adjustLacFunParams
CSEfundef: CSE: Begin traversing (wrapper) main
CSEfundef: CSE: Done traversing (wrapper) main
CSEfundef: CSE: Begin traversing (wrapper) prod
CSEfundef: CSE: Done traversing (wrapper) prod
CSEfundef: CSE: Begin traversing (wrapper) adjustLacFunParamsReshape
CSEfundef: CSE: Done traversing (wrapper) adjustLacFunParamsReshape
CSEfundef: CSE: Begin traversing (wrapper) eq
CSEfundef: CSE: Done traversing (wrapper) eq
CSEfundef: CSE: Begin traversing (wrapper) eq
CSEfundef: CSE: Done traversing (wrapper) eq
CSEfundef: CSE: Begin traversing (wrapper) eq
CSEfundef: CSE: Done traversing (wrapper) eq
CSEfundef: CSE: Begin traversing (wrapper) eq
CSEfundef: CSE: Done traversing (wrapper) eq
CSEfundef: CSE: Begin traversing (wrapper) eq
CSEfundef: CSE: Done traversing (wrapper) eq
CSEfundef: CSE: Begin traversing (wrapper) gridFiller
CSEfundef: CSE: Done traversing (wrapper) gridFiller
CSEfundef: CSE: Begin traversing (wrapper) zero
CSEfundef: CSE: Done traversing (wrapper) zero
CSEfundef: CSE: Begin traversing (wrapper) zero
CSEfundef: CSE: Done traversing (wrapper) zero
CSEfundef: CSE: Begin traversing (wrapper) zero
CSEfundef: CSE: Done traversing (wrapper) zero
CSEfundef: CSE: Begin traversing (wrapper) zero
CSEfundef: CSE: Done traversing (wrapper) zero
CSEfundef: CSE: Begin traversing (wrapper) zero
CSEfundef: CSE: Done traversing (wrapper) zero
CSEfundef: CSE: Begin traversing (wrapper) partitionMax
CSEfundef: CSE: Done traversing (wrapper) partitionMax
CSEfundef: CSE: Begin traversing function iota__Cond_0
CSEfundef: CSE: Done traversing function iota__Cond_0
I am not sure where to go from here. This unit test did work
properly some time ago.</pre>Clemens GrelckClemens Grelckhttps://gitlab.sac-home.org/sac-group/sac2c/-/issues/1975New TRAV mech fails on nested LOCALFNS2017-11-19T21:45:16ZRobert BerneckyNew TRAV mech fails on nested LOCALFNS| | |
| --- | --- |
| Bugzilla Link | [866](http://bugs.sac-home.org/show_bug.cgi?id=866) |
| Created on | Aug 30, 2011 19:13 |
| Resolution | INVALID |
| Resolved on | Aug 30, 2011 19:59 |
| Version | svn |
| OS | Linux |
| Architectu...| | |
| --- | --- |
| Bugzilla Link | [866](http://bugs.sac-home.org/show_bug.cgi?id=866) |
| Created on | Aug 30, 2011 19:13 |
| Resolution | INVALID |
| Resolved on | Aug 30, 2011 19:59 |
| Version | svn |
| OS | Linux |
| Architecture | PC |
| Attachments | [bug650Bivecyc.sac](/uploads/86d6a69ce2701f55dd6d8e99a860e684/bug650Bivecyc.sac) |
## Extended Description
<pre>Created an attachment (id=816)
source code to reproduce failure
I changed CWLE to use the new TRAV() mechanism, but
there's a small problem with it:
When it invokes a LACFUN, that LACFUN's FUNDEF_LOCALFNS
node is NULL, even when it has a LOCALFUN.
In this case, a COND fun invokes a LOOP() fun, but
said LOOP() fun does not appear in the COND()'s LOCALFNS
list (It's NULL).
cd ~/sac/testsuite/optimizations/cwle
sac2c -nocyc -noawlf -nowlf bug650Bivecyc.sac
will do the job for you.
sac2c -V
sac2c v1.00-beta (Haggis And Apple)
developer rev 17581:MODIFIED linux-gnu_x86_64
(Tue Aug 30 13:55:04 EDT 2011 by sac)</pre>Clemens GrelckClemens Grelckhttps://gitlab.sac-home.org/sac-group/sac2c/-/issues/1974New TRAV mech/AL breaks stdlib build2017-11-19T21:45:10ZRobert BerneckyNew TRAV mech/AL breaks stdlib build| | |
| --- | --- |
| Bugzilla Link | [865](http://bugs.sac-home.org/show_bug.cgi?id=865) |
| Created on | Aug 29, 2011 15:40 |
| Resolution | FIXED |
| Resolved on | Aug 30, 2011 15:54 |
| Version | svn |
| OS | Linux |
| Architecture...| | |
| --- | --- |
| Bugzilla Link | [865](http://bugs.sac-home.org/show_bug.cgi?id=865) |
| Created on | Aug 29, 2011 15:40 |
| Resolution | FIXED |
| Resolved on | Aug 30, 2011 15:54 |
| Version | svn |
| OS | Linux |
| Architecture | PC |
## Extended Description
<pre>Build #17578 gets in an inlining loop building the stdlib
Structures.sac. I got around it by backing off these documents:
svn update phase_sac2c.mac -r17570
svn update associative_law.c -r17566</pre>Clemens GrelckClemens Grelck