Known issues - Revision history

Pmelo at 10:20, 31 August 2026

2026-08-31T10:20:07Z

2026-08-31T07:56:05Z

Fix table overflow by removing leading white space on issue 11

← Older revision		Revision as of 07:56, 31 August 2026
Line 614:		Line 614:
	\| text = '''AOCC >= 4.0 does not produce runnable code when compiling without OpenMP-support''':		\| text = '''AOCC >= 4.0 does not produce runnable code when compiling without OpenMP-support''':
	The AOCC compiler version >= 4.0 apparently uses a more aggressive optimization on a particular symmetry routine (SGRGEN) when compiling '''without''' OpenMP-support. Thus, code produced using arch/makefile.include.aocc_ompi_aocl exits with:		The AOCC compiler version >= 4.0 apparently uses a more aggressive optimization on a particular symmetry routine (SGRGEN) when compiling '''without''' OpenMP-support. Thus, code produced using arch/makefile.include.aocc_ompi_aocl exits with:
	<code>VERY BAD NEWS! internal error in subroutine SGRGEN: Too many elements 49</code>		<code>VERY BAD NEWS! internal error in subroutine SGRGEN: Too many elements 49</code>
	Solution: adapt your makefile.include by adding <code>symlib.o</code> to the <code>OBJECTS_O1</code> line. This fixes the issue. The other options is to compile with openMP support (using arch/makefile.include.aocc_ompi_aocl_omp).		Solution: adapt your makefile.include by adding <code>symlib.o</code> to the <code>OBJECTS_O1</code> line. This fixes the issue. The other options is to compile with openMP support (using arch/makefile.include.aocc_ompi_aocl_omp).

2026-08-31T07:40:24Z

Add known issues for recent Wannier bug fixes

@@ Line 26: / Line 26: @@
 There is an inevitable gap in indices at the top of the issues list right now because IDs have now also been assigned to old issues that did not have any IDs before.
 --->
 {{KnownIssue
 | ID                  = 104

2026-08-20T19:42:57Z

← Older revision		Revision as of 19:42, 20 August 2026
Line 26:		Line 26:
	There is an inevitable gap in indices at the top of the issues list right now because IDs have now also been assigned to old issues that did not have any IDs before.		There is an inevitable gap in indices at the top of the issues list right now because IDs have now also been assigned to old issues that did not have any IDs before.
	--->		--->
			{{KnownIssue
			\| ID = 104
			\| resolved_in_version = -
			\| reported_in_version = 6.5.0
			\| date_added = 2026-08-20
			\| resolved_in_commit = 8c0c5ed5629da037630741efbe3ca36e40945035
			\| text =
			Combining {{TAG\|ELPH_SELFEN_GAPS}} or {{TAG\|ELPH_TRANSPORT}} with a restrictive {{TAG\|ELPH_SELFEN_BAND_START}}/{{TAG\|ELPH_SELFEN_BAND_STOP}} (or the {{TAG\|ELPH_SELFEN_BAND_START_KP}}/{{TAG\|ELPH_SELFEN_BAND_STOP_KP}} variants) can crash with an out-of-bounds array access if a relevant band falls outside the restricted band range.
			}}
	{{KnownIssue		{{KnownIssue
	\| ID = 103		\| ID = 103

2026-08-18T12:01:02Z

← Older revision		Revision as of 12:01, 18 August 2026
Line 33:		Line 33:
	\| resolved_in_commit = af1cc27f70a35d1577e9fa09519c52496e17da3e		\| resolved_in_commit = af1cc27f70a35d1577e9fa09519c52496e17da3e
	\| text =		\| text =
	With {{TAG\|LNOAUGXC}} = .TRUE., spin-polarized parallel runs on unevenly dividing FFT grids produce wrong potentials, and consequently also wrong total energies, forces, etc.		With {{TAG\|LNOAUGXC}} = .TRUE., spin-polarized parallel runs on unevenly dividing FFT grids produce wrong potentials, and consequently also wrong total energies, forces, etc. Note that {{TAG\|LNOAUGXC}} = .FALSE. by default.
	}}		}}
	{{KnownIssue		{{KnownIssue

2026-08-18T11:59:04Z

← Older revision		Revision as of 11:59, 18 August 2026
Line 26:		Line 26:
	There is an inevitable gap in indices at the top of the issues list right now because IDs have now also been assigned to old issues that did not have any IDs before.		There is an inevitable gap in indices at the top of the issues list right now because IDs have now also been assigned to old issues that did not have any IDs before.
	--->		--->
			{{KnownIssue
			\| ID = 103
			\| resolved_in_version = -
			\| reported_in_version = 6.4.3
			\| date_added = 2026-08-18
			\| resolved_in_commit = af1cc27f70a35d1577e9fa09519c52496e17da3e
			\| text =
			With {{TAG\|LNOAUGXC}} = .TRUE., spin-polarized parallel runs on unevenly dividing FFT grids produce wrong potentials, and consequently also wrong total energies, forces, etc.
			}}
	{{KnownIssue		{{KnownIssue
	\| ID = 102		\| ID = 102

2026-08-17T14:10:15Z

2026-08-17T14:05:10Z

2026-08-17T14:03:59Z

2026-08-17T14:01:53Z

@@ Line 26: / Line 26: @@
 There is an inevitable gap in indices at the top of the issues list right now because IDs have now also been assigned to old issues that did not have any IDs before.
 --->
 {{KnownIssue
 | ID                  = 101