Age | Commit message (Collapse) | Author | Files | Lines |
|
riscv64, armv7l, i586, x86_64 with gcc 13.1.0 in Riscv repository and
aarch64, armv7l, i586, x86_64 with gcc 9.2.0 in Unified repository should be built without build error.
With a previous commit, armv7l, i586, x86_64 in Riscv repo. got build
failed. So, remove %ifarch ~ %endif from a spec file.
Instead, add a patch checking gcc version and then applying the option if the version
is larger than 10.0.0.
Change-Id: Ie9d1dc7421eb91d116077b7abe6520d05af62836
Signed-off-by: Tae-Young Chung <ty83.chung@samsung.com>
|
|
The option is only supported from gcc-10, but currenlty
tizen supports gcc-9.2.0.
So, add the option when riscv64 which is built with gcc-13.
Change-Id: I70857c780a48146e03d2997cddc8bd41281f3e69
Signed-off-by: Tae-Young Chung <ty83.chung@samsung.com>
|
|
On gcc12, followling errors occur.
```
Error: Rank mismatch between actual argument at (1) and actual argument at (2) (scalar and rank-1)
``
To degrade the error to a warning, add fortran option.
Refer to https://gcc.gnu.org/onlinedocs/gfortran/Fortran-Dialect-Options.html#index-fallow-argument-mismatch
Change-Id: I2eaf2aa257f0a6e5b925a1b5e3fed124a53e0402
Signed-off-by: Tae-Young Chung <ty83.chung@samsung.com>
|
|
* Enable cblas package
* Remove unnecessary description
Signed-off-by: Sangjung woo <sangjung.woo@samsung.com>
Corrected subject format.
Signed-off-by: Inki Dae <inki.dae@samsung.com>
|
|
This patch newly added the RPM spec file to build RPM package.
Signed-off-by: Sangjung woo <sangjung.woo@samsung.com>
Corrected subject format.
Signed-off-by: Inki Dae <inki.dae@samsung.com>
|
|
|
|
Added and straightened Doxygen comments on BLAS parameters.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
fixing the definition of "info" in Aasen's
|
|
|
|
|
|
|
|
|
|
Update version number
from discussion with Vendors, they still would like to have the version
number inside the source file.
|
|
i.e. "A is TYPE array, dimension ( LDx, N )"
|
|
Some final line-break rearrangements that assist my parameter comment parser.
|
|
|
|
specifications containing NRHS to the 'dimension' line
This improves automatic parsing of array dimensions,
but creates overly long lines sometimes.
|
|
the array dimension line
|
|
avoid inconsistencies between parameter comments in S and D, C and Z
|
|
|
|
dimension (MAX(1,LWORK))
|
|
|
|
parentheses
|
|
Removing code line 243 which was useless.
Thanks to @vladimir-ch for pointing this out and @oamarques for approving the fix
|
|
more consistency of array dimensions in Fortran declarations and Doxygen comments
|
|
|
|
like in CLALSD
|
|
|
|
|
|
square matrices
This fixes commit 366f9b3541705231cfbb1724b221b8857ac768a8.
|
|
in leading dimensions
|
|
fix dimension specification in some comments
|
|
dimension specifications
|
|
|
|
analogous to DSGESV, DSPOSV
|
|
more cleaned up comments
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|