Skip to content
GitLab
Projects
Groups
Snippets
Help
Loading...
Help
Help
Support
Community forum
Keyboard shortcuts
?
Submit feedback
Contribute to GitLab
Sign in
Toggle navigation
elpa
Project overview
Project overview
Details
Activity
Releases
Repository
Repository
Files
Commits
Branches
Tags
Contributors
Graph
Compare
Issues
15
Issues
15
List
Boards
Labels
Service Desk
Milestones
Merge Requests
0
Merge Requests
0
Operations
Operations
Incidents
Environments
Analytics
Analytics
Repository
Value Stream
Wiki
Wiki
Members
Members
Collapse sidebar
Close sidebar
Activity
Graph
Create a new issue
Commits
Issue Boards
Open sidebar
elpa
elpa
Commits
8666cdd0
Commit
8666cdd0
authored
Apr 16, 2018
by
Andreas Marek
Browse files
Options
Browse Files
Download
Email Patches
Plain Diff
Fix some errors in SPARC64 6hv kernel
parent
1e9b07d2
Changes
1
Pipelines
1
Hide whitespace changes
Inline
Side-by-side
Showing
1 changed file
with
6 additions
and
6 deletions
+6
-6
src/elpa2/kernels/real_sse_6hv_template.c
src/elpa2/kernels/real_sse_6hv_template.c
+6
-6
No files found.
src/elpa2/kernels/real_sse_6hv_template.c
View file @
8666cdd0
...
...
@@ -792,10 +792,10 @@ __forceinline void hh_trafo_kernel_8_SPARC64_6hv_single(float* q, float* hh, int
#ifdef HAVE_SPARC64_SSE
#ifdef DOUBLE_PRECISION_REAL
h3
=
_mm_set_pd
(
hh
[(
ldh
*
2
)
+
nb
-
3
],
hh
[(
ldh
*
2
)
+
nb
-
3
];
h3
=
_mm_set_pd
(
hh
[(
ldh
*
2
)
+
nb
-
3
],
hh
[(
ldh
*
2
)
+
nb
-
3
]
)
;
#endif
#ifdef SINGLE_PRECISION_REAL
h3
=
_mm_set_ps
(
hh
[(
ldh
*
2
)
+
nb
-
3
],
hh
[(
ldh
*
2
)
+
nb
-
3
];
h3
=
_mm_set_ps
(
hh
[(
ldh
*
2
)
+
nb
-
3
],
hh
[(
ldh
*
2
)
+
nb
-
3
]
)
;
#endif
#endif
...
...
@@ -856,7 +856,7 @@ __forceinline void hh_trafo_kernel_8_SPARC64_6hv_single(float* q, float* hh, int
#endif
#ifdef HAVE_SPARC64_SSE
#ifdef DOUBLE_PRECISION_REAL
h1
=
_mm_set_pd
(
hh
[
nb
-
4
]
),
hh
[
nb
-
4
]
;
h1
=
_mm_set_pd
(
hh
[
nb
-
4
]
,
hh
[
nb
-
4
])
;
#endif
#ifdef SINGLE_PRECISION_REAL
h1
=
_mm_set_ps
(
hh
[
nb
-
4
],
hh
[
nb
-
4
]);
...
...
@@ -1274,7 +1274,7 @@ __forceinline void hh_trafo_kernel_8_SPARC64_6hv_single(float* q, float* hh, int
__SSE_DATATYPE
vs_4_6
=
_mm_set1_pd
(
scalarprods
[
13
]);
__SSE_DATATYPE
vs_5_6
=
_mm_set1_pd
(
scalarprods
[
14
]);
#endif
#ifdef HAVE_SPARC64_SSE
_INTRINSICS
#ifdef HAVE_SPARC64_SSE
__SSE_DATATYPE
vs_3_6
=
_mm_set_pd
(
scalarprods
[
12
],
scalarprods
[
12
]);
__SSE_DATATYPE
vs_4_6
=
_mm_set_pd
(
scalarprods
[
13
],
scalarprods
[
13
]);
__SSE_DATATYPE
vs_5_6
=
_mm_set_pd
(
scalarprods
[
14
],
scalarprods
[
14
]);
...
...
@@ -1644,10 +1644,10 @@ __forceinline void hh_trafo_kernel_8_SPARC64_6hv_single(float* q, float* hh, int
#endif
#ifdef HAVE_SPARC64_SSE
#ifdef DOUBLE_PRECISION_REAL
h6
=
_mm_set
1
_pd
(
hh
[(
ldh
*
5
)
+
5
],
hh
[(
ldh
*
5
)
+
5
]);
h6
=
_mm_set_pd
(
hh
[(
ldh
*
5
)
+
5
],
hh
[(
ldh
*
5
)
+
5
]);
#endif
#ifdef SINGLE_PRECISION_REAL
h6
=
_mm_set
1
_ps
(
hh
[(
ldh
*
5
)
+
5
],
hh
[(
ldh
*
5
)
+
5
]);
h6
=
_mm_set_ps
(
hh
[(
ldh
*
5
)
+
5
],
hh
[(
ldh
*
5
)
+
5
]);
#endif
#endif
...
...
Write
Preview
Markdown
is supported
0%
Try again
or
attach a new file
.
Attach a file
Cancel
You are about to add
0
people
to the discussion. Proceed with caution.
Finish editing this message first!
Cancel
Please
register
or
sign in
to comment