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
10
Issues
10
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
1e1812bc
Commit
1e1812bc
authored
Aug 18, 2016
by
Andreas Marek
Browse files
Options
Browse Files
Download
Email Patches
Plain Diff
Single precision kernel for AVX512 real block6
parent
5225392a
Changes
5
Pipelines
1
Expand all
Hide whitespace changes
Inline
Side-by-side
Showing
5 changed files
with
3491 additions
and
1366 deletions
+3491
-1366
.gitlab-ci.yml
.gitlab-ci.yml
+13
-0
Makefile.am
Makefile.am
+3
-3
src/elpa2_kernels/elpa2_kernels_real_avx512_4hv_single_precision.c
..._kernels/elpa2_kernels_real_avx512_4hv_single_precision.c
+4
-4
src/elpa2_kernels/elpa2_kernels_real_avx512_6hv_single_precision.c
..._kernels/elpa2_kernels_real_avx512_6hv_single_precision.c
+3411
-1299
src/mod_compute_hh_trafo_real.F90
src/mod_compute_hh_trafo_real.F90
+60
-60
No files found.
.gitlab-ci.yml
View file @
1e1812bc
...
...
@@ -2192,6 +2192,19 @@ intel-double-precision-mpi-noopenmp-ftimings-redirect-real-avx512_block6-complex
-
export LD_LIBRARY_PATH=$MKL_HOME/lib/intel64:$LD_LIBRARY_PATH
-
make check TEST_FLAGS='1000 500 128'
intel-single-precision-mpi-noopenmp-ftimings-redirect-real-avx512_block6-complex-avx512_block1-kernel-jobs
:
tags
:
-
KNL
script
:
-
./autogen.sh
-
./configure FC=mpiifort CC=mpiicc CFLAGS="-O3 -mtune=knl -axMIC-AVX512" FCFLAGS="-O3 -mtune=knl -axMIC-AVX512" SCALAPACK_FCFLAGS="-L$MKLROOT/lib/intel64 -lmkl_scalapack_lp64 -lmkl_intel_lp64 -lmkl_sequential -lmkl_core -lmkl_blacs_intelmpi_lp64 -lpthread -lm -I$MKLROOT/include/intel64/lp64" SCALAPACK_LDFLAGS="-L$MKLROOT/lib/intel64 -lmkl_scalapack_lp64 -lmkl_intel_lp64 -lmkl_sequential -lmkl_core -lmkl_blacs_intelmpi_lp64 -lpthread -lm -Wl,-rpath,$MKLROOT/lib/intel64" --with-real-avx512_block6-kernel-only --with-complex-avx512_block1-kernel-only --enable-single-precision
-
/home/elpa/wait_until_midnight.sh
-
make -j
8
-
export OMP_NUM_THREADS=1
-
export LD_LIBRARY_PATH=$MKL_HOME/lib/intel64:$LD_LIBRARY_PATH
-
make check TEST_FLAGS='1000 500 128'
intel-set-kernel-via-environment-variable-mpi-openmp-job
:
tags
:
-
cpu
...
...
Makefile.am
View file @
1e1812bc
...
...
@@ -204,9 +204,9 @@ endif
if
WITH_REAL_AVX512_BLOCK6_KERNEL
libelpa@SUFFIX@
_private_la_SOURCES
+=
src/elpa2_kernels/elpa2_kernels_real_avx512_6hv_double_precision.c
#
if WANT_SINGLE_PRECISION_REAL
#
libelpa@SUFFIX@_private_la_SOURCES += src/elpa2_kernels/elpa2_kernels_real_avx512_6hv_single_precision.c
#
endif
if
WANT_SINGLE_PRECISION_REAL
libelpa@SUFFIX@
_private_la_SOURCES
+=
src/elpa2_kernels/elpa2_kernels_real_avx512_6hv_single_precision.c
endif
endif
...
...
src/elpa2_kernels/elpa2_kernels_real_avx512_4hv_single_precision.c
View file @
1e1812bc
...
...
@@ -495,7 +495,7 @@ __forceinline void hh_trafo_kernel_64_AVX512_4hv_single(float* q, float* hh, int
q3
=
_mm512_NFMA_ps
(
x3
,
h1
,
q3
);
q3
=
_mm512_NFMA_ps
(
y3
,
h2
,
q3
);
q3
=
_mm512_NFMA_ps
(
z3
,
h3
,
q3
);
q3
=
_mm512_NFMA_p
d
(
w3
,
h4
,
q3
);
q3
=
_mm512_NFMA_p
s
(
w3
,
h4
,
q3
);
_mm512_store_ps
(
&
q
[(
i
*
ldq
)
+
32
],
q3
);
q4
=
_mm512_load_ps
(
&
q
[(
i
*
ldq
)
+
48
]);
...
...
@@ -932,7 +932,7 @@ __forceinline void hh_trafo_kernel_48_AVX512_4hv_single(float* q, float* hh, int
q3
=
_mm512_NFMA_ps
(
x3
,
h1
,
q3
);
q3
=
_mm512_NFMA_ps
(
y3
,
h2
,
q3
);
q3
=
_mm512_NFMA_ps
(
z3
,
h3
,
q3
);
q3
=
_mm512_NFMA_p
d
(
w3
,
h4
,
q3
);
q3
=
_mm512_NFMA_p
s
(
w3
,
h4
,
q3
);
_mm512_store_ps
(
&
q
[(
i
*
ldq
)
+
32
],
q3
);
// q4 = _mm512_load_ps(&q[(i*ldq)+48]);
...
...
@@ -1369,7 +1369,7 @@ __forceinline void hh_trafo_kernel_32_AVX512_4hv_single(float* q, float* hh, int
// q3 = _mm512_NFMA_ps(x3, h1, q3);
// q3 = _mm512_NFMA_ps(y3, h2, q3);
// q3 = _mm512_NFMA_ps(z3, h3, q3);
// q3 = _mm512_NFMA_p
d
(w3, h4, q3);
// q3 = _mm512_NFMA_p
s
(w3, h4, q3);
// _mm512_store_ps(&q[(i*ldq)+32],q3);
// q4 = _mm512_load_ps(&q[(i*ldq)+48]);
...
...
@@ -1806,7 +1806,7 @@ __forceinline void hh_trafo_kernel_16_AVX512_4hv_single(float* q, float* hh, int
// q3 = _mm512_NFMA_ps(x3, h1, q3);
// q3 = _mm512_NFMA_ps(y3, h2, q3);
// q3 = _mm512_NFMA_ps(z3, h3, q3);
// q3 = _mm512_NFMA_p
d
(w3, h4, q3);
// q3 = _mm512_NFMA_p
s
(w3, h4, q3);
// _mm512_store_ps(&q[(i*ldq)+32],q3);
// q4 = _mm512_load_ps(&q[(i*ldq)+48]);
...
...
src/elpa2_kernels/elpa2_kernels_real_avx512_6hv_single_precision.c
View file @
1e1812bc
This diff is collapsed.
Click to expand it.
src/mod_compute_hh_trafo_real.F90
View file @
1e1812bc
This diff is collapsed.
Click to expand it.
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