-
Notifications
You must be signed in to change notification settings - Fork 559
Commit
This commit does not belong to any branch on this repository, and may belong to a fork outside of the repository.
Add OpenBLAS patch for aarch64 (#987)
- Loading branch information
Showing
2 changed files
with
80 additions
and
0 deletions.
There are no files selected for viewing
40 changes: 40 additions & 0 deletions
40
...enBLAS32@0.3.9/bundled/patches/fix-initialization-to-zero-in-the-NEON-SGEMM_BETA-ke.patch
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| Original file line number | Diff line number | Diff line change |
|---|---|---|
| @@ -0,0 +1,40 @@ | ||
| From 1c64c55baca23e2a18b78a8391d83cfbc2cb56a7 Mon Sep 17 00:00:00 2001 | ||
| From: Martin Kroeker <martin@ruby.chemie.uni-freiburg.de> | ||
| Date: Tue, 31 Mar 2020 16:53:56 +0200 | ||
| Subject: [PATCH] fix initialization to zero in the NEON SGEMM_BETA kernel as | ||
| well | ||
|
|
||
| --- | ||
| kernel/arm64/sgemm_beta.S | 16 ++++++++-------- | ||
| 1 file changed, 8 insertions(+), 8 deletions(-) | ||
|
|
||
| diff --git a/kernel/arm64/sgemm_beta.S b/kernel/arm64/sgemm_beta.S | ||
| index a3b97e23..574485bc 100755 | ||
| --- a/kernel/arm64/sgemm_beta.S | ||
| +++ b/kernel/arm64/sgemm_beta.S | ||
| @@ -81,14 +81,14 @@ USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. | ||
| .endm | ||
|
|
||
| .macro INIT_ZERO | ||
| - fmul v0.4s, v0.4s, betaV0 | ||
| - fmul v1.4s, v1.4s, betaV0 | ||
| - fmul v2.4s, v2.4s, betaV0 | ||
| - fmul v3.4s, v3.4s, betaV0 | ||
| - fmul v4.4s, v4.4s, betaV0 | ||
| - fmul v5.4s, v5.4s, betaV0 | ||
| - fmul v6.4s, v6.4s, betaV0 | ||
| - fmul v7.4s, v7.4s, betaV0 | ||
| + movi v0.4s, #0x0 | ||
| + movi v1.4s, #0x0 | ||
| + movi v2.4s, #0x0 | ||
| + movi v3.4s, #0x0 | ||
| + movi v4.4s, #0x0 | ||
| + movi v5.4s, #0x0 | ||
| + movi v6.4s, #0x0 | ||
| + movi v7.4s, #0x0 | ||
| .endm | ||
|
|
||
| /************************************************************************************** | ||
| -- | ||
| 2.25.1 | ||
|
|
40 changes: 40 additions & 0 deletions
40
...OpenBLAS@0.3.9/bundled/patches/fix-initialization-to-zero-in-the-NEON-SGEMM_BETA-ke.patch
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| Original file line number | Diff line number | Diff line change |
|---|---|---|
| @@ -0,0 +1,40 @@ | ||
| From 1c64c55baca23e2a18b78a8391d83cfbc2cb56a7 Mon Sep 17 00:00:00 2001 | ||
| From: Martin Kroeker <martin@ruby.chemie.uni-freiburg.de> | ||
| Date: Tue, 31 Mar 2020 16:53:56 +0200 | ||
| Subject: [PATCH] fix initialization to zero in the NEON SGEMM_BETA kernel as | ||
| well | ||
|
|
||
| --- | ||
| kernel/arm64/sgemm_beta.S | 16 ++++++++-------- | ||
| 1 file changed, 8 insertions(+), 8 deletions(-) | ||
|
|
||
| diff --git a/kernel/arm64/sgemm_beta.S b/kernel/arm64/sgemm_beta.S | ||
| index a3b97e23..574485bc 100755 | ||
| --- a/kernel/arm64/sgemm_beta.S | ||
| +++ b/kernel/arm64/sgemm_beta.S | ||
| @@ -81,14 +81,14 @@ USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. | ||
| .endm | ||
|
|
||
| .macro INIT_ZERO | ||
| - fmul v0.4s, v0.4s, betaV0 | ||
| - fmul v1.4s, v1.4s, betaV0 | ||
| - fmul v2.4s, v2.4s, betaV0 | ||
| - fmul v3.4s, v3.4s, betaV0 | ||
| - fmul v4.4s, v4.4s, betaV0 | ||
| - fmul v5.4s, v5.4s, betaV0 | ||
| - fmul v6.4s, v6.4s, betaV0 | ||
| - fmul v7.4s, v7.4s, betaV0 | ||
| + movi v0.4s, #0x0 | ||
| + movi v1.4s, #0x0 | ||
| + movi v2.4s, #0x0 | ||
| + movi v3.4s, #0x0 | ||
| + movi v4.4s, #0x0 | ||
| + movi v5.4s, #0x0 | ||
| + movi v6.4s, #0x0 | ||
| + movi v7.4s, #0x0 | ||
| .endm | ||
|
|
||
| /************************************************************************************** | ||
| -- | ||
| 2.25.1 | ||
|
|