Skip to content
GitLab
Explore
Sign in
Register
Primary navigation
Search or go to…
Project
A
aom-rav1e
Manage
Activity
Members
Labels
Plan
Issues
Issue boards
Milestones
Wiki
Code
Merge requests
Repository
Branches
Commits
Tags
Repository graph
Compare revisions
Snippets
Build
Pipelines
Jobs
Pipeline schedules
Artifacts
Deploy
Releases
Container Registry
Model registry
Operate
Environments
Monitor
Incidents
Analyze
Value stream analytics
Contributor analytics
CI/CD analytics
Repository analytics
Model experiments
Help
Help
Support
GitLab documentation
Compare GitLab plans
Community forum
Contribute to GitLab
Provide feedback
Terms and privacy
Keyboard shortcuts
?
Snippets
Groups
Projects
Show more breadcrumbs
Xiph.Org
aom-rav1e
Commits
b85367a6
Commit
b85367a6
authored
11 years ago
by
Hangyu Kuang
Committed by
Gerrit Code Review
11 years ago
Browse files
Options
Downloads
Plain Diff
Merge "Optimise idct4x4: rearrange the instructions a bit to improve instruction scheduling."
parents
9655c2c7
610642c1
No related branches found
Branches containing commit
No related tags found
No related merge requests found
Changes
1
Hide whitespace changes
Inline
Side-by-side
Showing
1 changed file
vp9/common/arm/neon/vp9_short_idct4x4_add_neon.asm
+19
-12
19 additions, 12 deletions
vp9/common/arm/neon/vp9_short_idct4x4_add_neon.asm
with
19 additions
and
12 deletions
vp9/common/arm/neon/vp9_short_idct4x4_add_neon.asm
+
19
−
12
View file @
b85367a6
...
...
@@ -45,10 +45,6 @@
; cospi_24_64 = 6270 = 0x 187e
mov
r12
,
#
0x1800
add
r12
,
#
0x7e
; generate constant vectors
vdup.16
d20
,
r0
; ; replicate cospi_8_64
vdup.16
d21
,
r3
; ; replicate cospi_16_64
vdup.16
d22
,
r12
; ; replicate cospi_24_64
; transpose the input data
; 00 01 02 03 d16
...
...
@@ -57,6 +53,11 @@
; 30 31 32 33 d19
vtrn.16
d16
,
d17
vtrn.16
d18
,
d19
; generate constant vectors
vdup.16
d20
,
r0
; replicate cospi_8_64
vdup.16
d21
,
r3
; replicate cospi_16_64
; 00 10 02 12 d16
; 01 11 03 13 d17
; 20 30 22 32 d18
...
...
@@ -67,20 +68,24 @@
; 02 12 22 32 d18
; 03 13 23 33 d19
vdup.16
d22
,
r12
; replicate cospi_24_64
; do the transform on transposed rows
; stage 1
vadd.s16
d23
,
d16
,
d18
; (input[0] + input[2])
vsub.s16
d24
,
d16
,
d18
; (input[0] - input[2])
vmull.s16
q15
,
d17
,
d22
; input[1] * cospi_24_64
vmull.s16
q1
,
d17
,
d20
; input[1] * cospi_8_64
; (input[0] + input[2]) * cospi_16_64;
; (input[0] - input[2]) * cospi_16_64;
vadd.s16
d23
,
d16
,
d18
vsub.s16
d24
,
d16
,
d18
vmull.s16
q13
,
d23
,
d21
vmull.s16
q14
,
d24
,
d21
; input[1] * cospi_24_64 - input[3] * cospi_8_64;
; input[1] * cospi_8_64 + input[3] * cospi_24_64;
vmull.s16
q15
,
d17
,
d22
vmull.s16
q1
,
d17
,
d20
vmlsl.s16
q15
,
d19
,
d20
vmlal.s16
q1
,
d19
,
d22
...
...
@@ -119,17 +124,19 @@
; do the transform on columns
; stage 1
vadd.s16
d23
,
d16
,
d18
; (input[0] + input[2])
vsub.s16
d24
,
d16
,
d18
; (input[0] - input[2])
vmull.s16
q15
,
d17
,
d22
; input[1] * cospi_24_64
vmull.s16
q1
,
d17
,
d20
; input[1] * cospi_8_64
; (input[0] + input[2]) * cospi_16_64;
; (input[0] - input[2]) * cospi_16_64;
vadd.s16
d23
,
d16
,
d18
vsub.s16
d24
,
d16
,
d18
vmull.s16
q13
,
d23
,
d21
vmull.s16
q14
,
d24
,
d21
; input[1] * cospi_24_64 - input[3] * cospi_8_64;
; input[1] * cospi_8_64 + input[3] * cospi_24_64;
vmull.s16
q15
,
d17
,
d22
vmull.s16
q1
,
d17
,
d20
vmlsl.s16
q15
,
d19
,
d20
vmlal.s16
q1
,
d19
,
d22
...
...
This diff is collapsed.
Click to expand it.
Preview
0%
Loading
Try again
or
attach a new file
.
Cancel
You are about to add
0
people
to the discussion. Proceed with caution.
Finish editing this message first!
Save comment
Cancel
Please
register
or
sign in
to comment