Gitiles
Code Review
Sign In
LeafOS
/
LeafOS-Project
/
android_art
/
54f4fbd1a6834f06dc9b644b865423fdc03afb15
/
compiler
/
optimizing
/
loop_optimization.cc
54f4fbd
Remove MIPS support from Optimizing.
by Vladimir Marko
· 5 years ago
43f2f75
Revert^4 "Implement Dot Product Vectorization for x86"
by Alex Light
· 5 years ago
8e89500
Revert "Revert^2 "Implement Dot Product Vectorization for x86""
by Nicolas Geoffray
· 5 years ago
7cf5607
Revert^2 "Implement Dot Product Vectorization for x86"
by Vladimir Marko
· 5 years ago
b8c884e
Revert "Implement Dot Product Vectorization for x86"
by Vladimir Marko
· 5 years ago
4b7caee
Implement Dot Product Vectorization for x86
by Shalini Salomi Bodapati
· 5 years ago
0a51605
Revert "Make compiler/optimizing/ symbols hidden."
by Vladimir Marko
· 5 years ago
e272715
Make compiler/optimizing/ symbols hidden.
by Vladimir Marko
· 5 years ago
bac080b
Loop vectorizer should not remove instructions that can throw
by Georgia Kouveli
· 6 years ago
e65ade7
ART: Disable vectorization for debuggable graphs.
by Artem Serov
· 5 years ago
b45a435
Add AVX support for packed mul/div instructions.
by Shalini Salomi Bodapati
· 6 years ago
81d15be
Add AVX support for packed add/sub instructions on x86
by Shalini Salomi Bodapati
· 6 years ago
256c94b
Update induction ranges in superblock cloner.
by Nicolas Geoffray
· 6 years ago
aaac0e3
ART: ARM64: Support DotProd SIMD idiom.
by Artem Serov
· 6 years ago
18ba1da
ART: Implement loop full unrolling.
by Artem Serov
· 7 years ago
0e32908
ART: Refactor scalar loop optimizations.
by Artem Serov
· 7 years ago
a043111
Move instruction_set_ to CompilerOptions.
by Vladimir Marko
· 7 years ago
cf43fb6
ART: Enable scalar loop peeling and unrolling.
by Artem Serov
· 7 years ago
a8360cd
Perform rudimentary check on graph size for no-change assertions.
by Aart Bik
· 7 years ago
3f08e9b
Remove some SIMD recognition code.
by Aart Bik
· 7 years ago
2477320
Step 1 of 2: conditional passes.
by Aart Bik
· 7 years ago
6d05700
Run GVN earlier.
by Aart Bik
· 7 years ago
72411e6
ART: Implement scalar loop peeling.
by Artem Serov
· 7 years ago
5aac921
Enabled nested min-max SIMDization for narrower operands.
by Aart Bik
· 7 years ago
a3e2326
Revert "Revert "Refined add/sub analysis vis-a-vis SIMD idioms.""
by Nicolas Geoffray
· 7 years ago
7f31326
Revert "Refined add/sub analysis vis-a-vis SIMD idioms."
by Nicolas Geoffray
· 7 years ago
81a1f85
Refined add/sub analysis vis-a-vis SIMD idioms.
by Aart Bik
· 7 years ago
2a93c76
Merge "ART: Implement scalar loop unrolling."
by Aart Bik
· 7 years ago
121f203
ART: Implement scalar loop unrolling.
by Artem Serov
· 7 years ago
5a39276
Recognize signed saturation in single clipping.
by Aart Bik
· 7 years ago
1a38102
Minor cleanup of saturation arithmetic code.
by Aart Bik
· 7 years ago
29aa082
Vectorization of saturation arithmetic.
by Aart Bik
· 7 years ago
1f8d51b
Introduce MIN/MAX/ABS as HIR nodes.
by Aart Bik
· 7 years ago
3b2a595
Introduce ABS as HIR nodes (missing file).
by Aart Bik
· 7 years ago
66c158e
Clean up signed/unsigned in vectorizer.
by Aart Bik
· 7 years ago
2dd7b67
Fixed spilling bug (visible on ARM64): missed SIMD type.
by Aart Bik
· 7 years ago
2ca10eb
Refactored optimization passes setup.
by Aart Bik
· 7 years ago
38e380b
MIPS: Implement Sum-of-Abs-Differences
by Lena Djokic
· 7 years ago
2ffb703
cpplint: Cleanup errors
by Igor Murashkin
· 7 years ago
33bff25
ART: Make InstructionSet an enum class and add kLast.
by Vladimir Marko
· 7 years ago
13e6f29
Merge "Alignment optimizations in vectorizer."
by Aart Bik
· 7 years ago
38a3f21
Alignment optimizations in vectorizer.
by Aart Bik
· 7 years ago
e434c4f
MIPS: Basic SIMD reduction support.
by Lena Djokic
· 7 years ago
4d1a9d4
Improve sign and zero extension analysis.
by Aart Bik
· 7 years ago
89b8df0
Fix min/max SIMD reduction
by Goran Jakovljevic
· 7 years ago
6e9b137
ARM: Support SIMD reduction for 32-bit backend.
by Artem Serov
· 7 years ago
ca6fff8
ART: Use ScopedArenaAllocator for pass-local data.
by Vladimir Marko
· 7 years ago
46b6dbc
Try to preserve dex pc better in vector code.
by Aart Bik
· 7 years ago
d5d2f2c
ART: Introduce Uint8 compiler data type.
by Vladimir Marko
· 7 years ago
df011c3
Generalized zero/sign-ext analysis. Generalized SAD.
by Aart Bik
· 7 years ago
68ca702
Added SAD test. Generalized vector analysis of narrow type.
by Aart Bik
· 7 years ago
0ebe0d8
ART: Introduce compiler data type.
by Vladimir Marko
· 7 years ago
dbbac8f
Implement Sum-of-Abs-Differences idiom recognition.
by Aart Bik
· 7 years ago
521b50f
No unrolling for large loop bodies.
by Aart Bik
· 7 years ago
f26bb6c
ARM64: Tune SIMD loop unrolling factor heuristic.
by Artem Serov
· 7 years ago
b92cc33
Pass stats into the loop optimization phase.
by Aart Bik
· 7 years ago
21b8592
Added vectorization stats.
by Aart Bik
· 7 years ago
0148de4
Basic SIMD reduction support.
by Aart Bik
· 7 years ago
982334c
Revert "Basic SIMD reduction support."
by Nicolas Geoffray
· 7 years ago
cfa59b4
Basic SIMD reduction support.
by Aart Bik
· 7 years ago
a57b4ee
Revert "Basic SIMD reduction support."
by Aart Bik
· 7 years ago
9879d0e
Basic SIMD reduction support.
by Aart Bik
· 7 years ago
671e48a
Fix performance regression.
by Aart Bik
· 7 years ago
b29f684
Set basic framework for detecting reductions.
by Aart Bik
· 7 years ago
21c7e6f
ART: Fix SimplifyInduction for an instruction with HEnvironment.
by Artem Serov
· 7 years ago
8cf9cb3
ART: Include cleanup
by Andreas Gampe
· 8 years ago
51765b0
MIPS32: ART Vectorizer
by Lena Djokic
· 8 years ago
e128af5
Merge "ARM: ART Vectorizer (64-bit vectors)."
by Treehugger Robot
· 8 years ago
8f7c410
ARM: ART Vectorizer (64-bit vectors).
by Artem Serov
· 8 years ago
37dc4df
Improved subscript and data dependence analysis.
by Aart Bik
· 8 years ago
76754cc
Merge "Prevent loop optimization in debuggable mode."
by Aart Bik
· 8 years ago
f57c1ae
Prevent loop optimization in debuggable mode.
by Nicolas Geoffray
· 8 years ago
14a68b4
Unrolling and dynamic loop peeling framework in vectorizer.
by Aart Bik
· 8 years ago
1a0a519
Fix loop optimization in the presence of environment uses.
by Nicolas Geoffray
· 8 years ago
13600e9
Merge "MIPS64: Min/max vectorization support"
by Aart Bik
· 8 years ago
f45d61c
ART: Fix or disable some tidy warnings.
by Andreas Gampe
· 8 years ago
8fea1e1
MIPS64: Min/max vectorization support
by Goran Jakovljevic
· 8 years ago
6e92fb3
Pass through inputs beyond arguments in invoke.
by Aart Bik
· 8 years ago
19680d3
MIPS64: ART Vectorizer
by Goran Jakovljevic
· 8 years ago
304c8a5
Support for narrow operands in "dangerous" operations.
by Aart Bik
· 8 years ago
9231690
Revert "Support for narrow operands in "dangerous" operations."
by Nicolas Geoffray
· 8 years ago
636e870
Support for narrow operands in "dangerous" operations.
by Aart Bik
· 8 years ago
5f80500
Made idiom recognition more robust.
by Aart Bik
· 8 years ago
c8e93c7
Min/max SIMDization support.
by Aart Bik
· 8 years ago
50e20d5
Moved knowledge on masking shift operands to simplifier.
by Aart Bik
· 8 years ago
65ffd8e
Bug fix on shift that exceeds "lane width".
by Aart Bik
· 8 years ago
472821b
Enable string "array get" vectorization.
by Aart Bik
· 8 years ago
db14fcf
Pack booleans in the already existing bit field.
by Aart Bik
· 8 years ago
f3e61ee
Implement halving add idiom (with checker tests).
by Aart Bik
· 8 years ago
27fb1dc
Merge "Fixed bug on pending environment use of termination condition. With regression test."
by Aart Bik
· 8 years ago
d86c085
Fixed bug on pending environment use of termination condition.
by Aart Bik
· 8 years ago
4766f2c
Merge changes I1d4db176,Ifb931a99
by Aart Bik
· 8 years ago
3101e58
Fix bug in vectorization of charAt, with regression test
by Aart Bik
· 8 years ago
9899026
Add checker part of test, fix intrinsic copying
by Aart Bik
· 8 years ago
b31f91f
ARM64: Support vectorization for double and long.
by Artem Serov
· 8 years ago
d4bccf1
ARM64: Support 128-bit registers for SIMD.
by Artem Serov
· 8 years ago
fa76296
Fixed missing context while detecting unit strides.
by Aart Bik
· 8 years ago
24b905f
Fix a few comments in vectorization code that were incorrect or incomplete.
by Aart Bik
· 8 years ago
6bca1c6
Merge "Ensure environment is ready when populating loop."
by Aart Bik
· 8 years ago
b07d1bc
Ensure environment is ready when populating loop.
by Aart Bik
· 8 years ago
Next »