| 6d0bcdc8329dde4d3c83e95475bc670002f41309 |
|
17-Jan-2011 |
Henrik Smiding <henrik.smiding@stericsson.com> |
Add optimized version of memcpy for Cortex A9
Adds new code to memcpy function, optimized for Cortex A9.
Adds new ARM-only loop, for operations where source and
destination are aligned.
Copyright (C) ST-Ericsson SA 2010
Modified neon implementation to fit Cortex A9 cache line size,
for those running 32 bytes L2 cache line size.
Also split the implementation in aligned and unaligned access,
for those that allows unaligned memory access with Neon.
For totally aligned operations, arm-only code is used.
Change-Id: I95ebf6164cd6486b12a7e3e98e369db21e7e18d2
Author: Henrik Smiding henrik.smiding@stericsson.com for ST-Ericsson.
Signed-off-by: Christian Bejram <christian.bejram@stericsson.com>
|
| 08e72d0161e39e99ff1003bf1ce894f37d7b7eb4 |
|
07-May-2012 |
Prajakta Gudadhe <pgudadhe@nvidia.com> |
bionic: add support for non-NEON memcpy() on NEON SoCs
Some SoCs that support NEON nevertheless perform better with a non-NEON than a
NEON memcpy(). This patch adds build variable ARCH_ARM_USE_NON_NEON_MEMCPY,
which can be set in BoardConfig.mk. When ARCH_ARM_USE_NON_NEON_MEMCPY is
defined, we compile in the non-NEON optimized memcpy() even if the SoC supports
NEON.
Change-Id: Ia0e5bee6bad5880ffc5ff8f34a1382d567546cf9
|
| fe6338da9168330d44b409b2ee36103e8bfe6697 |
|
15-Sep-2010 |
Henrik Smiding <henrik.smiding@stericsson.com> |
Adjust memcpy for ARM Cortex A9 cache line size
ARM Cortex A8 use 64 bytes and ARM Cortex A9 use 32 bytes cache line
size.
The following patch:
Adds code to adjust memcpy cache line size to match A9 cache line
size.
Adds a flag to select between 32 bytes and 64 bytes cache line
size.
Copyright (C) ST-Ericsson SA 2010
Modified neon implementation to fit Cortex A9 cache line size
Author: Henrik Smiding henrik.smiding@stericsson.com for
ST-Ericsson.
Change-Id: I8a55946bfb074e6ec0a14805ed65f73fcd0984a3
Signed-off-by: Christian Bejram <christian.bejram@stericsson.com>
|
| 487b613e572160e80d0700e1bcd0e405420d14ea |
|
04-Oct-2011 |
Evgeniy Stepanov <eugenis@google.com> |
Use ENTRY and EXIT macros for strcmp, memcpy, atexit.
Without this change strcmp size is zero (not set), and it gets
ignored by Valgrind. Changes to memcpy and atexit don't affect the
generated binary in any way.
Change-Id: I05818cb5951f75901dc8c0eef02807a2e83a9231
|
| 420878c6908cf9c2862888477ec3f424a06cf172 |
|
16-Feb-2011 |
Kenny Root <kroot@google.com> |
Add function marks and size indications
Add a macro to annotate function end and start using both ENTRY and END
for each function. This allows valgrind (and presumably other debugging
tools) to use the debug symbols to trace the functions.
Change-Id: I5f09cef8e22fb356eb6f5cee952b031e567599b6
|
| ecede4022283650ee4c4ed117a792d6db0b929fd |
|
10-Mar-2010 |
Colin Cross <ccross@android.com> |
Only use NEON memcpy if __ARM_NEON__ is defined
Change-Id: I32e6b9385d46efeec15dee8e395a82eef24ba3ea
|
| 194d3fa048cf909ca592dd56fa538dc9cd3f5ddb |
|
13-Nov-2009 |
Jean-Baptiste Queru <jbq@google.com> |
eclair snapshot
|
| 199f9d923804d74e021dd80e48ec75c0a96dba77 |
|
28-Oct-2009 |
Mathias Agopian <mathias@google.com> |
Improve memcpy performance from 290 MiB/s to 340 MiB/s (17% improvment)
use 64 bytes cache lines, reduce the main loop to 64-bytes instead of
128 bytes and adjust the prefetch distance to the optimal value.
|
| fdc5c1f56f9d21034badb8e4b092c47098f19613 |
|
05-Oct-2009 |
David 'Digit' Turner <digit@google.com> |
Re-enable ARMv7 memcpy implementation.
Do not submit this patch before the one that modifies the Android emulator to
work-around a weird ARMv7 emulation issue. This is done to temporarily re-allow
the -user builds needed for QA.
|
| 1e40783d9ad29c88a4d547774e05080fc4327a2e |
|
03-Oct-2009 |
David 'Digit' Turner <digit@google.com> |
Fix armv7-user builds by disabling the ARMv7-optimized memcpy.
This is required to work-around some corny bugs in ARMv7 emulation.
The emulation itself is required to run the dex pre-optimization pass
for -user builds.
|
| ee223d02d96815c989b62043ff1237b1cd4e14b0 |
|
28-Sep-2009 |
Mathias Agopian <mathias@google.com> |
NEON optimized memcpy.
372 MB/s for large transfers, 440 MB/s for smaller ones down to 1KB. 130 MB/s for very small transfers ( < 32 bytes )
Performance is similar with non-congruent buffers.
|
| f355096a64b74c8e869527de55f7e908873e3128 |
|
27-Sep-2009 |
David 'Digit' Turner <digit@google.com> |
Remove NEON optimizations for memcpy
|
| bd192b470b69e00e9313680b70c5572a609e535d |
|
15-Sep-2009 |
Ben Cheng <bccheng@google.com> |
Add stack unwinding directives to assembly leaf functions.
So that the real culprit of native crashes can surface in the stack trace.
|
| 1bbc56cd227546cb155bb47721cdb717780a3400 |
|
26-Aug-2009 |
David 'Digit' Turner <digit@google.com> |
Neon-optimized versions of memcpy.
This optimization come from the external 0xdroid repository.
Original patch can be found here:
http://gitorious.org/0xdroid/bionic/commit/ebafe41c2c02f8c09a3c1d7746047083df180ac5
|
| 1dc9e472e19acfe6dc7f41e429236e7eef7ceda1 |
|
04-Mar-2009 |
The Android Open Source Project <initial-contribution@android.com> |
auto import from //depot/cupcake/@135843
|
| 1767f908af327fa388b1c66883760ad851267013 |
|
04-Mar-2009 |
The Android Open Source Project <initial-contribution@android.com> |
auto import from //depot/cupcake/@135843
|
| 4e468ed2eb86a2406e14f1eca82072ee501d05fd |
|
18-Dec-2008 |
The Android Open Source Project <initial-contribution@android.com> |
Code drop from //branches/cupcake/...@124589
|
| e0055e0f99732dbaea81fd1c7055efb7b506c221 |
|
04-Dec-2008 |
Jean-Baptiste Queru <jbq@google.com> |
Introduce a new ARM header file <machine/cpu-features.h>
Introduce a new header file containing ARM-specific feature
test macros (e.g. __ARM_HAVE_PAIR_LOAD_STORE corresponding
to ldrd/strd instructions). Also modify a few files in our
system to use the macros in order to build for ARMv4T.
|
| a27d2baa0c1a2ec70f47ea9199b1dd6762c8a349 |
|
21-Oct-2008 |
The Android Open Source Project <initial-contribution@android.com> |
Initial Contribution
|