| 2d1fdb26e458c4ddc04155c1d421bced3ba90cd0 |
|
29-May-2014 |
Stephen Hines <srhines@google.com> |
Update compiler-rt aosp/master for 3.5 (r209699) rebase.
Change-Id: I158a30186f0faea2e2400e9dfdd878db2eb40e90
/external/compiler-rt/SDKs/linux/usr/include/stdio.h
|
| d06f2fc03087e7cccda19bd550bf6fb3bc2f5ae7 |
|
23-May-2013 |
Bill Wendling <isanbard@gmail.com> |
Performance improvement.
Using fwrite and fread was very *very* slow. The resulting code was multiple
times slower than GCC's implementation of gcov. Replace the fwrite/fread system
with an mmap() version.
If the `.gcda' file doesn't exist, we (re)allocate a buffer that we write
into. That gets written to the `.gcda' file in one chunk. If the `.gcda' file
already exists, we simply mmap() the file, modify the mapped data, and use
msync() to write the contents out to disk. It's much easier than implementing
our own buffering scheme, and we don't have to use fwrite's and fread's
buffering.
For those who are numbers-oriented, here are some timings:
GCC Verison
-----------
`.gcda' files don't exist: 23s
`.gcda' files do exist: 14s
LLVM Version (before this change)
---------------------------------
`.gcda' files don't exist: 28s
`.gcda' files do exist: 28s
LLVM Version (with this change)
-------------------------------
`.gcda' files don't exist: 18s
`.gcda' files do exist: 4s
It's a win-win-win-win-lose-win-win scenario!
<rdar://problem/13466086>
git-svn-id: https://llvm.org/svn/llvm-project/compiler-rt/trunk@182563 91177308-0d34-0410-b5e6-96231b3b80d8
/external/compiler-rt/SDKs/linux/usr/include/stdio.h
|
| 2a46a60510948b5005d62025fbba307a0323ed8d |
|
15-May-2013 |
Bill Wendling <isanbard@gmail.com> |
Try to improve performance by using a read/write buffer instead of I/O.
The calls to fwrite/fread can be very expensive. GCC avoids this by using a
buffer to read and write from the file, thus limiting the number of fwrite/fread
calls.
<rdar://problem/13466086>
git-svn-id: https://llvm.org/svn/llvm-project/compiler-rt/trunk@181924 91177308-0d34-0410-b5e6-96231b3b80d8
/external/compiler-rt/SDKs/linux/usr/include/stdio.h
|
| 346e348af5740db3660c84490d2c1bbba7c570e5 |
|
04-May-2013 |
Bill Wendling <isanbard@gmail.com> |
I was wrong in my testing.
There isn't a speedup when using unbuffered I/O. It slows it down in fact.
git-svn-id: https://llvm.org/svn/llvm-project/compiler-rt/trunk@181060 91177308-0d34-0410-b5e6-96231b3b80d8
/external/compiler-rt/SDKs/linux/usr/include/stdio.h
|
| d83a6d8bafbf2afa3ebcf9e006bf0c7414f9b2dc |
|
03-May-2013 |
Bill Wendling <isanbard@gmail.com> |
Use unbuffered I/O. This reduces the runtime by about half. Our implementation is now only about 5 times slower than gcc's.
git-svn-id: https://llvm.org/svn/llvm-project/compiler-rt/trunk@180980 91177308-0d34-0410-b5e6-96231b3b80d8
/external/compiler-rt/SDKs/linux/usr/include/stdio.h
|
| 7ad8ae259389e8652a5476681ac01185105c3d68 |
|
17-Sep-2012 |
Benjamin Kramer <benny.kra@googlemail.com> |
Update linux stdio.h to support the recent additions to GCDAProfiling.
Hopefully unbreaks compiler-rt build on linux.
git-svn-id: https://llvm.org/svn/llvm-project/compiler-rt/trunk@164025 91177308-0d34-0410-b5e6-96231b3b80d8
/external/compiler-rt/SDKs/linux/usr/include/stdio.h
|
| 86f9994936336675dcc1454ede5857e8851080b6 |
|
08-Dec-2011 |
Daniel Dunbar <daniel@zuster.org> |
SDKs: Sketch an initial stub SDK for Linux, I believe this suffices for building
the main compiler-rt and profile modules, at least on x86.
git-svn-id: https://llvm.org/svn/llvm-project/compiler-rt/trunk@146131 91177308-0d34-0410-b5e6-96231b3b80d8
/external/compiler-rt/SDKs/linux/usr/include/stdio.h
|