Commit Graph

22 Commits

Author SHA1 Message Date
Elliott Hughes
625993dfbb Use VDSO for clock_gettime(2) and gettimeofday(2).
Bug: 15387103
Change-Id: Ifc3608ea65060c1dc38120b10b6e79874f182a36
2014-07-16 14:27:43 -07:00
Christopher Ferris
53531ccebb Make sure not to construct illegal property names.
Change-Id: I37624e69aec51efd4291f076fb87af3f35d33025
2014-07-15 19:26:28 -07:00
Brigid Smith
28417e6314 Added a __system_property_serial benchmark.
Change-Id: Ifc2116f26acb242f4dcdb6eefe128899a9dc63bb
2014-07-09 15:48:37 -07:00
Brigid Smith
a304476145 Added __system_propery_read benchmark.
Change-Id: Ic0e9118859a013aa04841de25822a81a7fc74b3c
2014-07-09 10:26:17 -07:00
Elliott Hughes
7086ad6919 Cache getpid.
In practice, with this implementation we never need to make a system call.
We get the main thread's tid (which is the same as our pid) back from
the set_tid_address system call we have to make during initialization.
A new pthread will have the same pid as its parent, and a fork child's
main (and only) thread will have a pid equal to its tid, which we get for
free from the kernel before clone returns.

The only time we'd actually have to make a getpid system call now is if
we take a signal during fork and the signal handler calls getpid. (That,
or we call getpid in the dynamic linker while it's still dealing with its
own relocations and hasn't even set up the main thread yet.)

Bug: 15387103
Change-Id: I6d4718ed0a5c912fc75b5f738c49a023dbed5189
2014-06-20 09:06:57 -07:00
Elliott Hughes
5ab51d0d1a Fix GCC 4.9 complaint about int/size_t mismatch.
Change-Id: Ib13857f98658dc8760cff15adb5829e8e140008a
2014-06-12 12:52:58 -07:00
Elliott Hughes
c0eed72cbf Size the benchmark name column appropriately.
Nobody likes misaligned output.

Change-Id: Ifbfd7beec755e91e204426a96c054d4ea20d3fae
2014-06-11 16:48:29 -07:00
Elliott Hughes
b27a840f4b Add __pure2 to a few more functions, most notably gettid and pthread_self.
Change-Id: I7eee9f26f45130038af09d8285782b07f70a996f
2014-06-11 16:32:10 -07:00
Elliott Hughes
7634db5a06 Add a couple more system call benchmarks.
Bug: 15387103
Change-Id: I13419ddf77d201fdbde4c784259c0cb0dcfb9a77
2014-06-09 18:35:21 -07:00
Serban Constantinescu
a147a1da5c AArch64: libm: Fix ARM64 fenv_t and refactor ARM64 libm implementation.
This patch fixes the ARM64 ABI for libm. fenv_t is now split in 32bit status
and 32bit control. This mirrors the AArch64 FPU control and status
registers (FPCR, FPSR).

The patch also refactors the libm implementation for ARM64 into a finer
grained control over the FPU registers.

Bionic-benchmarks has been expanded with 3 more benchmarks for floating
point operations. The new libm implementation for ARM64 performs better
over all the math benchmarks available.

Change-Id: I2a7f81d6b4e55c91f8a63a4c69614fc8b1bcf2db
Signed-off-by: Serban Constantinescu <serban.constantinescu@arm.com>
2014-06-09 13:52:28 -07:00
Elliott Hughes
d286796fce Turn on -Wunused and fix the mistakes it uncovers.
Change-Id: I023d2d8b547fbc21d4124bb7510d42b06a0dc501
2014-06-03 15:22:34 -07:00
Elliott Hughes
5d9a7ba0dc Avoid a system call in 'gettid'.
System calls can be pretty slow. This is mako, which has one of our
lowest latencies:

                       iterations      ns/op
  BM_unistd_getpid       10000000        209
  BM_unistd_gettid      200000000          8

Bug: 15297299 (kernel panic from too many gettid calls)
Bug: 15315766 (excessive gettid overhead in liblogd)
Change-Id: I49656c0fc5b5d092390264a59e4f2c0d8a8b1aeb
2014-06-02 10:32:55 -07:00
Dan Albert
7a39094465 Revert "Removes bionic's dependence on stlport"
This reverts commit 4d35da1df8.

Change-Id: Ie9a2620441edb55dbc5949e13a4b913ea4e66797
2014-05-19 23:46:51 +00:00
Dan Albert
4d35da1df8 Removes bionic's dependence on stlport
Only affects tests/benchmarks. Confirmed passing on arm.

Change-Id: If158e23e8cd06e7597ef6f3c9376f6799b467e62
2014-05-19 11:33:14 -07:00
Elliott Hughes
1728b23965 Switch to g_ for globals.
That's what the Google style guide recommends, and we're starting
to get a mix.

Change-Id: Ib0c53a890bb5deed5c679e887541a715faea91fc
2014-05-14 10:02:03 -07:00
Christopher Ferris
3347a79664 Add ability to run on host for x86/x86_64.
Change-Id: I063cb71ce0f7247cefacaa653c7b87b7b3e72025
2014-05-01 13:44:57 -07:00
Christopher Ferris
8b1ade5c0b Modify hard-coded directory.
Use the ANDROID_DATA environment variable instead of the hard-coded
directory for these benchmarks.

Change-Id: I00bae7b4a24e81e77fc8f52e1fe99f4d4918f520
2014-05-01 13:10:48 -07:00
Christopher Ferris
345b49a7c0 Build 32/64 bit versions of tests/benchmarks.
Change-Id: I4d146377aac6573a214dfaa5cab5cce8b10538b4
2014-04-22 10:42:12 -07:00
Elliott Hughes
02c78a3867 Reimplement isinf/isnan/fpclassify.
Also move isinf and isnan into libc like everyone else.

Also move fpclassify to libc like the BSDs (but unlike glibc). We need
this to be able to upgrade our float/double/long double parsing to gdtoa.

Also add some missing aliases. We now have all of:

  isnan, __isnan, isnanf, __isnanf, isnanl, __isnanl,
  isinf, __isinf, isinff, __isinff, isinfl, __isinfl,
  __fpclassify, __fpclassifyd, __fpclassifyf, __fpclassifyl.

Bug: 13469877
Change-Id: I407ffbac06c765a6c5fffda8106c37d7db04f27d
2014-04-14 14:35:47 -07:00
Elliott Hughes
b28e490b73 Add various benchmarks.
These are based on ones from system/extras/tests/bionic/.

Change-Id: I7b1ae15a2ca5d1031a6a511d97b88125c1770047
2014-03-11 11:20:52 -07:00
Serban Constantinescu
282e232e2a AArch64: Add fixes to bionic/tests
This patch adds minor fixes to the bionic unit tests.

Change-Id: Ie10f33c631ed6c10987923d678711d22931ddb05
Signed-off-by: Serban Constantinescu <serban.constantinescu@arm.com>
2013-12-16 17:40:30 -08:00
Colin Cross
bd3efbc9b5 bionic: move benchmarks out of tests directory
Change-Id: I4d054965198af22c9a9c821d1bc53f4e9ea01248
2013-12-10 18:24:30 -08:00