Commit Graph

7 Commits (36df75a0a903c8ba3e2e28eb7162c43f8dd5d8f0)

Author SHA1 Message Date
Richard Henderson 7ba7db9fa1 migration/xbzrle: Use i386 host/cpuinfo.h
Perform the function selection once, and only if CONFIG_AVX512_OPT
is enabled.  Centralize the selection to xbzrle.c, instead of
spreading the init across 3 files.

Remove xbzrle-bench.c.  The benefit of being able to benchmark
the different implementations is less important than not peeking
into the internals of the implementation.

Reviewed-by: Alex Bennée <alex.bennee@linaro.org>
Reviewed-by: Juan Quintela <quintela@redhat.com>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
2023-05-23 16:51:18 -07:00
Emilio Cota e3feb2cc22 util: import GTree as QTree
The only reason to add this implementation is to control the memory allocator
used. Some users (e.g. TCG) cannot work reliably in multi-threaded
environments (e.g. forking in user-mode) with GTree's allocator, GSlice.
See https://gitlab.com/qemu-project/qemu/-/issues/285 for details.

Importing GTree is a temporary workaround until GTree migrates away
from GSlice.

This implementation is identical to that in glib v2.75.0, except that
we don't import recent additions to the API nor deprecated API calls,
none of which are used in QEMU.

I've imported tests from glib and added a benchmark just to
make sure that performance is similar. Note: it cannot be identical
because (1) we are not using GSlice, (2) we use different compilation flags
(e.g. -fPIC) and (3) we're linking statically.

$ cat /proc/cpuinfo| grep 'model name' | head -1
model name      : AMD Ryzen 7 PRO 5850U with Radeon Graphics
$ echo '0' | sudo tee /sys/devices/system/cpu/cpufreq/boost
$ tests/bench/qtree-bench

 Tree         Op      32            1024            4096          131072         1048576
------------------------------------------------------------------------------------------------
GTree     Lookup   83.23           43.08           25.31           19.40           16.22
QTree     Lookup  113.42 (1.36x)   53.83 (1.25x)   28.38 (1.12x)   17.64 (0.91x)   13.04 (0.80x)
GTree     Insert   44.23           29.37           25.83           19.49           17.03
QTree     Insert   46.87 (1.06x)   25.62 (0.87x)   24.29 (0.94x)   16.83 (0.86x)   12.97 (0.76x)
GTree     Remove   53.27           35.15           31.43           24.64           16.70
QTree     Remove   57.32 (1.08x)   41.76 (1.19x)   38.37 (1.22x)   29.30 (1.19x)   15.07 (0.90x)
GTree  RemoveAll  135.44          127.52          126.72          120.11           64.34
QTree  RemoveAll  127.15 (0.94x)  110.37 (0.87x)  107.97 (0.85x)   97.13 (0.81x)   55.10 (0.86x)
GTree   Traverse  277.71          276.09          272.78          246.72           98.47
QTree   Traverse  370.33 (1.33x)  411.97 (1.49x)  400.23 (1.47x)  262.82 (1.07x)   78.52 (0.80x)
------------------------------------------------------------------------------------------------

As a sanity check, the same benchmark when Glib's version
is >= $glib_dropped_gslice_version (i.e. QTree == GTree):

 Tree         Op      32            1024            4096          131072         1048576
------------------------------------------------------------------------------------------------
GTree     Lookup   82.72           43.09           24.18           19.73           16.09
QTree     Lookup   81.82 (0.99x)   43.10 (1.00x)   24.20 (1.00x)   19.76 (1.00x)   16.26 (1.01x)
GTree     Insert   45.07           29.62           26.34           19.90           17.18
QTree     Insert   45.72 (1.01x)   29.60 (1.00x)   26.38 (1.00x)   19.71 (0.99x)   17.20 (1.00x)
GTree     Remove   54.48           35.36           31.77           24.97           16.95
QTree     Remove   54.46 (1.00x)   35.32 (1.00x)   31.77 (1.00x)   24.91 (1.00x)   17.15 (1.01x)
GTree  RemoveAll  140.68          127.36          125.43          121.45           68.20
QTree  RemoveAll  140.65 (1.00x)  127.64 (1.00x)  125.01 (1.00x)  121.73 (1.00x)   67.06 (0.98x)
GTree   Traverse  278.68          276.05          266.75          251.65          104.93
QTree   Traverse  278.31 (1.00x)  275.78 (1.00x)  266.42 (1.00x)  247.89 (0.99x)  104.58 (1.00x)
------------------------------------------------------------------------------------------------

Signed-off-by: Emilio Cota <cota@braap.org>
Message-Id: <20230205163758.416992-2-cota@braap.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
2023-03-28 15:23:10 -07:00
ling xu cc98c9fd5c Update bench-code for addressing CI problem
Unit test code is in test-xbzrle.c, and benchmark code is in xbzrle-bench.c
for performance benchmarking. we have modified xbzrle-bench.c to address
CI problem.

Signed-off-by: ling xu <ling1.xu@intel.com>
Co-authored-by: Zhou Zhao <zhou.zhao@intel.com>
Co-authored-by: Jun Jin <jun.i.jin@intel.com>
Reviewed-by: Juan Quintela <quintela@redhat.com>
Signed-off-by: Juan Quintela <quintela@redhat.com>
2023-02-11 16:51:09 +01:00
Markus Armbruster 66997c42e0 cleanup: Tweak and re-run return_directly.cocci
Tweak the semantic patch to drop redundant parenthesis around the
return expression.

Coccinelle drops a comment in hw/rdma/vmw/pvrdma_cmd.c; restored
manually.

Coccinelle messes up vmdk_co_create(), not sure why.  Change dropped,
will be done manually in the next commit.

Line breaks in target/avr/cpu.h and hw/rdma/vmw/pvrdma_cmd.c tidied up
manually.

Whitespace in tools/virtiofsd/fuse_lowlevel.c tidied up manually.

checkpatch.pl complains "return of an errno should typically be -ve"
two times for hw/9pfs/9p-synth.c.  Preexisting, the patch merely makes
it visible to checkpatch.pl.

Signed-off-by: Markus Armbruster <armbru@redhat.com>
Message-Id: <20221122134917.1217307-2-armbru@redhat.com>
Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Acked-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
2022-12-14 16:19:35 +01:00
Lei He 689309c4ac test/crypto: Add test suite for crypto akcipher
Add unit test and benchmark test for crypto akcipher.

Signed-off-by: lei he <helei.sig11@bytedance.com>
Signed-off-by: zhenwei pi <pizhenwei@bytedance.com>
Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>
2022-05-26 11:41:54 +01:00
Peter Maydell 5df022cf2e osdep: Move memalign-related functions to their own header
Move the various memalign-related functions out of osdep.h and into
their own header, which we include only where they are used.
While we're doing this, add some brief documentation comments.

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Message-id: 20220226180723.1706285-10-peter.maydell@linaro.org
2022-03-07 13:16:49 +00:00
Thomas Huth 3b472e71d5 tests: Move benchmarks into a separate folder
Make it clear that these files are related to benchmarks by moving
them into a new folder called "bench".

Message-Id: <20210312092238.79509-1-thuth@redhat.com>
Acked-by: Paolo Bonzini <pbonzini@redhat.com>
Signed-off-by: Thomas Huth <thuth@redhat.com>
2021-03-12 15:46:30 +01:00