use the LEA instruction more efficiently, since e.g. a + (b << 2) can be encoded
as one instruction. Even more importantly, it can constant-fold the
COPY_* enums together with the shifted negative constants, which also saves
some instructions. (We don't need it for LITERAL, since it happens to be 0.)
I am unsure why the compiler couldn't do this itself, but the theory is that
it cannot prove that len-1 and len-4 cannot underflow/wrap, and thus can't
do the optimization safely.
The gains are small but measurable; 0.5-1.0% over the BM_Z* benchmarks
(measured on Westmere, Sandy Bridge and Istanbul).
R=sanjay
git-svn-id: https://snappy.googlecode.com/svn/trunk@69 03e5f5b5-db94-4691-08a0-1a8bf15f6143
versions of libtool and automake on non-GNU platforms (e.g. Mac OS X).
R=sanjay
git-svn-id: https://snappy.googlecode.com/svn/trunk@68 03e5f5b5-db94-4691-08a0-1a8bf15f6143
it leaves the source in a state that's not appropriate for RawUncompress.
R=sanjay
git-svn-id: https://snappy.googlecode.com/svn/trunk@67 03e5f5b5-db94-4691-08a0-1a8bf15f6143
twice; it could cause division by zero in the unit test framework.
(We already had one fix for this in place, but it was incomplete.)
This could in theory happen on any system, since there are few guarantees
about gettimeofday(), but seems to only happen in practice on GNU/Hurd, where
gettimeofday() is cached and only updated ever so often.
R=sanjay
git-svn-id: https://snappy.googlecode.com/svn/trunk@65 03e5f5b5-db94-4691-08a0-1a8bf15f6143
apparently Debian still targets these by default, giving us segfaults on
armel.
R=sanjay
git-svn-id: https://snappy.googlecode.com/svn/trunk@64 03e5f5b5-db94-4691-08a0-1a8bf15f6143
Achieved by moving logging macro definitions to a test-only
header file, and by changing non-test code to use assert,
fprintf, and abort instead of LOG/CHECK macros.
R=sesse
git-svn-id: https://snappy.googlecode.com/svn/trunk@62 03e5f5b5-db94-4691-08a0-1a8bf15f6143
to 2 MB. This fixes issues with running the unit test on platforms with
little RAM (e.g. some ARM boards).
Also, reactivate the 2 MB test for 64-bit platforms; there's no good
reason why it shouldn't be.
R=sanjay
git-svn-id: https://snappy.googlecode.com/svn/trunk@58 03e5f5b5-db94-4691-08a0-1a8bf15f6143
warnings. There are still some in the unit test, but the main .cc file should
be clean. We haven't enabled -Wall for the default build, since the unit test
is still not clean.
This also fixes a real bug in the open-source implementation of
ReadFileToStringOrDie(); it would not detect errors correctly.
I had to go through some pains to avoid performance loss as the types
were changed; I think there might still be some with 32-bit if and only if LFS
is enabled (ie., size_t is 64-bit), but for regular 32-bit and 64-bit I can't
see any losses, and I've diffed the generated GCC assembler between the old and
new code without seeing any significant choices. If anything, it's ever so
slightly faster.
This may or may not enable compression of very large blocks (>2^32 bytes)
when size_t is 64-bit, but I haven't checked, and it is still not a supported
case.
git-svn-id: https://snappy.googlecode.com/svn/trunk@56 03e5f5b5-db94-4691-08a0-1a8bf15f6143
the current point, but there seems to be enough of a general interest in the
topic (cf. public bug #34).
R=csilvers,sanjay
git-svn-id: https://snappy.googlecode.com/svn/trunk@55 03e5f5b5-db94-4691-08a0-1a8bf15f6143
enum declarations, as it seems the Sun compiler does not like it.
Based on patch by Travis Vitek.
git-svn-id: https://snappy.googlecode.com/svn/trunk@48 03e5f5b5-db94-4691-08a0-1a8bf15f6143
instead of "4-byte offset" for the longest copies.
Also fix an inconsistency in the heading for section 2.2.3.
Both patches by Patrick Pelletier.
R=csilvers
git-svn-id: https://snappy.googlecode.com/svn/trunk@45 03e5f5b5-db94-4691-08a0-1a8bf15f6143
identical, even regarding cv-qualifiers.
This is required to work around a bug in the Solaris Studio C++ compiler
(it does not properly disregard cv-qualifiers when doing name mangling).
R=sanjay
git-svn-id: https://snappy.googlecode.com/svn/trunk@44 03e5f5b5-db94-4691-08a0-1a8bf15f6143
(I stumbled into this when changing the way we decompress literals.)
R=csilvers
Revision created by MOE tool push_codebase.
git-svn-id: https://snappy.googlecode.com/svn/trunk@43 03e5f5b5-db94-4691-08a0-1a8bf15f6143
Whenever we try to enter a copy fast-path, there is a certain cost in checking
that all the preconditions are in place, but it's normally offset by the fact
that we can usually take the cheaper path. However, in a certain path we've
already established that "avail < literal_length", which usually means that
either the available space is small, or the literal is big. Both will disqualify
us from taking the fast path, and thus we take the hit from the precondition
checking without gaining much from having a fast path. Thus, simply don't try
the fast path in this situation -- we're already on a slow path anyway
(one where we need to refill more data from the reader).
I'm a bit surprised at how much this gained; it could be that this path is
more common than I thought, or that the simpler structure somehow makes the
compiler happier. I haven't looked at the assembler, but it's a win across
the board on both Core 2, Core i7 and Opteron, at least for the cases we
typically care about. The gains seem to be the largest on Core i7, though.
Results from my Core i7 workstation:
Benchmark Time(ns) CPU(ns) Iterations
---------------------------------------------------
BM_UFlat/0 73337 73091 190996 1.3GB/s html [ +1.7%]
BM_UFlat/1 696379 693501 20173 965.5MB/s urls [ +2.7%]
BM_UFlat/2 9765 9734 1472135 12.1GB/s jpg [ +0.7%]
BM_UFlat/3 29720 29621 472973 3.0GB/s pdf [ +1.8%]
BM_UFlat/4 294636 293834 47782 1.3GB/s html4 [ +2.3%]
BM_UFlat/5 28399 28320 494700 828.5MB/s cp [ +3.5%]
BM_UFlat/6 12795 12760 1000000 833.3MB/s c [ +1.2%]
BM_UFlat/7 3984 3973 3526448 893.2MB/s lsp [ +5.7%]
BM_UFlat/8 991996 989322 14141 992.6MB/s xls [ +3.3%]
BM_UFlat/9 228620 227835 61404 636.6MB/s txt1 [ +4.0%]
BM_UFlat/10 197114 196494 72165 607.5MB/s txt2 [ +3.5%]
BM_UFlat/11 605240 603437 23217 674.4MB/s txt3 [ +3.7%]
BM_UFlat/12 804157 802016 17456 573.0MB/s txt4 [ +3.9%]
BM_UFlat/13 347860 346998 40346 1.4GB/s bin [ +1.2%]
BM_UFlat/14 44684 44559 315315 818.4MB/s sum [ +2.3%]
BM_UFlat/15 5120 5106 2739726 789.4MB/s man [ +3.3%]
BM_UFlat/16 76591 76355 183486 1.4GB/s pb [ +2.8%]
BM_UFlat/17 238564 237828 58824 739.1MB/s gaviota [ +1.6%]
BM_UValidate/0 42194 42060 333333 2.3GB/s html [ -0.1%]
BM_UValidate/1 433182 432005 32407 1.5GB/s urls [ -0.1%]
BM_UValidate/2 197 196 71428571 603.3GB/s jpg [ +0.5%]
BM_UValidate/3 14494 14462 972222 6.1GB/s pdf [ +0.5%]
BM_UValidate/4 168444 167836 83832 2.3GB/s html4 [ +0.1%]
R=jeff
Revision created by MOE tool push_codebase.
git-svn-id: https://snappy.googlecode.com/svn/trunk@42 03e5f5b5-db94-4691-08a0-1a8bf15f6143
Looking up into and decoding the values from char_table has long shown up as a
hotspot in the decompressor. While it turns out that it's hard to make a more
efficient decoder for the copy ops, the literals are simple enough that we can
decode them without needing a table lookup. (This means that 1/4 of the table
is now unused, although that in itself doesn't buy us anything.)
The gains are small, but definitely present; some tests win as much as 10%,
but 1-4% is more typical. These results are from Core i7, in 64-bit mode;
Core 2 and Opteron show similar results. (I've run with more iterations
than unusual to make sure the smaller gains don't drown entirely in noise.)
Benchmark Time(ns) CPU(ns) Iterations
---------------------------------------------------
BM_UFlat/0 74665 74428 182055 1.3GB/s html [ +3.1%]
BM_UFlat/1 714106 711997 19663 940.4MB/s urls [ +4.4%]
BM_UFlat/2 9820 9789 1427115 12.1GB/s jpg [ -1.2%]
BM_UFlat/3 30461 30380 465116 2.9GB/s pdf [ +0.8%]
BM_UFlat/4 301445 300568 46512 1.3GB/s html4 [ +2.2%]
BM_UFlat/5 29338 29263 479452 801.8MB/s cp [ +1.6%]
BM_UFlat/6 13004 12970 1000000 819.9MB/s c [ +2.1%]
BM_UFlat/7 4180 4168 3349282 851.4MB/s lsp [ +1.3%]
BM_UFlat/8 1026149 1024000 10000 959.0MB/s xls [+10.7%]
BM_UFlat/9 237441 236830 59072 612.4MB/s txt1 [ +0.3%]
BM_UFlat/10 203966 203298 69307 587.2MB/s txt2 [ +0.8%]
BM_UFlat/11 627230 625000 22400 651.2MB/s txt3 [ +0.7%]
BM_UFlat/12 836188 833979 16787 551.0MB/s txt4 [ +1.3%]
BM_UFlat/13 351904 350750 39886 1.4GB/s bin [ +3.8%]
BM_UFlat/14 45685 45562 308370 800.4MB/s sum [ +5.9%]
BM_UFlat/15 5286 5270 2656546 764.9MB/s man [ +1.5%]
BM_UFlat/16 78774 78544 178117 1.4GB/s pb [ +4.3%]
BM_UFlat/17 242270 241345 58091 728.3MB/s gaviota [ +1.2%]
BM_UValidate/0 42149 42000 333333 2.3GB/s html [ -3.0%]
BM_UValidate/1 432741 431303 32483 1.5GB/s urls [ +7.8%]
BM_UValidate/2 198 197 71428571 600.7GB/s jpg [+16.8%]
BM_UValidate/3 14560 14521 965517 6.1GB/s pdf [ -4.1%]
BM_UValidate/4 169065 168671 83832 2.3GB/s html4 [ -2.9%]
R=jeff
Revision created by MOE tool push_codebase.
git-svn-id: https://snappy.googlecode.com/svn/trunk@41 03e5f5b5-db94-4691-08a0-1a8bf15f6143
state of ip_ after decompression (or attempted decompresion) is
completely irrelevant, so we don't need the trailer.
Performance is, as expected, mostly flat -- there's a curious ~3-5%
loss in the "lsp" test, but that test case is so short it is hard to say
anything definitive about why (most likely, it's some sort of
unrelated effect).
R=jeff
git-svn-id: https://snappy.googlecode.com/svn/trunk@39 03e5f5b5-db94-4691-08a0-1a8bf15f6143
It is seemingly hard for the compiler to understand that ip_, the current input
pointer into the compressed data stream, can not alias on anything else, and
thus using it directly will incur memory traffic as it cannot be kept in a
register. The code already knew about this and cached it into a local
variable, but since Step() only decoded one tag, it had to move ip_ back into
place between every tag. This seems to have cost us a significant amount of
performance, so changing Step() into a function that decodes as much as it can
before it saves ip_ back and returns. (Note that Step() was already inlined,
so it is not the manual inlining that buys the performance here.)
The wins are about 3-6% for Core 2, 6-13% on Core i7 and 5-12% on Opteron
(for plain array-to-array decompression, in 64-bit opt mode).
There is a tiny difference in the behavior here; if an invalid literal is
encountered (ie., the writer refuses the Append() operation), ip_ will now
point to the byte past the tag byte, instead of where the literal was
originally thought to end. However, we don't use ip_ for anything after
DecompressAllTags() has returned, so this should not change external behavior
in any way.
Microbenchmark results for Core i7, 64-bit (Opteron results are similar):
Benchmark Time(ns) CPU(ns) Iterations
---------------------------------------------------
BM_UFlat/0 79134 79110 8835 1.2GB/s html [ +6.2%]
BM_UFlat/1 786126 786096 891 851.8MB/s urls [+10.0%]
BM_UFlat/2 9948 9948 69125 11.9GB/s jpg [ -1.3%]
BM_UFlat/3 31999 31998 21898 2.7GB/s pdf [ +6.5%]
BM_UFlat/4 318909 318829 2204 1.2GB/s html4 [ +6.5%]
BM_UFlat/5 31384 31390 22363 747.5MB/s cp [ +9.2%]
BM_UFlat/6 14037 14034 49858 757.7MB/s c [+10.6%]
BM_UFlat/7 4612 4612 151395 769.5MB/s lsp [ +9.5%]
BM_UFlat/8 1203174 1203007 582 816.3MB/s xls [+19.3%]
BM_UFlat/9 253869 253955 2757 571.1MB/s txt1 [+11.4%]
BM_UFlat/10 219292 219290 3194 544.4MB/s txt2 [+12.1%]
BM_UFlat/11 672135 672131 1000 605.5MB/s txt3 [+11.2%]
BM_UFlat/12 902512 902492 776 509.2MB/s txt4 [+12.5%]
BM_UFlat/13 372110 371998 1881 1.3GB/s bin [ +5.8%]
BM_UFlat/14 50407 50407 10000 723.5MB/s sum [+13.5%]
BM_UFlat/15 5699 5701 100000 707.2MB/s man [+12.4%]
BM_UFlat/16 83448 83424 8383 1.3GB/s pb [ +5.7%]
BM_UFlat/17 256958 256963 2723 684.1MB/s gaviota [ +7.9%]
BM_UValidate/0 42795 42796 16351 2.2GB/s html [+25.8%]
BM_UValidate/1 490672 490622 1427 1.3GB/s urls [+22.7%]
BM_UValidate/2 237 237 2950297 499.0GB/s jpg [+24.9%]
BM_UValidate/3 14610 14611 47901 6.0GB/s pdf [+26.8%]
BM_UValidate/4 171973 171990 4071 2.2GB/s html4 [+25.7%]
git-svn-id: https://snappy.googlecode.com/svn/trunk@38 03e5f5b5-db94-4691-08a0-1a8bf15f6143
This text is new, but an earlier version from Zeev Tarantov was used
as reference.
R=csilvers
DELTA=112 (111 added, 0 deleted, 1 changed)
Revision created by MOE tool push_codebase.
MOE_MIGRATION=1867
git-svn-id: https://snappy.googlecode.com/svn/trunk@36 03e5f5b5-db94-4691-08a0-1a8bf15f6143
not real time. Also, use nth_element instead of sort, since we
only need one element.
R=csilvers
DELTA=5 (3 added, 0 deleted, 2 changed)
Revision created by MOE tool push_codebase.
MOE_MIGRATION=1799
git-svn-id: https://snappy.googlecode.com/svn/trunk@35 03e5f5b5-db94-4691-08a0-1a8bf15f6143
properly cases where gettimeofday() can stand return the same
result twice (as sometimes on GNU/Hurd) or go backwards
(as when the user adjusts the clock). We avoid a division-by-zero,
and put a lower bound on the number of iterations -- the same
amount as we use to calibrate.
We should probably use CLOCK_MONOTONIC for platforms that support
it, to be robust against clock adjustments; we already use Windows'
monotonic timers. However, that's for a later changelist.
R=csilvers
DELTA=7 (5 added, 0 deleted, 2 changed)
Revision created by MOE tool push_codebase.
MOE_MIGRATION=1798
git-svn-id: https://snappy.googlecode.com/svn/trunk@34 03e5f5b5-db94-4691-08a0-1a8bf15f6143
libraries, not libsnappy.so (which doesn't need any such dependency).
R=csilvers
DELTA=20 (14 added, 0 deleted, 6 changed)
Revision created by MOE tool push_codebase.
MOE_MIGRATION=1710
git-svn-id: https://snappy.googlecode.com/svn/trunk@33 03e5f5b5-db94-4691-08a0-1a8bf15f6143
as MSVC doesn't include it. Replace with QueryPerformanceCounter(),
which is monotonic and probably reasonably high-resolution.
(Some machines have traditionally had bugs in QPC, but they should
be relatively rare these days, and there's really no much better
alternative that I know of.)
R=csilvers
DELTA=74 (55 added, 19 deleted, 0 changed)
Revision created by MOE tool push_codebase.
MOE_MIGRATION=1556
git-svn-id: https://snappy.googlecode.com/svn/trunk@31 03e5f5b5-db94-4691-08a0-1a8bf15f6143
we need for our own build system internally.
R=csilvers
DELTA=16 (13 added, 1 deleted, 2 changed)
Revision created by MOE tool push_codebase.
MOE_MIGRATION=1555
git-svn-id: https://snappy.googlecode.com/svn/trunk@30 03e5f5b5-db94-4691-08a0-1a8bf15f6143
so we won't pull in macro definitions of things like min() and max(),
which can conflict with <algorithm>.
R=csilvers
DELTA=1 (1 added, 0 deleted, 0 changed)
Revision created by MOE tool push_codebase.
MOE_MIGRATION=1485
git-svn-id: https://snappy.googlecode.com/svn/trunk@29 03e5f5b5-db94-4691-08a0-1a8bf15f6143
instead of getursage().
I thought I'd already committed this patch, so that the 1.0.1 release already
would have a Windows-compatible snappy_unittest, but I'd seemingly deleted it
instead, so this is a reconstruction.
R=csilvers
DELTA=43 (39 added, 3 deleted, 1 changed)
Revision created by MOE tool push_codebase.
MOE_MIGRATION=1295
git-svn-id: https://snappy.googlecode.com/svn/trunk@28 03e5f5b5-db94-4691-08a0-1a8bf15f6143
I've made a few changes since Martin's version; mostly style nits, but also
a semantic change -- most functions that return bool in the C++ version now
return an enum, to better match typical C (and zlib) semantics.
I've kept the copyright notice, since Martin is obviously the author here;
he has signed the contributor license agreement, though, so this should not
hinder Google's use in the future.
We'll need to update the libtool version number to match the added interface,
but as of http://www.gnu.org/software/libtool/manual/html_node/Updating-version-info.html
I'm going to wait until public release.
R=csilvers
DELTA=238 (233 added, 0 deleted, 5 changed)
Revision created by MOE tool push_codebase.
MOE_MIGRATION=1294
git-svn-id: https://snappy.googlecode.com/svn/trunk@27 03e5f5b5-db94-4691-08a0-1a8bf15f6143
The data compresses/decompresses slightly faster than the old data, and has
similar density.
R=lookingbill
DELTA=1 (0 added, 0 deleted, 1 changed)
Revision created by MOE tool push_codebase.
MOE_MIGRATION=1288
git-svn-id: https://snappy.googlecode.com/svn/trunk@26 03e5f5b5-db94-4691-08a0-1a8bf15f6143
Measure() loop. This gives all algorithms a small speed boost, except Snappy which
already didn't do reallocation (so the measurements were slightly biased in its
favor).
R=csilvers
DELTA=92 (69 added, 9 deleted, 14 changed)
Revision created by MOE tool push_codebase.
MOE_MIGRATION=1151
git-svn-id: https://snappy.googlecode.com/svn/trunk@24 03e5f5b5-db94-4691-08a0-1a8bf15f6143
the differences from the opensource code. Will make it easier
in the future to mix-and-match third-party code that uses
snappy with google code.
Currently, csearch shows that the only external user of
"namespace zippy" is some bigtable code that accesses
a TEST variable, which is temporarily kept in the zippy
namespace.
R=sesse
DELTA=123 (18 added, 3 deleted, 102 changed)
Revision created by MOE tool push_codebase.
MOE_MIGRATION=1150
git-svn-id: https://snappy.googlecode.com/svn/trunk@23 03e5f5b5-db94-4691-08a0-1a8bf15f6143
Replace the Apache 2.0 license header by the BSD-type license header;
somehow a lot of the files were missed in the last round.
R=dannyb,csilvers
DELTA=147 (74 added, 2 deleted, 71 changed)
Change on 2011-03-25 19:25:07-07:00 by sesse
Unbreak the build; the relicensing removed a bit too much (only comments
were intended, but I also accidentially removed some of the top lines of
the actual source).
Revision created by MOE tool push_codebase.
MOE_MIGRATION=1072
git-svn-id: https://snappy.googlecode.com/svn/trunk@21 03e5f5b5-db94-4691-08a0-1a8bf15f6143
