aboutsummaryrefslogtreecommitdiff
diff options
context:
space:
mode:
authorDenys Vlasenko <dvlasenk@redhat.com>2010-10-29 19:26:38 +0200
committerDenys Vlasenko <dvlasenk@redhat.com>2010-10-29 19:26:38 +0200
commit5d49b72f1a3a362957171610b722ca8ec3f49734 (patch)
treeefb4dbaaff13e8901a5c723c6b9945028570d837
parent1642d0271027a7021a0a69fda3ac7082f5b01314 (diff)
downloadbusybox-5d49b72f1a3a362957171610b722ca8ec3f49734.tar.gz
decompress_bunzip2: add profiling data to comment. no code changes
Signed-off-by: Denys Vlasenko <dvlasenk@redhat.com>
-rw-r--r--archival/libunarchive/decompress_bunzip2.c11
1 files changed, 11 insertions, 0 deletions
diff --git a/archival/libunarchive/decompress_bunzip2.c b/archival/libunarchive/decompress_bunzip2.c
index 00ec47c75..2fa0f898a 100644
--- a/archival/libunarchive/decompress_bunzip2.c
+++ b/archival/libunarchive/decompress_bunzip2.c
@@ -20,6 +20,17 @@
the Burrows-Wheeler transformation. Much of that time is delay
resulting from cache misses.
+ (2010 update by vda: profiled "bzcat <84mbyte.bz2 >/dev/null"
+ on x86-64 CPU with L2 > 1M: get_next_block is hotter than read_bunzip:
+ %time seconds calls function
+ 71.01 12.69 444 get_next_block
+ 28.65 5.12 93065 read_bunzip
+ 00.22 0.04 7736490 get_bits
+ 00.11 0.02 47 dealloc_bunzip
+ 00.00 0.00 93018 full_write
+ ...)
+
+
I would ask that anyone benefiting from this work, especially those
using it in commercial products, consider making a donation to my local
non-profit hospice organization (www.hospiceacadiana.com) in the name of