On Thu, 16 Aug 2012 17:06:47 +0200
Johannes Stezenbach [off-list ref] wrote:

Well, ~2x speedup on x86 is certainly a good achievement, but there
are more ARM based devices than there are PCs, and I guess many
embedded devices use lzo compressed kernels and file systems
while I'm not convinced many PCs rely on lzo in the kernel.

Keep in mind that a major user of LZO is the BTRFS filesystem, and I believe it
is used much more often on larger machines than on ARM, in fact it had problems
operating on ARM at all, until quite recently.

I know everyone's either busy or on vacation, but it would
be so cool if someone could test on a more modern ARM core,
with the userspace test code I posted it should be easy to do.

I have locked the Allwinner A10 CPU in my Mele A2000 to 60 MHz using cpufreq-set,
and ran your test. rnd.lzo is a 9 MB file from /dev/urandom compressed with lzo.
There doesn't seem to be a significant difference between all three variants.

# time for i in {1..20}; do old/unlzop < rnd.lzo >/dev/null ; done

real	0m11.353s
user	0m3.060s
sys	0m8.170s
# time for i in {1..20}; do new/unlzop < rnd.lzo >/dev/null ; done

real	0m11.416s
user	0m3.030s
sys	0m8.200s
# time for i in {1..20}; do test/unlzop < rnd.lzo >/dev/null ; done

real	0m11.310s
user	0m3.100s
sys	0m8.150s

-- 
With respect,
Roman

~~~~~~~~~~~~~~~~~~~~~~~~~~~
"Stallman had a printer,
with code he could not see.
So he began to tinker,
and set the software free."
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 198 bytes
Desc: not available
URL: <http://lists.infradead.org/pipermail/linux-arm-kernel/attachments/20120816/d1a5e749/attachment.sig>