This is the mail archive of the
libc-alpha@sourceware.org
mailing list for the glibc project.
Re: [PATCH v2] Single threaded stdio optimization
On 30/06/17 16:18, Torvald Riegel wrote:
> On Fri, 2017-06-30 at 13:15 +0100, Szabolcs Nagy wrote:
>> on a particular x86_64 cpu with particular loop count:
>>
>> time $orig/lib64/ld-2.25.90.so --library-path $orig/lib64 ./getchar
>> 5.89user 0.07system 0:05.98elapsed 99%CPU (0avgtext+0avgdata 2000maxresident)k
>> 0inputs+0outputs (0major+153minor)pagefaults 0swaps
>> time $stdio/lib64/ld-2.25.90.so --library-path $stdio/lib64 ./getchar
>> 2.66user 0.06system 0:02.73elapsed 99%CPU (0avgtext+0avgdata 2032maxresident)k
>> 0inputs+0outputs (0major+155minor)pagefaults 0swaps
>
> What's interesting here is that your high-level optimization is faster
> than doing the single-thread check in the low-level lock (x86 has it
> already in the low-level lock).
>
musl implemented stdio this way since 2011, which
is the reason for the 2x in getc/putc (i386):
http://www.etalabs.net/compare_libcs.html
we knew this already, but pushing a high level opt
in glibc is harder than doing target specific lll opt.
musl does not have to worry about fopencookie, libc
internal malloc interopsition, printf hooks, magic IO
flags, target specific hacks, so the optimization is
more obviously correct there, but now that the patch
only omits locks in specific functions the correctness
should be easier to verify.