Re: Official Git mirror?

[Óscar Fuentes]

Eli Zaretskii <eliz@gnu.org> writes:

>> But then the plan is failing, because the timings I posted show that
>> "smart" wins over "nosmart" even when a weak server is servicing a
>> mighty client over a fast network.
>
> Sometimes it indeed makes no significant difference, but sometimes it
> wins big time.  Observe:
>
>  bzr://bzr.savannah.gnu.org/emacs/trunk
>
>     real    45m4.820s
>     user    15m58.380s
>     sys     0m12.910s
>
>     Transferred: 540480KiB (199.9K/s r:540403K w:77K)
>
>   bzr branch nosmart+bzr://bzr.savannah.gnu.org/emacs/trunk
>
>
>     real    16m30.189s
>     user    15m22.090s
>     sys     0m14.560s
>
>     Transferred: 780914KiB (789.2K/s r:780640K w:275K)
>
> In the thread I mentioned on the Bazaar list, someone else also
> reported a huge speedup:
>
>> over a 3 Mbit/s connection:
>> 
>>    bzr://bzr.savannah.gnu.org/emacs/trunk
>>    6949.356  Transferred: 469739kB (67.6kB/s r:469659kB w:80kB)
>> 
>>    nosmart+bzr://bzr.savannah.gnu.org/emacs/trunk
>>    2919.117  Transferred: 524353kB (179.7kB/s r:524162kB w:191kB)
>
> That's almost 2 hours slashed to 48 minutes, an almost 3-fold speedup.
>
>> > The "smart" part is for sending less data, which is not going to win
>> > for the initial checkout.
>> 
>> You said on the other post that cloning time is network-bound. So being
>> smart and sending less data would be better.
>
> Not if "being smart" wastes CPU cycles on the server side and causes
> it to use the available bandwidth less efficiently.  See the network
> throughput figures above, reported by bzr on .bzr.log.

Eli, I re-quote what you said:

Oscar:
>> >> It seems that "nosmart" is used for compensating for servers with busy
>> >> CPUs
Eli:
>> > No, it's used to compensate for overly "smart" server when there's no
>> > win in being smart, because you need to send everything anyway.

You are contradicting yourself, because your data above is more evidence
supporting my hypotheses of nosmart being a tricky effective with
CPU-starved servers.

Besides, cloning from Savannah is almost twice as slow than from
Launchpad for me. If it is because network distance considerations
(pinging to it is 3 times slower than to Launchpad) or because the
server is CPU and/or bandwidth starved, I don't know. Your timings
comparing smart/nosmart seem to point to the CPU, though.