[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-ID: <4DEEB611.4050101@bredband.net>
Date: Wed, 08 Jun 2011 01:36:49 +0200
From: magnum <rawsmooth@...dband.net>
To: john-dev@...ts.openwall.com
Subject: Re: ETA for incremental
On 2011-06-06 22:30, magnum wrote:
> On 2011-06-06 21:52, jfoug wrote:
>>> From: magnum
>>> Good idea. As another alternative, I guess it would be easy to replace
>>> the "100.00%" with "DONE" instead, like this:
>>>
>>> guesses: 0 time: 0:00:00:06 DONE (Mon Jun 6 19:03:01 2011) c/s:23055M
>>
>> I like the above format a lot. I think this is more intuitive, but
>> to each
>> his own.
>>
>>> This will work in case we have no ETA too:
>>>
>>> guesses: 0 time: 0:00:00:06 DONE c/s:23055M
>>
>> Having the actual date/time of completion is very nice. I personally
>> think
>> it is very useful. Often, I run a batch (script) file with many smaller
>> runs in it. It is nice to track amount of time actually used, and
>> date they
>> completed.
>
> Ah, of course - we should output the ending time even if we did not
> have ETA during the run! Definitely doable.
A new patch is posted to the wiki, implementing the above. All modes
will output like this when finished:
guesses: 0 time: 0:00:00:06 DONE (Mon Jun 6 19:03:01 2011) c/s:23055M
Also, I found a bug in single.c that has been there all the time since
we got ETA. The hundredths was just a copy of the integer, so it was
0.00%, 1.01%, 42.42% and so on.
As an even worse hack than the current code (or maybe not) I played with
changing all get_progress functions to use a double* instead of a char*
and return a progress with lots of decimals (though still only shown
with two). I also circumvented the 32-bit overflow check if
sizeof(time_t) > 4. I got really good estimates after a *very* short
time (given a stable c/s rate). Also, I got disencouraging ETA's when
running slow formats, like "Apr 1 12:34:56 1231432" (i.e. 1.2 million
years from now) when running -inc:all :-)
That last stunt has little value but I could post it if anyone wants to
try it or polish it.
magnum
Powered by blists - more mailing lists
Confused about mailing lists and their use?
Read about mailing lists on Wikipedia
and check out these
guidelines on proper formatting of your messages.
