Re: progress and eta feature request

[Mag Gam]

On Sun, Jun 12, 2011 at 2:27 PM, Ole Tange <tange@gnu.org> wrote:
> On Fri, Jun 10, 2011 at 8:13 PM, Remik Ziemlinski
> <remik.ziemlinski@gmail.com> wrote:
>> First off, wonderful tool.  Thank you!
>
> You can show your appreciation by promoting GNU Parallel. You can:
>
> * Post the intro video http://www.youtube.com/watch?v=OpaiGYxkSuQ on
> your blog/Twitter/Facebook/Linkedin/user group/mailing lists
> * Write a review for your blog/Twitter/Facebook/Linkedin/user
> group/mailing lists
> * Help others use GNU Parallel when you see problems that can be
> solved using GNU Parallel
> * Request or build a package for your favourite distribution
> * Invite me to speak at your next conference (Contact http://ole.tange.dk)
>
> If GNU Parallel saves you money:
>
> * Donate (or convince your company to donate) to FSF 
> https://my.fsf.org/donate/
>
>> When I use the current progress indicator, it only shows how many tasks
>> completed, and the percentage always shows 100%.  Instead, I wish the
>> percentage reported the ratio of complete vs. incomplete tasks.
>
> I do not think that ratio would be very good: For 1000 jobs it would
> start out as 0/1000 and end with 1000/0.
Won't it be 1000/1000?


>
> I think a better measure is: incomplete vs. all or complete vs. all.
>
>> I also wish
>> it showed an estimated time remaining for all the tasks to finish; the
>> average time per task is a nice start already.
>
> For estimated time look at --eta.
>
> The 100% is the % of started jobs that are started on this computer.
> It changes if you use remote computers (-S).
>
> Currently --progress does not require GNU Parallel to read all
> arguments and I would like it to continue to be this way. However,
> --eta already requires reading all arguments, so it could be part of
> that. An ETA line looks like this:
>
> ETA: 4490s 99291left 0.02avg  local:2/9909/100%/0.1s
>
> We could change it to:
>
> ETA: 4490s 99291=90%left 0.02avg  local:2/9909/100%/0.1s
>
> Do have a better suggestion?
>
>
> /Ole
>
>