Strange mail bug (?)

Strange mail bug (?)

Since yesterday, I receive the benchmark results of some of my submissions more than once. Sometimes I receive mails from old submissions (i.e., I have submitted newer versions of the code since then). The benchmark appears to be re-running everytime (sometimes falling). E.g., this morning at 00:23 this night, I received the following mail:

Your submission from 2012-05-14 23:00:21 CET has been ran on our servers, here are the results :

error on a 40-cores HT machine :
Can't read file

This exact submission (23:00:21) ran 2 times successfully and and failed the last two time with this error (I have received 4 mails in total for 1 submission).

Am I the only one with this kind of problem?

8 posts / 0 new
Last post
For more complete information about compiler optimizations, see our Optimization Notice.

Only the newest submission will be stored. You will see "Can't read file" if you submit a new solution before the previous submission ran on the 40-cores machine.

No, the problem is more subtle than that: I always wait for a submission to be tested before submitting another zip file. The problem is that a single submission is often evaluated more than once.

It also happened to me this morning, but for me it may be the case underlined by bdrung. Everybody is testing and the servers are crowded I think, because there's a bigger delay in receiving results than in previous days.

That race between submissions should be solved.
If you send a new submission while another one is in the queue, it wil wait the end of its processing.While that former submission is in the queue, you can submit as many times as you want, only the latest submission will be processed next when your former submission will leave the queue.

Still, this last day will be stressful : results are slow to come back!And with this new difficult test in addition.

Yep, I've just received results for a submission at 10:46 that i thought it was overwritten (1h10m later). Besides that everyone want's to test, the thing that results are slow makes impatient people (like me) make multiple submissions, making the server more crowded :).LE: another one from 10:01 (2h later). Maybe somebody rerun my submissions with problems (can't read file)

I still receive random reports of past submissions (around 10AM CET), but new ones (after 12AM CET) don't come.I am not submiting to see if my time improved, but to test if I my solution passes the tests (I received "unexpected error" on 87 benchmark, probably due to high memory need when processing big sequences in the same time - we use O(N) memory and tried tu free everything).

Leave a Comment

Please sign in to add a comment. Not a member? Join today