Join PrimeGrid
Returning Participants
Community
Leader Boards
Results
Other
drummers-lowrise
|
Message boards :
Proth Prime Search :
PPS Mega statistics page incorrectly marking random rows as complete
Author |
Message |
|
As far as I understand, PPS Mega is currently testing 1201 ≤ k ≤ 9999 up to n = 3.6M. However, on the statistics page, if you scroll down to 1201, random rows below that are green, indicating they are 'complete', even though they are only at n = 3.33M.
It seems like the page is automatically highlighting a row as "complete" whenever there are 0 tasks in progress for that k value, even if the "tasks waiting" column is above zero. In fact if you wait a few hours and refresh the page, you'll find some rows go from green to white, and others change from white to green.
Is this intended, or a mistake? If it is a mistake, I would appreciate if it was corrected. Thanks.
____________
1 PPSE (+2 DC) & 5 SGS primes | |
|
Ravi FernandoProject administrator Volunteer tester Project scientist Send message
Joined: 21 Mar 19 Posts: 210 ID: 1108183 Credit: 13,047,881 RAC: 6,479
              
|
Where does it say that green means complete? I assumed that it was intended to mean "no tasks in progress". | |
|
Ravi FernandoProject administrator Volunteer tester Project scientist Send message
Joined: 21 Mar 19 Posts: 210 ID: 1108183 Credit: 13,047,881 RAC: 6,479
              
|
FWIW, I don't think the tables for different subprojects are necessarily intended to have a uniform "key". They were made at various times, most (I think) by JimB but maybe some before him. For example, green appears to mean "no tasks in progress or waiting" (but still not necessarily completed) in the PPSE table. And in the PPS-Sieve table, completed ranges are not even shown. | |
|
|
Right now there are two colors: White when tasks are in progress, and green otherwise. It should be easy enough to turn that into three colors: (1) when tasks are in progress, (2) when task are loaded or waiting but not in progress, and (3) when there are no current tasks (like for k < 1200 now).
Although it is not hard to understand the table as it is now, in my opinion.
I am not sure if some k can be so light-weight that occasionally no tasks exist in a "load" even though the 1200 < k < 10000.
/JeppeSN | |
|
Ravi FernandoProject administrator Volunteer tester Project scientist Send message
Joined: 21 Mar 19 Posts: 210 ID: 1108183 Credit: 13,047,881 RAC: 6,479
              
|
Currently there are three green k's for PPSE: 3125, 3375, and 4847. The first two have been tested (with both odd and even n) in the early stages of PPS-DIV, so they are done through n=3.322M. The last just has low weight. The same happens in the MEGA table, for example, for k=4861. | |
|
|
Currently there are three green k's for PPSE: 3125, 3375, and 4847. The first two have been tested (with both odd and even n) in the early stages of PPS-DIV, so they are done through n=3.322M. The last just has low weight. The same happens in the MEGA table, for example, for k=4861.
Yes, so you found an example of that! So once in a while, rows in the "active" interval 1200 < k < 10000 would get the color (3). There is probably nothing we could do about it. But we could still have three colors, if we wanted. /JeppeSN | |
|
|
Using 3 colours appears to be supported, as a light-blue is already used on the PSP and ESP status pages.
____________
1 PPSE (+2 DC) & 5 SGS primes | |
|
|
SR5 already has three colors:
/JeppeSN | |
|
streamVolunteer moderator Project administrator Volunteer developer Volunteer tester Send message
Joined: 1 Mar 14 Posts: 1022 ID: 301928 Credit: 543,195,188 RAC: 112
                        
|
We'll think if something can be improved here, but please understand that there is a major difference between PPS Mega and conjecture projects like SR5.
On conjecture project, it clear to find out is this k "active" or "inactive' - it becomes inactive when a prime is found, so no tasks in inactive state means that we've done with this k and it can be colored green. An active k without tasks is colored yellow.
In PPS Mega all k's are always active and running forever from server's point of view (technically, it ends where our sieving data ends, but it's at 12M - very, very, far in the future). k's from 3 to 1200 are not complete - it's a temporary admin's decision to not load new work into these k's. Some k's were temporary shared with DIV project. And so on, so forth. But server knows only two states: "have work" - "no work", and it not aware about all these little evil details.
We can color these lines yellow, if it makes more sense - to show that work on them be continued at any time. And use green color only for k's which will never be running again because physical limit of the range has been reached or prime has been found for conjecture.
| |
|
Message boards :
Proth Prime Search :
PPS Mega statistics page incorrectly marking random rows as complete |