mersenneforum.org Report Results Here
 Register FAQ Search Today's Posts Mark Forums Read

 2019-06-14, 22:19 #1134 WGJC3107     "Will Conway" May 2018 UK 10000002 Posts Do you think we could expand the OBD range at some point? I know it seems a while off but with the recent burst of activity there’s a possibility that we might run out of exponents. Taking 3321931967 up to 84 bits Last fiddled with by WGJC3107 on 2019-06-14 at 22:19
 2019-06-14, 22:30 #1135 James Heinrich     "James Heinrich" May 2004 ex-Northern Ontario 2×3×19×31 Posts For my site I have defined OBD range as between 3,321,928,097 and 3,321,935,999 I can easily adjust the upper limit if there is consensus to do so.
2019-06-15, 09:01   #1136
ET_
Banned

"Luigi"
Aug 2002
Team Italia

483410 Posts

Quote:
 Originally Posted by James Heinrich For my site I have defined OBD range as between 3,321,928,097 and 3,321,935,999 I can easily adjust the upper limit if there is consensus to do so.
After a talk with William (when the project was still alive), we decided not to extend the ranges because the low hanging fruits could distract users from the deep factorization that characterizes the search.

Right now, it seems that we users are quite oriented to the result, so we could well extend a bit the ranges.

I could put my GT 710 back to work presieving very low exponents in background (taking exponents up to M3,321,939,983 from 72 bits to 73 or 74) and leave them to people having more powerful boards, but lower than RTX 20xx

2019-06-15, 11:39   #1137
ET_
Banned

"Luigi"
Aug 2002
Team Italia

10010111000102 Posts

Quote:
 Originally Posted by ET_ After a talk with William (when the project was still alive), we decided not to extend the ranges because the low hanging fruits could distract users from the deep factorization that characterizes the search. Right now, it seems that we users are quite oriented to the result, so we could well extend a bit the ranges. I could put my GT 710 back to work presieving very low exponents in background (taking exponents up to M3,321,939,983 from 72 bits to 73 or 74) and leave them to people having more powerful boards, but lower than RTX 20xx
OK. Who is stepping over my toes? I found that some of the exponents external to the OBD range (namely at 72 bit) are now at 75. Someone is evidently doing the range I had previously proposed. I'm not angry with it, I just feel how stupid could be doing some work and after realize you have just wasted your cpu time

2019-06-15, 11:40   #1138
ET_
Banned

"Luigi"
Aug 2002
Team Italia

2·2,417 Posts

Quote:
 Originally Posted by ET_ OK. Who is stepping over my toes? I found that some of the exponents external to the OBD range (namely at 72 bit) are now at 75. Someone is evidently doing the range I had previously proposed. I'm not angry with it, I just feel how stupid could be doing some work and after realize you have just wasted your cpu time
No, correction. I AM angry.

 2019-06-15, 13:25 #1139 lavalamp     Oct 2007 Manchester, UK 55216 Posts I don't know which specific exponents you were checking, but I looked at the first few above the OBD range without factors and found TF work to 75 bits submitted in 2014. Additionally, I think that at one point James said somewhere that he was doing some low level bulk TFing, but I have no idea in what range or on what hardware.
2019-06-15, 14:01   #1140
James Heinrich

"James Heinrich"
May 2004
ex-Northern Ontario

353410 Posts

Quote:
 Originally Posted by ET_ I found that some of the exponents external to the OBD range (namely at 72 bit) are now at 75. Someone is evidently doing the range I had previously proposed.
The entire 1000M-4294M range (beyond PrimeNet and as high as mfkatc goes) is available for TF work at https://www.mersenne.ca/tf1G.php
Generally you won't have any conflict between OBD and breadth-first TF work since the OBD range is already TF'd so high. Outside that range, everything is TF'd to at least 67 and heading towards 71. If you look at the last-30-day report for the 3321M range
https://www.mersenne.ca/status/tf/0/30/3/332000
you can see that there has been ~50 high-level TF progress (83+) but also >1300 exponents outside OBD range being taken to 72.

OBD and tf1G pages share a common reservation system, so as long as you're reserving the exponents you want to work on, nobody should step on your toes. If you want to work to lower bitlevels outside the exponent range shown on the OBD page, please reserve the exponents you want on the tf1G page.

Quote:
 Originally Posted by lavalamp James said somewhere that he was doing some low level bulk TFing, but I have no idea in what range or on what hardware.
The 1000M-4294M bulk TF work was done in 2014-2015 taking the whole range to 64. Since then this range has been managed by the tf1G page. In 2018 I also bulk-TF'd the range of 4294M-10000M to 55 using my CPU, but since no GPU TF software can handle this range (yet) it is not available for reservation.

2019-06-15, 15:41   #1141
ET_
Banned

"Luigi"
Aug 2002
Team Italia

2·2,417 Posts

Quote:
 Originally Posted by James Heinrich OBD and tf1G pages share a common reservation system, so as long as you're reserving the exponents you want to work on, nobody should step on your toes. If you want to work to lower bitlevels outside the exponent range shown on the OBD page, please reserve the exponents you want on the tf1G page.
:doh:

I tried to reserve 1 exponent to take from 72 to 73 bits, namely 3321936811.
Using the following input:

3321936810
3321936812
73
72
1

and was given no exponent (no assignment found).

Then I tried to change 3321936812 with 3321941999.

Still no exponent reserved (no assignment found).

But from https://www.mersenne.ca/exponent/bro...001/3321939999 it seems that there are a lot of available exponents...
Attached Thumbnails

Last fiddled with by ET_ on 2019-06-15 at 15:45

2019-06-15, 16:14   #1142
James Heinrich

"James Heinrich"
May 2004
ex-Northern Ontario

2·3·19·31 Posts

Quote:
 Originally Posted by ET_ I tried to reserve 1 exponent to take from 72 to 73 bits, namely 3321936811. But from https://www.mersenne.ca/exponent/bro...001/3321939999 it seems that there are a lot of available exponents...
I'm not sure why you're seeing what you show in your screenshot because most of that range is already done to 74 (including M3321936811 you tried to reserve to 73.

As a quick summary for the 3321M range, the available exponent counts are currently:
Code:
+--------+------+---------+
| mrange | tf   | howmany |
+--------+------+---------+
|   3321 | fact |   23273 |
|   3321 |   67 |   18840 |
|   3321 |   69 |       9 |
|   3321 |   74 |      67 |
|   3321 |   75 |      29 |
|   3321 |   80 |       2 |
|   3321 |   83 |      39 |
|   3321 |   84 |      47 |
|   3321 |   85 |      25 |
|   3321 |   86 |       7 |
+--------+------+---------+
Including currently-assigned exponents and restricted to >1Gdigit:
Code:
+--------+------+---------+
| mrange | tf   | howmany |
+--------+------+---------+
|   3321 | fact |    1840 |
|   3321 |   72 |    1241 |
|   3321 |   74 |      67 |
|   3321 |   75 |      29 |
|   3321 |   83 |      43 |
|   3321 |   84 |      50 |
|   3321 |   85 |      26 |
|   3321 |   86 |       8 |
+--------+------+---------+
So there are a number of exponents at 72, but they're all currently assigned. They'll either be returned at a higher bitlevel within the next 10 days, or the assignment will expire and they'll become available again. I just checked and all 1241 of them have an assignment expiry date of 2019-06-24; it's possible they were inadvertently reserved by me when I was testing stuff yesterday, but I can't be certain of that so I don't want to summarily release the reservations and risk having duplicated work. Worst case, check back on June 25 and they should be available again then.
Attached Thumbnails

2019-06-15, 16:43   #1143
ET_
Banned

"Luigi"
Aug 2002
Team Italia

10010111000102 Posts

Quote:
 Originally Posted by James Heinrich So there are a number of exponents at 72, but they're all currently assigned. They'll either be returned at a higher bitlevel within the next 10 days, or the assignment will expire and they'll become available again. I just checked and all 1241 of them have an assignment expiry date of 2019-06-24; it's possible they were inadvertently reserved by me when I was testing stuff yesterday, but I can't be certain of that so I don't want to summarily release the reservations and risk having duplicated work. Worst case, check back on June 25 and they should be available again then.
I understand.

When I ran the project, once the exponent was assigned to a bitlevel, my page showed the exponent AT the RESERVED bitlevel, not the ACTUAL bitlevel, so that others could reserve the same exponent to a higher bitlevel.
If the reservation was completed the situation was ready for the next reservation, else the bitlevel of the relative exponent was rollbacked. As the users were few, we never had a hole to close, and even in that case I would have provided the patch.

2019-06-15, 23:14   #1144
James Heinrich

"James Heinrich"
May 2004
ex-Northern Ontario

67168 Posts

Quote:
 Originally Posted by lavalamp I'm still compiling a list so that appropriate credit can be applied, which I can either send in its current format for you to manually apply
And I'm still waiting for it.

 Similar Threads Thread Thread Starter Forum Replies Last Post petrw1 PrimeNet 5 2018-07-09 21:14 houding PrimeNet 15 2018-01-10 19:24 dcorbit Information & Answers 4 2012-08-13 21:53 James Heinrich Marin's Mersenne-aries 6 2004-05-27 05:47 remaker Software 4 2002-10-17 08:18

All times are UTC. The time now is 01:33.

Wed Dec 8 01:33:48 UTC 2021 up 137 days, 20:02, 0 users, load averages: 1.38, 1.45, 1.34