Upto 73 bits (79.000-79.001M)
 

More than two years ago (April 2005) [URL="http://www.mersenneforum.org/showpost.php?p=52781&postcount=143"]lycorn wrote[/URL]:
[quote=lycorn;52781]Now that version 24.11 is around, will somebody volunteer to factor a number to 73 bits, so we don´t have that "0" anymore? I would gladly do that, but haven´t got any AMD64 :down:[/quote]

I've taken 18 exponents in the range 79.000-79.001 and will go with them upto 73 bits (step by step). 65-bit limit should be reached in a few days. Results will reported here and sent to GW. :wink:

Done upto 65 bits
 

No factors upto 65 bits. TF for 18 numbers took about one day (in the previous post I used old Roll.Ave. established when a system had been overloaded). Hence, 66-bit should be ready on Sunday, and 67-bit - on Thursday, Nov. 8. And then results will be sent to GW. Next levels:
68-bit - Nov. 16; 69-bit - Dec. 2, 70-bit - Jan. 3, 2008, 71-bit in March, 72-bit in July and, finaly, 73-bit in March 2009 :down:. Maybe some factors will be found earlier, new CPU's will arrive, etc, so the above predictions are rather pesimistic. :cool:

66 bits
 

No factors upto 66 bits. :sad: Moreover, M79000093 has no factor to 2^67. The other 17 numbers should be tested in next 90 hours (or earlier if a factor will be found). :wink:

Are you trial factoring the whole 18 numbers to 73 bits? Man, you're brave!
Do you use a 64-bit client for that purpose? My original post was written having in mind the excellent perfomance of the 64 bit version of Prime95 at Trial Factoring. To trial factor numbers to such high limits it is highly recommended to use a machine with a really good performance. A 64-bit capable machine and a 64-bit OS (e.g Win XP-64) would do the job.

CPU
 

I have some free cycles at
[code]
vendor_id : GenuineIntel
cpu family : 15
model : 2
model name : Intel(R) Xeon(TM) CPU 2.40GHz
stepping : 5
cpu MHz : 2392.095
cache size : 512 KB
[/code]
under Linux (mprime is fantastic under Linux). I hope 2-3 factors will be found and I'll be able to use one or two more CPU's next year. Going bit by bit the first test to 73 bits should be finished in Aug. 2008 (using only this one comp). At present about 5.5 hours is necessary for one TF from 66 to 67 bits. I don't know if there are any changes in the algorithm for larger factors. If not, we have: 67->68 bits - 12 hours, 68->69 one day, and 16 days for 72->73 bits (31.75 days for 66 -> 73 bits). I have access to a dual core AMD Opteron 2GHz, which I can use later.

67 bits reached
 

No factors to 67 bits. The limit of 68 bits should be reached in 8-9 days (about 12 hours per exponent).

68 bits
 

The limit 68 bits has been reached. Two factors found:
M79000421 has a factor: 225004026300573568127
M79000697 has a factor: 153469874614532336401
Results mailed to George Woltman.
Results for 69 bits in about two weeks.

Next CPU
 

7 exponents TF'ed to 2^69 (w/o factors). The other 9 should be ready in a week (one number needs about 19-20 hours at present). The second comp has started TF to 70 bits. It's a bit slower and more loaded so at least two days per an exponent is needed. So this limit should be ready in about one month.

Veru rough estimation (with 2 CPU's) gives Aug 2008 as the end-date of this subsub...subproject.

69 bits
 

69 bits reached. No new factors. The 70-bit limit should be reached in a week since I may use another CPU for a while.

70 bits
 

70-bit limit reached (I've had acces to 4 CPU's for a week). No factors. Results sent to George W. The next step needs about 66 hours per number; 8 numbers per CPU (I'll use two of them) so about 22 days is needed.

4 numbers with no factors up to 2^71.