79084657409835 163 Greenbank (100000000000000)

35916314132685 164 Greenbank
72145055243415 164 Greenbank
13903212126195 164 Greenbank (100000000000000)

(All verified as primes, including the base 138 above).

[QUOTE=Greenbank]75783835007265 138 Greenbank (100000000000000)[/QUOTE]

Yesterday (post 135) i wrote:[quote=smh]Thanks for the update. I've found octoproths for all N below 151 except for 146 and 149.[/quote]

Actually now i have found octoproths for all n upto and including 150. I'll continue the current range as to make sure the ones i found are the first ones. I'll post all octoproths i found and to what bounds i searched in an hour or so when my last range finishes.

Sorry for the duplication, these things are quick to find (138 took me just under 40 minutes for that range from 0 to 10^14).

[EDIT] Also, there seems to be a bug in the program:-

./oc 138 75783835000000 75783836000000 finds no values
./oc 138 75783830000000 75783836000000 finds no values
./oc 138 75783800000000 75783836000000 finds no values
./oc 138 75783000000000 75783836000000 finds no values
./oc 138 75780000000000 75783836000000 finds no values

Only when I drop down to:-

$ ./oc 138 75700000000000 75783836000000
You can also find the k n values in results_octo.txt file ( These are 3-probable primes )
n=138, kmin=75700000000000, kmax=75783836000000
Starting the sieve...
75783835007265 138
The sieving is complete.
Time=5.32 sec.

Does it find that particular value.

(Notice the upper bound stays the same and I'm dropping the lower bound, despite it being lower than the known value).

[QUOTE=Greenbank]79084657409835 163 Greenbank (100000000000000)

35916314132685 164 Greenbank
72145055243415 164 Greenbank
13903212126195 164 Greenbank (100000000000000)

(All verified as primes, including the base 138 above).[/QUOTE]

I think that it would be better if you could upload results_octo.txt file to avoid a probably copy-paste error. Everybody can see what ranges have you finished and the timings. After this you can delete this results file.

I've calculated that octo program is using about 810 Kbytes of memory this means that if your computer's L2 cache is at least 1 Mbyte then all computations will be in fast L2 memory.

[QUOTE=Greenbank]
Also, there seems to be a bug in the program:-
[/QUOTE]

I'll see this on my computer. I think that this is a very small bug. Some redefinitions of kmin and kmax are in the program to use a special sieve.

[QUOTE=Greenbank]Sorry for the duplication, these things are quick to find (138 took me just under 40 minutes for that range from 0 to 10^14).[/QUOTE]

Don't worry, just wanted to warn you about the work i had already done.

Below are the bounds and all octoproths for the remaining numbers <151
[code]135 60000000000000 52324121395455
136 50000000000000 43732992858225, 49020296591265
137 20000000000000 16328957487255
138 100000000000000 75783835007265
139 10000000000000 2536458109485, 9744953313285
140 74481305417505 70641630562965
141 63000000000000 45672428687445, 52732393734195
143 20000000000000 9334074009855
144 2000000000000 1486857923775
145 98335464178875 74163464017005, 74300238381915,
146 365130319638525 316114654394385, 365130319638525
147 25640808056325 21668337395145
148 6000000000000 4767131022555
149 218016644008485 165652459923405, 169640719682595, 216874306279155, 218016644008485
150 40000000000000 34958329919835, 39172241106705[/code]

serious bugs
 

[QUOTE=R. Gerbicz]
I'll see this on my computer. I think that this is a very small bug. Some redefinitions of kmin and kmax are in the program to use a special sieve.[/QUOTE]

This isn't true, this happens every time if kmax/kmin is very close to 1
I've fixed this bug, this was a definition bug: an unsigned long int and unsigned long long int problem.

Looking the code I've also found two possible serious bugs, if the range>1.29*10^16, I've also corrected these bugs.

Sorry for these. Greenbank can you test my new code? Test this if kmax/kmin is very close to 1 and also if Range is large.

I suggest everyone to update octo program!!! Third line of the code indicates for you that this is a new version.

I don't think that there is any missed octoproth primes.

[QUOTE=R. Gerbicz]This isn't true, this happens every time if kmax/kmin is very close to 1
I've fixed this bug, this was a definition bug: an unsigned long int and unsigned long long int problem.

Looking the code I've also found two possible serious bugs, if the range>1.29*10^16, I've also corrected these bugs.

Sorry for these. Greenbank can you test my new code? Test this if kmax/kmin is very close to 1 and also if Range is large.

I suggest everyone to update octo program!!! Third line of the code indicates for you that this is a new version.

I don't think that there is any missed octoproth primes.[/QUOTE]

I still can't compile, so an executable is welcome.

This new software looks rather promising. Hopefully an executable will come out soon so we can all use it.

New octos
 

All found using R.Gerbicz's latest version of [I]octo[/I]

22354826339175 151 [50T]
16693766169885 152 [30T]
51618124858425 154 [60T]

I'll take care of the remaining n's 153-160.

BTW, here is an exe for Windows, but if you don't have cygwin installed on your machine you will need cygwin1.dll, 1.2M, available at [URL="http://www.cygwin.com"]http://www.cygwin.com[/URL] and cyggmp-3.dll, 158k, availabe at the same place but I don't know can you get only this one without installing the whole cygwin (?) I can mail the second one to interested parties, just let me know.

[QUOTE=Kosmaj]availabe at the same place but I don't know can you get only this one without installing the whole cygwin (?) I can mail the second one to interested parties, just let me know.[/QUOTE]

That's possible. ET_ on this forum: See [URL="http://home.earthlink.net/~elevensmooth/Billion.html#Download"]http://home.earthlink.net/~elevensmooth/Billion.html#Download[/URL]
He has done an exe file from his gmp source code ( his factor program) without using the
whole cygwin. Possibly he can try to make an exe file for octo without using the whole cygwin. Send him an email using mersenneforum. ( not a private message because his latest post was on 14 Dec 05, 07:53 AM )

[QUOTE=Dougy]This new software looks rather promising. Hopefully an executable will come out soon so we can all use it.[/QUOTE]

Possibly start a thread on Mersenneforum's Sofware forum :
[B]How can we make an exe file from gmp source code without installing the whole cygwin?[/B]
I don't think that's very difficult but I haven't used cygwin.

[QUOTE=R. Gerbicz]That's possible. ET_ on this forum: See [URL="http://home.earthlink.net/~elevensmooth/Billion.html#Download"]http://home.earthlink.net/~elevensmooth/Billion.html#Download[/URL]
He has done an exe file from his gmp source code ( his factor program) without using the
whole cygwin. Possibly he can try to make an exe file for octo without using the whole cygwin. Send him an email using mersenneforum. ( not a private message because his latest post was on 14 Dec 05, 07:53 AM )[/QUOTE]

ET_ didn't compile not using the whole cygwin. If you look carefully, you will see a link to the cygwin libraries (not the most recent ones) in [url]http://home.earthlink.net/~elevensmooth/cygwin_dll.zip[/url].

I don't think you can compile without cygwin. The best you could do is either only download the packages of interest (saves much space), or, if you're planning on not having to distribute the libraries, you can use MinGW and Msys, or even Cygwin itself! Using the option --mno-cygwin, you will tell the program to emulate a Mingw environment using its libraries, not cygwin's.

Edit:Cygwin1.dll is the main library and cyggmp-3 comes with gmp. With this information try to search in one of the [URL="ftp://mirrors.dotsrc.org/cygwin/release"]FTPs[/URL], where all of the files are stored.
P.S: Thanks for the executable!