Leyland Primes (x^y+y^x primes)

[rogue]

Originally Posted by pxp

Thank you. Fortunately I had written a note to myself from the last time I ran a .cpp program.

g++ xyyx.cpp -o xyyx

xyyx.cpp:60:56: warning: format specifies type 'unsigned long long *' but the argument has type 'long *'
      [-Wformat]
         if (sscanf(ptr, "%u %llu %u (%u,%u)", &index, &leylandNumber, &length, &x, &y) != 5)

The %llu and &leylandNumber were underlined. This was followed by a very similar warning ending in !=4 and finally a third one relating to an fprintf item containing the two offending variables. In spite of the warnings the created xyyx ran to create a list.html from a list.txt.

I can probably run this every time I update my a094133.txt document and share it here. A couple of minor issues: Christ van Willegen and Jens Kruse Andersen have lost their surnames and Göran Hemdal has lost the umlauted o (I assume that it is visible in the .txt version).

Try compiling with -m64. A long is 64-bits with -m64 but only 32-bits for -m32 and I assume that the default for the compiler is -m32.

I treat the input as ASCII, so it is likely losing the umlaut in his name. I'm not certain how to fix that.

The code assumes only two portions of the name. It would be easy to fix the parsing in the code to address that.

[pxp]

Originally Posted by rogue

Try compiling with -m64.

Yes, if only I knew how to do that. Is it really that important? What does my warnings-ignored converted document lack that a properly compiled conversion file would have?

[rogue]

Originally Posted by pxp

Yes, if only I knew how to do that. Is it really that important? What does my warnings-ignored converted document lack that a properly compiled conversion file would have?

You can change the datatype from "long" to "long long" to remove the warning that you showed.

[NorbSchneider]

Another new PRP:
452^50145+50145^452, 133142 digits.

[pxp]

Originally Posted by pxp

I have examined all Leyland numbers in the ten gaps between L(37614,265) <91148>, #1735, and L(40210,287) <98832> and found 117 new primes. That makes L(40210,287) #1862 and advances the index to L(40945,328), #1930.

I have examined all Leyland numbers in the gap between L(40945,328) <103013>, #1930, and L(41507,322) <104094> and found 14 new primes. That makes L(41507,322) #1945.

[pxp]

I have examined all Leyland numbers in the gap between L(148999,10) <149000> and L(149999,10) <150000> and found 14 new primes.

[pxp]

Originally Posted by pxp

That makes L(41507,322) #1945.

I have examined all Leyland numbers in the three gaps between L(41507,322) <104094>, #1945, and L(222748,3) <106278> and found 38 new primes. That makes L(222748,3) #1986.

[rogue]

29652^5083+5083^29652 is 3-PRP

This was an accidental find, in other words, unexpected. I had removed x with fewer than 50 terms from the main sieve to test separately because sieving is less efficient with x that have few terms.

[pxp]

It's actually nice to see another name in the recents. I am still ten days away from discovering this one. Be sure to claim it on PRPTop so that they will be up-to-date.

[rogue]

Originally Posted by pxp

It's actually nice to see another name in the recents. I am still ten days away from discovering this one. Be sure to claim it on PRPTop so that they will be up-to-date.

Submitted.

[pxp]

Originally Posted by pxp

I began to wonder if any of these L(x,10) is prime. I'm doing a run on a list that I didn't sieve particularly deeply and I can say that for x < 300000 the answer is none.

I finished that run up to 500000 and then decided to do from there to 10^6. I sieved that file to 10^12 resulting in 1309 candidates. The durations total to 293.3 days but distributed equally over 12 cores, 24.4 days. Of course I didn't have the pfgw durations when I started so I guessed at the distribution, resulting in some of the segments taking considerably longer than others. Sadly, no PRPs.