Unfortunately there is still a bug in the code. :no:

It is missing factors under certain unknown conditions.

[QUOTE=rogue;390847]Unfortunately there is still a bug in the code. :no:

It is missing factors under certain unknown conditions.[/QUOTE]

I'm still at a loss on this one. Sometimes I can run the code and get a nice OpenCL error at runtime, but I don't know why. The two are likely related. If I can track down the latter, the former might be fixed along with it.

As if I haven't stated this already, I recommend against anyone using MultiSieve. Not only will I not support it, it will not test n < b for either Cullen or Woodall sieves. For Cullens n can be less than b, but not for Woodalls so it is eliminating candidates that should not be eliminated.

The recommended method of building candidates for testing is as follows:

1) Run gwcsieve_smallp to to p = n.
2) Using the output from step 1, run gcwsieve for p > n.