anastasios-tampakis.net

I understand your point. I was just thinking that it will allow us to increase the Maximum mother blocks for focused search. Am I wrong about this too?

Do you think it will also be possible to implement a function that shows us how much time has passed since the last improvement. It will be of great help to know if we are stagnating and that it's time to make additional optimization settings.

Thanks.

Some more indication of the optimization progress is considered as an overall improvement, including that time since last improvement and a more insightful graph.
Going from 32 bit memory to 64 bit allows for some improvement to the various limits - which already works in a development 64-bit version like making a 51,20,4,4 or 25,15,6,6 wheels - but most operations have a logarithmic time and memory complexity which doesn't give much benefit from any extra available memory. In these wheel examples, no matter how much more memory we have available it does not offer any speedup benefit as it covers something like 0.001% of the possible search space.

Anyway, I am not against a 64bit version as long as I have addressed all the issues arising from that conversion. Some modules currently used are not 64-bit compatible so I have to implement alternative solutions to support such a version.

To put this simply, the only real benefit going from 32-bit to 64-bit in WG is the expansion of the v,k,t,m limits which is huge but this looks like to be the only area that can really benefit from this transition and the reason I work on that. But it will not offer any real advantage over the speedup, perhaps in very few additional wheels like if the 32-bit version it allows a speedup of 10% for a particular wheel, this wheel will probably operate at 100% speedup on the 64-bit version but this will be effective for a very limited set of wheels. All wheels that have a speedup of 1% or lower will certainly remain as such without any real speedup improvement.