Possible upcoming y-cruncher v0.8.x Tweaks

[_mat_]

41 minutes ago, Mysticial said:

How difficult would it be to change the 2.5b submission keys to keep everything consistent? /cc @Leeghoofd

I can give both of you the formula I used to generate the keys for all y-cruncher submissions - of all sizes.

It shouldn't be a problem, but we need to communicate it so the community uploads all sandbags before we switch keys.

Sounds good, let's use the same formula together to generate the keys.

[Leeghoofd]

Can you bring the HWBOT wrapper back online plz, need to check some things to crosscheck to help Mat

[Mysticial]

Figured I'd drop an update.

Right now I'm planning on v0.8.6 being the first version that will have additional settings. (So 2 versions away from now.) But instead of being fully configurable like the example in the opening post, it will likely be a list of built-in profiles that you can choose from.

So for example, the 22-ZN4 binary defaults to the 7950X profile. But it may have additional profiles for 8c Phoenix, 96c TR, or 128c Bergamo.

You will not be able to specify low-level things like the memory hierarchy or the algorithm parameter tables. At least not in this first iteration. We'll see how this goes first.

--------

Long story short, while I am able to"package all the editable settings in a way that can be imported from an external source, the problem is that there are way too many of them. And many of those settings are not human consumable - meaning that it's difficult to understand how the setting affects performance. For example, there are several very large tables of values that are not too dissimilar to weights of a neural network AI. So exposing them without also exposing the tools to generate (train) those weights is kind of useless.

 

Edited May 18 by Mysticial