[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [tlaplus] The effect of symmetry sets on TLC performance



On 02.12.2015 13:23, Ron Pressler wrote:
> Is there a way to improve the implementation(s) of permute()? A quick
> glance at the code showed what may possibly be way too many allocations
> (in tuples and record) due to what seems like an overly eager allocation
> policy (e.g. arrays are allocated before even a single permuted element
> is found). But as all that's required of the permutations is a
> fingerprint, perhaps a more substantial optimization may be possible.

Hi Ron,

those are interesting findings. Can you send me a (stripped down)
spec/model that exemplifies the performance problem?

Thanks
Markus