Has anybody used bigfloat or mpmath in #Python? I need to raise some floats to power (50+) & currently getting overflow error 34 —ideas? 
-
-
Replying to @o_guest
High powers are always suspect. Can you replace your variables by their logarithms? If you can, it's very probably the best choice.
2 replies 0 retweets 1 like -
The problem is not just overflow, but also precision loss in subsequent operations.
1 reply 0 retweets 1 like -
Replying to @khinsen
I'm using bigfloat though, so technically I think nothing lost.
1 reply 0 retweets 0 likes -
Replying to @o_guest
Bigfloat (Python wrapper for GNU MPFR) has fixed and finite precision. You choose it once and then stick to it.
2 replies 1 retweet 0 likes -
It's not adaptive precision. You get that in some so-called "exact real" libraries, at a huge performance cost.
1 reply 0 retweets 1 like -
Replying to @khinsen
precision is really not that important. I'm just calculating the weight in a Luce choice.
1 reply 0 retweets 0 likes -
Replying to @o_guest
At worst you can lose all precision, i.e. compute expensive random numbers. It's worth estimating how much precision you have left.
1 reply 0 retweets 1 like
ok I'll look into it. I saw your emails. Thanks.
Loading seems to be taking a while.
Twitter may be over capacity or experiencing a momentary hiccup. Try again or visit Twitter Status for more information.