[chemfp] chemfp 1.6.1 is available

[Andrew Dalke]

Hi all,

  I just released chemfp 1.6.1.

Up until now, there were specialized functions for 166-bit (really, 24-byte), 512-bit, 881-bit (really 112 byte), 1024-bit and 2048-bit fingerprint lengths. 

The new release adds specialized functions for all fingerprints which are a multiple of 8 bytes long, up to 1024-bit, and improves the performance of the implementation for fingerprints larger than 1024-bit which are also a multiple of 8 bytes long.

In one benchmark, small fingerprints (<256 bits) are about 20% faster, medium fingerprints (256 to 1024 bits) are about 10% faster, and larger fingerprints are a few percent faster.

I also added two new FingerprintArena methods. sample() randomly selects a subset of the fingerprints and returns them in a new arena. train_test_split() returns two randomly selected and disjoint subsets of the area, typically used as a training set and a test set.

Finally, I fixed a bug in fpcat where using --reorder would write the FPS header twice.


These fixes will be integrated in chemfp 3.4 within a few weeks.

				Andrew
				dalke at dalkescientific.com