--- Day changed Fri Dec 04 2015
01:03 -!- GAit [~GAit@2-230-161-158.ip202.fastwebnet.it] has joined #secp256k1
02:42 -!- GAit [~GAit@2-230-161-158.ip202.fastwebnet.it] has quit [Quit: Leaving.]
03:09 -!- GAit [~GAit@2-230-161-158.ip202.fastwebnet.it] has joined #secp256k1
04:17 -!- GAit [~GAit@2-230-161-158.ip202.fastwebnet.it] has quit [Quit: Leaving.]
04:19 -!- GAit [~GAit@2-230-161-158.ip202.fastwebnet.it] has joined #secp256k1
04:44 -!- GAit [~GAit@2-230-161-158.ip202.fastwebnet.it] has quit [Quit: Leaving.]
06:51 -!- jtimon [~quassel@74.29.134.37.dynamic.jazztel.es] has quit [Remote host closed the connection]
13:42 < gmaxwell> sipa: interesting, I have test_rand_bits failures; and considering that it's not boosted up by the count, I really shouldn't.
13:42 < gmaxwell> ./tests 1 937e62e54ceca1f14f1b147f6f4bb77b
13:43 < gmaxwell> that failure dies on a 24 bit test too, so a fairly high round count.
13:56 < gmaxwell> looping on it looks like dies in a few thousand tries. your one in a billion is uncorrected for multiple comparisons thoug the result is still surprising.
13:56 < gmaxwell> e.g. 6 mults, 27 shifts.
16:38 -!- fkhan [weechat@gateway/vpn/mullvad/x-ayqkrzoaatioinuf] has quit [Ping timeout: 250 seconds]
16:52 -!- fkhan [weechat@gateway/vpn/mullvad/x-hdxvkwzukfsqsuxc] has joined #secp256k1
21:29 < gmaxwell> accounting for multiple comparisons I expect one failure in 407612 test runs; which is much less often then I'm seeing errors.
22:09 < gmaxwell> sipa: dieharder is not impressed by secp256k1_rand_bits(31), assuming my bitpacking output isn't wrong.
22:09 < sipa> dieharder?
22:09 < gmaxwell> http://www.phy.duke.edu/~rgb/General/dieharder.php
22:35 < gmaxwell> looks like the 31st bit is always 0.
22:35 < sipa> well if you only ask for 31 bits, the highest bit will always be zero...
22:37 < sipa> hmm, it does a x >>= bits
22:37 < sipa> oh, but x is uint64_t, that should be well defined
22:40 < gmaxwell> yea, my bitpacker was just off by one.
22:49 < gmaxwell> with the fixed bitpacker it's passing. I still don't see the error in the test.
22:53 < gmaxwell> (other than the multiple comparisons; which doesn't explain the failure rate I'm seeing.)
23:19 -!- Luke-Jr [~luke-jr@unaffiliated/luke-jr] has quit [Ping timeout: 250 seconds]