1 2015-01-15 00:11:57 <midnightmagic> sipa: fewer for the osx build, unfortunately.
 2 2015-01-15 00:12:01 <midnightmagic> 
ACTION kicks server

 3 2015-01-15 00:13:11 <sipa> 
ACTION invites server

 4 2015-01-15 00:47:47 <phantomcircuit> so yeah that's a bit odd
 5 2015-01-15 00:47:55 <phantomcircuit> i was not at all expected that
 6 2015-01-15 00:48:04 <sipa> what did you not expect?
 7 2015-01-15 00:48:12 <gmaxwell> sipa: so yea, phantom was 35 minutes faster with sigchecks false vs a checkpoint set about 120 blocks before.
 8 2015-01-15 00:48:20 <sipa> huh
 9 2015-01-15 00:48:55 <phantomcircuit> perf results https://0bin.zertrin.org/paste/4d5df6f04c342c9964f49a565c630641ebf7b20c#cnCs7LiSe+Qnkcc2d04FWpkc1v7w06WsrV3mD9dz080=
10 2015-01-15 00:49:16 <gmaxwell> 2661 seconds with setting it false vs 4836 with checkpoint.
11 2015-01-15 00:49:43 <sipa> 12% of sha256?
12 2015-01-15 00:50:54 <phantomcircuit> btw this is git master
13 2015-01-15 00:50:55 <gmaxwell> this is with a huge dbcache too.
14 2015-01-15 00:52:58 <gmaxwell> I am unshocked by this profile. But sad that nothing totally stupid is standing out.
15 2015-01-15 00:53:38 <gmaxwell> phantomcircuit: any way to find out some more backtrace context from this profile? e.g. is their a particular spot where malloc is costing the most?
16 2015-01-15 01:06:18 <phantomcircuit> gmaxwell, excellent question
17 2015-01-15 01:06:20 <phantomcircuit> probably?
18 2015-01-15 01:07:01 <phantomcircuit> >.>
19 2015-01-15 01:07:06 <phantomcircuit> Selected -g but no callchain data. Did you call 'perf record' without -g?
20 2015-01-15 01:07:15 <phantomcircuit> 
ACTION grumbles and runs reindex again

21 2015-01-15 01:08:07 <op_mul> sounds like skylake processors might get a few percent speedup with sha256 instructions then
22 2015-01-15 01:09:19 <phantomcircuit> op_mul, this is an amazingly synthetic benchmark
23 2015-01-15 01:11:01 <op_mul> right
24 2015-01-15 01:11:11 <gmaxwell> op_mul: I expect that a lot of that sha256 is in the hashtrees, where we could probably get a 2x speedup using SSE2.
25 2015-01-15 01:36:36 <phantomcircuit> killed it early since im really just interested in the perf data
26 2015-01-15 01:38:25 <phantomcircuit> sigh
27 2015-01-15 01:38:26 <phantomcircuit> [ perf record: Captured and wrote 2545.375 MB bitcoind (~111209141 samples) ]
28 2015-01-15 01:38:33 <phantomcircuit> that's going to take fucking ages to finish
29 2015-01-15 01:50:38 <rondon29> greetings
30 2015-01-15 04:25:28 <gigawave> Im running Bitcoin Core/QT 0.9.3, lanuched for first time in about a month and im getting "No block source available", using unpnp with 16 active connections. Any ideas, been searching and tried a few things with no luck.
31 2015-01-15 04:26:52 <Dizzle> #bitcoin a better place to ask, technically. Off-hand I don't know the answer. But if you're on linux, there's a known bug with recent versions of openssl.
32 2015-01-15 04:28:06 <gigawave> On win blows. Thanks.