Commit Graph

12111 Commits

Author SHA1 Message Date
Steven Allen
63e2caae81 lint(lotus-sim): handle error 2021-06-22 15:06:44 -07:00
Steven Allen
87c306fd47 feat(lotus-sim): use current power instead of lookback
I'd _really_ like to use lookback, but don't have that when starting
from a snapshot.
2021-06-22 15:06:30 -07:00
Steven Allen
0b06de2bd3 fix(lotus-sim): unembed Node from Simulation
I wanted to expose the node's _fields_, but this also exposed the
methods. That got rather confusing.

(probably could have used a new type, but eh)

foo
2021-06-22 14:53:06 -07:00
Steven Allen
b7c36bc02c fix(lotus-sim): make NewNode take a repo
This is primarily for testing, so we can just pass an in-memory repo.
2021-06-21 11:32:19 -07:00
Steven Allen
c532b1b819 fix(lotus-sim): remove unused field and function 2021-06-21 11:25:40 -07:00
Steven Allen
80eba1069a feat(lotus-sim): NewNode to construct a node without a repo 2021-06-21 11:25:40 -07:00
Steven Allen
b5f9148748 build(lotus-sim): add a makefile target 2021-06-21 11:25:40 -07:00
Steven Allen
ffb63a93ff fix(lotus-sim): make 'fund' easier to understand 2021-06-21 11:25:40 -07:00
Steven Allen
d6abcff63c
fix(lotus-sim): apply code review from magik6k
Co-authored-by: Łukasz Magiera <magik6k@users.noreply.github.com>
2021-06-21 09:05:48 -07:00
Steven Allen
69a8a6bc0e fix(lotus-sim): lint 2021-06-18 15:44:37 -07:00
Jakub Sztandera
2fdf49e7da Add histogram and quantiles for message sizes
Resolves https://github.com/filecoin-project/lotus/issues/6513

Signed-off-by: Jakub Sztandera <kubuxu@protocol.ai>
2021-06-18 15:44:37 -07:00
Steven Allen
eb0a15faf0 fix(genesis): set initial balances to 0 2021-06-18 15:44:37 -07:00
Steven Allen
b30b5dd629 fix: move actors changes to template files 2021-06-18 15:44:37 -07:00
Steven Allen
eb2b706156 chore: fix lint errors in simulation 2021-06-18 15:44:37 -07:00
Steven Allen
2aedd82c72 fix(lotus-sim): correct window post batch sizes 2021-06-18 15:44:37 -07:00
Steven Allen
bc2698a988 fix(lotus-sim): simulate using realistic gas numbers
Previously, we assumed the network was "optimal". Now, we're using real
numbers.
2021-06-18 15:44:37 -07:00
Steven Allen
a26cd5a809 fix(lotus-sim): mark provecommit stage initialized 2021-06-18 15:44:37 -07:00
Steven Allen
f0d0b40bd3 fix(lotus-sim): debug log mock 2021-06-18 15:44:37 -07:00
Steven Allen
ce29a0ac17 fix(lotus-sim): initialize commit queue 2021-06-18 15:44:37 -07:00
Steven Allen
28e6fa5923 fix(lotus-sim): remove duplicate error handling 2021-06-18 15:44:37 -07:00
Steven Allen
ec5fab09a1 feat(lotus-sim): log failing proofs 2021-06-18 15:44:37 -07:00
Steven Allen
e41f0842b0 fix(lotus-sim): load prove-commits
(regression from refactor)
2021-06-18 15:44:37 -07:00
Steven Allen
af33d69357 fix(lotus-sim): don't close node on list 2021-06-18 15:44:37 -07:00
Steven Allen
885062f712 fix(lotus-sim): fix info state imports 2021-06-18 15:44:37 -07:00
Steven Allen
73ae1924bc feat(lotus-sim): state size command 2021-06-18 15:44:37 -07:00
Steven Allen
22267eb45d feat(lotus-sim): split info command file 2021-06-18 15:44:36 -07:00
Steven Allen
5d7b7ce5c1 feat(lotus-sim): allow profile info 2021-06-18 15:44:36 -07:00
Steven Allen
95cf577447 fix(lotus-sim): really cancel walk immediately 2021-06-18 15:44:36 -07:00
Steven Allen
8fffaa5c47 fix(lotus-sim): average over 2 days
There's too much noise per day.
2021-06-18 15:44:36 -07:00
Steven Allen
63178ce982 feat(lotus-sim): daily capacity growth 2021-06-18 15:44:36 -07:00
Steven Allen
0af7dcdedb fix(lotus-sim): rename power to capacity 2021-06-18 15:44:36 -07:00
Steven Allen
f9d2a23132 fix(lotus-sim): correctly handle cancellation in walk
1. Select order is not guaranteed, always check if the context has been
canceled explicitly.
2. Never close a work channel unless we're actually done. This can yield
out-of-order results due to buffering.
2021-06-18 15:44:36 -07:00
Steven Allen
ec3f969e9a feat(lotus-sim): allow walking back past the start 2021-06-18 15:44:36 -07:00
Steven Allen
f6043a0250 feat(lotus-sim): measure daily power growth 2021-06-18 15:44:36 -07:00
Steven Allen
52261fb814 refactor(lotus-sim): enterprise grade
While the previous version "worked", this version nicely separates out
the state for the separate stages. Hopefully, we'll be able to use this
to build different pipelines with different configs.
2021-06-18 15:44:36 -07:00
Steven Allen
8a215df46b fix(statetree): make StateTree.ForEach take layers into account
This likely isn't used anywhere, but this _should_ take layers into
account (and I kind of just assumed it did).
2021-06-18 15:44:36 -07:00
Jakub Sztandera
7dd58efb84 Add quantiles and histogram
Signed-off-by: Jakub Sztandera <kubuxu@protocol.ai>
2021-06-18 15:44:36 -07:00
Steven Allen
2721279e87 fix(lotus-sim): describe info commands 2021-06-18 15:44:36 -07:00
Steven Allen
985994cc0f feat(lotus-sim): add command for analyzing post stats over time 2021-06-18 15:44:36 -07:00
Steven Allen
707b3bf08a fix(lotus-sim): refuse to start simulation with no miners 2021-06-18 15:44:36 -07:00
Steven Allen
d551f2b4bd feat(lotus-sim): print duration info in days 2021-06-18 15:44:36 -07:00
Steven Allen
1df5445ed2 feat(lotus-sim): make walk parallel 2021-06-18 15:44:36 -07:00
Steven Allen
fbaffe86da fix(lotus-sim): return error from walk 2021-06-18 15:44:36 -07:00
Steven Allen
500fae6a52 fix(lotus-sim): less indentation in info 2021-06-18 15:44:36 -07:00
Steven Allen
ab59474c4c fix(lotus-sim): count single prove-commits when computing efficiency 2021-06-18 15:44:36 -07:00
Jakub Sztandera
3d3c26fa0c Add lookback limit
Signed-off-by: Jakub Sztandera <kubuxu@protocol.ai>
2021-06-18 15:44:36 -07:00
Steven Allen
68593ce995 fix(lotus-sim): obey context in walk 2021-06-18 15:44:36 -07:00
Jakub Sztandera
ca9eadd7c7 Add gas info command
Signed-off-by: Jakub Sztandera <kubuxu@protocol.ai>
2021-06-18 15:44:36 -07:00
Steven Allen
16449007ab fix(lotus-sim): fix funding error check 2021-06-18 15:44:35 -07:00
Steven Allen
8d734d81d9 fix(lotus-sim): log failed pre-commits and continue 2021-06-18 15:44:35 -07:00