Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

1279 gossiplb inform and decide have bugs - release branch #1349

Merged
Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
28 commits
Select commit Hold shift + click to select a range
2ebc0a9
#1279: gossiplb: Use best of multiple trials
nlslatt Feb 8, 2021
dccd64d
#1279: gossiplb: select unique nodes
nlslatt Feb 22, 2021
1493fcb
#1279: gossiplb: pass load instead of rank
nlslatt Feb 22, 2021
3c79020
#1279: gossiplb: init variable
lifflander Feb 8, 2021
c6b1674
#1279: gossiplb: improve debugging output
nlslatt Mar 2, 2021
ebb8cdd
#1279: gossiplb: corrected async informs
nlslatt Mar 2, 2021
0d75596
#1279: gossiplb: added sync informs that match lbaf
nlslatt Mar 2, 2021
f4759d6
#1279: gossiplb: cleanup
nlslatt Mar 2, 2021
773970a
#1279: gossiplb: order objects by object id
nlslatt Mar 2, 2021
28a07fe
#1279: gossiplb: add deterministic mode
nlslatt Mar 4, 2021
33d6f9f
#1279: gossiplb: fix loads that ended up in microseconds
nlslatt Mar 5, 2021
fa3332a
#1279: gossiplb: add object ordering options
nlslatt Mar 5, 2021
e699d07
#1279: gossiplb: cleanup
nlslatt Mar 5, 2021
c4539e2
#1279: gossiplb: increase precision of imbalance
nlslatt Mar 6, 2021
4d9f0af
#1279: gossiplb: add options for cmf
nlslatt Mar 6, 2021
30be5fa
#1279: gossiplb: tune print verbosity
nlslatt Mar 8, 2021
14b756a
#1279: gossiplb: clean up redundant code
nlslatt Mar 8, 2021
3b362f7
#1279: gossiplb: minimize reductions when not debugging
nlslatt Mar 8, 2021
629a6bd
#1279: gossiplb: fix bug in cmf computation
nlslatt Mar 15, 2021
d93455f
#1279: gossiplb: change default options
nlslatt Mar 15, 2021
30bad34
#1279: gossiplb: fix missing allowed key
nlslatt Mar 16, 2021
ba5e59c
#1279: gossiplb: prevent informs from being received early
nlslatt Mar 17, 2021
d73d9d3
#1279: gossiplb: roll back to best iter
nlslatt Mar 17, 2021
aeef3ca
#1279: gossiplb: add option to target long pole load instead of avg
nlslatt Mar 17, 2021
11c164f
#1279: gossiplb: name barrier to prevent hang
nlslatt Mar 23, 2021
a102728
#1279: gossiplb: cleanup
nlslatt Mar 26, 2021
f1821c0
#1279: gossiplb: document lb args
nlslatt Mar 26, 2021
db3cf59
#1279: gossiplb: clean up sync round number
nlslatt Mar 26, 2021
File filter

Filter by extension

Filter by extension

Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
50 changes: 50 additions & 0 deletions src/vt/vrt/collection/balance/gossiplb/gossip_msg.h
Original file line number Diff line number Diff line change
Expand Up @@ -81,6 +81,30 @@ struct GossipMsg : vt::Message {
NodeLoadType node_load_ = {};
};

struct GossipMsgAsync : GossipMsg {
using MessageParentType = GossipMsg;

GossipMsgAsync() = default;
GossipMsgAsync(
NodeType in_from_node, NodeLoadType const& in_node_load, int round
)
: GossipMsg(in_from_node, in_node_load), round_(round)
{ }

uint8_t getRound() const {
return round_;
}

template <typename SerializerT>
void serialize(SerializerT& s) {
MessageParentType::serialize(s);
s | round_;
}

private:
int round_;
};

struct LazyMigrationMsg : vt::Message {
using ObjsType = std::unordered_map<lb::BaseLB::ObjIDType, lb::BaseLB::LoadType>;

Expand All @@ -106,6 +130,32 @@ struct LazyMigrationMsg : vt::Message {
ObjsType objs_ = {};
};

struct RejectionStats {
RejectionStats() = default;
RejectionStats(int n_rejected, int n_transfers)
: n_rejected_(n_rejected), n_transfers_(n_transfers) { }

friend RejectionStats operator+(RejectionStats a1, RejectionStats const& a2) {
a1.n_rejected_ += a2.n_rejected_;
a1.n_transfers_ += a2.n_transfers_;

return a1;
}

int n_rejected_ = 0;
int n_transfers_ = 0;
};

struct GossipRejectionStatsMsg : collective::ReduceTMsg<RejectionStats> {
GossipRejectionStatsMsg() = default;
GossipRejectionStatsMsg(int n_rejected, int n_transfers)
: ReduceTMsg<RejectionStats>(RejectionStats(n_rejected, n_transfers))
{ }
GossipRejectionStatsMsg(RejectionStats&& rs)
: ReduceTMsg<RejectionStats>(std::move(rs))
{ }
};

}}}} /* end namespace vt::vrt::collection::balance */

#endif /*INCLUDED_VT_VRT_COLLECTION_BALANCE_GOSSIPLB_GOSSIP_MSG_H*/
Loading