Skip to content

Commit 5209952

Browse files
committed
feat(transfer_engine): improve topology logging messages
- Update auto-discovery completion log message for clarity - Add topology details logging after discovery - Log custom topology details during transport installation Signed-off-by: staryxchen <[email protected]>
1 parent d41df94 commit 5209952

File tree

2 files changed

+6
-5
lines changed

2 files changed

+6
-5
lines changed

mooncake-transfer-engine/src/topology.cpp

Lines changed: 3 additions & 4 deletions
Original file line numberDiff line numberDiff line change
@@ -203,10 +203,9 @@ static std::vector<TopologyEntry> discoverCudaTopology(
203203
avail_hca.push_back(hca.name);
204204
}
205205
}
206-
topology.push_back(
207-
TopologyEntry{.name = "cuda:" + std::to_string(i),
208-
.preferred_hca = std::move(preferred_hca),
209-
.avail_hca = std::move(avail_hca)});
206+
topology.push_back(TopologyEntry{.name = "cuda:" + std::to_string(i),
207+
.preferred_hca = preferred_hca,
208+
.avail_hca = avail_hca});
210209
topology.push_back(
211210
TopologyEntry{.name = "gpu:" + std::to_string(i),
212211
.preferred_hca = std::move(preferred_hca),

mooncake-transfer-engine/src/transfer_engine.cpp

Lines changed: 3 additions & 1 deletion
Original file line numberDiff line numberDiff line change
@@ -150,8 +150,9 @@ int TransferEngine::init(const std::string &metadata_conn_string,
150150
} else {
151151
local_topology_->discover(filter_);
152152
}
153-
LOG(INFO) << "Topology discovery complete. Found "
153+
LOG(INFO) << "Auto-discovering topology complete: Found "
154154
<< local_topology_->getHcaList().size() << " HCAs.";
155+
LOG(INFO) << "Topology:\n" << local_topology_->toString();
155156

156157
#ifdef USE_MNNVL
157158
if (local_topology_->getHcaList().size() > 0 &&
@@ -215,6 +216,7 @@ Transport *TransferEngine::installTransport(const std::string &proto,
215216
LOG(ERROR) << "Failed to parse NIC priority matrix";
216217
return nullptr;
217218
}
219+
LOG(INFO) << "Install transport " << proto << " with custom topology:\n" << local_topology_->toString();
218220
}
219221

220222
transport = multi_transports_->installTransport(proto, local_topology_);

0 commit comments

Comments
 (0)