Skip to content

Distinguishing between replica rank and group rank across the project (#181) #187

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.

Sign up for GitHub

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

Merged
merged 2 commits into from
May 6, 2025
Merged

Distinguishing between replica rank and group rank across the project (#181) #187

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
9ab316d
Distinguishing between replica rank and group rank across the project…
WarrenZhu050413 May 3, 2025
840cd31
lint
d4l3k May 5, 2025
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
10 changes: 5 additions & 5 deletions proto/torchft.proto
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -73,7 +73,7 @@ service LighthouseService {
}

message ManagerQuorumRequest {
int64 rank = 1;
int64 group_rank = 1;
int64 step = 2;
string checkpoint_metadata = 3;
bool shrink_only = 4;
Expand All @@ -84,12 +84,12 @@ message ManagerQuorumRequest {
message ManagerQuorumResponse {
int64 quorum_id = 1;
string recover_src_manager_address = 2;
optional int64 recover_src_rank = 3;
repeated int64 recover_dst_ranks = 4;
optional int64 recover_src_replica_rank = 3;
repeated int64 recover_dst_replica_ranks = 4;
string store_address = 5;
// These are information for the replicas which are at the max step.
int64 max_step = 6;
optional int64 max_rank = 7;
optional int64 max_replica_rank = 7;
int64 max_world_size = 8;
// These are information for all replicas including behind replicas.
int64 replica_rank = 9;
Expand All @@ -108,7 +108,7 @@ message CheckpointMetadataResponse {

message ShouldCommitRequest {
bool should_commit = 1;
int64 rank = 2;
int64 group_rank = 2;
int64 step = 3;
}
message ShouldCommitResponse {
Expand Down
26 changes: 13 additions & 13 deletions src/lib.rs
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -172,7 +172,7 @@ impl ManagerClient {
fn _quorum(
&self,
py: Python<'_>,
rank: i64,
group_rank: i64,
step: i64,
checkpoint_metadata: String,
shrink_only: bool,
Expand All @@ -182,7 +182,7 @@ impl ManagerClient {
) -> Result<QuorumResult, StatusError> {
py.allow_threads(move || {
let mut request = tonic::Request::new(ManagerQuorumRequest {
rank: rank,
group_rank: group_rank,
step: step,
checkpoint_metadata: checkpoint_metadata,
shrink_only: shrink_only,
Expand All @@ -201,11 +201,11 @@ impl ManagerClient {
replica_rank: resp.replica_rank,
replica_world_size: resp.replica_world_size,
recover_src_manager_address: resp.recover_src_manager_address,
recover_src_rank: resp.recover_src_rank,
recover_dst_ranks: resp.recover_dst_ranks,
recover_src_replica_rank: resp.recover_src_replica_rank,
recover_dst_replica_ranks: resp.recover_dst_replica_ranks,
store_address: resp.store_address,
max_step: resp.max_step,
max_rank: resp.max_rank,
max_replica_rank: resp.max_replica_rank,
max_world_size: resp.max_world_size,
heal: resp.heal,
})
Expand Down Expand Up @@ -250,14 +250,14 @@ impl ManagerClient {
fn should_commit(
&self,
py: Python<'_>,
rank: i64,
group_rank: i64,
step: i64,
should_commit: bool,
timeout: Duration,
) -> Result<bool, StatusError> {
py.allow_threads(move || {
let mut request = tonic::Request::new(ShouldCommitRequest {
rank: rank,
group_rank: group_rank,
step: step,
should_commit: should_commit,
});
Expand All @@ -281,11 +281,11 @@ struct QuorumResult {
replica_rank: i64,
replica_world_size: i64,
recover_src_manager_address: String,
recover_src_rank: Option<i64>,
recover_dst_ranks: Vec<i64>,
recover_src_replica_rank: Option<i64>,
recover_dst_replica_ranks: Vec<i64>,
store_address: String,
max_step: i64,
max_rank: Option<i64>,
max_replica_rank: Option<i64>,
max_world_size: i64,
heal: bool,
}
Expand All @@ -299,11 +299,11 @@ impl QuorumResult {
replica_rank: 0,
replica_world_size: 1,
recover_src_manager_address: "".to_string(),
recover_src_rank: None,
recover_dst_ranks: Vec::new(),
recover_src_replica_rank: None,
recover_dst_replica_ranks: Vec::new(),
store_address: "".to_string(),
max_step: 0,
max_rank: None,
max_replica_rank: None,
max_world_size: 1,
heal: false,
}
Expand Down
Loading