Commit 02761ab1 authored by zhuwenwen's avatar zhuwenwen
Browse files

fix pp1 rank error

parent 17b375d6
......@@ -420,10 +420,11 @@ class P2pNcclConnector(KVConnectorBase_V1):
if (self.multiple_machines):
ip_second = self.get_ip_value(ip)
if (self.pp_size == 1):
self.p2p_nccl_engine.send_tensor(request_id + "#" + layer_name,
kv_cache, remote_address)
self.p2p_nccl_engine.send_tensor(request_id + "#" + layer_name,
kv_cache, str(ip_second) + ":" + str(port + self._rank + 8))
if self._rank < 8:
self.p2p_nccl_engine.send_tensor(request_id + "#" + layer_name,
kv_cache, remote_address)
self.p2p_nccl_engine.send_tensor(request_id + "#" + layer_name,
kv_cache, str(ip_second) + ":" + str(port + self._rank + 8))
elif (self.pp_size == 2):
if (pp_rank == 0):
self.p2p_nccl_engine.send_tensor(request_id + "#" + layer_name,
......
Markdown is supported
0% or .
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment