Merge branch 'vijay/ordering_perf_fix' into 'main'

reordering perf fix

See merge request ADLR/megatron-lm!154

megatron/checkpointing.py

@@ -101,7 +101,7 @@ def save_checkpoint(iteration, model, optimizer, lr_scheduler):
         # Arguments, iteration, and model.
         state_dict = {}
         state_dict['args'] = args
-        state_dict['checkpoint_version'] = 1.0
+        state_dict['checkpoint_version'] = 2.0
         state_dict['iteration'] = iteration
         state_dict['model'] = model.state_dict_for_save_checkpoint()
 

megatron/model/transformer.py

@@ -172,17 +172,30 @@ class ParallelSelfAttention(MegatronModule):
             init_method=output_layer_init_method,
             skip_bias_add=True)
 
-    def _transpose_last_dim(self, mixed_layer):
-        """[s, b, 3 * hp] -->(view) [s, b, 3, hp] -->(tranpose)
-        [s, b, hp, 3] -->(view) [s, b, 3 * hp] """
-
+    def _transpose_last_dim(self, mixed_layer, num_splits, num_splits_first):
         input_shape = mixed_layer.size();
-        last_dim = input_shape[-1]
-        assert last_dim % 3 == 0, "expected QKV dimension"
-        last_dim_split = last_dim // 3
+        if num_splits_first:
+            """[s, b, num_splits * np * hn] 
+            -->(view) [s, b, num_splits, np, hn] 
+            -->(tranpose) [s, b, np, num_splits, hn] 
+            -->(view) [s, b, np * num_splits * hn] """
 
             intermediate_shape = input_shape[:-1] +\
-            (3, last_dim_split)
+                (num_splits, self.num_attention_heads_per_partition,
+                 self.hidden_size_per_attention_head)
+
+            mixed_layer = mixed_layer.view(*intermediate_shape)
+            mixed_layer = mixed_layer.transpose(-2, -3).contiguous()
+        else:
+            """[s, b, np * hn * num_splits] 
+            -->(view) [s, b, np, hn, num_splits] 
+            -->(tranpose) [s, b, np, num_splits, hn] 
+            -->(view) [s, b, np * num_splits * hn] """
+
+            intermediate_shape = input_shape[:-1] +\
+                (self.num_attention_heads_per_partition,
+                 self.hidden_size_per_attention_head, num_splits)
+
             mixed_layer = mixed_layer.view(*intermediate_shape)
             mixed_layer = mixed_layer.transpose(-1, -2).contiguous()
         mixed_layer = mixed_layer.view(*input_shape)
@@ -197,25 +210,28 @@ class ParallelSelfAttention(MegatronModule):
         # Query, Key, and Value
         # =====================
 
-        # Attention heads [sq, b, hp] --> [sq, b, hp * 3]
+        # Attention heads [sq, b, h] --> [sq, b, (np * 3 * hn)]
         mixed_x_layer, _ = self.query_key_value(hidden_states)
 
         checkpoint_version = get_checkpoint_version()
-        if checkpoint_version is not None and \
-           checkpoint_version == 0:
-            # [sq, b, 3 * hp] --> [sq, b, hp * 3]
-            mixed_x_layer = self._transpose_last_dim(mixed_x_layer)
-
-        # [sq, b, hp * 3] --> [sq, b, np, hn, 3]  
+        if checkpoint_version is not None:
+           if checkpoint_version == 0:
+               # [s, b, (3 * np * hn)] --> [s, b, (np * 3 * hn)]
+               mixed_x_layer = self._transpose_last_dim(mixed_x_layer, 3, True)
+           elif checkpoint_version == 1.0:
+               # [s, b, (np * hn * 3)] --> [s, b, (np * 3 * hn)]
+               mixed_x_layer = self._transpose_last_dim(mixed_x_layer, 3, False)
+
+        # [sq, b, (np * 3 * hn)] --> [sq, b, np, 3 * hn]
         new_tensor_shape = mixed_x_layer.size()[:-1] + \
             (self.num_attention_heads_per_partition,
-             self.hidden_size_per_attention_head, 3)
+             3 * self.hidden_size_per_attention_head)
         mixed_x_layer = mixed_x_layer.view(*new_tensor_shape)
 
-        # [sq, b, np, hn, 3] --> 3 [sq, b, np, hn]
-        query_layer = mixed_x_layer[:,:,:,:,0]
-        key_layer = mixed_x_layer[:,:,:,:,1]
-        value_layer = mixed_x_layer[:,:,:,:,2]
+        # [sq, b, np, 3 * hn] --> 3 [sq, b, np, hn]
+        (query_layer,
+         key_layer,
+         value_layer) = mpu.split_tensor_along_last_dim(mixed_x_layer, 3)
 
         # ==================================
         # Adjust key and value for inference