XPK Start: Tue Apr 21 02:22:49 UTC 2026 2026-04-21 02:22:53.377893: E external/local_xla/xla/stream_executor/cuda/cuda_fft.cc:467] Unable to register cuFFT factory: Attempting to register factory for plugin cuFFT when one has already been registered WARNING: All log messages before absl::InitializeLog() is called are written to STDERR E0000 00:00:1776738173.390721 10 cuda_dnn.cc:8579] Unable to register cuDNN factory: Attempting to register factory for plugin cuDNN when one has already been registered E0000 00:00:1776738173.394462 10 cuda_blas.cc:1407] Unable to register cuBLAS factory: Attempting to register factory for plugin cuBLAS when one has already been registered W0000 00:00:1776738173.405813 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. W0000 00:00:1776738173.405840 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. W0000 00:00:1776738173.405843 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. W0000 00:00:1776738173.405845 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. 2026-04-21 02:23:12.487114: E external/local_xla/xla/stream_executor/cuda/cuda_platform.cc:51] failed call to cuInit: INTERNAL: CUDA error: Failed call to cuInit: UNKNOWN ERROR (303) I0421 02:23:12.999870 139640734226240 max_utils.py:273] Attempting to initialize the jax distributed system... INFO:2026-04-21 02:23:22,043:jax._src.distributed:140: Starting JAX distributed service on [::]:8482 I0421 02:23:22.043308 139640734226240 distributed.py:140] Starting JAX distributed service on [::]:8482 INFO:2026-04-21 02:23:22,045:jax._src.distributed:157: Connecting to JAX distributed service on mt-16-zarr3-false-sv-ddy9s-slice-job-0-0.mt-16-zarr3-false-sv-ddy9s:8482 I0421 02:23:22.045587 139640734226240 distributed.py:157] Connecting to JAX distributed service on mt-16-zarr3-false-sv-ddy9s-slice-job-0-0.mt-16-zarr3-false-sv-ddy9s:8482 I0421 02:23:23.395179 139640734226240 max_utils.py:284] Jax distributed system initialized! I0421 02:23:30.550477 139640734226240 max_utils.py:800] System Information: Jax Version: 0.8.1 I0421 02:23:30.550581 139640734226240 max_utils.py:801] System Information: Jaxlib Version: 0.8.1 I0421 02:23:30.550622 139640734226240 max_utils.py:802] System Information: Jax Backend: PJRT C API TFRT TPU v6 lite Built on Nov 12 2025 14:16:36 (1762985796) cl/831091709 I0421 02:23:30.550656 139640734226240 train_utils.py:347] WARNING: Sequence packing is essentially ignored for synthetic data. Please use a real dataset to use sequence packing. I0421 02:23:31.168697 139640734226240 maxtext_utils.py:1398] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1) I0421 02:23:31.168992 139640734226240 checkpointing.py:677] Setting up checkpoint logger... I0421 02:23:31.169048 139640734226240 checkpointing.py:233] Creating checkpoint manager with ocdbt=True and zarr3=False I0421 02:23:31.169092 139640734226240 pytree_checkpoint_handler.py:589] save_device_host_concurrent_bytes=None I0421 02:23:31.169451 139640734226240 base_pytree_checkpoint_handler.py:415] Created BasePyTreeCheckpointHandler: use_ocdbt=True, use_zarr3=False, pytree_metadata_options=PyTreeMetadataOptions(support_rich_types=False), array_metadata_store=<orbax.checkpoint._src.metadata.array_metadata_store.Store object at 0x7f0018779130>, enable_pinned_host_transfer=False, save_concurrent_bytes: 96000000000 (89.4 GiB), restore_concurrent_bytes: 96000000000 (89.4 GiB) I0421 02:23:34.301450 139640734226240 checkpointing.py:265] Enabling policy for fixed interval checkpointing. I0421 02:23:34.301793 139640734226240 checkpoint_manager.py:709] [process=1][thread=MainThread] CheckpointManager init: checkpointers=None, item_names=('items',), item_handlers={'items': <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7eebd81c1880>}, handler_registry=None I0421 02:23:34.302036 139640734226240 composite_checkpoint_handler.py:237] Deferred registration for item: "items". Adding handler `<orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7eebd81c1880>` for item "items" and save args `<class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeSaveArgs'>` and restore args `<class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>` to `_handler_registry`. I0421 02:23:34.302088 139640734226240 composite_checkpoint_handler.py:237] Deferred registration for item: "metrics". Adding handler `<orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7eebc07735f0>` for item "metrics" and save args `<class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>` and restore args `<class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>` to `_handler_registry`. I0421 02:23:34.302128 139640734226240 composite_checkpoint_handler.py:505] Initialized registry DefaultCheckpointHandlerRegistry({('items', <class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeSaveArgs'>): <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7eebd81c1880>, ('items', <class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>): <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7eebd81c1880>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7eebc07735f0>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7eebc07735f0>}). I0421 02:23:34.302629 139640734226240 abstract_checkpointer.py:35] orbax-checkpoint version: 0.11.33 I0421 02:23:34.302708 139640734226240 async_checkpointer.py:177] [process=1][thread=MainThread] Using barrier_sync_fn: <function get_barrier_sync_fn.<locals>._fn at 0x7eebc01a0040> timeout: 600 secs and primary_host=0 for async checkpoint writes I0421 02:23:36.643530 139640734226240 checkpoint_manager.py:1818] Found 0 checkpoint steps in gs://lance-maxtext/linen_ckpt_xpk_test_pipeline_scan_nnx_20260421_005327/linen_xpk_test_pipeline_scan_nnx_20260421_005327_16_zarr3_false/checkpoints I0421 02:23:37.080663 139640734226240 checkpoint_manager.py:929] [process=1][thread=MainThread] CheckpointManager created, primary_host=0, CheckpointManagerOptions=CheckpointManagerOptions(save_interval_steps=1, max_to_keep=None, keep_time_interval=None, keep_period=None, should_keep_fn=None, best_fn=None, best_mode='max', keep_checkpoints_without_metrics=True, step_prefix=None, step_format_fixed_length=None, step_name_format=None, create=True, cleanup_tmp_directories=False, save_on_steps=frozenset(), single_host_load_and_broadcast=False, todelete_subdir=None, todelete_full_path=None, enable_background_delete=False, read_only=False, enable_async_checkpointing=True, async_options=None, multiprocessing_options=MultiprocessingOptions(primary_host=0, active_processes=None, barrier_sync_key_prefix=None), should_save_fn=None, file_options=FileOptions(path_permission_mode=None), save_root_metadata=True, temporary_path_class=None, save_decision_policy=FixedIntervalPolicy(interval=5), preservation_policy=LatestN(n=None), prevent_write_metrics=False, enable_should_save_is_saving_in_progress_check=True, enable_per_process_directory_creation=False, lightweight_initialize=False), root_directory=gs://lance-maxtext/linen_ckpt_xpk_test_pipeline_scan_nnx_20260421_005327/linen_xpk_test_pipeline_scan_nnx_20260421_005327_16_zarr3_false/checkpoints: <orbax.checkpoint.checkpoint_manager.CheckpointManager object at 0x7eebc076d3d0> I0421 02:23:37.080835 139640734226240 checkpointing.py:301] Checkpoint manager created! I0421 02:23:37.988302 139640734226240 nnx_wrappers.py:453] Unknown Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed'). I0421 02:23:37.988418 139640734226240 nnx_wrappers.py:453] Unknown Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0421 02:23:38.357792 139640734226240 attentions.py:1084] attentions/inputs_q Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0421 02:23:38.357888 139640734226240 attentions.py:1084] attentions/inputs_q Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0421 02:23:38.373735 139640734226240 attentions.py:1085] attentions/inputs_kv Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0421 02:23:38.373796 139640734226240 attentions.py:1085] attentions/inputs_kv Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0421 02:23:38.396415 139640734226240 attentions.py:1150] attentions/query Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0421 02:23:38.396481 139640734226240 attentions.py:1150] attentions/query Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0421 02:23:38.412349 139640734226240 attentions.py:1151] attentions/key Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0421 02:23:38.412421 139640734226240 attentions.py:1151] attentions/key Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0421 02:23:38.431086 139640734226240 attentions.py:1152] attentions/value Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0421 02:23:38.431192 139640734226240 attentions.py:1152] attentions/value Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0421 02:23:38.458885 139640734226240 attentions.py:1193] attentions/out Logical: bfloat16[32,2048,16,128].................................... ('activation_batch', 'activation_attn_length', 'activation_heads', 'activation_kv'). I0421 02:23:38.458961 139640734226240 attentions.py:1193] attentions/out Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0421 02:23:38.478501 139640734226240 linears.py:525] linears/x Logical: bfloat16[32,2048,7168]...................................... ('activation_batch', 'activation_length', 'activation_mlp'). I0421 02:23:38.478565 139640734226240 linears.py:525] linears/x Physical: bfloat16[32,2048,7168]...................................... ('fsdp', None, None). I0421 02:23:38.669096 139640734226240 checkpointing.py:577] checkpoint manager exists so trying to load this run's existing checkpoint I0421 02:23:38.669205 139640734226240 checkpointing.py:665] No existing checkpoints found, not restoring checkpoint. fsdp: 32 I0421 02:23:40.108535 139640734226240 maxtext_utils.py:1501] params/params/decoder/decoder_norm/scale Shape: float32[2048] Logical: PartitionSpec('norm',) Physical: (None,) I0421 02:23:40.108672 139640734226240 maxtext_utils.py:1501] params/params/decoder/layers/mlp/wi_0/kernel Shape: float32[2048,16,7168] Logical: PartitionSpec('embed', 'layers', 'mlp') Physical: ('fsdp', None, None) I0421 02:23:40.108726 139640734226240 maxtext_utils.py:1501] params/params/decoder/layers/mlp/wi_1/kernel Shape: float32[2048,16,7168] Logical: PartitionSpec('embed', 'layers', 'mlp') Physical: ('fsdp', None, None) I0421 02:23:40.108787 139640734226240 maxtext_utils.py:1501] params/params/decoder/layers/mlp/wo/kernel Shape: float32[7168,16,2048] Logical: PartitionSpec('mlp', 'layers', 'embed') Physical: (None, None, 'fsdp') I0421 02:23:40.108841 139640734226240 maxtext_utils.py:1501] params/params/decoder/layers/post_self_attention_layer_norm/scale Shape: float32[2048,16] Logical: PartitionSpec('norm', 'layers') Physical: (None, None) I0421 02:23:40.108881 139640734226240 maxtext_utils.py:1501] params/params/decoder/layers/pre_self_attention_layer_norm/scale Shape: float32[2048,16] Logical: PartitionSpec('norm', 'layers') Physical: (None, None) I0421 02:23:40.108935 139640734226240 maxtext_utils.py:1501] params/params/decoder/layers/self_attention/key/kernel Shape: float32[2048,16,16,128] Logical: PartitionSpec('embed', 'layers', 'kv_heads', 'kv_head_dim') Physical: ('fsdp', None, None, None) I0421 02:23:40.108990 139640734226240 maxtext_utils.py:1501] params/params/decoder/layers/self_attention/out/kernel Shape: float32[16,16,128,2048] Logical: PartitionSpec('heads', 'layers', 'kv', 'embed') Physical: (None, None, None, 'fsdp') I0421 02:23:40.109034 139640734226240 maxtext_utils.py:1501] params/params/decoder/layers/self_attention/query/kernel Shape: float32[2048,16,16,128] Logical: PartitionSpec('embed', 'layers', 'q_heads', 'kv') Physical: ('fsdp', None, None, None) I0421 02:23:40.109071 139640734226240 maxtext_utils.py:1501] params/params/decoder/layers/self_attention/value/kernel Shape: float32[2048,16,16,128] Logical: PartitionSpec('embed', 'layers', 'kv_heads', 'kv_head_dim') Physical: ('fsdp', None, None, None) I0421 02:23:40.109121 139640734226240 maxtext_utils.py:1501] params/params/decoder/logits_dense/kernel Shape: float32[2048,32000] Logical: PartitionSpec('embed', 'vocab') Physical: ('fsdp', None) I0421 02:23:40.109177 139640734226240 maxtext_utils.py:1501] params/params/token_embedder/embedding Shape: float32[32000,2048] Logical: PartitionSpec('vocab', 'embed') Physical: (None, 'fsdp') I0421 02:23:40.617881 139640734226240 train.py:156] train/xent Logical: float32[32,2048]............................................ ('activation_embed_and_logits_batch', 'activation_length'). I0421 02:23:40.617983 139640734226240 train.py:156] train/xent Physical: float32[32,2048]............................................ ('fsdp', None). I0421 02:23:40.633060 139640734226240 train.py:163] train/z_loss Logical: float32[32,2048]............................................ ('activation_embed_and_logits_batch', 'activation_length'). I0421 02:23:40.633121 139640734226240 train.py:163] train/z_loss Physical: float32[32,2048]............................................ ('fsdp', None). I0421 02:23:51.282921 139640734226240 max_utils.py:791] Total memory size: 1.5 GB, Output size: 0.4 GB, Temp size: 1.1 GB, Argument size: 0.4 GB, Host temp size: 0.0 GB. I0421 02:23:51.283791 139640734226240 metric_logger.py:289] number parameters: 1.104 billion I0421 02:24:02.457964 139640734226240 checkpointing.py:772] Waiting for step 0 to finish before checkpoint... I0421 02:24:02.611490 139640734226240 checkpointing.py:776] Waited 0.15351438522338867 seconds for step 0 to finish before starting checkpointing. I0421 02:24:02.614198 139640734226240 checkpoint_manager.py:2013] [process=1][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0421 02:24:02.615856 139640734226240 checkpoint_manager.py:1518] [process=1] Saving checkpoint at step 0 I0421 02:24:02.618768 139640734226240 async_checkpointer.py:452] [process=1] Started async saving checkpoint to gs://lance-maxtext/linen_ckpt_xpk_test_pipeline_scan_nnx_20260421_005327/linen_xpk_test_pipeline_scan_nnx_20260421_005327_16_zarr3_false/checkpoints/0. I0421 02:24:04.039124 139640734226240 signaling_client.py:364] Using JaxDistributedSignalingClient I0421 02:24:04.040060 139640734226240 jax_array_handlers.py:358] Scheduling D2H of 39 prioritized jax.Array. I0421 02:24:04.040115 139640734226240 replica_slices.py:424] Transferring arrays to host memory with options: use_replica_parallel=True, min_slice_bytes_for_replica_parallel=None, max_replicas_for_replica_parallel=None, enable_pinned_host_transfer=False I0421 02:24:04.292273 139640734226240 base_pytree_checkpoint_handler.py:153] [process=1][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.253228s I0421 02:24:04.292460 139640734226240 base_pytree_checkpoint_handler.py:129] [process=1] /jax/checkpoint/write/blocking_gbytes_per_sec: 3.010 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.5124797821044922 s) (per-host) I0421 02:24:04.292514 139640734226240 base_pytree_checkpoint_handler.py:737] [process=1][thread=MainThread] Initiated Pytree async_save. Time taken: 0.512543s (batch_requests_ready=0.255690s, total_serialization_initiated=0.256783s, others=0.000071s) I0421 02:24:04.292604 139640734226240 composite_checkpoint_handler.py:715] [process=1][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.516651s (all_items=0.000019s, per_item={'items': '0.00001907'}, temp_paths=0.516632) I0421 02:24:04.293358 139517977683712 async_checkpointer.py:79] [process=1][thread=async_save] Background save thread started. I0421 02:24:04.293526 139640734226240 async_checkpointer.py:561] Finished blocking save. Time taken: 1.677605s. Continuing background save to gs://lance-maxtext/linen_ckpt_xpk_test_pipeline_scan_nnx_20260421_005327/linen_xpk_test_pipeline_scan_nnx_20260421_005327_16_zarr3_false/checkpoints/0. I0421 02:24:04.758911 139640734226240 checkpoint_manager.py:1566] [process=1][thread=MainThread][step=0] Starting CheckpointManager Save Finalize thread=save_finalize I0421 02:24:04.759295 139514186016512 async_checkpointer.py:265] [process=1][thread=save_finalize] Waiting for background save thread=async_save. I0421 02:24:04.759487 139640734226240 standard_logger.py:34] {'step': 0, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_test_pipeline_scan_nnx_20260421_005327/linen_xpk_test_pipeline_scan_nnx_20260421_005327_16_zarr3_false/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776738242.6141813, 'wait_for_prev_duration_secs': 6.341934204101562e-05, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776738242.6158984, 'checkpointer_blocking_duration_secs': 1.6777002811431885, 'get_old_steps_start_time': 1776738244.2936184, 'get_old_steps_duration_secs': 2.2172927856445312e-05, 'checkpoint_manager_blocking_start_time': 1776738242.6121006, 'checkpoint_manager_blocking_duration_secs': 2.147345781326294} I0421 02:24:04.759594 139640734226240 checkpointing.py:408] Started an asynchronous checkpoint save for step 0 I0421 02:24:04.759644 139640734226240 max_utils.py:750] Memstats: After params initialized: I0421 02:24:04.759699 139640734226240 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_2(process=1,(2,0,0,0)) I0421 02:24:04.759733 139640734226240 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_3(process=1,(3,0,0,0)) I0421 02:24:04.759760 139640734226240 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_6(process=1,(2,1,0,0)) I0421 02:24:04.759784 139640734226240 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_7(process=1,(3,1,0,0)) I0421 02:24:05.028884 139640734226240 metric_logger.py:185] completed step: 0, seconds: 11.174, TFLOP/s/device: 1.216, Tokens/s/device: 183.282, total_weights: 65536, loss: 10.874 I0421 02:24:05.208652 139640734226240 metric_logger.py:185] completed step: 1, seconds: 2.570, TFLOP/s/device: 5.287, Tokens/s/device: 796.988, total_weights: 65536, loss: 10.874 I0421 02:24:05.600974 139640734226240 metric_logger.py:185] completed step: 2, seconds: 0.033, TFLOP/s/device: 416.068, Tokens/s/device: 62714.356, total_weights: 65536, loss: 10.560 I0421 02:24:05.748316 139640734226240 metric_logger.py:185] completed step: 3, seconds: 0.393, TFLOP/s/device: 34.586, Tokens/s/device: 5213.186, total_weights: 65536, loss: 9.980 I0421 02:24:05.753013 139640734226240 checkpointing.py:772] Waiting for step 5 to finish before checkpoint... I0421 02:24:06.043082 139640734226240 checkpointing.py:776] Waited 0.2900390625 seconds for step 5 to finish before starting checkpointing. I0421 02:24:06.045950 139640734226240 checkpoint_manager.py:2024] [process=1][thread=MainThread][step=0][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0421 02:24:08.189666 2642 google_auth_provider.cc:181] Running on GCE, using service account 562977990677-compute@developer.gserviceaccount.com I0421 02:24:10.535385 139517986076416 array_metadata_store.py:203] [process=1][thread=array_type_handler] Wrote 39 array_metadata.ArrayMetadata to gs://lance-maxtext/linen_ckpt_xpk_test_pipeline_scan_nnx_20260421_005327/linen_xpk_test_pipeline_scan_nnx_20260421_005327_16_zarr3_false/checkpoints/0/items/array_metadatas/process_1 I0421 02:24:46.050826 139517977683712 base_pytree_checkpoint_handler.py:129] [process=1] /jax/checkpoint/write/gbytes_per_sec: 37.369 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 42.270803451538086 s) (per-host) I0421 02:24:46.050954 139517977683712 async_checkpointer.py:90] [process=1][thread=async_save] 3 Handler Commit operations completed. Time taken: 41.757471s. I0421 02:24:54.486711 139517977683712 async_checkpointer.py:144] [process=1][thread=async_save] Background save thread done. Time taken: 50.193212s. I0421 02:24:54.487043 139514186016512 async_checkpointer.py:273] [process=1][thread=save_finalize] Done with waiting for background save thread=async_save. I0421 02:24:54.487190 139514186016512 async_checkpointer.py:283] [process=1][thread=save_finalize] No errors found in background save thread=async_save. I0421 02:24:54.487250 139514186016512 checkpoint_manager.py:2133] [process=1][thread=save_finalize][step=0] CheckpointManager Save Finalize is syncing with other hosts... I0421 02:24:54.488939 139514186016512 checkpoint_manager.py:2142] [process=1][thread=save_finalize][step=0] CheckpointManager Save Finalize is done on all hosts. I0421 02:24:54.489102 139640734226240 checkpoint_manager.py:2036] [process=1][thread=MainThread][step=0][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=0. W0421 02:24:54.489230 139640734226240 checkpoint_manager.py:1458] Waiting for previous save to complete took 48.443288 seconds. If this number is high, consider checkpointing less frequently. I0421 02:24:54.491296 139640734226240 checkpoint_manager.py:1518] [process=1] Saving checkpoint at step 5 I0421 02:24:54.495305 139640734226240 async_checkpointer.py:452] [process=1] Started async saving checkpoint to gs://lance-maxtext/linen_ckpt_xpk_test_pipeline_scan_nnx_20260421_005327/linen_xpk_test_pipeline_scan_nnx_20260421_005327_16_zarr3_false/checkpoints/5. I0421 02:24:55.460136 139640734226240 jax_array_handlers.py:358] Scheduling D2H of 39 prioritized jax.Array. I0421 02:24:55.460239 139640734226240 replica_slices.py:424] Transferring arrays to host memory with options: use_replica_parallel=True, min_slice_bytes_for_replica_parallel=None, max_replicas_for_replica_parallel=None, enable_pinned_host_transfer=False I0421 02:24:55.498973 139640734226240 base_pytree_checkpoint_handler.py:153] [process=1][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.040016s I0421 02:24:55.499148 139640734226240 base_pytree_checkpoint_handler.py:129] [process=1] /jax/checkpoint/write/blocking_gbytes_per_sec: 5.206 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.2962825298309326 s) (per-host) I0421 02:24:55.499202 139640734226240 base_pytree_checkpoint_handler.py:737] [process=1][thread=MainThread] Initiated Pytree async_save. Time taken: 0.296346s (batch_requests_ready=0.254417s, total_serialization_initiated=0.041858s, others=0.000071s) I0421 02:24:55.499288 139640734226240 composite_checkpoint_handler.py:715] [process=1][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.301229s (all_items=0.000016s, per_item={'items': '0.00001574'}, temp_paths=0.301213) I0421 02:24:55.500199 139517986076416 async_checkpointer.py:79] [process=1][thread=async_save] Background save thread started. I0421 02:24:55.500360 139640734226240 async_checkpointer.py:561] Finished blocking save. Time taken: 1.008991s. Continuing background save to gs://lance-maxtext/linen_ckpt_xpk_test_pipeline_scan_nnx_20260421_005327/linen_xpk_test_pipeline_scan_nnx_20260421_005327_16_zarr3_false/checkpoints/5. I0421 02:24:55.946175 139640734226240 checkpoint_manager.py:1566] [process=1][thread=MainThread][step=5] Starting CheckpointManager Save Finalize thread=save_finalize I0421 02:24:55.946583 139514186016512 async_checkpointer.py:265] [process=1][thread=save_finalize] Waiting for background save thread=async_save. I0421 02:24:55.946756 139640734226240 standard_logger.py:34] {'step': 5, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_test_pipeline_scan_nnx_20260421_005327/linen_xpk_test_pipeline_scan_nnx_20260421_005327_16_zarr3_false/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776738246.045917, 'wait_for_prev_duration_secs': 48.44328784942627, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776738294.4913354, 'checkpointer_blocking_duration_secs': 1.0091440677642822, 'get_old_steps_start_time': 1776738295.5004985, 'get_old_steps_duration_secs': 2.5033950805664062e-05, 'checkpoint_manager_blocking_start_time': 1776738246.0437467, 'checkpoint_manager_blocking_duration_secs': 49.90297245979309} I0421 02:24:55.946873 139640734226240 checkpointing.py:408] Started an asynchronous checkpoint save for step 5 I0421 02:24:55.947449 139640734226240 metric_logger.py:185] completed step: 4, seconds: 0.152, TFLOP/s/device: 89.120, Tokens/s/device: 13433.208, total_weights: 65536, loss: 9.460 I0421 02:24:55.953342 139640734226240 metric_logger.py:185] completed step: 5, seconds: 0.147, TFLOP/s/device: 92.246, Tokens/s/device: 13904.259, total_weights: 65536, loss: 8.959 I0421 02:25:15.825735 139640734226240 metric_logger.py:185] completed step: 6, seconds: 50.199, TFLOP/s/device: 0.271, Tokens/s/device: 40.797, total_weights: 65536, loss: 8.469 I0421 02:25:15.973240 139640734226240 metric_logger.py:185] completed step: 7, seconds: 19.726, TFLOP/s/device: 0.689, Tokens/s/device: 103.822, total_weights: 65536, loss: 8.003 I0421 02:25:16.120765 139640734226240 metric_logger.py:185] completed step: 8, seconds: 0.152, TFLOP/s/device: 89.261, Tokens/s/device: 13454.388, total_weights: 65536, loss: 7.572 I0421 02:25:16.267973 139640734226240 checkpointing.py:772] Waiting for step 9 to finish before checkpoint... I0421 02:25:16.268654 139640734226240 checkpointing.py:776] Waited 0.0007011890411376953 seconds for step 9 to finish before starting checkpointing. I0421 02:25:16.270715 139640734226240 checkpoint_manager.py:2024] [process=1][thread=MainThread][step=5][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0421 02:25:21.044074 139514211194624 array_metadata_store.py:203] [process=1][thread=array_type_handler] Wrote 39 array_metadata.ArrayMetadata to gs://lance-maxtext/linen_ckpt_xpk_test_pipeline_scan_nnx_20260421_005327/linen_xpk_test_pipeline_scan_nnx_20260421_005327_16_zarr3_false/checkpoints/5/items/array_metadatas/process_1 I0421 02:25:56.690418 139517986076416 base_pytree_checkpoint_handler.py:129] [process=1] /jax/checkpoint/write/gbytes_per_sec: 25.690 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 61.487502574920654 s) (per-host) I0421 02:25:56.690537 139517986076416 async_checkpointer.py:90] [process=1][thread=async_save] 3 Handler Commit operations completed. Time taken: 61.190227s. I0421 02:26:05.148391 139517986076416 async_checkpointer.py:144] [process=1][thread=async_save] Background save thread done. Time taken: 69.648066s. I0421 02:26:05.148690 139514186016512 async_checkpointer.py:273] [process=1][thread=save_finalize] Done with waiting for background save thread=async_save. I0421 02:26:05.148827 139514186016512 async_checkpointer.py:283] [process=1][thread=save_finalize] No errors found in background save thread=async_save. I0421 02:26:05.148875 139514186016512 checkpoint_manager.py:2133] [process=1][thread=save_finalize][step=5] CheckpointManager Save Finalize is syncing with other hosts... I0421 02:26:05.151226 139514186016512 checkpoint_manager.py:2142] [process=1][thread=save_finalize][step=5] CheckpointManager Save Finalize is done on all hosts. I0421 02:26:05.151431 139640734226240 checkpoint_manager.py:2036] [process=1][thread=MainThread][step=5][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=5. W0421 02:26:05.151567 139640734226240 checkpoint_manager.py:1458] Waiting for previous save to complete took 48.880863 seconds. If this number is high, consider checkpointing less frequently. I0421 02:26:05.153315 139640734226240 checkpoint_manager.py:1518] [process=1] Saving checkpoint at step 9 I0421 02:26:05.156828 139640734226240 async_checkpointer.py:452] [process=1] Started async saving checkpoint to gs://lance-maxtext/linen_ckpt_xpk_test_pipeline_scan_nnx_20260421_005327/linen_xpk_test_pipeline_scan_nnx_20260421_005327_16_zarr3_false/checkpoints/9. I0421 02:26:06.126791 139640734226240 jax_array_handlers.py:358] Scheduling D2H of 39 prioritized jax.Array. I0421 02:26:06.126885 139640734226240 replica_slices.py:424] Transferring arrays to host memory with options: use_replica_parallel=True, min_slice_bytes_for_replica_parallel=None, max_replicas_for_replica_parallel=None, enable_pinned_host_transfer=False I0421 02:26:06.167726 139640734226240 base_pytree_checkpoint_handler.py:153] [process=1][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.041912s I0421 02:26:06.167896 139640734226240 base_pytree_checkpoint_handler.py:129] [process=1] /jax/checkpoint/write/blocking_gbytes_per_sec: 5.263 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.293076753616333 s) (per-host) I0421 02:26:06.167943 139640734226240 base_pytree_checkpoint_handler.py:737] [process=1][thread=MainThread] Initiated Pytree async_save. Time taken: 0.293134s (batch_requests_ready=0.249150s, total_serialization_initiated=0.043921s, others=0.000063s) I0421 02:26:06.168030 139640734226240 composite_checkpoint_handler.py:715] [process=1][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.297289s (all_items=0.000011s, per_item={'items': '0.00001097'}, temp_paths=0.297278) I0421 02:26:06.168840 139514709174016 async_checkpointer.py:79] [process=1][thread=async_save] Background save thread started. I0421 02:26:06.168995 139640734226240 async_checkpointer.py:561] Finished blocking save. Time taken: 1.015614s. Continuing background save to gs://lance-maxtext/linen_ckpt_xpk_test_pipeline_scan_nnx_20260421_005327/linen_xpk_test_pipeline_scan_nnx_20260421_005327_16_zarr3_false/checkpoints/9. I0421 02:26:07.039093 139640734226240 checkpoint_manager.py:1566] [process=1][thread=MainThread][step=9] Starting CheckpointManager Save Finalize thread=save_finalize I0421 02:26:07.039486 139514186016512 async_checkpointer.py:265] [process=1][thread=save_finalize] Waiting for background save thread=async_save. I0421 02:26:07.039658 139640734226240 standard_logger.py:34] {'step': 9, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_test_pipeline_scan_nnx_20260421_005327/linen_xpk_test_pipeline_scan_nnx_20260421_005327_16_zarr3_false/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776738316.2706883, 'wait_for_prev_duration_secs': 48.880863428115845, 'time_between_consecutive_saves_sec': 21.781707763671875, 'checkpointer_blocking_start_time': 1776738365.1533546, 'checkpointer_blocking_duration_secs': 1.0157461166381836, 'get_old_steps_start_time': 1776738366.1691127, 'get_old_steps_duration_secs': 2.1457672119140625e-05, 'checkpoint_manager_blocking_start_time': 1776738316.268877, 'checkpoint_manager_blocking_duration_secs': 50.770747423172} I0421 02:26:07.039767 139640734226240 checkpointing.py:408] Started an asynchronous checkpoint save for step 9 I0421 02:26:07.039807 139640734226240 checkpoint_manager.py:2024] [process=1][thread=MainThread][step=9][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0421 02:26:13.496528 139517977683712 array_metadata_store.py:203] [process=1][thread=array_type_handler] Wrote 39 array_metadata.ArrayMetadata to gs://lance-maxtext/linen_ckpt_xpk_test_pipeline_scan_nnx_20260421_005327/linen_xpk_test_pipeline_scan_nnx_20260421_005327_16_zarr3_false/checkpoints/9/items/array_metadatas/process_1 I0421 02:26:48.799313 139514709174016 base_pytree_checkpoint_handler.py:129] [process=1] /jax/checkpoint/write/gbytes_per_sec: 36.799 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 42.92444825172424 s) (per-host) I0421 02:26:48.799441 139514709174016 async_checkpointer.py:90] [process=1][thread=async_save] 3 Handler Commit operations completed. Time taken: 42.630486s. I0421 02:26:57.569154 139514709174016 async_checkpointer.py:144] [process=1][thread=async_save] Background save thread done. Time taken: 51.400185s. I0421 02:26:57.569486 139514186016512 async_checkpointer.py:273] [process=1][thread=save_finalize] Done with waiting for background save thread=async_save. I0421 02:26:57.569611 139514186016512 async_checkpointer.py:283] [process=1][thread=save_finalize] No errors found in background save thread=async_save. I0421 02:26:57.569656 139514186016512 checkpoint_manager.py:2133] [process=1][thread=save_finalize][step=9] CheckpointManager Save Finalize is syncing with other hosts... I0421 02:26:57.571148 139514186016512 checkpoint_manager.py:2142] [process=1][thread=save_finalize][step=9] CheckpointManager Save Finalize is done on all hosts. I0421 02:26:57.571316 139640734226240 checkpoint_manager.py:2036] [process=1][thread=MainThread][step=9][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=9. I0421 02:26:57.571478 139640734226240 checkpoint_manager.py:2013] [process=1][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0421 02:26:57.572119 139640734226240 metric_logger.py:185] completed step: 9, seconds: 0.147, TFLOP/s/device: 92.141, Tokens/s/device: 13888.512, total_weights: 65536, loss: 7.181 Per train step: Total TFLOPs: 13.59 split as 93.93% learnable weight flops and 6.07% attention flops XPK End: Tue Apr 21 02:27:10 UTC 2026 EXIT_CODE=0