XPK Start: Sun Apr 19 10:25:28 UTC 2026 2026-04-19 10:25:32.219070: 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 00:00:1776594332.231690 10 cuda_dnn.cc:8579] Unable to register cuDNN factory: Attempting to register factory for plugin cuDNN when one has already been registered E0000 00:00:1776594332.235379 10 cuda_blas.cc:1407] Unable to register cuBLAS factory: Attempting to register factory for plugin cuBLAS when one has already been registered W0000 00:00:1776594332.246773 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. W0000 00:00:1776594332.246793 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. W0000 00:00:1776594332.246796 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. W0000 00:00:1776594332.246798 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. 2026-04-19 10:25:51.544216: 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) I0419 10:25:52.065008 133223426119488 max_utils.py:273] Attempting to initialize the jax distributed system... INFO:2026-04-19 10:26:01,106:jax._src.distributed:140: Starting JAX distributed service on [::]:8482 I0419 10:26:01.106412 133223426119488 distributed.py:140] Starting JAX distributed service on [::]:8482 INFO:2026-04-19 10:26:01,108:jax._src.distributed:157: Connecting to JAX distributed service on mt-16-zarr3-false-sv-wqlsq-slice-job-0-0.mt-16-zarr3-false-sv-wqlsq:8482 I0419 10:26:01.108706 133223426119488 distributed.py:157] Connecting to JAX distributed service on mt-16-zarr3-false-sv-wqlsq-slice-job-0-0.mt-16-zarr3-false-sv-wqlsq:8482 I0419 10:26:02.601595 133223426119488 max_utils.py:284] Jax distributed system initialized! I0419 10:26:08.827280 133223426119488 max_utils.py:800] System Information: Jax Version: 0.8.1 I0419 10:26:08.827390 133223426119488 max_utils.py:801] System Information: Jaxlib Version: 0.8.1 I0419 10:26:08.827431 133223426119488 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 I0419 10:26:08.827465 133223426119488 train_utils.py:347] WARNING: Sequence packing is essentially ignored for synthetic data. Please use a real dataset to use sequence packing. I0419 10:26:09.453244 133223426119488 maxtext_utils.py:1398] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1) I0419 10:26:09.453551 133223426119488 checkpointing.py:677] Setting up checkpoint logger... I0419 10:26:09.453625 133223426119488 checkpointing.py:233] Creating checkpoint manager with ocdbt=True and zarr3=False I0419 10:26:09.453688 133223426119488 pytree_checkpoint_handler.py:589] save_device_host_concurrent_bytes=None I0419 10:26:09.454075 133223426119488 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 0x7929f8979070>, enable_pinned_host_transfer=False, save_concurrent_bytes: 96000000000 (89.4 GiB), restore_concurrent_bytes: 96000000000 (89.4 GiB) I0419 10:26:12.397199 133223426119488 checkpointing.py:265] Enabling policy for fixed interval checkpointing. I0419 10:26:12.397552 133223426119488 checkpoint_manager.py:709] [process=7][thread=MainThread] CheckpointManager init: checkpointers=None, item_names=('items',), item_handlers={'items': <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7916183a5040>}, handler_registry=None I0419 10:26:12.397800 133223426119488 composite_checkpoint_handler.py:237] Deferred registration for item: "items". Adding handler `<orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7916183a5040>` 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`. I0419 10:26:12.397850 133223426119488 composite_checkpoint_handler.py:237] Deferred registration for item: "metrics". Adding handler `<orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7916183a9040>` 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`. I0419 10:26:12.397887 133223426119488 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 0x7916183a5040>, ('items', <class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>): <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7916183a5040>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7916183a9040>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7916183a9040>}). I0419 10:26:12.400520 133223426119488 abstract_checkpointer.py:35] orbax-checkpoint version: 0.11.33 I0419 10:26:12.400618 133223426119488 async_checkpointer.py:177] [process=7][thread=MainThread] Using barrier_sync_fn: <function get_barrier_sync_fn.<locals>._fn at 0x7929e5ddb2e0> timeout: 600 secs and primary_host=0 for async checkpoint writes I0419 10:26:13.817641 133223426119488 checkpoint_manager.py:1818] Found 0 checkpoint steps in gs://lance-maxtext/linen_ckpt_xpk_test_pipeline_scan_nnx_20260419_085454/linen_xpk_test_pipeline_scan_nnx_20260419_085454_16_zarr3_false/checkpoints I0419 10:26:14.251251 133223426119488 checkpoint_manager.py:929] [process=7][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_20260419_085454/linen_xpk_test_pipeline_scan_nnx_20260419_085454_16_zarr3_false/checkpoints: <orbax.checkpoint.checkpoint_manager.CheckpointManager object at 0x791650129e50> I0419 10:26:14.251423 133223426119488 checkpointing.py:301] Checkpoint manager created! I0419 10:26:15.145087 133223426119488 nnx_wrappers.py:453] Unknown Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed'). I0419 10:26:15.145195 133223426119488 nnx_wrappers.py:453] Unknown Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0419 10:26:15.513658 133223426119488 attentions.py:1084] attentions/inputs_q Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0419 10:26:15.513752 133223426119488 attentions.py:1084] attentions/inputs_q Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0419 10:26:15.529672 133223426119488 attentions.py:1085] attentions/inputs_kv Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0419 10:26:15.529729 133223426119488 attentions.py:1085] attentions/inputs_kv Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0419 10:26:15.553266 133223426119488 attentions.py:1150] attentions/query Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0419 10:26:15.553337 133223426119488 attentions.py:1150] attentions/query Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0419 10:26:15.569420 133223426119488 attentions.py:1151] attentions/key Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0419 10:26:15.569481 133223426119488 attentions.py:1151] attentions/key Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0419 10:26:15.585554 133223426119488 attentions.py:1152] attentions/value Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0419 10:26:15.585615 133223426119488 attentions.py:1152] attentions/value Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0419 10:26:15.609973 133223426119488 attentions.py:1193] attentions/out Logical: bfloat16[32,2048,16,128].................................... ('activation_batch', 'activation_attn_length', 'activation_heads', 'activation_kv'). I0419 10:26:15.610047 133223426119488 attentions.py:1193] attentions/out Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0419 10:26:15.629930 133223426119488 linears.py:525] linears/x Logical: bfloat16[32,2048,7168]...................................... ('activation_batch', 'activation_length', 'activation_mlp'). I0419 10:26:15.629997 133223426119488 linears.py:525] linears/x Physical: bfloat16[32,2048,7168]...................................... ('fsdp', None, None). I0419 10:26:15.821388 133223426119488 checkpointing.py:577] checkpoint manager exists so trying to load this run's existing checkpoint I0419 10:26:15.821499 133223426119488 checkpointing.py:665] No existing checkpoints found, not restoring checkpoint. fsdp: 32 I0419 10:26:17.245383 133223426119488 maxtext_utils.py:1501] params/params/decoder/decoder_norm/scale Shape: float32[2048] Logical: PartitionSpec('norm',) Physical: (None,) I0419 10:26:17.245517 133223426119488 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) I0419 10:26:17.245586 133223426119488 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) I0419 10:26:17.245653 133223426119488 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') I0419 10:26:17.245706 133223426119488 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) I0419 10:26:17.245747 133223426119488 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) I0419 10:26:17.245800 133223426119488 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) I0419 10:26:17.245854 133223426119488 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') I0419 10:26:17.245919 133223426119488 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) I0419 10:26:17.245958 133223426119488 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) I0419 10:26:17.246009 133223426119488 maxtext_utils.py:1501] params/params/decoder/logits_dense/kernel Shape: float32[2048,32000] Logical: PartitionSpec('embed', 'vocab') Physical: ('fsdp', None) I0419 10:26:17.246057 133223426119488 maxtext_utils.py:1501] params/params/token_embedder/embedding Shape: float32[32000,2048] Logical: PartitionSpec('vocab', 'embed') Physical: (None, 'fsdp') I0419 10:26:17.753556 133223426119488 train.py:156] train/xent Logical: float32[32,2048]............................................ ('activation_embed_and_logits_batch', 'activation_length'). I0419 10:26:17.753668 133223426119488 train.py:156] train/xent Physical: float32[32,2048]............................................ ('fsdp', None). I0419 10:26:17.768842 133223426119488 train.py:163] train/z_loss Logical: float32[32,2048]............................................ ('activation_embed_and_logits_batch', 'activation_length'). I0419 10:26:17.768904 133223426119488 train.py:163] train/z_loss Physical: float32[32,2048]............................................ ('fsdp', None). I0419 10:26:28.471073 133223426119488 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. I0419 10:26:28.472010 133223426119488 metric_logger.py:289] number parameters: 1.104 billion I0419 10:26:39.438436 133223426119488 checkpointing.py:772] Waiting for step 0 to finish before checkpoint... I0419 10:26:39.605113 133223426119488 checkpointing.py:776] Waited 0.16666245460510254 seconds for step 0 to finish before starting checkpointing. I0419 10:26:39.607598 133223426119488 checkpoint_manager.py:2013] [process=7][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0419 10:26:39.609402 133223426119488 checkpoint_manager.py:1518] [process=7] Saving checkpoint at step 0 I0419 10:26:39.612152 133223426119488 async_checkpointer.py:452] [process=7] Started async saving checkpoint to gs://lance-maxtext/linen_ckpt_xpk_test_pipeline_scan_nnx_20260419_085454/linen_xpk_test_pipeline_scan_nnx_20260419_085454_16_zarr3_false/checkpoints/0. I0419 10:26:40.628941 133223426119488 signaling_client.py:364] Using JaxDistributedSignalingClient I0419 10:26:40.629841 133223426119488 jax_array_handlers.py:358] Scheduling D2H of 39 prioritized jax.Array. I0419 10:26:40.629898 133223426119488 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 I0419 10:26:40.885208 133223426119488 base_pytree_checkpoint_handler.py:153] [process=7][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.256350s I0419 10:26:40.885377 133223426119488 base_pytree_checkpoint_handler.py:129] [process=7] /jax/checkpoint/write/blocking_gbytes_per_sec: 2.754 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.5601577758789062 s) (per-host) I0419 10:26:40.885427 133223426119488 base_pytree_checkpoint_handler.py:737] [process=7][thread=MainThread] Initiated Pytree async_save. Time taken: 0.560216s (batch_requests_ready=0.300097s, total_serialization_initiated=0.260054s, others=0.000066s) I0419 10:26:40.885524 133223426119488 composite_checkpoint_handler.py:715] [process=7][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.564528s (all_items=0.000019s, per_item={'items': '0.00001931'}, temp_paths=0.564509) I0419 10:26:40.886439 133104031221504 async_checkpointer.py:79] [process=7][thread=async_save] Background save thread started. I0419 10:26:40.886634 133223426119488 async_checkpointer.py:561] Finished blocking save. Time taken: 1.277162s. Continuing background save to gs://lance-maxtext/linen_ckpt_xpk_test_pipeline_scan_nnx_20260419_085454/linen_xpk_test_pipeline_scan_nnx_20260419_085454_16_zarr3_false/checkpoints/0. I0419 10:26:41.324016 133223426119488 checkpoint_manager.py:1566] [process=7][thread=MainThread][step=0] Starting CheckpointManager Save Finalize thread=save_finalize I0419 10:26:41.324461 133099190990592 async_checkpointer.py:265] [process=7][thread=save_finalize] Waiting for background save thread=async_save. I0419 10:26:41.324648 133223426119488 standard_logger.py:34] {'step': 0, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_test_pipeline_scan_nnx_20260419_085454/linen_xpk_test_pipeline_scan_nnx_20260419_085454_16_zarr3_false/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776594399.6075807, 'wait_for_prev_duration_secs': 6.127357482910156e-05, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776594399.6094408, 'checkpointer_blocking_duration_secs': 1.277298927307129, 'get_old_steps_start_time': 1776594400.8867602, 'get_old_steps_duration_secs': 2.5272369384765625e-05, 'checkpoint_manager_blocking_start_time': 1776594399.6056638, 'checkpoint_manager_blocking_duration_secs': 1.7189433574676514} I0419 10:26:41.324859 133223426119488 checkpointing.py:408] Started an asynchronous checkpoint save for step 0 I0419 10:26:41.324914 133223426119488 max_utils.py:750] Memstats: After params initialized: I0419 10:26:41.324970 133223426119488 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_26(process=7,(2,6,0,0)) I0419 10:26:41.325006 133223426119488 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_27(process=7,(3,6,0,0)) I0419 10:26:41.325045 133223426119488 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_30(process=7,(2,7,0,0)) I0419 10:26:41.325086 133223426119488 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_31(process=7,(3,7,0,0)) I0419 10:26:41.590943 133223426119488 metric_logger.py:185] completed step: 0, seconds: 10.966, TFLOP/s/device: 1.239, Tokens/s/device: 186.754, total_weights: 65536, loss: 10.874 I0419 10:26:41.766327 133223426119488 metric_logger.py:185] completed step: 1, seconds: 2.151, TFLOP/s/device: 6.316, Tokens/s/device: 952.032, total_weights: 65536, loss: 10.874 I0419 10:26:42.157318 133223426119488 metric_logger.py:185] completed step: 2, seconds: 0.028, TFLOP/s/device: 481.728, Tokens/s/device: 72611.239, total_weights: 65536, loss: 10.560 I0419 10:26:42.304875 133223426119488 metric_logger.py:185] completed step: 3, seconds: 0.392, TFLOP/s/device: 34.689, Tokens/s/device: 5228.772, total_weights: 65536, loss: 9.980 I0419 10:26:42.309587 133223426119488 checkpointing.py:772] Waiting for step 5 to finish before checkpoint... I0419 10:26:42.599433 133223426119488 checkpointing.py:776] Waited 0.28981614112854004 seconds for step 5 to finish before starting checkpointing. I0419 10:26:42.602212 133223426119488 checkpoint_manager.py:2024] [process=7][thread=MainThread][step=0][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0419 10:26:44.166743 2555 google_auth_provider.cc:181] Running on GCE, using service account 562977990677-compute@developer.gserviceaccount.com I0419 10:26:46.487033 133104039614208 array_metadata_store.py:203] [process=7][thread=array_type_handler] Wrote 39 array_metadata.ArrayMetadata to gs://lance-maxtext/linen_ckpt_xpk_test_pipeline_scan_nnx_20260419_085454/linen_xpk_test_pipeline_scan_nnx_20260419_085454_16_zarr3_false/checkpoints/0/items/array_metadatas/process_7 I0419 10:27:22.933285 133104031221504 base_pytree_checkpoint_handler.py:129] [process=7] /jax/checkpoint/write/gbytes_per_sec: 37.073 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 42.60802388191223 s) (per-host) I0419 10:27:22.933414 133104031221504 async_checkpointer.py:90] [process=7][thread=async_save] 3 Handler Commit operations completed. Time taken: 42.046845s. I0419 10:27:31.077082 133104031221504 async_checkpointer.py:144] [process=7][thread=async_save] Background save thread done. Time taken: 50.190499s. I0419 10:27:31.077413 133099190990592 async_checkpointer.py:273] [process=7][thread=save_finalize] Done with waiting for background save thread=async_save. I0419 10:27:31.077561 133099190990592 async_checkpointer.py:283] [process=7][thread=save_finalize] No errors found in background save thread=async_save. I0419 10:27:31.077611 133099190990592 checkpoint_manager.py:2133] [process=7][thread=save_finalize][step=0] CheckpointManager Save Finalize is syncing with other hosts... I0419 10:27:31.079206 133099190990592 checkpoint_manager.py:2142] [process=7][thread=save_finalize][step=0] CheckpointManager Save Finalize is done on all hosts. I0419 10:27:31.079391 133223426119488 checkpoint_manager.py:2036] [process=7][thread=MainThread][step=0][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=0. W0419 10:27:31.079522 133223426119488 checkpoint_manager.py:1458] Waiting for previous save to complete took 48.477323 seconds. If this number is high, consider checkpointing less frequently. I0419 10:27:31.082089 133223426119488 checkpoint_manager.py:1518] [process=7] Saving checkpoint at step 5 I0419 10:27:31.086377 133223426119488 async_checkpointer.py:452] [process=7] Started async saving checkpoint to gs://lance-maxtext/linen_ckpt_xpk_test_pipeline_scan_nnx_20260419_085454/linen_xpk_test_pipeline_scan_nnx_20260419_085454_16_zarr3_false/checkpoints/5. I0419 10:27:32.037762 133223426119488 jax_array_handlers.py:358] Scheduling D2H of 39 prioritized jax.Array. I0419 10:27:32.037862 133223426119488 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 I0419 10:27:32.087606 133223426119488 base_pytree_checkpoint_handler.py:153] [process=7][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.051091s I0419 10:27:32.087753 133223426119488 base_pytree_checkpoint_handler.py:129] [process=7] /jax/checkpoint/write/blocking_gbytes_per_sec: 5.051 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.3053913116455078 s) (per-host) I0419 10:27:32.087803 133223426119488 base_pytree_checkpoint_handler.py:737] [process=7][thread=MainThread] Initiated Pytree async_save. Time taken: 0.305450s (batch_requests_ready=0.252480s, total_serialization_initiated=0.052906s, others=0.000064s) I0419 10:27:32.087885 133223426119488 composite_checkpoint_handler.py:715] [process=7][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.310389s (all_items=0.000015s, per_item={'items': '0.00001550'}, temp_paths=0.310374) I0419 10:27:32.088754 133104039614208 async_checkpointer.py:79] [process=7][thread=async_save] Background save thread started. I0419 10:27:32.088885 133223426119488 async_checkpointer.py:561] Finished blocking save. Time taken: 1.006724s. Continuing background save to gs://lance-maxtext/linen_ckpt_xpk_test_pipeline_scan_nnx_20260419_085454/linen_xpk_test_pipeline_scan_nnx_20260419_085454_16_zarr3_false/checkpoints/5. I0419 10:27:32.516904 133223426119488 checkpoint_manager.py:1566] [process=7][thread=MainThread][step=5] Starting CheckpointManager Save Finalize thread=save_finalize I0419 10:27:32.517273 133099190990592 async_checkpointer.py:265] [process=7][thread=save_finalize] Waiting for background save thread=async_save. I0419 10:27:32.517452 133223426119488 standard_logger.py:34] {'step': 5, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_test_pipeline_scan_nnx_20260419_085454/linen_xpk_test_pipeline_scan_nnx_20260419_085454_16_zarr3_false/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776594402.602175, 'wait_for_prev_duration_secs': 48.477323055267334, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776594451.0821285, 'checkpointer_blocking_duration_secs': 1.0069372653961182, 'get_old_steps_start_time': 1776594452.0890841, 'get_old_steps_duration_secs': 2.7179718017578125e-05, 'checkpoint_manager_blocking_start_time': 1776594402.6001272, 'checkpoint_manager_blocking_duration_secs': 49.91728758811951} I0419 10:27:32.517575 133223426119488 checkpointing.py:408] Started an asynchronous checkpoint save for step 5 I0419 10:27:32.518174 133223426119488 metric_logger.py:185] completed step: 4, seconds: 0.152, TFLOP/s/device: 89.112, Tokens/s/device: 13431.974, total_weights: 65536, loss: 9.460 I0419 10:27:32.524069 133223426119488 metric_logger.py:185] completed step: 5, seconds: 0.147, TFLOP/s/device: 92.126, Tokens/s/device: 13886.252, total_weights: 65536, loss: 8.959 I0419 10:27:52.273216 133223426119488 metric_logger.py:185] completed step: 6, seconds: 50.213, TFLOP/s/device: 0.271, Tokens/s/device: 40.786, total_weights: 65536, loss: 8.469 I0419 10:27:52.420975 133223426119488 metric_logger.py:185] completed step: 7, seconds: 19.603, TFLOP/s/device: 0.693, Tokens/s/device: 104.473, total_weights: 65536, loss: 8.003 I0419 10:27:52.568342 133223426119488 metric_logger.py:185] completed step: 8, seconds: 0.152, TFLOP/s/device: 89.421, Tokens/s/device: 13478.561, total_weights: 65536, loss: 7.572 I0419 10:27:52.715719 133223426119488 checkpointing.py:772] Waiting for step 9 to finish before checkpoint... I0419 10:27:52.716368 133223426119488 checkpointing.py:776] Waited 0.0006694793701171875 seconds for step 9 to finish before starting checkpointing. I0419 10:27:52.718279 133223426119488 checkpoint_manager.py:2024] [process=7][thread=MainThread][step=5][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0419 10:27:56.603622 133099178387200 array_metadata_store.py:203] [process=7][thread=array_type_handler] Wrote 39 array_metadata.ArrayMetadata to gs://lance-maxtext/linen_ckpt_xpk_test_pipeline_scan_nnx_20260419_085454/linen_xpk_test_pipeline_scan_nnx_20260419_085454_16_zarr3_false/checkpoints/5/items/array_metadatas/process_7 I0419 10:28:32.337611 133104039614208 base_pytree_checkpoint_handler.py:129] [process=7] /jax/checkpoint/write/gbytes_per_sec: 26.085 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 60.55521583557129 s) (per-host) I0419 10:28:32.337722 133104039614208 async_checkpointer.py:90] [process=7][thread=async_save] 3 Handler Commit operations completed. Time taken: 60.248752s. I0419 10:28:40.538124 133104039614208 async_checkpointer.py:144] [process=7][thread=async_save] Background save thread done. Time taken: 68.449139s. I0419 10:28:40.538419 133099190990592 async_checkpointer.py:273] [process=7][thread=save_finalize] Done with waiting for background save thread=async_save. I0419 10:28:40.538567 133099190990592 async_checkpointer.py:283] [process=7][thread=save_finalize] No errors found in background save thread=async_save. I0419 10:28:40.538625 133099190990592 checkpoint_manager.py:2133] [process=7][thread=save_finalize][step=5] CheckpointManager Save Finalize is syncing with other hosts... I0419 10:28:40.541898 133099190990592 checkpoint_manager.py:2142] [process=7][thread=save_finalize][step=5] CheckpointManager Save Finalize is done on all hosts. I0419 10:28:40.542080 133223426119488 checkpoint_manager.py:2036] [process=7][thread=MainThread][step=5][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=5. W0419 10:28:40.542216 133223426119488 checkpoint_manager.py:1458] Waiting for previous save to complete took 47.823950 seconds. If this number is high, consider checkpointing less frequently. I0419 10:28:40.543859 133223426119488 checkpoint_manager.py:1518] [process=7] Saving checkpoint at step 9 I0419 10:28:40.547291 133223426119488 async_checkpointer.py:452] [process=7] Started async saving checkpoint to gs://lance-maxtext/linen_ckpt_xpk_test_pipeline_scan_nnx_20260419_085454/linen_xpk_test_pipeline_scan_nnx_20260419_085454_16_zarr3_false/checkpoints/9. I0419 10:28:41.079905 133223426119488 jax_array_handlers.py:358] Scheduling D2H of 39 prioritized jax.Array. I0419 10:28:41.080001 133223426119488 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 I0419 10:28:41.128268 133223426119488 base_pytree_checkpoint_handler.py:153] [process=7][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.049433s I0419 10:28:41.128437 133223426119488 base_pytree_checkpoint_handler.py:129] [process=7] /jax/checkpoint/write/blocking_gbytes_per_sec: 5.004 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.30828189849853516 s) (per-host) I0419 10:28:41.128485 133223426119488 base_pytree_checkpoint_handler.py:737] [process=7][thread=MainThread] Initiated Pytree async_save. Time taken: 0.308339s (batch_requests_ready=0.256833s, total_serialization_initiated=0.051444s, others=0.000062s) I0419 10:28:41.128584 133223426119488 composite_checkpoint_handler.py:715] [process=7][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.312478s (all_items=0.000011s, per_item={'items': '0.00001144'}, temp_paths=0.312467) I0419 10:28:41.129444 133099744646912 async_checkpointer.py:79] [process=7][thread=async_save] Background save thread started. I0419 10:28:41.129622 133223426119488 async_checkpointer.py:561] Finished blocking save. Time taken: 0.585690s. Continuing background save to gs://lance-maxtext/linen_ckpt_xpk_test_pipeline_scan_nnx_20260419_085454/linen_xpk_test_pipeline_scan_nnx_20260419_085454_16_zarr3_false/checkpoints/9. I0419 10:28:41.140648 133223426119488 checkpoint_manager.py:1566] [process=7][thread=MainThread][step=9] Starting CheckpointManager Save Finalize thread=save_finalize I0419 10:28:41.140889 133099190990592 async_checkpointer.py:265] [process=7][thread=save_finalize] Waiting for background save thread=async_save. I0419 10:28:41.141017 133223426119488 standard_logger.py:34] {'step': 9, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_test_pipeline_scan_nnx_20260419_085454/linen_xpk_test_pipeline_scan_nnx_20260419_085454_16_zarr3_false/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776594472.7182508, 'wait_for_prev_duration_secs': 47.82394981384277, 'time_between_consecutive_saves_sec': 21.639004945755005, 'checkpointer_blocking_start_time': 1776594520.5438993, 'checkpointer_blocking_duration_secs': 0.5858306884765625, 'get_old_steps_start_time': 1776594521.1297424, 'get_old_steps_duration_secs': 2.1219253540039062e-05, 'checkpoint_manager_blocking_start_time': 1776594472.7166204, 'checkpoint_manager_blocking_duration_secs': 48.42436504364014} I0419 10:28:41.141123 133223426119488 checkpointing.py:408] Started an asynchronous checkpoint save for step 9 I0419 10:28:41.141165 133223426119488 checkpoint_manager.py:2024] [process=7][thread=MainThread][step=9][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0419 10:28:45.789121 133104031221504 array_metadata_store.py:203] [process=7][thread=array_type_handler] Wrote 39 array_metadata.ArrayMetadata to gs://lance-maxtext/linen_ckpt_xpk_test_pipeline_scan_nnx_20260419_085454/linen_xpk_test_pipeline_scan_nnx_20260419_085454_16_zarr3_false/checkpoints/9/items/array_metadatas/process_7 I0419 10:29:23.103691 133099744646912 base_pytree_checkpoint_handler.py:129] [process=7] /jax/checkpoint/write/gbytes_per_sec: 37.357 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 42.283488273620605 s) (per-host) I0419 10:29:23.103811 133099744646912 async_checkpointer.py:90] [process=7][thread=async_save] 3 Handler Commit operations completed. Time taken: 41.974240s. I0419 10:29:31.127695 133099744646912 async_checkpointer.py:144] [process=7][thread=async_save] Background save thread done. Time taken: 49.998098s. I0419 10:29:31.127907 133099190990592 async_checkpointer.py:273] [process=7][thread=save_finalize] Done with waiting for background save thread=async_save. I0419 10:29:31.127971 133099190990592 async_checkpointer.py:283] [process=7][thread=save_finalize] No errors found in background save thread=async_save. I0419 10:29:31.128018 133099190990592 checkpoint_manager.py:2133] [process=7][thread=save_finalize][step=9] CheckpointManager Save Finalize is syncing with other hosts... I0419 10:29:31.129807 133099190990592 checkpoint_manager.py:2142] [process=7][thread=save_finalize][step=9] CheckpointManager Save Finalize is done on all hosts. I0419 10:29:31.129980 133223426119488 checkpoint_manager.py:2036] [process=7][thread=MainThread][step=9][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=9. I0419 10:29:31.130130 133223426119488 checkpoint_manager.py:2013] [process=7][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0419 10:29:31.130774 133223426119488 metric_logger.py:185] completed step: 9, seconds: 0.148, TFLOP/s/device: 91.948, Tokens/s/device: 13859.470, 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: Sun Apr 19 10:29:48 UTC 2026 EXIT_CODE=0