XPK Start: Tue Apr 21 01:06:18 UTC 2026 2026-04-21 01:06:22.748557: 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 00:00:1776733582.760956 10 cuda_dnn.cc:8579] Unable to register cuDNN factory: Attempting to register factory for plugin cuDNN when one has already been registered E0000 00:00:1776733582.764723 10 cuda_blas.cc:1407] Unable to register cuBLAS factory: Attempting to register factory for plugin cuBLAS when one has already been registered W0000 00:00:1776733582.775817 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. W0000 00:00:1776733582.775834 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. W0000 00:00:1776733582.775836 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. W0000 00:00:1776733582.775839 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. 2026-04-21 01:06:41.913466: 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 01:06:42.434719 132451369105216 max_utils.py:273] Attempting to initialize the jax distributed system... INFO:2026-04-21 01:06:51,476:jax._src.distributed:140: Starting JAX distributed service on [::]:8482 I0421 01:06:51.476060 132451369105216 distributed.py:140] Starting JAX distributed service on [::]:8482 INFO:2026-04-21 01:06:51,478:jax._src.distributed:157: Connecting to JAX distributed service on mt-06-grad-accum-mmnf4-slice-job-0-0.mt-06-grad-accum-mmnf4:8482 I0421 01:06:51.478457 132451369105216 distributed.py:157] Connecting to JAX distributed service on mt-06-grad-accum-mmnf4-slice-job-0-0.mt-06-grad-accum-mmnf4:8482 I0421 01:06:53.001260 132451369105216 max_utils.py:284] Jax distributed system initialized! I0421 01:06:59.698421 132451369105216 max_utils.py:800] System Information: Jax Version: 0.8.1 I0421 01:06:59.698528 132451369105216 max_utils.py:801] System Information: Jaxlib Version: 0.8.1 I0421 01:06:59.698569 132451369105216 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 01:06:59.698604 132451369105216 train_utils.py:347] WARNING: Sequence packing is essentially ignored for synthetic data. Please use a real dataset to use sequence packing. I0421 01:07:00.321312 132451369105216 maxtext_utils.py:1398] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1) I0421 01:07:00.321600 132451369105216 checkpointing.py:677] Setting up checkpoint logger... I0421 01:07:00.321656 132451369105216 checkpointing.py:233] Creating checkpoint manager with ocdbt=True and zarr3=True I0421 01:07:00.321726 132451369105216 pytree_checkpoint_handler.py:589] save_device_host_concurrent_bytes=None I0421 01:07:00.322073 132451369105216 base_pytree_checkpoint_handler.py:415] Created BasePyTreeCheckpointHandler: use_ocdbt=True, use_zarr3=True, pytree_metadata_options=PyTreeMetadataOptions(support_rich_types=False), array_metadata_store=<orbax.checkpoint._src.metadata.array_metadata_store.Store object at 0x787635f74e60>, enable_pinned_host_transfer=False, save_concurrent_bytes: 96000000000 (89.4 GiB), restore_concurrent_bytes: 96000000000 (89.4 GiB) I0421 01:07:03.567093 132451369105216 checkpointing.py:265] Enabling policy for fixed interval checkpointing. I0421 01:07:03.567332 132451369105216 checkpoint_manager.py:709] [process=4][thread=MainThread] CheckpointManager init: checkpointers=None, item_names=('items',), item_handlers={'items': <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7863ce8a80e0>}, handler_registry=None I0421 01:07:03.567565 132451369105216 composite_checkpoint_handler.py:237] Deferred registration for item: "items". Adding handler `<orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7863ce8a80e0>` 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 01:07:03.567613 132451369105216 composite_checkpoint_handler.py:237] Deferred registration for item: "metrics". Adding handler `<orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7863ce8a8cb0>` 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 01:07:03.567649 132451369105216 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 0x7863ce8a80e0>, ('items', <class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>): <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7863ce8a80e0>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7863ce8a8cb0>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7863ce8a8cb0>}). I0421 01:07:03.570314 132451369105216 abstract_checkpointer.py:35] orbax-checkpoint version: 0.11.33 I0421 01:07:03.570397 132451369105216 async_checkpointer.py:177] [process=4][thread=MainThread] Using barrier_sync_fn: <function get_barrier_sync_fn.<locals>._fn at 0x78761c3fb2e0> timeout: 600 secs and primary_host=0 for async checkpoint writes I0421 01:07:05.159851 132451369105216 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_06_grad_accum/checkpoints I0421 01:07:05.191650 132451369105216 checkpoint_manager.py:929] [process=4][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=10), 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_06_grad_accum/checkpoints: <orbax.checkpoint.checkpoint_manager.CheckpointManager object at 0x7863ce8ab980> I0421 01:07:05.191805 132451369105216 checkpointing.py:301] Checkpoint manager created! I0421 01:07:06.867896 132451369105216 nnx_wrappers.py:453] Unknown Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed'). I0421 01:07:06.868004 132451369105216 nnx_wrappers.py:453] Unknown Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0421 01:07:07.238139 132451369105216 attentions.py:1084] attentions/inputs_q Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0421 01:07:07.238240 132451369105216 attentions.py:1084] attentions/inputs_q Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0421 01:07:07.254057 132451369105216 attentions.py:1085] attentions/inputs_kv Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0421 01:07:07.254119 132451369105216 attentions.py:1085] attentions/inputs_kv Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0421 01:07:07.277363 132451369105216 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 01:07:07.277443 132451369105216 attentions.py:1150] attentions/query Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0421 01:07:07.293216 132451369105216 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 01:07:07.293278 132451369105216 attentions.py:1151] attentions/key Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0421 01:07:07.308960 132451369105216 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 01:07:07.309019 132451369105216 attentions.py:1152] attentions/value Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0421 01:07:07.333032 132451369105216 attentions.py:1193] attentions/out Logical: bfloat16[32,2048,16,128].................................... ('activation_batch', 'activation_attn_length', 'activation_heads', 'activation_kv'). I0421 01:07:07.333103 132451369105216 attentions.py:1193] attentions/out Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0421 01:07:07.352466 132451369105216 linears.py:525] linears/x Logical: bfloat16[32,2048,7168]...................................... ('activation_batch', 'activation_length', 'activation_mlp'). I0421 01:07:07.352530 132451369105216 linears.py:525] linears/x Physical: bfloat16[32,2048,7168]...................................... ('fsdp', None, None). I0421 01:07:07.545747 132451369105216 checkpointing.py:577] checkpoint manager exists so trying to load this run's existing checkpoint I0421 01:07:07.545857 132451369105216 checkpointing.py:665] No existing checkpoints found, not restoring checkpoint. fsdp: 32 I0421 01:07:08.940274 132451369105216 maxtext_utils.py:1501] params/params/decoder/decoder_norm/scale Shape: float32[2048] Logical: PartitionSpec('norm',) Physical: (None,) I0421 01:07:08.940405 132451369105216 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 01:07:08.940460 132451369105216 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 01:07:08.940519 132451369105216 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 01:07:08.940573 132451369105216 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 01:07:08.940612 132451369105216 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 01:07:08.940665 132451369105216 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 01:07:08.940730 132451369105216 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 01:07:08.940774 132451369105216 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 01:07:08.940814 132451369105216 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 01:07:08.940865 132451369105216 maxtext_utils.py:1501] params/params/decoder/logits_dense/kernel Shape: float32[2048,32000] Logical: PartitionSpec('embed', 'vocab') Physical: ('fsdp', None) I0421 01:07:08.940913 132451369105216 maxtext_utils.py:1501] params/params/token_embedder/embedding Shape: float32[32000,2048] Logical: PartitionSpec('vocab', 'embed') Physical: (None, 'fsdp') I0421 01:07:08.964351 132451369105216 gradient_accumulation.py:68] gradient_accumulation/inputs Logical: float32[2048]............................................... Unknown. I0421 01:07:08.964422 132451369105216 gradient_accumulation.py:68] gradient_accumulation/inputs Physical: float32[2048]............................................... (None,). I0421 01:07:08.978822 132451369105216 gradient_accumulation.py:68] gradient_accumulation/inputs Logical: float32[2048,16,7168]....................................... Unknown. I0421 01:07:08.978879 132451369105216 gradient_accumulation.py:68] gradient_accumulation/inputs Physical: float32[2048,16,7168]....................................... ('fsdp', None, None). I0421 01:07:09.007028 132451369105216 gradient_accumulation.py:68] gradient_accumulation/inputs Logical: float32[7168,16,2048]....................................... Unknown. I0421 01:07:09.007094 132451369105216 gradient_accumulation.py:68] gradient_accumulation/inputs Physical: float32[7168,16,2048]....................................... (None, None, 'fsdp'). I0421 01:07:09.021193 132451369105216 gradient_accumulation.py:68] gradient_accumulation/inputs Logical: float32[2048,16]............................................ Unknown. I0421 01:07:09.021252 132451369105216 gradient_accumulation.py:68] gradient_accumulation/inputs Physical: float32[2048,16]............................................ (None, None). I0421 01:07:09.049328 132451369105216 gradient_accumulation.py:68] gradient_accumulation/inputs Logical: float32[2048,16,16,128]..................................... Unknown. I0421 01:07:09.049391 132451369105216 gradient_accumulation.py:68] gradient_accumulation/inputs Physical: float32[2048,16,16,128]..................................... ('fsdp', None, None, None). I0421 01:07:09.063557 132451369105216 gradient_accumulation.py:68] gradient_accumulation/inputs Logical: float32[16,16,128,2048]..................................... Unknown. I0421 01:07:09.063612 132451369105216 gradient_accumulation.py:68] gradient_accumulation/inputs Physical: float32[16,16,128,2048]..................................... (None, None, None, 'fsdp'). I0421 01:07:09.105834 132451369105216 gradient_accumulation.py:68] gradient_accumulation/inputs Logical: float32[2048,32000]......................................... Unknown. I0421 01:07:09.105904 132451369105216 gradient_accumulation.py:68] gradient_accumulation/inputs Physical: float32[2048,32000]......................................... ('fsdp', None). I0421 01:07:09.119982 132451369105216 gradient_accumulation.py:68] gradient_accumulation/inputs Logical: float32[32000,2048]......................................... Unknown. I0421 01:07:09.120037 132451369105216 gradient_accumulation.py:68] gradient_accumulation/inputs Physical: float32[32000,2048]......................................... (None, 'fsdp'). I0421 01:07:09.774708 132451369105216 train.py:156] train/xent Logical: float32[32,2048]............................................ ('activation_embed_and_logits_batch', 'activation_length'). I0421 01:07:09.774799 132451369105216 train.py:156] train/xent Physical: float32[32,2048]............................................ ('fsdp', None). I0421 01:07:09.790227 132451369105216 train.py:163] train/z_loss Logical: float32[32,2048]............................................ ('activation_embed_and_logits_batch', 'activation_length'). I0421 01:07:09.790286 132451369105216 train.py:163] train/z_loss Physical: float32[32,2048]............................................ ('fsdp', None). I0421 01:07:20.958754 132451369105216 max_utils.py:791] Total memory size: 1.7 GB, Output size: 0.4 GB, Temp size: 1.3 GB, Argument size: 0.4 GB, Host temp size: 0.0 GB. I0421 01:07:20.959651 132451369105216 metric_logger.py:289] number parameters: 1.104 billion I0421 01:07:22.998897 132451369105216 checkpointing.py:772] Waiting for step 0 to finish before checkpoint... I0421 01:07:33.413304 132451369105216 checkpointing.py:776] Waited 10.414388418197632 seconds for step 0 to finish before starting checkpointing. I0421 01:07:33.416082 132451369105216 checkpoint_manager.py:2013] [process=4][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0421 01:07:33.417875 132451369105216 checkpoint_manager.py:1518] [process=4] Saving checkpoint at step 0 I0421 01:07:33.420614 132451369105216 async_checkpointer.py:452] [process=4] 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_06_grad_accum/checkpoints/0. I0421 01:07:34.875288 132451369105216 signaling_client.py:364] Using JaxDistributedSignalingClient I0421 01:07:34.876373 132451369105216 jax_array_handlers.py:358] Scheduling D2H of 39 prioritized jax.Array. I0421 01:07:34.876430 132451369105216 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 01:07:35.117374 132451369105216 base_pytree_checkpoint_handler.py:153] [process=4][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.242165s I0421 01:07:35.117542 132451369105216 base_pytree_checkpoint_handler.py:129] [process=4] /jax/checkpoint/write/blocking_gbytes_per_sec: 1.617 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.9541909694671631 s) (per-host) I0421 01:07:35.117592 132451369105216 base_pytree_checkpoint_handler.py:737] [process=4][thread=MainThread] Initiated Pytree async_save. Time taken: 0.954252s (batch_requests_ready=0.708566s, total_serialization_initiated=0.245618s, others=0.000068s) I0421 01:07:35.117679 132451369105216 composite_checkpoint_handler.py:715] [process=4][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.958172s (all_items=0.000020s, per_item={'items': '0.00002003'}, temp_paths=0.958152) I0421 01:07:35.118625 132327782991616 async_checkpointer.py:79] [process=4][thread=async_save] Background save thread started. I0421 01:07:35.118763 132451369105216 async_checkpointer.py:561] Finished blocking save. Time taken: 1.700816s. Continuing background save to gs://lance-maxtext/linen_ckpt_xpk_test_pipeline_scan_nnx_20260421_005327/linen_xpk_test_pipeline_scan_nnx_20260421_005327_06_grad_accum/checkpoints/0. I0421 01:07:35.120786 132451369105216 checkpoint_manager.py:1566] [process=4][thread=MainThread][step=0] Starting CheckpointManager Save Finalize thread=save_finalize I0421 01:07:35.121083 132324570158848 async_checkpointer.py:265] [process=4][thread=save_finalize] Waiting for background save thread=async_save. I0421 01:07:35.121238 132451369105216 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_06_grad_accum/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776733653.416064, 'wait_for_prev_duration_secs': 6.4849853515625e-05, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776733653.417915, 'checkpointer_blocking_duration_secs': 1.701024055480957, 'get_old_steps_start_time': 1776733655.1189618, 'get_old_steps_duration_secs': 2.5033950805664062e-05, 'checkpoint_manager_blocking_start_time': 1776733653.413833, 'checkpoint_manager_blocking_duration_secs': 1.7073664665222168} I0421 01:07:35.121343 132451369105216 checkpointing.py:408] Started an asynchronous checkpoint save for step 0 I0421 01:07:35.121393 132451369105216 max_utils.py:750] Memstats: After params initialized: I0421 01:07:35.121446 132451369105216 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_16(process=4,(0,4,0,0)) I0421 01:07:35.121480 132451369105216 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_17(process=4,(1,4,0,0)) I0421 01:07:35.121507 132451369105216 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_20(process=4,(0,5,0,0)) I0421 01:07:35.121532 132451369105216 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_21(process=4,(1,5,0,0)) I0421 01:07:35.381536 132451369105216 metric_logger.py:185] completed step: 0, seconds: 2.039, TFLOP/s/device: 26.653, Tokens/s/device: 4017.435, total_weights: 262144, loss: 10.877 I0421 01:07:35.995434 132451369105216 metric_logger.py:185] completed step: 1, seconds: 12.381, TFLOP/s/device: 4.390, Tokens/s/device: 661.643, total_weights: 262144, loss: 10.877 I0421 01:07:36.573802 132451369105216 metric_logger.py:185] completed step: 2, seconds: 0.036, TFLOP/s/device: 1502.087, Tokens/s/device: 226410.923, total_weights: 262144, loss: 10.563 I0421 01:07:37.152058 132451369105216 metric_logger.py:185] completed step: 3, seconds: 0.584, TFLOP/s/device: 93.102, Tokens/s/device: 14033.405, total_weights: 262144, loss: 10.272 I0421 01:07:37.900731 2368 google_auth_provider.cc:181] Running on GCE, using service account 562977990677-compute@developer.gserviceaccount.com I0421 01:07:38.308644 132451369105216 metric_logger.py:185] completed step: 4, seconds: 0.578, TFLOP/s/device: 93.995, Tokens/s/device: 14168.010, total_weights: 262144, loss: 10.022 I0421 01:07:38.314296 132451369105216 metric_logger.py:185] completed step: 5, seconds: 0.578, TFLOP/s/device: 93.996, Tokens/s/device: 14168.157, total_weights: 262144, loss: 9.820 I0421 01:07:39.782634 132327791384320 array_metadata_store.py:203] [process=4][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_06_grad_accum/checkpoints/0/items/array_metadatas/process_4 I0421 01:08:12.962636 132327782991616 base_pytree_checkpoint_handler.py:129] [process=4] /jax/checkpoint/write/gbytes_per_sec: 40.712 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 38.79924440383911 s) (per-host) I0421 01:08:12.962764 132327782991616 async_checkpointer.py:90] [process=4][thread=async_save] 3 Handler Commit operations completed. Time taken: 37.843922s. I0421 01:08:16.285091 132451369105216 metric_logger.py:185] completed step: 6, seconds: 1.157, TFLOP/s/device: 46.974, Tokens/s/device: 7080.368, total_weights: 262144, loss: 9.667 I0421 01:08:16.863355 132451369105216 metric_logger.py:185] completed step: 7, seconds: 37.393, TFLOP/s/device: 1.453, Tokens/s/device: 219.077, total_weights: 262144, loss: 9.561 I0421 01:08:17.441266 132451369105216 metric_logger.py:185] completed step: 8, seconds: 0.583, TFLOP/s/device: 93.218, Tokens/s/device: 14050.759, total_weights: 262144, loss: 9.496 I0421 01:08:18.019312 132451369105216 checkpointing.py:772] Waiting for step 9 to finish before checkpoint... I0421 01:08:18.019971 132451369105216 checkpointing.py:776] Waited 0.0006783008575439453 seconds for step 9 to finish before starting checkpointing. I0421 01:08:18.022091 132451369105216 checkpoint_manager.py:2024] [process=4][thread=MainThread][step=0][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0421 01:08:56.902977 132327782991616 async_checkpointer.py:144] [process=4][thread=async_save] Background save thread done. Time taken: 81.784119s. I0421 01:08:56.903323 132324570158848 async_checkpointer.py:273] [process=4][thread=save_finalize] Done with waiting for background save thread=async_save. I0421 01:08:56.903448 132324570158848 async_checkpointer.py:283] [process=4][thread=save_finalize] No errors found in background save thread=async_save. I0421 01:08:56.903499 132324570158848 checkpoint_manager.py:2133] [process=4][thread=save_finalize][step=0] CheckpointManager Save Finalize is syncing with other hosts... I0421 01:08:56.907649 132324570158848 checkpoint_manager.py:2142] [process=4][thread=save_finalize][step=0] CheckpointManager Save Finalize is done on all hosts. I0421 01:08:56.907824 132451369105216 checkpoint_manager.py:2036] [process=4][thread=MainThread][step=0][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=0. W0421 01:08:56.907966 132451369105216 checkpoint_manager.py:1458] Waiting for previous save to complete took 38.885881 seconds. If this number is high, consider checkpointing less frequently. I0421 01:08:56.909881 132451369105216 checkpoint_manager.py:1518] [process=4] Saving checkpoint at step 9 I0421 01:08:56.913474 132451369105216 async_checkpointer.py:452] [process=4] 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_06_grad_accum/checkpoints/9. I0421 01:08:57.886584 132451369105216 jax_array_handlers.py:358] Scheduling D2H of 39 prioritized jax.Array. I0421 01:08:57.886681 132451369105216 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 01:08:57.929246 132451369105216 base_pytree_checkpoint_handler.py:153] [process=4][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.043608s I0421 01:08:57.929409 132451369105216 base_pytree_checkpoint_handler.py:129] [process=4] /jax/checkpoint/write/blocking_gbytes_per_sec: 5.067 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.3044414520263672 s) (per-host) I0421 01:08:57.929464 132451369105216 base_pytree_checkpoint_handler.py:737] [process=4][thread=MainThread] Initiated Pytree async_save. Time taken: 0.304505s (batch_requests_ready=0.258949s, total_serialization_initiated=0.045487s, others=0.000070s) I0421 01:08:57.929554 132451369105216 composite_checkpoint_handler.py:715] [process=4][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.308471s (all_items=0.000016s, per_item={'items': '0.00001645'}, temp_paths=0.308455) I0421 01:08:57.930485 132324561766144 async_checkpointer.py:79] [process=4][thread=async_save] Background save thread started. I0421 01:08:57.930626 132451369105216 async_checkpointer.py:561] Finished blocking save. Time taken: 1.020667s. Continuing background save to gs://lance-maxtext/linen_ckpt_xpk_test_pipeline_scan_nnx_20260421_005327/linen_xpk_test_pipeline_scan_nnx_20260421_005327_06_grad_accum/checkpoints/9. I0421 01:08:58.374053 132451369105216 checkpoint_manager.py:1566] [process=4][thread=MainThread][step=9] Starting CheckpointManager Save Finalize thread=save_finalize I0421 01:08:58.374421 132324570158848 async_checkpointer.py:265] [process=4][thread=save_finalize] Waiting for background save thread=async_save. I0421 01:08:58.374607 132451369105216 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_06_grad_accum/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776733698.0220616, 'wait_for_prev_duration_secs': 38.88588070869446, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776733736.909919, 'checkpointer_blocking_duration_secs': 1.02089524269104, 'get_old_steps_start_time': 1776733737.9308362, 'get_old_steps_duration_secs': 2.5510787963867188e-05, 'checkpoint_manager_blocking_start_time': 1776733698.0202112, 'checkpoint_manager_blocking_duration_secs': 40.35435605049133} I0421 01:08:58.374723 132451369105216 checkpointing.py:408] Started an asynchronous checkpoint save for step 9 I0421 01:08:58.374769 132451369105216 checkpoint_manager.py:2024] [process=4][thread=MainThread][step=9][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0421 01:09:05.636406 132327791384320 array_metadata_store.py:203] [process=4][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_06_grad_accum/checkpoints/9/items/array_metadatas/process_4 I0421 01:09:41.395312 132324561766144 base_pytree_checkpoint_handler.py:129] [process=4] /jax/checkpoint/write/gbytes_per_sec: 36.088 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 43.770310163497925 s) (per-host) I0421 01:09:41.395431 132324561766144 async_checkpointer.py:90] [process=4][thread=async_save] 3 Handler Commit operations completed. Time taken: 43.464711s. I0421 01:09:49.787054 132324561766144 async_checkpointer.py:144] [process=4][thread=async_save] Background save thread done. Time taken: 51.856319s. I0421 01:09:49.787355 132324570158848 async_checkpointer.py:273] [process=4][thread=save_finalize] Done with waiting for background save thread=async_save. I0421 01:09:49.787473 132324570158848 async_checkpointer.py:283] [process=4][thread=save_finalize] No errors found in background save thread=async_save. I0421 01:09:49.787519 132324570158848 checkpoint_manager.py:2133] [process=4][thread=save_finalize][step=9] CheckpointManager Save Finalize is syncing with other hosts... I0421 01:09:49.789972 132324570158848 checkpoint_manager.py:2142] [process=4][thread=save_finalize][step=9] CheckpointManager Save Finalize is done on all hosts. I0421 01:09:49.790160 132451369105216 checkpoint_manager.py:2036] [process=4][thread=MainThread][step=9][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=9. I0421 01:09:49.790297 132451369105216 checkpoint_manager.py:2013] [process=4][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0421 01:09:49.790908 132451369105216 metric_logger.py:185] completed step: 9, seconds: 0.578, TFLOP/s/device: 94.019, Tokens/s/device: 14171.564, total_weights: 262144, loss: 9.457 Per train step: Total TFLOPs: 54.35 split as 93.93% learnable weight flops and 6.07% attention flops XPK End: Tue Apr 21 01:10:00 UTC 2026 EXIT_CODE=0