XPK Start: Tue Apr 21 01:00:22 UTC 2026 2026-04-21 01:00:26.174566: 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 00:00:1776733226.187028 10 cuda_dnn.cc:8579] Unable to register cuDNN factory: Attempting to register factory for plugin cuDNN when one has already been registered E0000 00:00:1776733226.190705 10 cuda_blas.cc:1407] Unable to register cuBLAS factory: Attempting to register factory for plugin cuBLAS when one has already been registered W0000 00:00:1776733226.201792 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. W0000 00:00:1776733226.201810 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. W0000 00:00:1776733226.201812 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. W0000 00:00:1776733226.201814 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. 2026-04-21 01:00:45.281741: 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 01:00:45.794456 135989502220096 max_utils.py:273] Attempting to initialize the jax distributed system... INFO:2026-04-21 01:00:54,834:jax._src.distributed:140: Starting JAX distributed service on [::]:8482 I0421 01:00:54.834907 135989502220096 distributed.py:140] Starting JAX distributed service on [::]:8482 INFO:2026-04-21 01:00:54,837:jax._src.distributed:157: Connecting to JAX distributed service on mt-04-int8-t4dsb-slice-job-0-0.mt-04-int8-t4dsb:8482 I0421 01:00:54.837203 135989502220096 distributed.py:157] Connecting to JAX distributed service on mt-04-int8-t4dsb-slice-job-0-0.mt-04-int8-t4dsb:8482 I0421 01:00:56.183633 135989502220096 max_utils.py:284] Jax distributed system initialized! I0421 01:01:02.087350 135989502220096 max_utils.py:800] System Information: Jax Version: 0.8.1 I0421 01:01:02.087457 135989502220096 max_utils.py:801] System Information: Jaxlib Version: 0.8.1 I0421 01:01:02.087498 135989502220096 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 01:01:02.087533 135989502220096 train_utils.py:347] WARNING: Sequence packing is essentially ignored for synthetic data. Please use a real dataset to use sequence packing. I0421 01:01:02.702760 135989502220096 maxtext_utils.py:1398] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1) I0421 01:01:02.703167 135989502220096 checkpointing.py:677] Setting up checkpoint logger... I0421 01:01:02.703225 135989502220096 checkpointing.py:233] Creating checkpoint manager with ocdbt=True and zarr3=True I0421 01:01:02.703270 135989502220096 pytree_checkpoint_handler.py:589] save_device_host_concurrent_bytes=None I0421 01:01:02.703620 135989502220096 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 0x7bae0c174bc0>, enable_pinned_host_transfer=False, save_concurrent_bytes: 96000000000 (89.4 GiB), restore_concurrent_bytes: 96000000000 (89.4 GiB) I0421 01:01:05.517687 135989502220096 checkpointing.py:265] Enabling policy for fixed interval checkpointing. I0421 01:01:05.518074 135989502220096 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 0x7b9b6a327e30>}, handler_registry=None I0421 01:01:05.518306 135989502220096 composite_checkpoint_handler.py:237] Deferred registration for item: "items". Adding handler `<orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7b9b6a327e30>` 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 01:01:05.518357 135989502220096 composite_checkpoint_handler.py:237] Deferred registration for item: "metrics". Adding handler `<orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7b9b6a3beab0>` 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 01:01:05.518395 135989502220096 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 0x7b9b6a327e30>, ('items', <class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>): <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7b9b6a327e30>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7b9b6a3beab0>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7b9b6a3beab0>}). I0421 01:01:05.518915 135989502220096 abstract_checkpointer.py:35] orbax-checkpoint version: 0.11.33 I0421 01:01:05.518995 135989502220096 async_checkpointer.py:177] [process=4][thread=MainThread] Using barrier_sync_fn: <function get_barrier_sync_fn.<locals>._fn at 0x7b9a7c1c0040> timeout: 600 secs and primary_host=0 for async checkpoint writes I0421 01:01:06.704261 135989502220096 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_04_int8/checkpoints I0421 01:01:07.148055 135989502220096 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_04_int8/checkpoints: <orbax.checkpoint.checkpoint_manager.CheckpointManager object at 0x7b9b6a3bf470> I0421 01:01:07.148233 135989502220096 checkpointing.py:301] Checkpoint manager created! I0421 01:01:08.347894 135989502220096 nnx_wrappers.py:453] Unknown Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed'). I0421 01:01:08.348004 135989502220096 nnx_wrappers.py:453] Unknown Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0421 01:01:08.723622 135989502220096 attentions.py:1084] attentions/inputs_q Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0421 01:01:08.723749 135989502220096 attentions.py:1084] attentions/inputs_q Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0421 01:01:08.741474 135989502220096 attentions.py:1085] attentions/inputs_kv Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0421 01:01:08.741537 135989502220096 attentions.py:1085] attentions/inputs_kv Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0421 01:01:08.819369 135989502220096 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 01:01:08.819466 135989502220096 attentions.py:1150] attentions/query Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0421 01:01:08.835020 135989502220096 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 01:01:08.835085 135989502220096 attentions.py:1151] attentions/key Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0421 01:01:08.850673 135989502220096 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 01:01:08.850744 135989502220096 attentions.py:1152] attentions/value Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0421 01:01:08.874535 135989502220096 attentions.py:1193] attentions/out Logical: bfloat16[32,2048,16,128].................................... ('activation_batch', 'activation_attn_length', 'activation_heads', 'activation_kv'). I0421 01:01:08.874604 135989502220096 attentions.py:1193] attentions/out Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0421 01:01:08.957498 135989502220096 linears.py:525] linears/x Logical: bfloat16[32,2048,7168]...................................... ('activation_batch', 'activation_length', 'activation_mlp'). I0421 01:01:08.957591 135989502220096 linears.py:525] linears/x Physical: bfloat16[32,2048,7168]...................................... ('fsdp', None, None). I0421 01:01:09.563310 135989502220096 checkpointing.py:577] checkpoint manager exists so trying to load this run's existing checkpoint I0421 01:01:09.563429 135989502220096 checkpointing.py:665] No existing checkpoints found, not restoring checkpoint. fsdp: 32 I0421 01:01:12.200514 135989502220096 maxtext_utils.py:1501] params/params/decoder/decoder_norm/scale Shape: float32[2048] Logical: PartitionSpec('norm',) Physical: (None,) I0421 01:01:12.200650 135989502220096 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 01:01:12.200729 135989502220096 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 01:01:12.200798 135989502220096 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 01:01:12.200852 135989502220096 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 01:01:12.200892 135989502220096 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 01:01:12.200945 135989502220096 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 01:01:12.200997 135989502220096 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 01:01:12.201037 135989502220096 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 01:01:12.201074 135989502220096 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 01:01:12.201121 135989502220096 maxtext_utils.py:1501] params/params/decoder/logits_dense/kernel Shape: float32[2048,32000] Logical: PartitionSpec('embed', 'vocab') Physical: ('fsdp', None) I0421 01:01:12.201168 135989502220096 maxtext_utils.py:1501] params/params/token_embedder/embedding Shape: float32[32000,2048] Logical: PartitionSpec('vocab', 'embed') Physical: (None, 'fsdp') I0421 01:01:13.904006 135989502220096 train.py:156] train/xent Logical: float32[32,2048]............................................ ('activation_embed_and_logits_batch', 'activation_length'). I0421 01:01:13.904112 135989502220096 train.py:156] train/xent Physical: float32[32,2048]............................................ ('fsdp', None). I0421 01:01:13.918839 135989502220096 train.py:163] train/z_loss Logical: float32[32,2048]............................................ ('activation_embed_and_logits_batch', 'activation_length'). I0421 01:01:13.918899 135989502220096 train.py:163] train/z_loss Physical: float32[32,2048]............................................ ('fsdp', None). I0421 01:01:27.727654 135989502220096 max_utils.py:791] Total memory size: 1.8 GB, Output size: 0.4 GB, Temp size: 1.4 GB, Argument size: 0.4 GB, Host temp size: 0.0 GB. I0421 01:01:27.728586 135989502220096 metric_logger.py:289] number parameters: 1.104 billion I0421 01:01:43.081961 135989502220096 checkpointing.py:772] Waiting for step 0 to finish before checkpoint... I0421 01:01:43.321874 135989502220096 checkpointing.py:776] Waited 0.23989629745483398 seconds for step 0 to finish before starting checkpointing. I0421 01:01:43.324464 135989502220096 checkpoint_manager.py:2013] [process=4][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0421 01:01:43.326565 135989502220096 checkpoint_manager.py:1518] [process=4] Saving checkpoint at step 0 I0421 01:01:43.329188 135989502220096 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_04_int8/checkpoints/0. I0421 01:01:44.780970 135989502220096 signaling_client.py:364] Using JaxDistributedSignalingClient I0421 01:01:44.781972 135989502220096 jax_array_handlers.py:358] Scheduling D2H of 39 prioritized jax.Array. I0421 01:01:44.782030 135989502220096 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 01:01:45.027049 135989502220096 base_pytree_checkpoint_handler.py:153] [process=4][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.246162s I0421 01:01:45.027234 135989502220096 base_pytree_checkpoint_handler.py:129] [process=4] /jax/checkpoint/write/blocking_gbytes_per_sec: 2.886 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.5345571041107178 s) (per-host) I0421 01:01:45.027291 135989502220096 base_pytree_checkpoint_handler.py:737] [process=4][thread=MainThread] Initiated Pytree async_save. Time taken: 0.534624s (batch_requests_ready=0.284449s, total_serialization_initiated=0.250099s, others=0.000075s) I0421 01:01:45.027384 135989502220096 composite_checkpoint_handler.py:715] [process=4][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.539174s (all_items=0.000020s, per_item={'items': '0.00002003'}, temp_paths=0.539154) I0421 01:01:45.028365 135870585747200 async_checkpointer.py:79] [process=4][thread=async_save] Background save thread started. I0421 01:01:45.028538 135989502220096 async_checkpointer.py:561] Finished blocking save. Time taken: 1.701905s. Continuing background save to gs://lance-maxtext/linen_ckpt_xpk_test_pipeline_scan_nnx_20260421_005327/linen_xpk_test_pipeline_scan_nnx_20260421_005327_04_int8/checkpoints/0. I0421 01:01:45.872298 135989502220096 checkpoint_manager.py:1566] [process=4][thread=MainThread][step=0] Starting CheckpointManager Save Finalize thread=save_finalize I0421 01:01:45.872679 135860393592576 async_checkpointer.py:265] [process=4][thread=save_finalize] Waiting for background save thread=async_save. I0421 01:01:45.872869 135989502220096 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_04_int8/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776733303.324445, 'wait_for_prev_duration_secs': 6.461143493652344e-05, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776733303.3266032, 'checkpointer_blocking_duration_secs': 1.702040433883667, 'get_old_steps_start_time': 1776733305.0286624, 'get_old_steps_duration_secs': 2.5987625122070312e-05, 'checkpoint_manager_blocking_start_time': 1776733303.3224993, 'checkpoint_manager_blocking_duration_secs': 2.5503275394439697} I0421 01:01:45.872985 135989502220096 checkpointing.py:408] Started an asynchronous checkpoint save for step 0 I0421 01:01:45.873038 135989502220096 max_utils.py:750] Memstats: After params initialized: I0421 01:01:45.873094 135989502220096 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_16(process=4,(0,4,0,0)) I0421 01:01:45.873127 135989502220096 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_17(process=4,(1,4,0,0)) I0421 01:01:45.873155 135989502220096 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_20(process=4,(0,5,0,0)) I0421 01:01:45.873184 135989502220096 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_21(process=4,(1,5,0,0)) I0421 01:01:46.144786 135989502220096 metric_logger.py:185] completed step: 0, seconds: 15.353, TFLOP/s/device: 0.885, Tokens/s/device: 133.392, total_weights: 65536, loss: 10.874 I0421 01:01:46.389860 135989502220096 metric_logger.py:185] completed step: 1, seconds: 3.061, TFLOP/s/device: 4.438, Tokens/s/device: 668.965, total_weights: 65536, loss: 10.874 I0421 01:01:47.224144 135989502220096 metric_logger.py:185] completed step: 2, seconds: 0.016, TFLOP/s/device: 834.180, Tokens/s/device: 125736.739, total_weights: 65536, loss: 10.262 I0421 01:01:47.453662 135989502220096 metric_logger.py:185] completed step: 3, seconds: 0.835, TFLOP/s/device: 16.273, Tokens/s/device: 2452.888, total_weights: 65536, loss: 9.731 I0421 01:01:47.913179 135989502220096 metric_logger.py:185] completed step: 4, seconds: 0.234, TFLOP/s/device: 57.964, Tokens/s/device: 8736.978, total_weights: 65536, loss: 9.272 I0421 01:01:47.918988 135989502220096 metric_logger.py:185] completed step: 5, seconds: 0.230, TFLOP/s/device: 59.197, Tokens/s/device: 8922.853, total_weights: 65536, loss: 8.887 I0421 01:02:14.480565 135989502220096 metric_logger.py:185] completed step: 6, seconds: 0.460, TFLOP/s/device: 29.559, Tokens/s/device: 4455.506, total_weights: 65536, loss: 8.588 I0421 01:02:14.710228 135989502220096 metric_logger.py:185] completed step: 7, seconds: 26.334, TFLOP/s/device: 0.516, Tokens/s/device: 77.771, total_weights: 65536, loss: 8.380 I0421 01:02:14.939506 135989502220096 metric_logger.py:185] completed step: 8, seconds: 0.234, TFLOP/s/device: 58.186, Tokens/s/device: 8770.502, total_weights: 65536, loss: 8.251 I0421 01:02:15.168839 135989502220096 checkpointing.py:772] Waiting for step 9 to finish before checkpoint... I0421 01:02:15.169524 135989502220096 checkpointing.py:776] Waited 0.0007071495056152344 seconds for step 9 to finish before starting checkpointing. I0421 01:02:15.171500 135989502220096 checkpoint_manager.py:2024] [process=4][thread=MainThread][step=0][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0421 01:02:18.104505 2746 google_auth_provider.cc:181] Running on GCE, using service account 562977990677-compute@developer.gserviceaccount.com I0421 01:02:21.261464 135870594139904 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_04_int8/checkpoints/0/items/array_metadatas/process_4 I0421 01:02:47.387735 135870585747200 base_pytree_checkpoint_handler.py:129] [process=4] /jax/checkpoint/write/gbytes_per_sec: 25.115 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 62.89501667022705 s) (per-host) I0421 01:02:47.387843 135870585747200 async_checkpointer.py:90] [process=4][thread=async_save] 3 Handler Commit operations completed. Time taken: 62.359358s. I0421 01:02:55.829242 135870585747200 async_checkpointer.py:144] [process=4][thread=async_save] Background save thread done. Time taken: 70.800732s. I0421 01:02:55.829454 135860393592576 async_checkpointer.py:273] [process=4][thread=save_finalize] Done with waiting for background save thread=async_save. I0421 01:02:55.829519 135860393592576 async_checkpointer.py:283] [process=4][thread=save_finalize] No errors found in background save thread=async_save. I0421 01:02:55.829579 135860393592576 checkpoint_manager.py:2133] [process=4][thread=save_finalize][step=0] CheckpointManager Save Finalize is syncing with other hosts... I0421 01:02:55.832191 135860393592576 checkpoint_manager.py:2142] [process=4][thread=save_finalize][step=0] CheckpointManager Save Finalize is done on all hosts. I0421 01:02:55.832377 135989502220096 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 01:02:55.832515 135989502220096 checkpoint_manager.py:1458] Waiting for previous save to complete took 40.661018 seconds. If this number is high, consider checkpointing less frequently. I0421 01:02:55.834366 135989502220096 checkpoint_manager.py:1518] [process=4] Saving checkpoint at step 9 I0421 01:02:55.837920 135989502220096 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_04_int8/checkpoints/9. I0421 01:02:57.674433 135989502220096 jax_array_handlers.py:358] Scheduling D2H of 39 prioritized jax.Array. I0421 01:02:57.674536 135989502220096 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 01:02:57.720057 135989502220096 base_pytree_checkpoint_handler.py:153] [process=4][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.046964s I0421 01:02:57.720232 135989502220096 base_pytree_checkpoint_handler.py:129] [process=4] /jax/checkpoint/write/blocking_gbytes_per_sec: 1.332 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 1.1576869487762451 s) (per-host) I0421 01:02:57.720283 135989502220096 base_pytree_checkpoint_handler.py:737] [process=4][thread=MainThread] Initiated Pytree async_save. Time taken: 1.157747s (batch_requests_ready=1.108750s, total_serialization_initiated=0.048930s, others=0.000068s) I0421 01:02:57.720371 135989502220096 composite_checkpoint_handler.py:715] [process=4][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 1.161876s (all_items=0.000015s, per_item={'items': '0.00001526'}, temp_paths=1.161861) I0421 01:02:57.721335 135864151688960 async_checkpointer.py:79] [process=4][thread=async_save] Background save thread started. I0421 01:02:57.721497 135989502220096 async_checkpointer.py:561] Finished blocking save. Time taken: 1.887059s. Continuing background save to gs://lance-maxtext/linen_ckpt_xpk_test_pipeline_scan_nnx_20260421_005327/linen_xpk_test_pipeline_scan_nnx_20260421_005327_04_int8/checkpoints/9. I0421 01:02:57.837684 135989502220096 checkpoint_manager.py:1566] [process=4][thread=MainThread][step=9] Starting CheckpointManager Save Finalize thread=save_finalize I0421 01:02:57.838057 135860393592576 async_checkpointer.py:265] [process=4][thread=save_finalize] Waiting for background save thread=async_save. I0421 01:02:57.838211 135989502220096 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_04_int8/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776733335.1714733, 'wait_for_prev_duration_secs': 40.66101789474487, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776733375.8344054, 'checkpointer_blocking_duration_secs': 1.8871984481811523, 'get_old_steps_start_time': 1776733377.7216227, 'get_old_steps_duration_secs': 2.6702880859375e-05, 'checkpoint_manager_blocking_start_time': 1776733335.1697693, 'checkpoint_manager_blocking_duration_secs': 42.66840434074402} I0421 01:02:57.838369 135989502220096 checkpointing.py:408] Started an asynchronous checkpoint save for step 9 I0421 01:02:57.838418 135989502220096 checkpoint_manager.py:2024] [process=4][thread=MainThread][step=9][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0421 01:03:04.106527 135870594139904 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_04_int8/checkpoints/9/items/array_metadatas/process_4 I0421 01:03:40.534421 135864151688960 base_pytree_checkpoint_handler.py:129] [process=4] /jax/checkpoint/write/gbytes_per_sec: 35.923 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 43.97183299064636 s) (per-host) I0421 01:03:40.534547 135864151688960 async_checkpointer.py:90] [process=4][thread=async_save] 3 Handler Commit operations completed. Time taken: 42.813097s. I0421 01:03:50.391681 135864151688960 async_checkpointer.py:144] [process=4][thread=async_save] Background save thread done. Time taken: 52.670217s. I0421 01:03:50.392006 135860393592576 async_checkpointer.py:273] [process=4][thread=save_finalize] Done with waiting for background save thread=async_save. I0421 01:03:50.392123 135860393592576 async_checkpointer.py:283] [process=4][thread=save_finalize] No errors found in background save thread=async_save. I0421 01:03:50.392180 135860393592576 checkpoint_manager.py:2133] [process=4][thread=save_finalize][step=9] CheckpointManager Save Finalize is syncing with other hosts... I0421 01:03:50.393645 135860393592576 checkpoint_manager.py:2142] [process=4][thread=save_finalize][step=9] CheckpointManager Save Finalize is done on all hosts. I0421 01:03:50.393843 135989502220096 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 01:03:50.393985 135989502220096 checkpoint_manager.py:2013] [process=4][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0421 01:03:50.394672 135989502220096 metric_logger.py:185] completed step: 9, seconds: 0.230, TFLOP/s/device: 59.151, Tokens/s/device: 8915.822, total_weights: 65536, loss: 8.175 Per train step: Total TFLOPs: 13.59 split as 93.93% learnable weight flops and 6.07% attention flops XPK End: Tue Apr 21 01:04:02 UTC 2026 EXIT_CODE=0