MaxView

Timing Test Reference Updates
← Back to run

Log Summary

XPK Start: Sun Apr 19 09:08:12 UTC 2026
2026-04-19 09:08:16.980963: 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 00:00:1776589696.993530      10 cuda_dnn.cc:8579] Unable to register cuDNN factory: Attempting to register factory for plugin cuDNN when one has already been registered
E0000 00:00:1776589696.997283      10 cuda_blas.cc:1407] Unable to register cuBLAS factory: Attempting to register factory for plugin cuBLAS when one has already been registered
W0000 00:00:1776589697.008587      10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once.
W0000 00:00:1776589697.008605      10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once.
W0000 00:00:1776589697.008607      10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once.
W0000 00:00:1776589697.008609      10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once.
2026-04-19 09:08:36.058232: 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 09:08:36.568255 138570881185600 max_utils.py:273] Attempting to initialize the jax distributed system...
INFO:2026-04-19 09:08:45,607:jax._src.distributed:140: Starting JAX distributed service on [::]:8482
I0419 09:08:45.607895 138570881185600 distributed.py:140] Starting JAX distributed service on [::]:8482
INFO:2026-04-19 09:08:45,610:jax._src.distributed:157: Connecting to JAX distributed service on mt-05-fp8-o9lys-slice-job-0-0.mt-05-fp8-o9lys:8482
I0419 09:08:45.610202 138570881185600 distributed.py:157] Connecting to JAX distributed service on mt-05-fp8-o9lys-slice-job-0-0.mt-05-fp8-o9lys:8482
I0419 09:08:47.262533 138570881185600 max_utils.py:284] Jax distributed system initialized!
I0419 09:08:54.125870 138570881185600 max_utils.py:800] System Information: Jax Version: 0.8.1
I0419 09:08:54.125997 138570881185600 max_utils.py:801] System Information: Jaxlib Version: 0.8.1
I0419 09:08:54.126044 138570881185600 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 09:08:54.126080 138570881185600 train_utils.py:347] WARNING: Sequence packing is essentially ignored for synthetic data. Please use a real dataset to use sequence packing.
I0419 09:08:54.802915 138570881185600 maxtext_utils.py:1398] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1)
I0419 09:08:54.803229 138570881185600 checkpointing.py:677] Setting up checkpoint logger...
I0419 09:08:54.803286 138570881185600 checkpointing.py:233] Creating checkpoint manager with ocdbt=True and zarr3=True
I0419 09:08:54.803330 138570881185600 pytree_checkpoint_handler.py:589] save_device_host_concurrent_bytes=None
I0419 09:08:54.803681 138570881185600 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 0x7e0710563c20>, enable_pinned_host_transfer=False, save_concurrent_bytes: 96000000000 (89.4 GiB), restore_concurrent_bytes: 96000000000 (89.4 GiB)
I0419 09:08:57.514712 138570881185600 checkpointing.py:265] Enabling policy for fixed interval checkpointing.
I0419 09:08:57.515091 138570881185600 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 0x7df3fda9be00>}, handler_registry=None
I0419 09:08:57.515331 138570881185600 composite_checkpoint_handler.py:237] Deferred registration for item: "items". Adding handler `<orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7df3fda9be00>` 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 09:08:57.515380 138570881185600 composite_checkpoint_handler.py:237] Deferred registration for item: "metrics". Adding handler `<orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7df3fda02ba0>` 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 09:08:57.515417 138570881185600 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 0x7df3fda9be00>, ('items', <class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>): <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7df3fda9be00>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7df3fda02ba0>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7df3fda02ba0>}).
I0419 09:08:57.518047 138570881185600 abstract_checkpointer.py:35] orbax-checkpoint version: 0.11.33
I0419 09:08:57.518131 138570881185600 async_checkpointer.py:177] [process=4][thread=MainThread] Using barrier_sync_fn: <function get_barrier_sync_fn.<locals>._fn at 0x7e06ec1d72e0> timeout: 600 secs and primary_host=0 for async checkpoint writes
I0419 09:08:59.226549 138570881185600 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_05_fp8/checkpoints
I0419 09:08:59.738237 138570881185600 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_20260419_085454/linen_xpk_test_pipeline_scan_nnx_20260419_085454_05_fp8/checkpoints: <orbax.checkpoint.checkpoint_manager.CheckpointManager object at 0x7df3fda00f20>
I0419 09:08:59.738412 138570881185600 checkpointing.py:301] Checkpoint manager created!
I0419 09:09:00.840302 138570881185600 nnx_wrappers.py:453] Unknown Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed').
I0419 09:09:00.840409 138570881185600 nnx_wrappers.py:453] Unknown Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None).
I0419 09:09:01.216519 138570881185600 attentions.py:1084] attentions/inputs_q Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed').
I0419 09:09:01.216615 138570881185600 attentions.py:1084] attentions/inputs_q Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None).
I0419 09:09:01.232243 138570881185600 attentions.py:1085] attentions/inputs_kv Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed').
I0419 09:09:01.232306 138570881185600 attentions.py:1085] attentions/inputs_kv Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None).
I0419 09:09:01.290857 138570881185600 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 09:09:01.290951 138570881185600 attentions.py:1150] attentions/query Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None).
I0419 09:09:01.306638 138570881185600 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 09:09:01.306699 138570881185600 attentions.py:1151] attentions/key Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None).
I0419 09:09:01.322251 138570881185600 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 09:09:01.322315 138570881185600 attentions.py:1152] attentions/value Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None).
I0419 09:09:01.346158 138570881185600 attentions.py:1193] attentions/out Logical: bfloat16[32,2048,16,128].................................... ('activation_batch', 'activation_attn_length', 'activation_heads', 'activation_kv').
I0419 09:09:01.346230 138570881185600 attentions.py:1193] attentions/out Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None).
I0419 09:09:01.405398 138570881185600 linears.py:525] linears/x Logical: bfloat16[32,2048,7168]...................................... ('activation_batch', 'activation_length', 'activation_mlp').
I0419 09:09:01.405483 138570881185600 linears.py:525] linears/x Physical: bfloat16[32,2048,7168]...................................... ('fsdp', None, None).
I0419 09:09:01.867565 138570881185600 checkpointing.py:577] checkpoint manager exists so trying to load this run's existing checkpoint
I0419 09:09:01.867682 138570881185600 checkpointing.py:665] No existing checkpoints found, not restoring checkpoint.
fsdp: 32
I0419 09:09:03.907149 138570881185600 maxtext_utils.py:1501]  params/_overwrite_with_gradient/decoder/layers/mlp/wi_0/Fp8DirectDotGeneralOp_0/input_amax_history
    Shape:     float32[16,1024]
    Logical:   PartitionSpec()
    Physical:  ()
I0419 09:09:03.907277 138570881185600 maxtext_utils.py:1501]  params/_overwrite_with_gradient/decoder/layers/mlp/wi_0/Fp8DirectDotGeneralOp_0/input_scale
    Shape:     float32[16,1]
    Logical:   PartitionSpec()
    Physical:  ()
I0419 09:09:03.907325 138570881185600 maxtext_utils.py:1501]  params/_overwrite_with_gradient/decoder/layers/mlp/wi_0/Fp8DirectDotGeneralOp_0/kernel_amax_history
    Shape:     float32[16,1024]
    Logical:   PartitionSpec()
    Physical:  ()
I0419 09:09:03.907364 138570881185600 maxtext_utils.py:1501]  params/_overwrite_with_gradient/decoder/layers/mlp/wi_0/Fp8DirectDotGeneralOp_0/kernel_scale
    Shape:     float32[16,1]
    Logical:   PartitionSpec()
    Physical:  ()
I0419 09:09:03.907399 138570881185600 maxtext_utils.py:1501]  params/_overwrite_with_gradient/decoder/layers/mlp/wi_0/Fp8DirectDotGeneralOp_0/output_grad_amax_history
    Shape:     float32[16,1024]
    Logical:   PartitionSpec()
    Physical:  ()
I0419 09:09:03.907435 138570881185600 maxtext_utils.py:1501]  params/_overwrite_with_gradient/decoder/layers/mlp/wi_0/Fp8DirectDotGeneralOp_0/output_grad_scale
    Shape:     float32[16,1]
    Logical:   PartitionSpec()
    Physical:  ()
I0419 09:09:03.907467 138570881185600 maxtext_utils.py:1501]  params/_overwrite_with_gradient/decoder/layers/mlp/wi_1/Fp8DirectDotGeneralOp_0/input_amax_history
    Shape:     float32[16,1024]
    Logical:   PartitionSpec()
    Physical:  ()
I0419 09:09:03.907496 138570881185600 maxtext_utils.py:1501]  params/_overwrite_with_gradient/decoder/layers/mlp/wi_1/Fp8DirectDotGeneralOp_0/input_scale
    Shape:     float32[16,1]
    Logical:   PartitionSpec()
    Physical:  ()
I0419 09:09:03.907526 138570881185600 maxtext_utils.py:1501]  params/_overwrite_with_gradient/decoder/layers/mlp/wi_1/Fp8DirectDotGeneralOp_0/kernel_amax_history
    Shape:     float32[16,1024]
    Logical:   PartitionSpec()
    Physical:  ()
I0419 09:09:03.907554 138570881185600 maxtext_utils.py:1501]  params/_overwrite_with_gradient/decoder/layers/mlp/wi_1/Fp8DirectDotGeneralOp_0/kernel_scale
    Shape:     float32[16,1]
    Logical:   PartitionSpec()
    Physical:  ()
I0419 09:09:03.907582 138570881185600 maxtext_utils.py:1501]  params/_overwrite_with_gradient/decoder/layers/mlp/wi_1/Fp8DirectDotGeneralOp_0/output_grad_amax_history
    Shape:     float32[16,1024]
    Logical:   PartitionSpec()
    Physical:  ()
I0419 09:09:03.907612 138570881185600 maxtext_utils.py:1501]  params/_overwrite_with_gradient/decoder/layers/mlp/wi_1/Fp8DirectDotGeneralOp_0/output_grad_scale
    Shape:     float32[16,1]
    Logical:   PartitionSpec()
    Physical:  ()
I0419 09:09:03.907641 138570881185600 maxtext_utils.py:1501]  params/_overwrite_with_gradient/decoder/layers/mlp/wo/Fp8DirectDotGeneralOp_0/input_amax_history
    Shape:     float32[16,1024]
    Logical:   PartitionSpec()
    Physical:  ()
I0419 09:09:03.907671 138570881185600 maxtext_utils.py:1501]  params/_overwrite_with_gradient/decoder/layers/mlp/wo/Fp8DirectDotGeneralOp_0/input_scale
    Shape:     float32[16,1]
    Logical:   PartitionSpec()
    Physical:  ()
I0419 09:09:03.907700 138570881185600 maxtext_utils.py:1501]  params/_overwrite_with_gradient/decoder/layers/mlp/wo/Fp8DirectDotGeneralOp_0/kernel_amax_history
    Shape:     float32[16,1024]
    Logical:   PartitionSpec()
    Physical:  ()
I0419 09:09:03.907729 138570881185600 maxtext_utils.py:1501]  params/_overwrite_with_gradient/decoder/layers/mlp/wo/Fp8DirectDotGeneralOp_0/kernel_scale
    Shape:     float32[16,1]
    Logical:   PartitionSpec()
    Physical:  ()
I0419 09:09:03.907760 138570881185600 maxtext_utils.py:1501]  params/_overwrite_with_gradient/decoder/layers/mlp/wo/Fp8DirectDotGeneralOp_0/output_grad_amax_history
    Shape:     float32[16,1024]
    Logical:   PartitionSpec()
    Physical:  ()
I0419 09:09:03.907790 138570881185600 maxtext_utils.py:1501]  params/_overwrite_with_gradient/decoder/layers/mlp/wo/Fp8DirectDotGeneralOp_0/output_grad_scale
    Shape:     float32[16,1]
    Logical:   PartitionSpec()
    Physical:  ()
I0419 09:09:03.907818 138570881185600 maxtext_utils.py:1501]  params/_overwrite_with_gradient/decoder/layers/self_attention/key/Fp8DirectDotGeneralOp_0/input_amax_history
    Shape:     float32[16,1024]
    Logical:   PartitionSpec()
    Physical:  ()
I0419 09:09:03.907846 138570881185600 maxtext_utils.py:1501]  params/_overwrite_with_gradient/decoder/layers/self_attention/key/Fp8DirectDotGeneralOp_0/input_scale
    Shape:     float32[16,1]
    Logical:   PartitionSpec()
    Physical:  ()
I0419 09:09:03.907873 138570881185600 maxtext_utils.py:1501]  params/_overwrite_with_gradient/decoder/layers/self_attention/key/Fp8DirectDotGeneralOp_0/kernel_amax_history
    Shape:     float32[16,1024]
    Logical:   PartitionSpec()
    Physical:  ()
I0419 09:09:03.907899 138570881185600 maxtext_utils.py:1501]  params/_overwrite_with_gradient/decoder/layers/self_attention/key/Fp8DirectDotGeneralOp_0/kernel_scale
    Shape:     float32[16,1]
    Logical:   PartitionSpec()
    Physical:  ()
I0419 09:09:03.907927 138570881185600 maxtext_utils.py:1501]  params/_overwrite_with_gradient/decoder/layers/self_attention/key/Fp8DirectDotGeneralOp_0/output_grad_amax_history
    Shape:     float32[16,1024]
    Logical:   PartitionSpec()
    Physical:  ()
I0419 09:09:03.907982 138570881185600 maxtext_utils.py:1501]  params/_overwrite_with_gradient/decoder/layers/self_attention/key/Fp8DirectDotGeneralOp_0/output_grad_scale
    Shape:     float32[16,1]
    Logical:   PartitionSpec()
    Physical:  ()
I0419 09:09:03.908022 138570881185600 maxtext_utils.py:1501]  params/_overwrite_with_gradient/decoder/layers/self_attention/out/Fp8DirectDotGeneralOp_0/input_amax_history
    Shape:     float32[16,1024]
    Logical:   PartitionSpec()
    Physical:  ()
I0419 09:09:03.908051 138570881185600 maxtext_utils.py:1501]  params/_overwrite_with_gradient/decoder/layers/self_attention/out/Fp8DirectDotGeneralOp_0/input_scale
    Shape:     float32[16,1]
    Logical:   PartitionSpec()
    Physical:  ()
I0419 09:09:03.908079 138570881185600 maxtext_utils.py:1501]  params/_overwrite_with_gradient/decoder/layers/self_attention/out/Fp8DirectDotGeneralOp_0/kernel_amax_history
    Shape:     float32[16,1024]
    Logical:   PartitionSpec()
    Physical:  ()
I0419 09:09:03.908108 138570881185600 maxtext_utils.py:1501]  params/_overwrite_with_gradient/decoder/layers/self_attention/out/Fp8DirectDotGeneralOp_0/kernel_scale
    Shape:     float32[16,1]
    Logical:   PartitionSpec()
    Physical:  ()
I0419 09:09:03.908136 138570881185600 maxtext_utils.py:1501]  params/_overwrite_with_gradient/decoder/layers/self_attention/out/Fp8DirectDotGeneralOp_0/output_grad_amax_history
    Shape:     float32[16,1024]
    Logical:   PartitionSpec()
    Physical:  ()
I0419 09:09:03.908163 138570881185600 maxtext_utils.py:1501]  params/_overwrite_with_gradient/decoder/layers/self_attention/out/Fp8DirectDotGeneralOp_0/output_grad_scale
    Shape:     float32[16,1]
    Logical:   PartitionSpec()
    Physical:  ()
I0419 09:09:03.908189 138570881185600 maxtext_utils.py:1501]  params/_overwrite_with_gradient/decoder/layers/self_attention/query/Fp8DirectDotGeneralOp_0/input_amax_history
    Shape:     float32[16,1024]
    Logical:   PartitionSpec()
    Physical:  ()
I0419 09:09:03.908216 138570881185600 maxtext_utils.py:1501]  params/_overwrite_with_gradient/decoder/layers/self_attention/query/Fp8DirectDotGeneralOp_0/input_scale
    Shape:     float32[16,1]
    Logical:   PartitionSpec()
    Physical:  ()
I0419 09:09:03.908243 138570881185600 maxtext_utils.py:1501]  params/_overwrite_with_gradient/decoder/layers/self_attention/query/Fp8DirectDotGeneralOp_0/kernel_amax_history
    Shape:     float32[16,1024]
    Logical:   PartitionSpec()
    Physical:  ()
I0419 09:09:03.908269 138570881185600 maxtext_utils.py:1501]  params/_overwrite_with_gradient/decoder/layers/self_attention/query/Fp8DirectDotGeneralOp_0/kernel_scale
    Shape:     float32[16,1]
    Logical:   PartitionSpec()
    Physical:  ()
I0419 09:09:03.908296 138570881185600 maxtext_utils.py:1501]  params/_overwrite_with_gradient/decoder/layers/self_attention/query/Fp8DirectDotGeneralOp_0/output_grad_amax_history
    Shape:     float32[16,1024]
    Logical:   PartitionSpec()
    Physical:  ()
I0419 09:09:03.908322 138570881185600 maxtext_utils.py:1501]  params/_overwrite_with_gradient/decoder/layers/self_attention/query/Fp8DirectDotGeneralOp_0/output_grad_scale
    Shape:     float32[16,1]
    Logical:   PartitionSpec()
    Physical:  ()
I0419 09:09:03.908349 138570881185600 maxtext_utils.py:1501]  params/_overwrite_with_gradient/decoder/layers/self_attention/value/Fp8DirectDotGeneralOp_0/input_amax_history
    Shape:     float32[16,1024]
    Logical:   PartitionSpec()
    Physical:  ()
I0419 09:09:03.908390 138570881185600 maxtext_utils.py:1501]  params/_overwrite_with_gradient/decoder/layers/self_attention/value/Fp8DirectDotGeneralOp_0/input_scale
    Shape:     float32[16,1]
    Logical:   PartitionSpec()
    Physical:  ()
I0419 09:09:03.908423 138570881185600 maxtext_utils.py:1501]  params/_overwrite_with_gradient/decoder/layers/self_attention/value/Fp8DirectDotGeneralOp_0/kernel_amax_history
    Shape:     float32[16,1024]
    Logical:   PartitionSpec()
    Physical:  ()
I0419 09:09:03.908451 138570881185600 maxtext_utils.py:1501]  params/_overwrite_with_gradient/decoder/layers/self_attention/value/Fp8DirectDotGeneralOp_0/kernel_scale
    Shape:     float32[16,1]
    Logical:   PartitionSpec()
    Physical:  ()
I0419 09:09:03.908484 138570881185600 maxtext_utils.py:1501]  params/_overwrite_with_gradient/decoder/layers/self_attention/value/Fp8DirectDotGeneralOp_0/output_grad_amax_history
    Shape:     float32[16,1024]
    Logical:   PartitionSpec()
    Physical:  ()
I0419 09:09:03.908512 138570881185600 maxtext_utils.py:1501]  params/_overwrite_with_gradient/decoder/layers/self_attention/value/Fp8DirectDotGeneralOp_0/output_grad_scale
    Shape:     float32[16,1]
    Logical:   PartitionSpec()
    Physical:  ()
I0419 09:09:03.908563 138570881185600 maxtext_utils.py:1501]  params/params/decoder/decoder_norm/scale
    Shape:     float32[2048]
    Logical:   PartitionSpec('norm',)
    Physical:  (None,)
I0419 09:09:03.908619 138570881185600 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 09:09:03.908657 138570881185600 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 09:09:03.908705 138570881185600 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 09:09:03.908751 138570881185600 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 09:09:03.908785 138570881185600 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 09:09:03.908832 138570881185600 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 09:09:03.908877 138570881185600 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 09:09:03.908913 138570881185600 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 09:09:03.909002 138570881185600 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 09:09:03.909089 138570881185600 maxtext_utils.py:1501]  params/params/decoder/logits_dense/kernel
    Shape:     float32[2048,32000]
    Logical:   PartitionSpec('embed', 'vocab')
    Physical:  ('fsdp', None)

I0419 09:09:03.909148 138570881185600 maxtext_utils.py:1501]  params/params/token_embedder/embedding
    Shape:     float32[32000,2048]
    Logical:   PartitionSpec('vocab', 'embed')
    Physical:  (None, 'fsdp')
I0419 09:09:05.681148 138570881185600 train.py:156] train/xent Logical: float32[32,2048]............................................ ('activation_embed_and_logits_batch', 'activation_length').
I0419 09:09:05.681243 138570881185600 train.py:156] train/xent Physical: float32[32,2048]............................................ ('fsdp', None).
I0419 09:09:05.696092 138570881185600 train.py:163] train/z_loss Logical: float32[32,2048]............................................ ('activation_embed_and_logits_batch', 'activation_length').
I0419 09:09:05.696153 138570881185600 train.py:163] train/z_loss Physical: float32[32,2048]............................................ ('fsdp', None).
I0419 09:09:18.773150 138570881185600 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 09:09:18.774204 138570881185600 metric_logger.py:289] number parameters: 1.105 billion
I0419 09:09:33.215782 138570881185600 checkpointing.py:772] Waiting for step 0 to finish before checkpoint...
I0419 09:09:33.460000 138570881185600 checkpointing.py:776] Waited 0.24419736862182617 seconds for step 0 to finish before starting checkpointing.
I0419 09:09:33.462663 138570881185600 checkpoint_manager.py:2013] [process=4][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning.
I0419 09:09:33.464574 138570881185600 checkpoint_manager.py:1518] [process=4] Saving checkpoint at step 0
I0419 09:09:33.467405 138570881185600 async_checkpointer.py:452] [process=4] 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_05_fp8/checkpoints/0.
I0419 09:09:34.459261 138570881185600 signaling_client.py:364] Using JaxDistributedSignalingClient
I0419 09:09:34.460499 138570881185600 jax_array_handlers.py:358] Scheduling D2H of 81 prioritized jax.Array.
I0419 09:09:34.460559 138570881185600 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 09:09:34.823082 138570881185600 base_pytree_checkpoint_handler.py:153] [process=4][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.363945s
I0419 09:09:34.823246 138570881185600 base_pytree_checkpoint_handler.py:129] [process=4] /jax/checkpoint/write/blocking_gbytes_per_sec: 2.501 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.6167466640472412 s) (per-host)
I0419 09:09:34.823293 138570881185600 base_pytree_checkpoint_handler.py:737] [process=4][thread=MainThread] Initiated Pytree async_save. Time taken: 0.616803s (batch_requests_ready=0.247687s, total_serialization_initiated=0.369051s, others=0.000065s)
I0419 09:09:34.823392 138570881185600 composite_checkpoint_handler.py:715] [process=4][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.620937s (all_items=0.000020s, per_item={'items': '0.00002003'}, temp_paths=0.620917)
I0419 09:09:34.824294 138448027494144 async_checkpointer.py:79] [process=4][thread=async_save] Background save thread started.
I0419 09:09:34.824464 138570881185600 async_checkpointer.py:561] Finished blocking save. Time taken: 1.359823s. Continuing background save to gs://lance-maxtext/linen_ckpt_xpk_test_pipeline_scan_nnx_20260419_085454/linen_xpk_test_pipeline_scan_nnx_20260419_085454_05_fp8/checkpoints/0.
I0419 09:09:35.279509 138570881185600 checkpoint_manager.py:1566] [process=4][thread=MainThread][step=0] Starting CheckpointManager Save Finalize thread=save_finalize
I0419 09:09:35.279884 138441190782720 async_checkpointer.py:265] [process=4][thread=save_finalize] Waiting for background save thread=async_save.
I0419 09:09:35.280086 138570881185600 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_05_fp8/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776589773.4626458, 'wait_for_prev_duration_secs': 6.175041198730469e-05, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776589773.4646132, 'checkpointer_blocking_duration_secs': 1.359954595565796, 'get_old_steps_start_time': 1776589774.8245876, 'get_old_steps_duration_secs': 2.5510787963867188e-05, 'checkpoint_manager_blocking_start_time': 1776589773.4607778, 'checkpoint_manager_blocking_duration_secs': 1.8192667961120605}
I0419 09:09:35.280200 138570881185600 checkpointing.py:408] Started an asynchronous checkpoint save for step 0
I0419 09:09:35.280254 138570881185600 max_utils.py:750] 
Memstats: After params initialized:
I0419 09:09:35.280312 138570881185600 max_utils.py:756] 	Using (GB) 0.45 / 31.25 (1.440000%) on TPU_16(process=4,(0,4,0,0))
I0419 09:09:35.280347 138570881185600 max_utils.py:756] 	Using (GB) 0.45 / 31.25 (1.440000%) on TPU_17(process=4,(1,4,0,0))
I0419 09:09:35.280376 138570881185600 max_utils.py:756] 	Using (GB) 0.45 / 31.25 (1.440000%) on TPU_20(process=4,(0,5,0,0))
I0419 09:09:35.280401 138570881185600 max_utils.py:756] 	Using (GB) 0.45 / 31.25 (1.440000%) on TPU_21(process=4,(1,5,0,0))
I0419 09:09:35.545837 138570881185600 metric_logger.py:185] completed step: 0, seconds: 14.441, TFLOP/s/device: 0.941, Tokens/s/device: 141.814, total_weights: 65536, loss: 10.874
I0419 09:09:35.723007 138570881185600 metric_logger.py:185] completed step: 1, seconds: 2.328, TFLOP/s/device: 5.836, Tokens/s/device: 879.667, total_weights: 65536, loss: 10.874
I0419 09:09:36.130380 138570881185600 metric_logger.py:185] completed step: 2, seconds: 0.020, TFLOP/s/device: 671.268, Tokens/s/device: 101180.772, total_weights: 65536, loss: 10.816
I0419 09:09:36.288550 138570881185600 metric_logger.py:185] completed step: 3, seconds: 0.408, TFLOP/s/device: 33.324, Tokens/s/device: 5022.956, total_weights: 65536, loss: 10.431
I0419 09:09:36.605765 138570881185600 metric_logger.py:185] completed step: 4, seconds: 0.163, TFLOP/s/device: 83.180, Tokens/s/device: 12537.803, total_weights: 65536, loss: 9.992
I0419 09:09:36.611871 138570881185600 metric_logger.py:185] completed step: 5, seconds: 0.158, TFLOP/s/device: 85.975, Tokens/s/device: 12959.073, total_weights: 65536, loss: 9.549
I0419 09:09:56.665903    2726 google_auth_provider.cc:181] Running on GCE, using service account 562977990677-compute@developer.gserviceaccount.com
I0419 09:09:56.853344 138570881185600 metric_logger.py:185] completed step: 6, seconds: 0.318, TFLOP/s/device: 42.787, Tokens/s/device: 6449.297, total_weights: 65536, loss: 9.111
I0419 09:09:57.011599 138570881185600 metric_logger.py:185] completed step: 7, seconds: 20.079, TFLOP/s/device: 0.677, Tokens/s/device: 101.998, total_weights: 65536, loss: 8.685
I0419 09:09:57.169702 138570881185600 metric_logger.py:185] completed step: 8, seconds: 0.169, TFLOP/s/device: 80.631, Tokens/s/device: 12153.654, total_weights: 65536, loss: 8.281
I0419 09:09:57.174690 138570881185600 checkpointing.py:772] Waiting for step 10 to finish before checkpoint...
I0419 09:09:57.486382 138570881185600 checkpointing.py:776] Waited 0.3116581439971924 seconds for step 10 to finish before starting checkpointing.
I0419 09:09:57.489093 138570881185600 checkpoint_manager.py:2024] [process=4][thread=MainThread][step=0][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete.
I0419 09:09:58.514986 138448035886848 array_metadata_store.py:203] [process=4][thread=array_type_handler] Wrote 81 array_metadata.ArrayMetadata to gs://lance-maxtext/linen_ckpt_xpk_test_pipeline_scan_nnx_20260419_085454/linen_xpk_test_pipeline_scan_nnx_20260419_085454_05_fp8/checkpoints/0/items/array_metadatas/process_4
I0419 09:10:13.908863 138448027494144 base_pytree_checkpoint_handler.py:129] [process=4] /jax/checkpoint/write/gbytes_per_sec: 39.790 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 39.70232677459717 s) (per-host)
I0419 09:10:13.908988 138448027494144 async_checkpointer.py:90] [process=4][thread=async_save] 3 Handler Commit operations completed. Time taken: 39.084581s.
I0419 09:10:22.149784 138448027494144 async_checkpointer.py:144] [process=4][thread=async_save] Background save thread done. Time taken: 47.325360s.
I0419 09:10:22.150073 138441190782720 async_checkpointer.py:273] [process=4][thread=save_finalize] Done with waiting for background save thread=async_save.
I0419 09:10:22.150202 138441190782720 async_checkpointer.py:283] [process=4][thread=save_finalize] No errors found in background save thread=async_save.
I0419 09:10:22.150254 138441190782720 checkpoint_manager.py:2133] [process=4][thread=save_finalize][step=0] CheckpointManager Save Finalize is syncing with other hosts...
I0419 09:10:22.151897 138441190782720 checkpoint_manager.py:2142] [process=4][thread=save_finalize][step=0] CheckpointManager Save Finalize is done on all hosts.
I0419 09:10:22.152091 138570881185600 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.
W0419 09:10:22.152219 138570881185600 checkpoint_manager.py:1458] Waiting for previous save to complete took 24.663133 seconds. If this number is high, consider checkpointing less frequently.
I0419 09:10:22.154110 138570881185600 checkpoint_manager.py:1518] [process=4] Saving checkpoint at step 10
I0419 09:10:22.157469 138570881185600 async_checkpointer.py:452] [process=4] 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_05_fp8/checkpoints/10.
I0419 09:10:23.126567 138570881185600 jax_array_handlers.py:358] Scheduling D2H of 81 prioritized jax.Array.
I0419 09:10:23.126666 138570881185600 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 09:10:23.186091 138570881185600 base_pytree_checkpoint_handler.py:153] [process=4][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.060643s
I0419 09:10:23.186256 138570881185600 base_pytree_checkpoint_handler.py:129] [process=4] /jax/checkpoint/write/blocking_gbytes_per_sec: 4.903 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.3146815299987793 s) (per-host)
I0419 09:10:23.186307 138570881185600 base_pytree_checkpoint_handler.py:737] [process=4][thread=MainThread] Initiated Pytree async_save. Time taken: 0.314742s (batch_requests_ready=0.250629s, total_serialization_initiated=0.064045s, others=0.000068s)
I0419 09:10:23.186405 138570881185600 composite_checkpoint_handler.py:715] [process=4][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.318842s (all_items=0.000015s, per_item={'items': '0.00001526'}, temp_paths=0.318827)
I0419 09:10:23.187240 138439605348096 async_checkpointer.py:79] [process=4][thread=async_save] Background save thread started.
I0419 09:10:23.187408 138570881185600 async_checkpointer.py:561] Finished blocking save. Time taken: 1.033222s. Continuing background save to gs://lance-maxtext/linen_ckpt_xpk_test_pipeline_scan_nnx_20260419_085454/linen_xpk_test_pipeline_scan_nnx_20260419_085454_05_fp8/checkpoints/10.
I0419 09:10:23.201360 138570881185600 checkpoint_manager.py:1566] [process=4][thread=MainThread][step=10] Starting CheckpointManager Save Finalize thread=save_finalize
I0419 09:10:23.201662 138441190782720 async_checkpointer.py:265] [process=4][thread=save_finalize] Waiting for background save thread=async_save.
I0419 09:10:23.201831 138570881185600 standard_logger.py:34] {'step': 10, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_test_pipeline_scan_nnx_20260419_085454/linen_xpk_test_pipeline_scan_nnx_20260419_085454_05_fp8/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776589797.4890637, 'wait_for_prev_duration_secs': 24.663132905960083, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776589822.1541483, 'checkpointer_blocking_duration_secs': 1.0333852767944336, 'get_old_steps_start_time': 1776589823.1875515, 'get_old_steps_duration_secs': 2.574920654296875e-05, 'checkpoint_manager_blocking_start_time': 1776589797.4873035, 'checkpoint_manager_blocking_duration_secs': 25.714492082595825}
I0419 09:10:23.201974 138570881185600 checkpointing.py:408] Started an asynchronous checkpoint save for step 10
I0419 09:10:23.202387 138570881185600 metric_logger.py:185] completed step: 9, seconds: 0.158, TFLOP/s/device: 85.927, Tokens/s/device: 12951.861, total_weights: 65536, loss: 7.908
I0419 09:10:23.209375 138570881185600 metric_logger.py:185] completed step: 10, seconds: 0.158, TFLOP/s/device: 85.953, Tokens/s/device: 12955.711, total_weights: 65536, loss: 7.568
I0419 09:10:23.366935 138570881185600 metric_logger.py:185] completed step: 11, seconds: 26.033, TFLOP/s/device: 0.522, Tokens/s/device: 78.670, total_weights: 65536, loss: 7.288
I0419 09:10:23.525027 138570881185600 metric_logger.py:185] completed step: 12, seconds: 0.006, TFLOP/s/device: 2228.129, Tokens/s/device: 335847.819, total_weights: 65536, loss: 7.032
I0419 09:10:23.683244 138570881185600 metric_logger.py:185] completed step: 13, seconds: 0.158, TFLOP/s/device: 85.731, Tokens/s/device: 12922.277, total_weights: 65536, loss: 6.815
I0419 09:10:23.841460 138570881185600 metric_logger.py:185] completed step: 14, seconds: 0.158, TFLOP/s/device: 85.850, Tokens/s/device: 12940.240, total_weights: 65536, loss: 6.635
I0419 09:10:23.999649 138570881185600 metric_logger.py:185] completed step: 15, seconds: 0.158, TFLOP/s/device: 85.961, Tokens/s/device: 12956.941, total_weights: 65536, loss: 6.492
I0419 09:10:24.157835 138570881185600 metric_logger.py:185] completed step: 16, seconds: 0.158, TFLOP/s/device: 85.854, Tokens/s/device: 12940.812, total_weights: 65536, loss: 6.380
I0419 09:10:24.316287 138570881185600 metric_logger.py:185] completed step: 17, seconds: 0.158, TFLOP/s/device: 85.946, Tokens/s/device: 12954.728, total_weights: 65536, loss: 6.293
I0419 09:10:24.474337 138570881185600 metric_logger.py:185] completed step: 18, seconds: 0.158, TFLOP/s/device: 85.859, Tokens/s/device: 12941.630, total_weights: 65536, loss: 6.223
I0419 09:10:24.632222 138570881185600 checkpointing.py:772] Waiting for step 19 to finish before checkpoint...
I0419 09:10:24.633304 138570881185600 checkpointing.py:776] Waited 0.0010993480682373047 seconds for step 19 to finish before starting checkpointing.
I0419 09:10:24.635248 138570881185600 checkpoint_manager.py:2024] [process=4][thread=MainThread][step=10][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete.
I0419 09:10:28.384821 138441719260928 array_metadata_store.py:203] [process=4][thread=array_type_handler] Wrote 81 array_metadata.ArrayMetadata to gs://lance-maxtext/linen_ckpt_xpk_test_pipeline_scan_nnx_20260419_085454/linen_xpk_test_pipeline_scan_nnx_20260419_085454_05_fp8/checkpoints/10/items/array_metadatas/process_4
I0419 09:11:04.603105 138439605348096 base_pytree_checkpoint_handler.py:129] [process=4] /jax/checkpoint/write/gbytes_per_sec: 37.855 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 41.731486797332764 s) (per-host)
I0419 09:11:04.603228 138439605348096 async_checkpointer.py:90] [process=4][thread=async_save] 3 Handler Commit operations completed. Time taken: 41.415873s.
I0419 09:11:13.211492 138439605348096 async_checkpointer.py:144] [process=4][thread=async_save] Background save thread done. Time taken: 50.024120s.
I0419 09:11:13.211804 138441190782720 async_checkpointer.py:273] [process=4][thread=save_finalize] Done with waiting for background save thread=async_save.
I0419 09:11:13.211925 138441190782720 async_checkpointer.py:283] [process=4][thread=save_finalize] No errors found in background save thread=async_save.
I0419 09:11:13.211994 138441190782720 checkpoint_manager.py:2133] [process=4][thread=save_finalize][step=10] CheckpointManager Save Finalize is syncing with other hosts...
I0419 09:11:13.213320 138441190782720 checkpoint_manager.py:2142] [process=4][thread=save_finalize][step=10] CheckpointManager Save Finalize is done on all hosts.
I0419 09:11:13.213518 138570881185600 checkpoint_manager.py:2036] [process=4][thread=MainThread][step=10][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=10.
W0419 09:11:13.213636 138570881185600 checkpoint_manager.py:1458] Waiting for previous save to complete took 48.578398 seconds. If this number is high, consider checkpointing less frequently.
I0419 09:11:13.215109 138570881185600 checkpoint_manager.py:1518] [process=4] Saving checkpoint at step 19
I0419 09:11:13.218575 138570881185600 async_checkpointer.py:452] [process=4] 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_05_fp8/checkpoints/19.
I0419 09:11:14.164991 138570881185600 jax_array_handlers.py:358] Scheduling D2H of 81 prioritized jax.Array.
I0419 09:11:14.165091 138570881185600 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 09:11:14.224373 138570881185600 base_pytree_checkpoint_handler.py:153] [process=4][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.060645s
I0419 09:11:14.224534 138570881185600 base_pytree_checkpoint_handler.py:129] [process=4] /jax/checkpoint/write/blocking_gbytes_per_sec: 4.922 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.313457727432251 s) (per-host)
I0419 09:11:14.224585 138570881185600 base_pytree_checkpoint_handler.py:737] [process=4][thread=MainThread] Initiated Pytree async_save. Time taken: 0.313519s (batch_requests_ready=0.249477s, total_serialization_initiated=0.063976s, others=0.000065s)
I0419 09:11:14.224707 138570881185600 composite_checkpoint_handler.py:715] [process=4][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.317649s (all_items=0.000011s, per_item={'items': '0.00001144'}, temp_paths=0.317638)
I0419 09:11:14.225650 138441719260928 async_checkpointer.py:79] [process=4][thread=async_save] Background save thread started.
I0419 09:11:14.225809 138570881185600 async_checkpointer.py:561] Finished blocking save. Time taken: 1.010630s. Continuing background save to gs://lance-maxtext/linen_ckpt_xpk_test_pipeline_scan_nnx_20260419_085454/linen_xpk_test_pipeline_scan_nnx_20260419_085454_05_fp8/checkpoints/19.
I0419 09:11:14.663342 138570881185600 checkpoint_manager.py:1566] [process=4][thread=MainThread][step=19] Starting CheckpointManager Save Finalize thread=save_finalize
I0419 09:11:14.663701 138441190782720 async_checkpointer.py:265] [process=4][thread=save_finalize] Waiting for background save thread=async_save.
I0419 09:11:14.663867 138570881185600 standard_logger.py:34] {'step': 19, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_test_pipeline_scan_nnx_20260419_085454/linen_xpk_test_pipeline_scan_nnx_20260419_085454_05_fp8/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776589824.6352217, 'wait_for_prev_duration_secs': 48.57839798927307, 'time_between_consecutive_saves_sec': 2.483268976211548, 'checkpointer_blocking_start_time': 1776589873.2151492, 'checkpointer_blocking_duration_secs': 1.0107667446136475, 'get_old_steps_start_time': 1776589874.2259312, 'get_old_steps_duration_secs': 3.218650817871094e-05, 'checkpoint_manager_blocking_start_time': 1776589824.6336234, 'checkpoint_manager_blocking_duration_secs': 50.03021168708801}
I0419 09:11:14.663995 138570881185600 checkpointing.py:408] Started an asynchronous checkpoint save for step 19
I0419 09:11:14.664038 138570881185600 checkpoint_manager.py:2024] [process=4][thread=MainThread][step=19][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete.
I0419 09:11:19.810820 138441207568128 array_metadata_store.py:203] [process=4][thread=array_type_handler] Wrote 81 array_metadata.ArrayMetadata to gs://lance-maxtext/linen_ckpt_xpk_test_pipeline_scan_nnx_20260419_085454/linen_xpk_test_pipeline_scan_nnx_20260419_085454_05_fp8/checkpoints/19/items/array_metadatas/process_4
I0419 09:11:55.150171 138441719260928 base_pytree_checkpoint_handler.py:129] [process=4] /jax/checkpoint/write/gbytes_per_sec: 38.307 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 41.2390513420105 s) (per-host)
I0419 09:11:55.150294 138441719260928 async_checkpointer.py:90] [process=4][thread=async_save] 3 Handler Commit operations completed. Time taken: 40.924533s.
I0419 09:12:04.491114 138441719260928 async_checkpointer.py:144] [process=4][thread=async_save] Background save thread done. Time taken: 50.265337s.
I0419 09:12:04.491397 138441190782720 async_checkpointer.py:273] [process=4][thread=save_finalize] Done with waiting for background save thread=async_save.
I0419 09:12:04.491523 138441190782720 async_checkpointer.py:283] [process=4][thread=save_finalize] No errors found in background save thread=async_save.
I0419 09:12:04.491566 138441190782720 checkpoint_manager.py:2133] [process=4][thread=save_finalize][step=19] CheckpointManager Save Finalize is syncing with other hosts...
I0419 09:12:04.493035 138441190782720 checkpoint_manager.py:2142] [process=4][thread=save_finalize][step=19] CheckpointManager Save Finalize is done on all hosts.
I0419 09:12:04.493195 138570881185600 checkpoint_manager.py:2036] [process=4][thread=MainThread][step=19][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=19.
I0419 09:12:04.493338 138570881185600 checkpoint_manager.py:2013] [process=4][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning.
I0419 09:12:04.494000 138570881185600 metric_logger.py:185] completed step: 19, seconds: 0.158, TFLOP/s/device: 85.772, Tokens/s/device: 12928.477, total_weights: 65536, loss: 6.167
Per train step:
 Total TFLOPs: 13.59 
 split as 93.93% learnable weight flops and 6.07% attention flops
XPK End: Sun Apr 19 09:12:13 UTC 2026
EXIT_CODE=0