Skip to content

Commit f6a9fcb

Browse files
authored
init v6e mxla nightly (GoogleCloudPlatform#769)
1 parent c0cd731 commit f6a9fcb

2 files changed

Lines changed: 53 additions & 0 deletions

File tree

dags/common/vm_resource.py

Lines changed: 7 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -263,6 +263,13 @@ class XpkClusters:
263263
project=Project.TPU_PROD_ENV_MULTIPOD.value,
264264
zone=Zone.EUROPE_WEST4_B.value,
265265
)
266+
TPU_V6E_8_CLUSTER = XpkClusterConfig(
267+
name="bodaborg-v6e-8-yucmhab-c",
268+
device_version=TpuVersion.TRILLIUM,
269+
core_count=8,
270+
project=Project.TPU_PROD_ENV_ONE_VM.value,
271+
zone=Zone.US_EAST5_B.value,
272+
)
266273
TPU_V6E_256_CLUSTER = XpkClusterConfig(
267274
name="bodaborg-v6e-256",
268275
device_version=TpuVersion.TRILLIUM,

dags/multipod/mxla_maxtext_nightly_gke.py

Lines changed: 46 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -46,6 +46,7 @@
4646
group_id="Quarantine", dag=dag, prefix_group_id=False
4747
)
4848

49+
# v5p tests
4950
maxtext_nightly_1slice_v5p_8 = gke_config.get_gke_maxtext_nightly_config(
5051
cluster=XpkClusters.TPU_V5P_8_CLUSTER,
5152
time_out_in_min=60,
@@ -81,9 +82,54 @@
8182
test_owner=test_owner.RAYMOND_Z,
8283
).run_with_quarantine(quarantine_task_group)
8384

85+
# v6e tests
86+
maxtext_nightly_1slice_v6e_8 = gke_config.get_gke_maxtext_nightly_config(
87+
cluster=XpkClusters.TPU_V6E_8_CLUSTER,
88+
time_out_in_min=60,
89+
test_name=default_test_name,
90+
docker_image=jax_nightly_image.value,
91+
test_owner=test_owner.RISHABH_B,
92+
).run_with_quarantine(quarantine_task_group)
93+
94+
maxtext_nightly_2slice_v6e_8 = gke_config.get_gke_maxtext_nightly_config(
95+
num_slices=2,
96+
cluster=XpkClusters.TPU_V6E_8_CLUSTER,
97+
time_out_in_min=60,
98+
test_name=default_test_name,
99+
docker_image=jax_nightly_image.value,
100+
test_owner=test_owner.RISHABH_B,
101+
).run_with_quarantine(quarantine_task_group)
102+
103+
maxtext_nightly_4slice_v6e_8 = gke_config.get_gke_maxtext_nightly_config(
104+
num_slices=4,
105+
cluster=XpkClusters.TPU_V6E_8_CLUSTER,
106+
time_out_in_min=60,
107+
test_name=default_test_name,
108+
docker_image=jax_nightly_image.value,
109+
test_owner=test_owner.RISHABH_B,
110+
).run_with_quarantine(quarantine_task_group)
111+
112+
maxtext_nightly_8slice_v6e_8 = gke_config.get_gke_maxtext_nightly_config(
113+
num_slices=8,
114+
cluster=XpkClusters.TPU_V6E_8_CLUSTER,
115+
time_out_in_min=60,
116+
test_name=default_test_name,
117+
docker_image=jax_nightly_image.value,
118+
test_owner=test_owner.RISHABH_B,
119+
).run_with_quarantine(quarantine_task_group)
120+
121+
# Define dependencies for v5p tests to run sequentially
84122
(
85123
maxtext_nightly_1slice_v5p_8
86124
>> maxtext_nightly_2slice_v5p_8
87125
>> maxtext_nightly_4slice_v5p_8
88126
>> maxtext_nightly_8slice_v5p_8
89127
)
128+
129+
# Define dependencies for v6e tests to run sequentially
130+
(
131+
maxtext_nightly_1slice_v6e_8
132+
>> maxtext_nightly_2slice_v6e_8
133+
>> maxtext_nightly_4slice_v6e_8
134+
>> maxtext_nightly_8slice_v6e_8
135+
)

0 commit comments

Comments
 (0)