Colin
@colinstuckless.bsky.social
Level 2 seems to be least I/O intensive, and more CPU/GPU bound based on our results running on local device with SSD vs AVD on SMB shares. It does not appear as sensitive to I/O latency as Level 1 and Level 3.
March 27, 2025 at 1:32 AM
Level 2 seems to be least I/O intensive, and more CPU/GPU bound based on our results running on local device with SSD vs AVD on SMB shares. It does not appear as sensitive to I/O latency as Level 1 and Level 3.
Here's how the I/O seems to breakdown between the Level 1/2/3 tests in our PAT runs.
March 26, 2025 at 1:11 PM
Here's how the I/O seems to breakdown between the Level 1/2/3 tests in our PAT runs.
The best marks so far belonged to:
L1 - Qumulo just under 6m
L2 - Qumulo just under 8m
L3 - Azure Files Metadata Cache just under 25m
Is there a doc that describes differences in I/O pattern between the tests? We are more interested in the I/O at the moment than CPU/GPU which is constant on AVD.
L1 - Qumulo just under 6m
L2 - Qumulo just under 8m
L3 - Azure Files Metadata Cache just under 25m
Is there a doc that describes differences in I/O pattern between the tests? We are more interested in the I/O at the moment than CPU/GPU which is constant on AVD.
March 23, 2025 at 10:26 PM
The best marks so far belonged to:
L1 - Qumulo just under 6m
L2 - Qumulo just under 8m
L3 - Azure Files Metadata Cache just under 25m
Is there a doc that describes differences in I/O pattern between the tests? We are more interested in the I/O at the moment than CPU/GPU which is constant on AVD.
L1 - Qumulo just under 6m
L2 - Qumulo just under 8m
L3 - Azure Files Metadata Cache just under 25m
Is there a doc that describes differences in I/O pattern between the tests? We are more interested in the I/O at the moment than CPU/GPU which is constant on AVD.