Increase cool_down_time_seconds, warm_up_time_seconds and ntttcp execution time If ncpu > 64. by SRIKKANTH · Pull Request #4528 · microsoft/lisa

SRIKKANTH · 2026-06-12T17:50:52Z

If ncpu > 64, increase cool_down_time_seconds and warm_up_time_seconds to 5, and set ntttcp execution time to 120s.

On large VMs (>64 CPUs) the default 10s ntttcp run is too short for throughput to stabilize across all queues, leading to noisy results. Use a 120s run time on these SKUs and keep the default 10s elsewhere to avoid lengthening every test.

Description

Related Issue

Type of Change

Checklist

Description is filled in above
No credentials, secrets, or internal details are included
Peer review requested (if not, add required peer reviewers after raising PR)
Tests executed and results posted below

Test Validation

Key Test Cases:

Impacted LISA Features:

Tested Azure Marketplace Images:

Canonical 0001-com-ubuntu-server-jammy 22_04-lts-gen2 latest

Test Results

Image	VM Size	Result
Canonical 0001-com-ubuntu-server-jammy 22_04-lts-gen2 latest	Standard_L160ias_v5	PASSED / FAILED / SKIPPED

…s to 5, and set ntttcp execution time to 120s. On large VMs (>64 CPUs) the default 10s ntttcp run is too short for throughput to stabilize across all queues, leading to noisy results. Use a 120s run time on these SKUs and keep the default 10s elsewhere to avoid lengthening every test.

Copilot

Pull request overview

This PR updates the ntttcp performance helper to use longer warm-up/cool-down periods and a longer ntttcp duration on large VMs, aiming to reduce throughput noise on high core-count SKUs.

Changes:

Detect “large VM” based on CPU count and increase ntttcp duration from 10s to 120s when above the threshold.
Increase ntttcp warm-up and cool-down times from 1s to 5s on those larger VMs.
Plumb the warm-up/cool-down/run-time parameters into the ntttcp client/server invocations.

+        client_core_count = client.tools[Lscpu].get_core_count()
+        run_time_seconds = 120 if client_core_count > 64 else 10
+        warm_up_time_seconds = 5 if client_core_count > 64 else 1
+        cool_down_time_seconds = 5 if client_core_count > 64 else 1


SRIKKANTH · 2026-06-18T12:44:40Z

                    server_result = server_ntttcp.run_as_server_async(
                        server_nic_name,
                        server_ip=(
                            server.internal_address
                            if isinstance(server.os, BSD)
                            else ""
                        ),
                        ports_count=num_threads_p,
                        buffer_size=buffer_size,
                        dev_differentiator=dev_differentiator,
                        udp_mode=udp_mode,
                        no_debug_log=True,
+                        cool_down_time_seconds=cool_down_time_seconds,
+                        warm_up_time_seconds=warm_up_time_seconds,
                    )


github-actions · 2026-06-12T19:52:16Z

⚠️ AI Test Selection — cancelled

65 test case(s) selected (view run)

Marketplace image: canonical 0001-com-ubuntu-server-jammy 22_04-lts-gen2 latest

	Count
✅ Passed	12
❌ Failed	0
⏭️ Skipped	15
Total	27

Test case details

Test Case	Status	Time (s)	Message
perf_tcp_iperf_passthrough_host_guest (lisa_0_35)	⏭️ SKIPPED	0.000	requires [azure] but VM supports [cloud-hypervisor, hyperv]
perf_udp_iperf_passthrough_host_guest (lisa_0_36)	⏭️ SKIPPED	0.000	requires [azure] but VM supports [cloud-hypervisor, hyperv]
perf_tcp_single_pps_passthrough_host_guest (lisa_0_37)	⏭️ SKIPPED	0.000	requires [azure] but VM supports [cloud-hypervisor, hyperv]
perf_tcp_max_pps_passthrough_host_guest (lisa_0_38)	⏭️ SKIPPED	0.000	requires [azure] but VM supports [cloud-hypervisor, hyperv]
perf_tcp_ntttcp_passthrough_host_guest (lisa_0_39)	⏭️ SKIPPED	0.000	requires [azure] but VM supports [cloud-hypervisor, hyperv]
perf_udp_1k_ntttcp_passthrough_host_guest (lisa_0_40)	⏭️ SKIPPED	0.000	requires [azure] but VM supports [cloud-hypervisor, hyperv]
perf_tcp_iperf_passthrough_two_guest (lisa_0_41)	⏭️ SKIPPED	0.000	requires [azure] but VM supports [cloud-hypervisor, hyperv]
perf_udp_iperf_passthrough_two_guest (lisa_0_42)	⏭️ SKIPPED	0.000	requires [azure] but VM supports [cloud-hypervisor, hyperv]
perf_tcp_single_pps_passthrough_two_guest (lisa_0_43)	⏭️ SKIPPED	0.000	requires [azure] but VM supports [cloud-hypervisor, hyperv]
perf_tcp_max_pps_passthrough_two_guest (lisa_0_44)	⏭️ SKIPPED	0.000	requires [azure] but VM supports [cloud-hypervisor, hyperv]
perf_tcp_ntttcp_passthrough_two_guest (lisa_0_45)	⏭️ SKIPPED	0.000	requires [azure] but VM supports [cloud-hypervisor, hyperv]
perf_udp_1k_ntttcp_passthrough_two_guest (lisa_0_46)	⏭️ SKIPPED	0.000	requires [azure] but VM supports [cloud-hypervisor, hyperv]
perf_nvme_io_uring (lisa_0_1)	✅ PASSED	5344.889
perf_epoll (lisa_0_11)	✅ PASSED	187.676
perf_messaging (lisa_0_10)	✅ PASSED	2111.893
perf_nested_hyperv_storage_singledisk (lisa_0_4)	⏭️ SKIPPED	0.000	check skipped: OS type mismatch: ["requires [<class 'lisa.operating_system.Windows'>] but VM supports [<class 'lisa.oper
perf_nested_kvm_storage_singledisk (lisa_0_2)	✅ PASSED	5642.951
perf_nested_hyperv_storage_multidisk (lisa_0_5)	⏭️ SKIPPED	0.000	check skipped: OS type mismatch: ["requires [<class 'lisa.operating_system.Windows'>] but VM supports [<class 'lisa.oper
perf_nested_hyperv_ntttcp_different_l1_nat (lisa_0_8)	⏭️ SKIPPED	0.000	check skipped: OS type mismatch: ["requires [<class 'lisa.operating_system.Windows'>] but VM supports [<class 'lisa.oper
perf_resource_disk_1024k (lisa_0_55)	✅ PASSED	1622.366
perf_resource_disk_4k (lisa_0_56)	✅ PASSED	4417.976
perf_ultra_datadisks_1024k (lisa_0_48)	✅ PASSED	1059.565
perf_storage_generic_fio_test (lisa_0_64)	✅ PASSED	4485.996
perf_ultra_datadisks_4k (lisa_0_47)	✅ PASSED	4376.831
perf_tcp_latency_sriov (lisa_0_13)	✅ PASSED	193.401
perf_tcp_single_pps_sriov (lisa_0_15)	✅ PASSED	179.962
perf_tcp_max_pps_sriov (lisa_0_17)	✅ PASSED	147.524

github-actions · 2026-06-14T14:39:00Z

⚠️ AI Test Selection — cancelled

65 test case(s) selected (view run)

Marketplace image: canonical 0001-com-ubuntu-server-jammy 22_04-lts-gen2 latest

	Count
✅ Passed	12
❌ Failed	0
⏭️ Skipped	15
Total	27

Test case details

Test Case	Status	Time (s)	Message
perf_tcp_iperf_passthrough_host_guest (lisa_0_35)	⏭️ SKIPPED	0.000	requires [azure] but VM supports [hyperv, cloud-hypervisor]
perf_udp_iperf_passthrough_host_guest (lisa_0_36)	⏭️ SKIPPED	0.000	requires [azure] but VM supports [hyperv, cloud-hypervisor]
perf_tcp_single_pps_passthrough_host_guest (lisa_0_37)	⏭️ SKIPPED	0.000	requires [azure] but VM supports [hyperv, cloud-hypervisor]
perf_tcp_max_pps_passthrough_host_guest (lisa_0_38)	⏭️ SKIPPED	0.000	requires [azure] but VM supports [hyperv, cloud-hypervisor]
perf_tcp_ntttcp_passthrough_host_guest (lisa_0_39)	⏭️ SKIPPED	0.000	requires [azure] but VM supports [hyperv, cloud-hypervisor]
perf_udp_1k_ntttcp_passthrough_host_guest (lisa_0_40)	⏭️ SKIPPED	0.000	requires [azure] but VM supports [hyperv, cloud-hypervisor]
perf_tcp_iperf_passthrough_two_guest (lisa_0_41)	⏭️ SKIPPED	0.000	requires [azure] but VM supports [hyperv, cloud-hypervisor]
perf_udp_iperf_passthrough_two_guest (lisa_0_42)	⏭️ SKIPPED	0.000	requires [azure] but VM supports [hyperv, cloud-hypervisor]
perf_tcp_single_pps_passthrough_two_guest (lisa_0_43)	⏭️ SKIPPED	0.000	requires [azure] but VM supports [hyperv, cloud-hypervisor]
perf_tcp_max_pps_passthrough_two_guest (lisa_0_44)	⏭️ SKIPPED	0.000	requires [azure] but VM supports [hyperv, cloud-hypervisor]
perf_tcp_ntttcp_passthrough_two_guest (lisa_0_45)	⏭️ SKIPPED	0.000	requires [azure] but VM supports [hyperv, cloud-hypervisor]
perf_udp_1k_ntttcp_passthrough_two_guest (lisa_0_46)	⏭️ SKIPPED	0.000	requires [azure] but VM supports [hyperv, cloud-hypervisor]
perf_nested_hyperv_storage_singledisk (lisa_0_4)	⏭️ SKIPPED	0.000	check skipped: OS type mismatch: ["requires [<class 'lisa.operating_system.Windows'>] but VM supports [<class 'lisa.oper
perf_nested_kvm_storage_singledisk (lisa_0_2)	✅ PASSED	5472.640
perf_nested_hyperv_storage_multidisk (lisa_0_5)	⏭️ SKIPPED	0.000	check skipped: OS type mismatch: ["requires [<class 'lisa.operating_system.Windows'>] but VM supports [<class 'lisa.oper
perf_nested_hyperv_ntttcp_different_l1_nat (lisa_0_8)	⏭️ SKIPPED	0.000	check skipped: OS type mismatch: ["requires [<class 'lisa.operating_system.Windows'>] but VM supports [<class 'lisa.oper
perf_nvme_io_uring (lisa_0_1)	✅ PASSED	5338.601
perf_epoll (lisa_0_11)	✅ PASSED	194.906
perf_messaging (lisa_0_10)	✅ PASSED	2038.656
perf_resource_disk_1024k (lisa_0_55)	✅ PASSED	1613.694
perf_resource_disk_4k (lisa_0_56)	✅ PASSED	4421.172
perf_ultra_datadisks_1024k (lisa_0_48)	✅ PASSED	1066.630
perf_storage_generic_fio_test (lisa_0_64)	✅ PASSED	4424.029
perf_ultra_datadisks_4k (lisa_0_47)	✅ PASSED	4378.893
perf_tcp_latency_sriov (lisa_0_13)	✅ PASSED	195.285
perf_tcp_single_pps_sriov (lisa_0_15)	✅ PASSED	178.924
perf_tcp_max_pps_sriov (lisa_0_17)	✅ PASSED	147.048

LiliDeng · 2026-06-16T01:47:02Z

+        run_time_seconds = 120 if client_core_count > 64 else 10
+        warm_up_time_seconds = 5 if client_core_count > 64 else 1
+        cool_down_time_seconds = 5 if client_core_count > 64 else 1
        perf_ntttcp_message_list: List[


Please also change the for

lisa/lisa/microsoft/testsuites/performance/common.py

Line 567 in a150939

run_time_seconds=10,

Copilot

Pull request overview

Copilot reviewed 1 out of 1 changed files in this pull request and generated 2 comments.

+        client_core_count = client.tools[Lscpu].get_thread_count()
+        run_time_seconds = 120 if client_core_count > 64 else 10
+        warm_up_time_seconds = 5 if client_core_count > 64 else 1
+        cool_down_time_seconds = 5 if client_core_count > 64 else 1


                    client_lagscope_process = client_lagscope.run_as_client_async(
                        server_ip=server.internal_address,
                        ping_count=0,
-                        run_time_seconds=10,
+                        run_time_seconds=(
+                            run_time_seconds
+                            + cool_down_time_seconds
+                            + warm_up_time_seconds
+                        ),


github-actions · 2026-06-18T14:46:27Z

⚠️ AI Test Selection — cancelled

65 test case(s) selected (view run)

Marketplace image: canonical 0001-com-ubuntu-server-jammy 22_04-lts-gen2 latest

	Count
✅ Passed	14
❌ Failed	1
⏭️ Skipped	15
Total	30

Test case details

Test Case	Status	Time (s)	Message
perf_tcp_iperf_passthrough_host_guest (lisa_0_35)	⏭️ SKIPPED	0.000	requires [azure] but VM supports [cloud-hypervisor, hyperv]
perf_udp_iperf_passthrough_host_guest (lisa_0_36)	⏭️ SKIPPED	0.000	requires [azure] but VM supports [cloud-hypervisor, hyperv]
perf_tcp_single_pps_passthrough_host_guest (lisa_0_37)	⏭️ SKIPPED	0.000	requires [azure] but VM supports [cloud-hypervisor, hyperv]
perf_tcp_max_pps_passthrough_host_guest (lisa_0_38)	⏭️ SKIPPED	0.000	requires [azure] but VM supports [cloud-hypervisor, hyperv]
perf_tcp_ntttcp_passthrough_host_guest (lisa_0_39)	⏭️ SKIPPED	0.000	requires [azure] but VM supports [cloud-hypervisor, hyperv]
perf_udp_1k_ntttcp_passthrough_host_guest (lisa_0_40)	⏭️ SKIPPED	0.000	requires [azure] but VM supports [cloud-hypervisor, hyperv]
perf_tcp_iperf_passthrough_two_guest (lisa_0_41)	⏭️ SKIPPED	0.000	requires [azure] but VM supports [cloud-hypervisor, hyperv]
perf_udp_iperf_passthrough_two_guest (lisa_0_42)	⏭️ SKIPPED	0.000	requires [azure] but VM supports [cloud-hypervisor, hyperv]
perf_tcp_single_pps_passthrough_two_guest (lisa_0_43)	⏭️ SKIPPED	0.000	requires [azure] but VM supports [cloud-hypervisor, hyperv]
perf_tcp_max_pps_passthrough_two_guest (lisa_0_44)	⏭️ SKIPPED	0.000	requires [azure] but VM supports [cloud-hypervisor, hyperv]
perf_tcp_ntttcp_passthrough_two_guest (lisa_0_45)	⏭️ SKIPPED	0.000	requires [azure] but VM supports [cloud-hypervisor, hyperv]
perf_udp_1k_ntttcp_passthrough_two_guest (lisa_0_46)	⏭️ SKIPPED	0.000	requires [azure] but VM supports [cloud-hypervisor, hyperv]
perf_epoll (lisa_0_11)	✅ PASSED	195.315
perf_messaging (lisa_0_10)	✅ PASSED	2094.082
perf_nvme_io_uring (lisa_0_1)	✅ PASSED	5346.331
perf_nested_hyperv_storage_singledisk (lisa_0_4)	⏭️ SKIPPED	0.000	check skipped: OS type mismatch: ["requires [<class 'lisa.operating_system.Windows'>] but VM supports [<class 'lisa.oper
perf_nested_kvm_storage_singledisk (lisa_0_2)	✅ PASSED	5480.947
perf_nested_hyperv_storage_multidisk (lisa_0_5)	⏭️ SKIPPED	0.000	check skipped: OS type mismatch: ["requires [<class 'lisa.operating_system.Windows'>] but VM supports [<class 'lisa.oper
perf_nested_hyperv_ntttcp_different_l1_nat (lisa_0_8)	⏭️ SKIPPED	0.000	check skipped: OS type mismatch: ["requires [<class 'lisa.operating_system.Windows'>] but VM supports [<class 'lisa.oper
perf_resource_disk_1024k (lisa_0_55)	✅ PASSED	1623.212
perf_resource_disk_4k (lisa_0_56)	✅ PASSED	4420.231
perf_ultra_datadisks_1024k (lisa_0_48)	✅ PASSED	1069.842
perf_premiumv2_datadisks_1024k (lisa_0_50)	❌ FAILED	0.000	deployment cannot connect to TCP port: [20.171.8.241:22], error code: 110
perf_storage_generic_fio_test (lisa_0_64)	✅ PASSED	4473.677
perf_ultra_datadisks_4k (lisa_0_47)	✅ PASSED	4381.924
perf_tcp_latency_sriov (lisa_0_13)	✅ PASSED	212.386
perf_tcp_single_pps_sriov (lisa_0_15)	✅ PASSED	179.826
perf_tcp_max_pps_sriov (lisa_0_17)	✅ PASSED	149.758
perf_tcp_latency_synthetic (lisa_0_12)	✅ PASSED	321.915
perf_tcp_ntttcp_128_connections_synthetic (lisa_0_18)	✅ PASSED	58.529

SRIKKANTH requested a review from lubaihua33 as a code owner June 12, 2026 17:50

Copilot AI review requested due to automatic review settings June 12, 2026 17:50

SRIKKANTH requested review from LiliDeng and johnsongeorge-w as code owners June 12, 2026 17:50

Copilot started reviewing on behalf of SRIKKANTH June 12, 2026 17:51 View session

SRIKKANTH requested a review from umfranci June 12, 2026 17:53

Copilot AI reviewed Jun 12, 2026

View reviewed changes

copilot comment fixes

a150939

LiliDeng reviewed Jun 16, 2026

View reviewed changes

Update common.py

eed085f

Copilot AI review requested due to automatic review settings June 18, 2026 12:43

Copilot started reviewing on behalf of SRIKKANTH June 18, 2026 12:45 View session

Copilot AI reviewed Jun 18, 2026

View reviewed changes

LiliDeng approved these changes Jun 25, 2026

View reviewed changes

LiliDeng merged commit 0b805fd into main Jun 25, 2026
61 of 62 checks passed

LiliDeng deleted the smyakam/26_05_05/ntttcp_duration_increase branch June 25, 2026 15:06

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Uh oh!

Increase cool_down_time_seconds, warm_up_time_seconds and ntttcp execution time If ncpu > 64.#4528

Increase cool_down_time_seconds, warm_up_time_seconds and ntttcp execution time If ncpu > 64.#4528
LiliDeng merged 3 commits into
mainfrom
smyakam/26_05_05/ntttcp_duration_increase

SRIKKANTH commented Jun 12, 2026

Uh oh!

Copilot AI left a comment

Uh oh!

SRIKKANTH Jun 18, 2026

Uh oh!

github-actions Bot commented Jun 12, 2026

Uh oh!

github-actions Bot commented Jun 14, 2026

Uh oh!

LiliDeng Jun 16, 2026

Uh oh!

SRIKKANTH Jun 18, 2026

Uh oh!

Copilot AI left a comment

Uh oh!

github-actions Bot commented Jun 18, 2026

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

3 participants

Uh oh!

Conversation

SRIKKANTH commented Jun 12, 2026

Description

Related Issue

Type of Change

Checklist

Test Validation

Test Results

Uh oh!

Copilot AI left a comment

Choose a reason for hiding this comment

Pull request overview

Uh oh!

SRIKKANTH Jun 18, 2026

Choose a reason for hiding this comment

Uh oh!

github-actions Bot commented Jun 12, 2026

⚠️ AI Test Selection — cancelled

Uh oh!

github-actions Bot commented Jun 14, 2026

⚠️ AI Test Selection — cancelled

Uh oh!

LiliDeng Jun 16, 2026

Choose a reason for hiding this comment

Uh oh!

SRIKKANTH Jun 18, 2026

Choose a reason for hiding this comment

Uh oh!

Copilot AI left a comment

Choose a reason for hiding this comment

Pull request overview

Uh oh!

github-actions Bot commented Jun 18, 2026

⚠️ AI Test Selection — cancelled

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

3 participants