Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

[Misc] Ensure Llumlet main thread exits on Engine.Step errors #38

Merged
merged 4 commits into from
Oct 10, 2024

Conversation

KuilongCui
Copy link
Contributor

@KuilongCui KuilongCui commented Sep 13, 2024

When an error occurs in the engine.step thread, the main thread will not exit.
In the pr, catch the exception and exit the main thread to release the resources.

Copy link

github-actions bot commented Oct 8, 2024

migration_size 8.00 MB 16.00 MB 24.00 MB 32.00 MB 40.00 MB 48.00 MB 56.00 MB 64.00 MB 72.00 MB 80.00 MB 88.00 MB 96.00 MB 104.00 MB 112.00 MB 120.00 MB 128.00 MB 136.00 MB 144.00 MB 152.00 MB 160.00 MB 168.00 MB 176.00 MB 184.00 MB 192.00 MB 200.00 MB 224.00 MB 232.00 MB 240.00 MB 312.00 MB 328.00 MB 352.00 MB 384.00 MB 416.00 MB 424.00 MB 464.00 MB 472.00 MB 480.00 MB 536.00 MB 744.00 MB 912.00 MB
rpc_speed(GB/s) 1.03 1.50 1.76 1.86 2.00 2.07 2.14 2.13 2.18 2.26 2.35 2.38 2.34 2.30 2.40 2.41 2.43 2.46 2.53 2.47 2.43 2.37 2.56 2.54 2.39 2.34 2.41 2.41 2.85 2.68 2.83 3.00 2.73 2.77 3.22 3.01 2.91 3.27 3.12 3.19
migration_size 8.00 MB 16.00 MB 24.00 MB 32.00 MB 40.00 MB 48.00 MB 56.00 MB 64.00 MB 72.00 MB 80.00 MB 88.00 MB 96.00 MB 104.00 MB 112.00 MB 120.00 MB 128.00 MB 136.00 MB 144.00 MB 160.00 MB 168.00 MB 176.00 MB 184.00 MB 192.00 MB 224.00 MB 232.00 MB 240.00 MB 264.00 MB 312.00 MB 416.00 MB 448.00 MB 464.00 MB 480.00 MB 536.00 MB 560.00 MB 728.00 MB
gloo_speed(GB/s) 0.92 1.52 1.98 2.16 2.36 2.53 2.72 2.84 2.58 3.13 3.05 2.89 2.57 3.15 2.90 2.84 2.53 2.27 2.41 2.97 2.27 2.53 2.57 2.77 2.41 2.52 2.33 2.16 2.66 2.34 1.36 2.64 1.73 2.40 2.90
migration_size 8.00 MB 16.00 MB 24.00 MB 32.00 MB 40.00 MB 48.00 MB 56.00 MB 64.00 MB 72.00 MB 80.00 MB 88.00 MB 96.00 MB 104.00 MB 112.00 MB 120.00 MB 128.00 MB 136.00 MB 144.00 MB 152.00 MB 160.00 MB 168.00 MB 176.00 MB 184.00 MB 192.00 MB 200.00 MB 208.00 MB 232.00 MB 264.00 MB 272.00 MB 304.00 MB 312.00 MB 320.00 MB 344.00 MB 416.00 MB 424.00 MB 472.00 MB 480.00 MB 488.00 MB 560.00 MB
nccl_speed(GB/s) 0.18 0.51 0.73 0.87 1.13 1.30 1.34 1.43 1.80 2.08 2.53 2.40 2.08 2.31 2.62 2.46 2.82 3.13 3.34 3.97 4.06 3.55 3.19 4.83 3.79 3.28 3.90 1.81 3.13 3.19 3.50 4.42 3.65 4.88 5.41 1.98 5.04 5.65 5.03

Copy link

github-actions bot commented Oct 8, 2024

prefill p25 p50 p75 p95 p99 mean
latency(ms) 32512.50 88367.50 176054.25 268331.50 298296.04 107189.26
decode p25 p50 p75 p95 p99 mean
latency(ms) 54.04 60.94 83.13 153.21 423.63 80.33

Copy link

github-actions bot commented Oct 8, 2024

migration_size 8.00 MB 16.00 MB 24.00 MB 32.00 MB 40.00 MB 48.00 MB 56.00 MB 64.00 MB 72.00 MB 80.00 MB 88.00 MB 96.00 MB 104.00 MB 112.00 MB 120.00 MB 128.00 MB 136.00 MB 144.00 MB 152.00 MB 160.00 MB 168.00 MB 176.00 MB 184.00 MB 192.00 MB 200.00 MB 208.00 MB 224.00 MB 232.00 MB 240.00 MB 248.00 MB 280.00 MB 312.00 MB 416.00 MB 432.00 MB 464.00 MB 480.00 MB 536.00 MB
rpc_speed(GB/s) 1.03 1.51 1.76 1.85 2.00 2.06 2.13 2.11 2.24 2.29 2.29 2.30 2.36 2.36 2.43 2.43 2.43 2.40 2.35 2.47 2.49 2.31 2.46 2.47 2.41 2.65 2.66 2.39 2.53 2.61 2.53 2.71 2.94 2.95 2.83 2.93 3.10
migration_size 8.00 MB 16.00 MB 24.00 MB 32.00 MB 40.00 MB 48.00 MB 56.00 MB 64.00 MB 72.00 MB 80.00 MB 88.00 MB 96.00 MB 104.00 MB 112.00 MB 120.00 MB 128.00 MB 136.00 MB 144.00 MB 152.00 MB 160.00 MB 168.00 MB 176.00 MB 184.00 MB 192.00 MB 200.00 MB 232.00 MB 312.00 MB 320.00 MB 416.00 MB 464.00 MB 480.00 MB 560.00 MB
gloo_speed(GB/s) 0.93 1.59 1.93 2.40 2.42 2.63 2.80 2.94 2.88 2.76 2.79 2.90 2.56 2.89 2.46 2.67 2.37 2.09 2.71 2.31 2.73 1.87 2.44 2.74 1.66 2.49 2.26 2.85 2.47 1.82 2.72 2.64
migration_size 8.00 MB 16.00 MB 24.00 MB 32.00 MB 40.00 MB 48.00 MB 56.00 MB 64.00 MB 72.00 MB 80.00 MB 88.00 MB 96.00 MB 104.00 MB 112.00 MB 120.00 MB 128.00 MB 136.00 MB 144.00 MB 152.00 MB 160.00 MB 168.00 MB 176.00 MB 184.00 MB 192.00 MB 200.00 MB 208.00 MB 232.00 MB 312.00 MB 320.00 MB 416.00 MB 424.00 MB 432.00 MB 472.00 MB 480.00 MB 536.00 MB 912.00 MB
nccl_speed(GB/s) 0.19 0.48 0.67 0.91 1.08 1.23 1.35 1.60 1.67 2.01 2.11 2.10 2.17 2.49 2.28 2.46 2.58 2.99 2.90 3.61 3.10 3.39 3.27 5.31 3.47 3.60 3.72 3.96 5.96 6.61 5.78 6.00 6.41 5.56 4.04 4.31

Copy link

github-actions bot commented Oct 8, 2024

prefill p25 p50 p75 p95 p99 mean
latency(ms) 30338.75 109947.50 176418.25 241808.35 246791.02 109480.60
decode p25 p50 p75 p95 p99 mean
latency(ms) 53.71 60.69 79.91 172.13 552.59 82.87

Copy link

github-actions bot commented Oct 8, 2024

prefill p25 p50 p75 p95 p99 mean
latency(ms) 35336.75 100567.50 176278.00 252971.65 301798.24 109331.58
decode p25 p50 p75 p95 p99 mean
latency(ms) 54.27 61.31 78.15 165.67 458.58 80.54

Copy link

github-actions bot commented Oct 8, 2024

migration_size 8.00 MB 16.00 MB 24.00 MB 32.00 MB 40.00 MB 48.00 MB 56.00 MB 64.00 MB 72.00 MB 80.00 MB 88.00 MB 96.00 MB 104.00 MB 112.00 MB 120.00 MB 128.00 MB 136.00 MB 144.00 MB 152.00 MB 160.00 MB 168.00 MB 176.00 MB 184.00 MB 192.00 MB 200.00 MB 208.00 MB 224.00 MB 232.00 MB 280.00 MB 312.00 MB 416.00 MB 424.00 MB 432.00 MB 472.00 MB 536.00 MB
rpc_speed(GB/s) 1.03 1.49 1.76 1.89 2.00 2.03 2.13 2.08 2.17 2.27 2.22 2.26 2.31 2.38 2.37 2.40 2.38 2.45 2.38 2.42 2.53 2.46 2.50 2.26 2.46 2.55 2.63 2.66 2.65 2.69 2.89 2.84 2.82 2.95 3.13
migration_size 8.00 MB 16.00 MB 24.00 MB 32.00 MB 40.00 MB 48.00 MB 56.00 MB 64.00 MB 72.00 MB 80.00 MB 88.00 MB 96.00 MB 104.00 MB 112.00 MB 120.00 MB 128.00 MB 136.00 MB 144.00 MB 152.00 MB 160.00 MB 168.00 MB 176.00 MB 184.00 MB 192.00 MB 200.00 MB 208.00 MB 224.00 MB 232.00 MB 280.00 MB 312.00 MB 328.00 MB 384.00 MB 416.00 MB 480.00 MB 488.00 MB 544.00 MB 568.00 MB 720.00 MB
gloo_speed(GB/s) 0.97 1.60 2.04 2.28 2.47 2.64 2.67 2.70 2.73 2.64 2.89 2.79 2.82 2.60 3.00 2.63 2.42 2.24 2.51 2.96 2.06 2.80 2.97 2.98 2.58 2.61 3.00 2.66 1.94 1.95 2.50 2.92 2.57 2.47 2.25 2.74 2.74 2.47
migration_size 8.00 MB 16.00 MB 24.00 MB 32.00 MB 40.00 MB 48.00 MB 56.00 MB 64.00 MB 72.00 MB 80.00 MB 88.00 MB 96.00 MB 104.00 MB 112.00 MB 120.00 MB 128.00 MB 136.00 MB 144.00 MB 152.00 MB 160.00 MB 168.00 MB 176.00 MB 184.00 MB 192.00 MB 216.00 MB 224.00 MB 240.00 MB 248.00 MB 312.00 MB 416.00 MB 424.00 MB 464.00 MB 480.00 MB 536.00 MB 560.00 MB 576.00 MB
nccl_speed(GB/s) 0.18 0.50 0.63 0.97 1.04 1.20 1.52 1.39 1.95 2.11 1.92 2.18 2.18 2.47 2.80 2.67 3.07 2.42 2.94 3.13 3.29 2.76 3.89 3.80 3.49 3.65 5.78 3.71 3.37 5.58 5.31 5.28 5.85 3.90 4.20 4.33

Copy link
Contributor

@ZeldaHuang ZeldaHuang left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

We should add some tests to ensure resource(gpu&&all threads in llumlet) is properly released when engine crashed.

Copy link

github-actions bot commented Oct 8, 2024

migration_size 8.00 MB 16.00 MB 24.00 MB 32.00 MB 40.00 MB 48.00 MB 56.00 MB 64.00 MB 72.00 MB 80.00 MB 88.00 MB 96.00 MB 104.00 MB 112.00 MB 120.00 MB 128.00 MB 136.00 MB 144.00 MB 152.00 MB 160.00 MB 168.00 MB 176.00 MB 184.00 MB 192.00 MB 200.00 MB 208.00 MB 224.00 MB 232.00 MB 312.00 MB 320.00 MB 416.00 MB 440.00 MB 472.00 MB 488.00 MB 544.00 MB 560.00 MB 696.00 MB 912.00 MB
rpc_speed(GB/s) 1.04 1.54 1.75 1.90 1.99 2.11 2.14 2.15 2.24 2.28 2.36 2.34 2.36 2.36 2.44 2.44 2.43 2.50 2.52 2.49 2.28 2.40 2.41 2.71 2.61 2.68 2.52 2.55 2.64 2.79 3.03 3.11 3.04 3.05 3.27 3.21 3.24 3.27
migration_size 8.00 MB 16.00 MB 24.00 MB 32.00 MB 40.00 MB 48.00 MB 56.00 MB 64.00 MB 72.00 MB 80.00 MB 88.00 MB 96.00 MB 104.00 MB 112.00 MB 120.00 MB 128.00 MB 136.00 MB 144.00 MB 152.00 MB 160.00 MB 168.00 MB 176.00 MB 184.00 MB 192.00 MB 200.00 MB 208.00 MB 240.00 MB 256.00 MB 280.00 MB 296.00 MB 312.00 MB 416.00 MB 424.00 MB 480.00 MB 536.00 MB 656.00 MB 688.00 MB
gloo_speed(GB/s) 0.92 1.53 1.81 2.13 2.39 2.78 2.78 2.81 2.95 2.75 3.03 3.09 3.01 2.52 2.50 2.65 2.63 2.10 2.47 2.55 1.26 2.51 2.41 2.48 2.34 2.61 2.65 3.08 2.37 2.78 2.07 2.67 2.76 2.74 2.64 2.65 2.87
migration_size 8.00 MB 16.00 MB 24.00 MB 32.00 MB 40.00 MB 48.00 MB 56.00 MB 64.00 MB 72.00 MB 80.00 MB 88.00 MB 96.00 MB 104.00 MB 112.00 MB 120.00 MB 128.00 MB 136.00 MB 144.00 MB 152.00 MB 160.00 MB 168.00 MB 176.00 MB 184.00 MB 192.00 MB 232.00 MB 240.00 MB 256.00 MB 264.00 MB 280.00 MB 312.00 MB 336.00 MB 384.00 MB 416.00 MB 424.00 MB 480.00 MB 576.00 MB
nccl_speed(GB/s) 0.18 0.45 0.69 0.98 1.08 1.33 1.36 1.53 1.78 1.77 1.80 2.41 2.39 2.63 2.40 2.52 2.97 2.69 3.55 3.61 4.35 3.65 3.51 3.72 3.75 3.94 4.28 2.54 2.56 3.85 7.75 4.75 5.16 5.30 6.97 4.75

Copy link

github-actions bot commented Oct 8, 2024

prefill p25 p50 p75 p95 p99 mean
latency(ms) 30862.25 119005.00 199547.75 245687.35 247074.30 115053.90
decode p25 p50 p75 p95 p99 mean
latency(ms) 53.19 61.07 78.50 132.22 328.09 74.25

Copy link

github-actions bot commented Oct 9, 2024

migration_size 8.00 MB 16.00 MB 24.00 MB 32.00 MB 40.00 MB 48.00 MB 56.00 MB 64.00 MB 72.00 MB 80.00 MB 88.00 MB 96.00 MB 104.00 MB 112.00 MB 120.00 MB 128.00 MB 136.00 MB 144.00 MB 152.00 MB 160.00 MB 168.00 MB 176.00 MB 184.00 MB 192.00 MB 200.00 MB 208.00 MB 240.00 MB 248.00 MB 312.00 MB 320.00 MB 344.00 MB 352.00 MB 384.00 MB 416.00 MB 424.00 MB 464.00 MB 472.00 MB 488.00 MB 568.00 MB
rpc_speed(GB/s) 1.05 1.56 1.75 1.89 2.00 2.07 2.13 2.10 2.23 2.26 2.22 2.30 2.30 2.37 2.42 2.43 2.47 2.36 2.50 2.46 2.45 2.16 2.28 2.42 2.24 2.45 2.45 2.72 2.66 2.73 2.77 2.79 2.75 2.98 2.87 3.04 3.04 3.04 3.03
migration_size 8.00 MB 16.00 MB 24.00 MB 32.00 MB 40.00 MB 48.00 MB 56.00 MB 64.00 MB 72.00 MB 80.00 MB 88.00 MB 96.00 MB 104.00 MB 112.00 MB 120.00 MB 128.00 MB 136.00 MB 144.00 MB 152.00 MB 160.00 MB 168.00 MB 176.00 MB 192.00 MB 200.00 MB 224.00 MB 288.00 MB 312.00 MB 320.00 MB 416.00 MB 448.00 MB 456.00 MB 472.00 MB 480.00 MB 704.00 MB
gloo_speed(GB/s) 0.95 1.45 1.91 2.16 2.44 2.47 2.75 2.73 2.83 2.98 2.91 3.27 2.81 3.01 2.79 2.79 2.33 2.47 2.64 2.34 2.23 2.37 3.00 2.73 2.51 2.53 2.22 2.66 2.81 2.33 2.68 2.59 2.77 2.11
migration_size 8.00 MB 16.00 MB 24.00 MB 32.00 MB 40.00 MB 48.00 MB 56.00 MB 64.00 MB 72.00 MB 80.00 MB 88.00 MB 96.00 MB 104.00 MB 112.00 MB 120.00 MB 128.00 MB 136.00 MB 144.00 MB 152.00 MB 160.00 MB 168.00 MB 176.00 MB 184.00 MB 192.00 MB 200.00 MB 208.00 MB 216.00 MB 232.00 MB 280.00 MB 288.00 MB 312.00 MB 320.00 MB 400.00 MB 416.00 MB 424.00 MB 440.00 MB 480.00 MB 536.00 MB 560.00 MB
nccl_speed(GB/s) 0.19 0.45 0.69 0.89 1.18 1.25 1.54 1.89 1.59 2.09 2.37 2.13 2.20 2.04 2.81 2.86 2.79 3.39 3.29 3.21 2.96 4.66 4.67 3.78 3.82 3.60 3.92 4.89 2.35 2.87 4.50 3.67 5.67 5.83 5.35 4.32 5.40 4.20 4.42

Copy link

github-actions bot commented Oct 9, 2024

prefill p25 p50 p75 p95 p99 mean
latency(ms) 19375.00 88817.50 228663.75 279214.45 298182.10 119057.91
decode p25 p50 p75 p95 p99 mean
latency(ms) 52.87 60.12 76.84 138.21 318.36 74.00

Copy link
Collaborator

@zhypku zhypku left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

other changes look good

Copy link

github-actions bot commented Oct 9, 2024

prefill p25 p50 p75 p95 p99 mean
latency(ms) 50817.00 110316.50 179232.25 231337.15 233864.04 113701.17
decode p25 p50 p75 p95 p99 mean
latency(ms) 52.56 57.94 71.17 119.16 312.32 71.29

Copy link

github-actions bot commented Oct 9, 2024

migration_size 8.00 MB 16.00 MB 24.00 MB 32.00 MB 40.00 MB 48.00 MB 56.00 MB 64.00 MB 72.00 MB 80.00 MB 88.00 MB 96.00 MB 104.00 MB 112.00 MB 120.00 MB 128.00 MB 136.00 MB 144.00 MB 152.00 MB 160.00 MB 168.00 MB 176.00 MB 184.00 MB 192.00 MB 200.00 MB 216.00 MB 224.00 MB 232.00 MB 264.00 MB 280.00 MB 296.00 MB 312.00 MB 344.00 MB 416.00 MB 576.00 MB 632.00 MB
rpc_speed(GB/s) 1.05 1.57 1.81 1.97 2.06 2.18 2.18 2.26 2.34 2.35 2.41 2.36 2.46 2.57 2.55 2.51 2.51 2.37 2.57 2.54 2.60 2.69 2.51 2.56 2.45 2.72 2.61 2.74 2.59 2.73 2.73 3.02 2.79 2.89 3.19 3.23
migration_size 8.00 MB 16.00 MB 24.00 MB 32.00 MB 40.00 MB 48.00 MB 56.00 MB 64.00 MB 72.00 MB 80.00 MB 88.00 MB 96.00 MB 104.00 MB 112.00 MB 120.00 MB 128.00 MB 136.00 MB 144.00 MB 152.00 MB 160.00 MB 168.00 MB 176.00 MB 184.00 MB 192.00 MB 200.00 MB 208.00 MB 240.00 MB 272.00 MB 280.00 MB 312.00 MB 320.00 MB 376.00 MB 416.00 MB 424.00 MB 480.00 MB 544.00 MB 552.00 MB 560.00 MB 912.00 MB
gloo_speed(GB/s) 0.95 1.52 1.96 2.25 2.35 2.46 2.67 2.72 2.83 2.90 3.11 2.97 2.77 3.09 3.01 2.73 2.04 2.59 2.62 2.79 1.60 2.57 2.49 2.05 2.44 3.03 2.69 2.34 1.54 2.52 2.52 2.70 2.65 2.95 2.65 2.67 2.71 2.50 2.30
migration_size 8.00 MB 16.00 MB 24.00 MB 32.00 MB 40.00 MB 48.00 MB 56.00 MB 64.00 MB 72.00 MB 80.00 MB 88.00 MB 96.00 MB 104.00 MB 112.00 MB 120.00 MB 128.00 MB 136.00 MB 144.00 MB 152.00 MB 160.00 MB 168.00 MB 176.00 MB 184.00 MB 192.00 MB 200.00 MB 216.00 MB 248.00 MB 272.00 MB 312.00 MB 320.00 MB 416.00 MB 424.00 MB 440.00 MB 488.00 MB 560.00 MB 912.00 MB
nccl_speed(GB/s) 0.21 0.50 0.61 0.88 1.05 1.40 1.39 1.66 1.93 1.91 1.91 2.14 2.29 2.08 2.80 2.52 2.54 2.84 3.04 3.41 3.60 3.69 3.22 3.55 4.46 4.05 3.42 2.54 3.96 5.85 5.87 5.61 6.37 5.91 4.11 4.70

Copy link

github-actions bot commented Oct 9, 2024

migration_size 8.00 MB 16.00 MB 24.00 MB 32.00 MB 40.00 MB 48.00 MB 56.00 MB 64.00 MB 72.00 MB 80.00 MB 88.00 MB 96.00 MB 104.00 MB 112.00 MB 120.00 MB 128.00 MB 136.00 MB 144.00 MB 152.00 MB 160.00 MB 168.00 MB 176.00 MB 184.00 MB 192.00 MB 200.00 MB 208.00 MB 216.00 MB 232.00 MB 312.00 MB 320.00 MB 416.00 MB 424.00 MB 432.00 MB 448.00 MB 480.00 MB 536.00 MB 544.00 MB 912.00 MB
rpc_speed(GB/s) 1.02 1.51 1.75 1.85 1.98 2.07 2.11 2.08 2.18 2.20 2.27 2.28 2.33 2.32 2.33 2.48 2.49 2.36 2.50 2.49 2.34 2.50 2.40 2.53 2.58 2.58 2.68 2.50 2.70 2.39 2.84 2.94 2.99 2.83 2.91 3.33 3.04 3.20
migration_size 8.00 MB 16.00 MB 24.00 MB 32.00 MB 40.00 MB 48.00 MB 56.00 MB 64.00 MB 72.00 MB 80.00 MB 88.00 MB 96.00 MB 104.00 MB 112.00 MB 120.00 MB 128.00 MB 136.00 MB 144.00 MB 152.00 MB 160.00 MB 168.00 MB 176.00 MB 192.00 MB 200.00 MB 240.00 MB 280.00 MB 288.00 MB 304.00 MB 312.00 MB 320.00 MB 416.00 MB 424.00 MB 440.00 MB 480.00 MB 488.00 MB 496.00 MB 536.00 MB 544.00 MB 560.00 MB 920.00 MB
gloo_speed(GB/s) 0.92 1.48 1.98 2.22 2.21 2.28 2.58 2.63 2.58 2.84 3.20 2.76 3.29 2.58 3.01 2.40 2.07 2.54 2.43 2.61 0.55 2.45 2.62 2.60 2.64 2.66 2.44 2.72 2.48 2.66 2.81 2.54 2.30 2.91 2.71 2.62 2.70 2.66 2.84 2.72
migration_size 8.00 MB 16.00 MB 24.00 MB 32.00 MB 40.00 MB 48.00 MB 56.00 MB 64.00 MB 72.00 MB 80.00 MB 88.00 MB 96.00 MB 104.00 MB 112.00 MB 120.00 MB 128.00 MB 136.00 MB 144.00 MB 152.00 MB 160.00 MB 168.00 MB 176.00 MB 184.00 MB 192.00 MB 200.00 MB 208.00 MB 224.00 MB 232.00 MB 280.00 MB 296.00 MB 312.00 MB 320.00 MB 328.00 MB 360.00 MB 392.00 MB 416.00 MB 424.00 MB 432.00 MB 480.00 MB 488.00 MB 536.00 MB
nccl_speed(GB/s) 0.19 0.44 0.65 0.83 1.11 1.24 1.63 1.49 1.87 1.84 2.10 2.07 2.06 2.89 2.53 2.59 4.33 2.73 3.16 3.29 3.26 3.26 3.64 3.71 3.34 4.27 4.01 3.82 2.07 2.61 4.10 3.49 4.83 5.08 3.87 5.33 4.26 5.28 5.86 6.09 4.36

Copy link

github-actions bot commented Oct 9, 2024

prefill p25 p50 p75 p95 p99 mean
latency(ms) 31121.75 105756.50 171415.25 241300.70 282416.16 110177.19
decode p25 p50 p75 p95 p99 mean
latency(ms) 53.90 59.69 76.68 131.91 228.24 73.39

Copy link

migration_size 8.00 MB 16.00 MB 24.00 MB 32.00 MB 40.00 MB 48.00 MB 56.00 MB 64.00 MB 72.00 MB 80.00 MB 88.00 MB 96.00 MB 104.00 MB 112.00 MB 120.00 MB 128.00 MB 136.00 MB 144.00 MB 152.00 MB 160.00 MB 168.00 MB 176.00 MB 184.00 MB 192.00 MB 200.00 MB 208.00 MB 216.00 MB 232.00 MB 312.00 MB 320.00 MB 416.00 MB 432.00 MB 472.00 MB 480.00 MB 536.00 MB 560.00 MB 696.00 MB
rpc_speed(GB/s) 1.03 1.51 1.74 1.90 2.01 2.09 2.14 2.19 2.26 2.32 2.33 2.40 2.39 2.42 2.35 2.43 2.47 2.44 2.51 2.51 2.45 2.38 2.43 2.62 2.58 2.57 2.71 2.66 2.71 2.63 3.20 3.06 3.19 3.29 3.23 3.10 3.32
migration_size 8.00 MB 16.00 MB 24.00 MB 32.00 MB 40.00 MB 48.00 MB 56.00 MB 64.00 MB 72.00 MB 80.00 MB 88.00 MB 96.00 MB 104.00 MB 112.00 MB 120.00 MB 128.00 MB 136.00 MB 144.00 MB 152.00 MB 160.00 MB 168.00 MB 176.00 MB 184.00 MB 192.00 MB 200.00 MB 216.00 MB 232.00 MB 240.00 MB 312.00 MB 416.00 MB 424.00 MB 480.00 MB 536.00 MB
gloo_speed(GB/s) 0.94 1.46 1.86 2.10 2.21 2.51 2.50 2.70 2.75 2.74 2.78 3.37 3.31 3.15 3.20 2.69 2.73 2.22 2.50 2.66 2.10 2.72 2.53 2.00 2.64 1.09 2.54 2.55 2.22 2.68 2.68 2.52 2.68
migration_size 8.00 MB 16.00 MB 24.00 MB 32.00 MB 40.00 MB 48.00 MB 56.00 MB 64.00 MB 72.00 MB 80.00 MB 88.00 MB 96.00 MB 104.00 MB 112.00 MB 120.00 MB 128.00 MB 136.00 MB 144.00 MB 152.00 MB 160.00 MB 168.00 MB 176.00 MB 184.00 MB 200.00 MB 208.00 MB 216.00 MB 232.00 MB 280.00 MB 312.00 MB 320.00 MB 392.00 MB 416.00 MB 424.00 MB 440.00 MB 472.00 MB 480.00 MB 560.00 MB 696.00 MB
nccl_speed(GB/s) 0.20 0.48 0.72 0.87 1.29 1.53 1.40 1.46 1.70 1.59 2.21 1.92 2.27 2.52 2.38 2.61 2.83 2.66 3.10 3.43 3.62 3.28 3.84 2.75 5.28 3.51 5.06 3.01 4.12 4.99 5.40 5.22 6.22 5.54 6.02 5.67 4.55 4.78

Copy link

prefill p25 p50 p75 p95 p99 mean
latency(ms) 23362.50 100729.50 225398.25 250282.40 253712.22 118267.43
decode p25 p50 p75 p95 p99 mean
latency(ms) 52.42 58.94 73.91 134.73 359.49 73.59

@KuilongCui KuilongCui changed the title [Misc] Ensure Llumlet Main Thread Exits on Engine.Step Errors [Misc] Ensure Llumlet main thread exits on Engine.Step errors Oct 10, 2024
Copy link

migration_size 8.00 MB 16.00 MB 24.00 MB 32.00 MB 40.00 MB 48.00 MB 56.00 MB 64.00 MB 72.00 MB 80.00 MB 88.00 MB 96.00 MB 104.00 MB 112.00 MB 120.00 MB 128.00 MB 136.00 MB 144.00 MB 152.00 MB 160.00 MB 168.00 MB 176.00 MB 184.00 MB 192.00 MB 200.00 MB 232.00 MB 312.00 MB 320.00 MB 384.00 MB 392.00 MB 416.00 MB 424.00 MB 448.00 MB 472.00 MB 488.00 MB 544.00 MB
rpc_speed(GB/s) 1.02 1.51 1.71 1.88 2.00 2.06 2.11 2.16 2.18 2.23 2.28 2.28 2.30 2.37 2.45 2.47 2.45 2.36 2.51 2.50 2.59 2.36 2.48 2.55 2.48 2.54 2.70 2.74 2.79 2.85 3.11 2.98 2.99 2.99 2.89 3.14
migration_size 8.00 MB 16.00 MB 24.00 MB 32.00 MB 40.00 MB 48.00 MB 56.00 MB 64.00 MB 72.00 MB 80.00 MB 88.00 MB 96.00 MB 104.00 MB 112.00 MB 120.00 MB 128.00 MB 136.00 MB 144.00 MB 152.00 MB 160.00 MB 168.00 MB 176.00 MB 184.00 MB 192.00 MB 200.00 MB 208.00 MB 232.00 MB 256.00 MB 296.00 MB 312.00 MB 328.00 MB 384.00 MB 416.00 MB 424.00 MB 432.00 MB 464.00 MB 480.00 MB
gloo_speed(GB/s) 0.92 1.62 2.06 2.21 2.45 2.37 2.86 2.94 2.87 2.66 3.40 2.75 2.75 3.05 3.35 2.32 2.22 3.01 2.61 2.78 1.77 2.65 1.21 2.77 2.36 2.73 2.00 2.41 2.73 2.10 2.51 2.57 2.57 2.64 2.56 2.71 2.68
migration_size 8.00 MB 16.00 MB 24.00 MB 32.00 MB 40.00 MB 48.00 MB 56.00 MB 64.00 MB 72.00 MB 80.00 MB 88.00 MB 96.00 MB 104.00 MB 112.00 MB 120.00 MB 128.00 MB 136.00 MB 144.00 MB 152.00 MB 160.00 MB 168.00 MB 176.00 MB 184.00 MB 192.00 MB 200.00 MB 208.00 MB 232.00 MB 304.00 MB 312.00 MB 320.00 MB 344.00 MB 376.00 MB 416.00 MB 432.00 MB 464.00 MB 480.00 MB 496.00 MB 536.00 MB 544.00 MB 552.00 MB 560.00 MB 696.00 MB 728.00 MB
nccl_speed(GB/s) 0.18 0.44 0.65 0.96 1.27 1.25 1.50 1.55 1.66 1.89 1.74 2.21 2.28 2.43 2.80 2.89 2.74 2.38 3.76 3.39 3.72 3.51 3.24 3.34 3.54 4.02 3.63 4.48 4.07 2.85 5.28 4.84 4.50 7.55 5.30 6.34 6.03 4.13 4.30 4.00 5.43 4.93 5.35

Copy link

prefill p25 p50 p75 p95 p99 mean
latency(ms) 27639.50 113264.50 209162.25 236989.10 240008.17 114394.47
decode p25 p50 p75 p95 p99 mean
latency(ms) 52.67 59.12 73.90 125.34 336.13 73.30

@KuilongCui KuilongCui merged commit fc5ecee into main Oct 10, 2024
14 checks passed
@KuilongCui KuilongCui deleted the exception branch October 10, 2024 07:13
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

Successfully merging this pull request may close these issues.

4 participants