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

[Core] Use ray queue to put request output tokens back to the api server #41

Open
wants to merge 3 commits into
base: main
Choose a base branch
from

Conversation

KuilongCui
Copy link
Contributor

No description provided.

Copy link

prefill p25 p50 p75 p95 p99 mean
latency(ms) 21530.75 76643.50 184948.75 299576.50 303088.43 109563.64
decode p25 p50 p75 p95 p99 mean
latency(ms) 53.56 58.74 71.55 129.07 261.07 70.65

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 280.00 MB 312.00 MB 320.00 MB 416.00 MB 424.00 MB 448.00 MB 464.00 MB 472.00 MB 480.00 MB 536.00 MB 568.00 MB 696.00 MB 912.00 MB
rpc_speed(GB/s) 1.04 1.56 1.78 1.88 1.98 2.08 2.19 2.14 2.24 2.24 2.32 2.37 2.38 2.45 2.46 2.43 2.47 2.49 2.45 2.51 2.51 2.31 2.60 2.65 2.39 2.70 2.72 2.73 2.61 2.89 2.89 3.01 2.92 3.12 3.05 2.93 3.10 3.21 3.29
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 240.00 MB 264.00 MB 280.00 MB 288.00 MB 312.00 MB 320.00 MB 352.00 MB 384.00 MB 416.00 MB 424.00 MB 480.00 MB 536.00 MB 560.00 MB
gloo_speed(GB/s) 0.92 1.55 2.02 2.23 2.38 2.47 2.80 2.60 2.59 3.13 2.87 3.01 2.88 2.97 3.20 2.75 2.28 2.98 2.64 2.60 1.55 2.79 2.53 1.30 2.72 2.56 2.77 3.03 2.78 2.48 2.73 2.12 2.49 2.53 2.51 2.55 2.55 2.55 2.68 2.69
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 232.00 MB 256.00 MB 288.00 MB 312.00 MB 368.00 MB 416.00 MB 424.00 MB 480.00 MB 536.00 MB 552.00 MB
nccl_speed(GB/s) 0.19 0.48 0.61 0.90 1.10 1.22 1.52 1.42 1.82 2.02 2.00 2.44 2.32 2.89 2.57 2.48 2.81 2.89 3.28 3.70 3.48 3.95 3.11 5.03 3.46 4.06 6.27 2.74 3.59 5.09 4.88 5.59 6.34 3.88 4.05

Copy link

prefill p25 p50 p75 p95 p99 mean
latency(ms) 23239.50 80667.00 214905.50 266603.00 296250.51 112550.03
decode p25 p50 p75 p95 p99 mean
latency(ms) 53.20 59.40 78.21 156.72 434.27 79.37

@KuilongCui KuilongCui changed the title [Core] Add rayqueue [Core] Use ray queue to put request output tokens back to the api server (#28) Sep 20, 2024
@KuilongCui KuilongCui changed the title [Core] Use ray queue to put request output tokens back to the api server (#28) [Core] Use ray queue to put request output tokens back to the api server Sep 20, 2024
@KuilongCui KuilongCui requested review from s5u13b and ZeldaHuang and removed request for s5u13b September 20, 2024 08:47
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 224.00 MB 232.00 MB 240.00 MB 312.00 MB 320.00 MB 384.00 MB 416.00 MB 424.00 MB 432.00 MB 440.00 MB 472.00 MB 912.00 MB
rpc_speed(GB/s) 1.02 1.49 1.72 1.83 1.94 2.02 2.13 2.15 2.17 2.25 2.25 2.28 2.24 2.40 2.52 2.45 2.48 2.54 2.39 2.47 2.36 2.33 2.40 2.56 2.43 2.61 2.61 2.43 2.65 2.51 2.75 2.82 2.92 3.01 2.93 3.08 2.66 3.05
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 280.00 MB 312.00 MB 416.00 MB 424.00 MB 464.00 MB 480.00 MB 488.00 MB 560.00 MB
gloo_speed(GB/s) 0.90 1.59 1.95 2.09 2.38 2.46 2.73 2.83 2.57 3.06 2.99 3.17 2.65 3.07 3.45 2.62 2.73 2.29 2.57 2.32 1.58 2.63 2.90 2.65 2.62 2.59 2.57 2.21 3.00 2.62 2.83 2.19 1.57 2.60
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 232.00 MB 312.00 MB 320.00 MB 328.00 MB 392.00 MB 416.00 MB 464.00 MB 480.00 MB 536.00 MB 912.00 MB
nccl_speed(GB/s) 0.21 0.47 0.76 1.03 1.17 1.30 1.41 1.54 1.94 1.85 1.90 1.85 2.16 2.39 2.68 2.19 2.70 2.81 2.99 3.68 3.85 4.87 3.28 4.33 3.71 4.36 4.90 5.71 4.66 6.12 5.72 4.23 4.44

Copy link

prefill p25 p50 p75 p95 p99 mean
latency(ms) 18943.00 94628.50 211077.25 253174.80 293724.47 110821.18
decode p25 p50 p75 p95 p99 mean
latency(ms) 53.70 59.67 75.57 133.51 434.39 75.13

Copy link

prefill p25 p50 p75 p95 p99 mean
latency(ms) 31160.00 119095.50 194684.75 237867.70 248418.27 112714.30
decode p25 p50 p75 p95 p99 mean
latency(ms) 52.00 57.65 72.33 140.56 418.93 76.01

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 352.00 MB 416.00 MB 424.00 MB 432.00 MB 464.00 MB 472.00 MB 480.00 MB
rpc_speed(GB/s) 1.04 1.53 1.71 1.90 1.98 2.08 2.14 2.18 2.23 2.34 2.25 2.39 2.39 2.41 2.43 2.46 2.48 2.43 2.46 2.55 2.59 2.50 2.44 2.59 2.82 2.47 2.81 2.39 2.95 2.93 3.04 2.92 3.20 3.13 2.98
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 240.00 MB 264.00 MB 288.00 MB 312.00 MB 328.00 MB 344.00 MB 416.00 MB 472.00 MB 480.00 MB 696.00 MB
gloo_speed(GB/s) 0.95 1.48 1.89 2.19 2.26 2.60 2.53 2.70 2.70 2.98 3.12 2.89 3.19 3.33 2.64 2.64 2.63 1.85 2.61 2.70 1.94 2.07 2.18 2.99 3.00 2.55 1.24 2.88 0.90 2.58 2.88 2.30 2.64 2.63 2.71 2.72 1.92 2.66
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 240.00 MB 248.00 MB 264.00 MB 312.00 MB 320.00 MB 416.00 MB 424.00 MB 472.00 MB 480.00 MB 536.00 MB 560.00 MB
nccl_speed(GB/s) 0.19 0.48 0.63 0.87 1.00 1.15 1.46 1.61 1.91 1.70 1.90 2.20 2.22 2.52 2.45 2.57 2.88 2.62 2.86 2.95 3.47 3.66 3.51 3.25 3.75 4.23 4.30 4.23 4.40 2.35 5.40 4.90 5.23 5.37 7.53 5.09 3.75 3.76

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.

1 participant