mudassiruddin21 commited on
Commit
e3ff911
·
verified ·
1 Parent(s): cba2e02

Run r2: logs and results

Browse files
runs/r2/engine_logs/20261118-160650-eval-r1.log CHANGED
@@ -247,3 +247,37 @@ INFO: 127.0.0.1:29179 - "POST /v1/completions HTTP/1.1" 200 OK
247
  INFO: 127.0.0.1:24970 - "GET /metrics HTTP/1.1" 200 OK
248
  INFO: 127.0.0.1:63359 - "GET /metrics HTTP/1.1" 200 OK
249
  INFO: 127.0.0.1:20803 - "POST /v1/completions HTTP/1.1" 200 OK
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
247
  INFO: 127.0.0.1:24970 - "GET /metrics HTTP/1.1" 200 OK
248
  INFO: 127.0.0.1:63359 - "GET /metrics HTTP/1.1" 200 OK
249
  INFO: 127.0.0.1:20803 - "POST /v1/completions HTTP/1.1" 200 OK
250
+ (APIServer pid=20653) INFO 11-18 16:12:56 [loggers.py:320] Engine 000: Avg prompt throughput: 27.3 tokens/s, Avg generation throughput: 67.8 tokens/s, Running: 1 reqs, Waiting: 0 reqs, GPU KV cache usage: 2.2%
251
+ (APIServer pid=20653) INFO 11-18 16:12:56 [metrics.py:120] SpecDecoding metrics: Mean acceptance length: 3.75, Accepted throughput: 49.69 tokens/s, Drafted throughput: 54.29 tokens/s, Accepted: 497 tokens, Drafted: 543 tokens, Per-position acceptance rate: 0.961, 0.917, 0.867, Avg Draft acceptance rate: 91.5%
252
+ INFO: 127.0.0.1:53076 - "POST /v1/completions HTTP/1.1" 200 OK
253
+ INFO: 127.0.0.1:41014 - "GET /metrics HTTP/1.1" 200 OK
254
+ INFO: 127.0.0.1:30960 - "GET /metrics HTTP/1.1" 200 OK
255
+ INFO: 127.0.0.1:65408 - "POST /v1/completions HTTP/1.1" 200 OK
256
+ (APIServer pid=20653) INFO 11-18 16:13:06 [loggers.py:320] Engine 000: Avg prompt throughput: 16.3 tokens/s, Avg generation throughput: 65.2 tokens/s, Running: 1 reqs, Waiting: 0 reqs, GPU KV cache usage: 2.2%
257
+ (APIServer pid=20653) INFO 11-18 16:13:06 [metrics.py:120] SpecDecoding metrics: Mean acceptance length: 3.60, Accepted throughput: 47.09 tokens/s, Drafted throughput: 54.29 tokens/s, Accepted: 471 tokens, Drafted: 543 tokens, Per-position acceptance rate: 0.961, 0.862, 0.779, Avg Draft acceptance rate: 86.7%
258
+ (APIServer pid=20653) INFO 11-18 16:13:16 [loggers.py:320] Engine 000: Avg prompt throughput: 0.0 tokens/s, Avg generation throughput: 42.6 tokens/s, Running: 1 reqs, Waiting: 0 reqs, GPU KV cache usage: 2.2%
259
+ (APIServer pid=20653) INFO 11-18 16:13:16 [metrics.py:120] SpecDecoding metrics: Mean acceptance length: 2.32, Accepted throughput: 24.20 tokens/s, Drafted throughput: 55.20 tokens/s, Accepted: 242 tokens, Drafted: 552 tokens, Per-position acceptance rate: 0.717, 0.391, 0.207, Avg Draft acceptance rate: 43.8%
260
+ INFO: 127.0.0.1:40164 - "POST /v1/completions HTTP/1.1" 200 OK
261
+ (APIServer pid=20653) INFO 11-18 16:13:26 [loggers.py:320] Engine 000: Avg prompt throughput: 3.8 tokens/s, Avg generation throughput: 43.7 tokens/s, Running: 1 reqs, Waiting: 0 reqs, GPU KV cache usage: 2.2%
262
+ (APIServer pid=20653) INFO 11-18 16:13:26 [metrics.py:120] SpecDecoding metrics: Mean acceptance length: 2.39, Accepted throughput: 25.49 tokens/s, Drafted throughput: 54.89 tokens/s, Accepted: 255 tokens, Drafted: 549 tokens, Per-position acceptance rate: 0.710, 0.426, 0.257, Avg Draft acceptance rate: 46.4%
263
+ (APIServer pid=20653) INFO 11-18 16:13:36 [loggers.py:320] Engine 000: Avg prompt throughput: 0.0 tokens/s, Avg generation throughput: 44.5 tokens/s, Running: 1 reqs, Waiting: 0 reqs, GPU KV cache usage: 2.2%
264
+ (APIServer pid=20653) INFO 11-18 16:13:36 [metrics.py:120] SpecDecoding metrics: Mean acceptance length: 2.41, Accepted throughput: 26.00 tokens/s, Drafted throughput: 55.49 tokens/s, Accepted: 260 tokens, Drafted: 555 tokens, Per-position acceptance rate: 0.768, 0.416, 0.222, Avg Draft acceptance rate: 46.8%
265
+ INFO: 127.0.0.1:19667 - "GET /metrics HTTP/1.1" 200 OK
266
+ (APIServer pid=20653) INFO 11-18 16:13:45 [launcher.py:128] [shutdown] API server: shutdown triggered
267
+ (APIServer pid=20653) INFO 11-18 16:13:45 [launcher.py:144] [shutdown] API server: stopping engine client mode=abort timeout=0s
268
+ (APIServer pid=20653) INFO 11-18 16:13:45 [core_client.py:762] [shutdown] MPClient: start timeout=0s
269
+ (APIServer pid=20653) INFO 11-18 16:13:45 [core_client.py:764] [shutdown] MPClient: stopping engine manager
270
+ (APIServer pid=20653) INFO 11-18 16:13:45 [utils.py:638] [shutdown] Process manager: send sigterm to process EngineCore
271
+ (APIServer pid=20653) WARNING 11-18 16:13:45 [utils.py:658] [shutdown] Process manager: force killing remaining processes count=1
272
+ (APIServer pid=20653) WARNING 11-18 16:13:45 [utils.py:663] [shutdown] Process manager: force killing remaining process EngineCore pid 21001
273
+ (APIServer pid=20653) INFO 11-18 16:13:45 [core_client.py:766] [shutdown] MPClient: engine manager stopped
274
+ (APIServer pid=20653) INFO 11-18 16:13:45 [core_client.py:767] [shutdown] MPClient: cleaning up background resources
275
+ (APIServer pid=20653) INFO 11-18 16:13:45 [core_client.py:769] [shutdown] MPClient: complete
276
+ (APIServer pid=20653) INFO 11-18 16:13:45 [launcher.py:153] [shutdown] API server: engine client stopped
277
+ (APIServer pid=20653) INFO 11-18 16:13:45 [launcher.py:156] [shutdown] API server: signalling HTTP server shutdown
278
+ (APIServer pid=20653) INFO 11-18 16:13:45 [launcher.py:178] [shutdown] API server: shutting down FastAPI HTTP server
279
+ INFO: Shutting down
280
+ INFO: Waiting for application shutdown.
281
+ INFO: Application shutdown complete.
282
+ /workspace/emc/uv-python/cpython-3.12.13-linux-x86_64-gnu/lib/python3.12/multiprocessing/resource_tracker.py:279: UserWarning: resource_tracker: There appear to be 1 leaked semaphore objects to clean up at shutdown
283
+ warnings.warn('resource_tracker: There appear to be %d '
runs/r2/kit_results/20261118-161345-results.json ADDED
@@ -0,0 +1,99 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "method": "mtp",
3
+ "prompts": 8,
4
+ "repeats": 1,
5
+ "acceptance_length": 2.8757894736842102,
6
+ "acceptance_by_domain": {
7
+ "chat": 2.9710982658959537,
8
+ "code": 3.7848837209302326,
9
+ "math": 3.6769230769230767,
10
+ "longform": 2.390449438202247
11
+ },
12
+ "tpot_median_s": 0.015276186888992942,
13
+ "ttft_median_s": 0.10750844149970362,
14
+ "memory_bytes": 607177728,
15
+ "weight_bytes": 540068864,
16
+ "kv_cache_bytes": 67108864,
17
+ "decode_total_s": 77.94211177399984,
18
+ "num_speculative_tokens": 3,
19
+ "max_model_len": 16384,
20
+ "engine": {
21
+ "vllm_version": "0.31.0",
22
+ "vllm_version_pinned": "0.31.0",
23
+ "image": "pip:vllm==0.31.0",
24
+ "matches_pin": true
25
+ },
26
+ "per_prompt": [
27
+ {
28
+ "id": "chat-0000",
29
+ "domain": "chat",
30
+ "tokens": 512,
31
+ "ttft_s": 0.11134,
32
+ "total_s": 10.87529,
33
+ "tpot_s": 0.021064
34
+ },
35
+ {
36
+ "id": "chat-0001",
37
+ "domain": "chat",
38
+ "tokens": 512,
39
+ "ttft_s": 0.05799,
40
+ "total_s": 8.01027,
41
+ "tpot_s": 0.015562
42
+ },
43
+ {
44
+ "id": "code-0000",
45
+ "domain": "code",
46
+ "tokens": 139,
47
+ "ttft_s": 0.06155,
48
+ "total_s": 1.95345,
49
+ "tpot_s": 0.013709
50
+ },
51
+ {
52
+ "id": "code-0001",
53
+ "domain": "code",
54
+ "tokens": 512,
55
+ "ttft_s": 0.12115,
56
+ "total_s": 7.53887,
57
+ "tpot_s": 0.014516
58
+ },
59
+ {
60
+ "id": "math-0000",
61
+ "domain": "math",
62
+ "tokens": 384,
63
+ "ttft_s": 0.10368,
64
+ "total_s": 5.67368,
65
+ "tpot_s": 0.014543
66
+ },
67
+ {
68
+ "id": "math-0001",
69
+ "domain": "math",
70
+ "tokens": 333,
71
+ "ttft_s": 0.11574,
72
+ "total_s": 5.09248,
73
+ "tpot_s": 0.01499
74
+ },
75
+ {
76
+ "id": "longform-0000",
77
+ "domain": "longform",
78
+ "tokens": 805,
79
+ "ttft_s": 0.10182,
80
+ "total_s": 18.79314,
81
+ "tpot_s": 0.023248
82
+ },
83
+ {
84
+ "id": "longform-0001",
85
+ "domain": "longform",
86
+ "tokens": 895,
87
+ "ttft_s": 0.11159,
88
+ "total_s": 20.00494,
89
+ "tpot_s": 0.022252
90
+ }
91
+ ],
92
+ "score": {
93
+ "score": 110.85,
94
+ "ranked": true,
95
+ "s_acceptance": 1.0407,
96
+ "s_memory": 1.5,
97
+ "s_tpot": 0.8526
98
+ }
99
+ }
runs/r2/logs/pipeline.log CHANGED
@@ -41,3 +41,4 @@
41
  [2026-11-18 16:06:40] done probe (22s)
42
  [2026-11-18 16:06:41] start smoke_fp8
43
  [2026-11-18 16:07:03] done prompts (56s)
 
 
41
  [2026-11-18 16:06:40] done probe (22s)
42
  [2026-11-18 16:06:41] start smoke_fp8
43
  [2026-11-18 16:07:03] done prompts (56s)
44
+ [2026-11-18 16:13:46] done smoke_fp8 (425s)
runs/r2/logs/pipeline.out CHANGED
@@ -41,3 +41,4 @@
41
  [2026-11-18 16:06:40] done probe (22s)
42
  [2026-11-18 16:06:41] start smoke_fp8
43
  [2026-11-18 16:07:03] done prompts (56s)
 
 
41
  [2026-11-18 16:06:40] done probe (22s)
42
  [2026-11-18 16:06:41] start smoke_fp8
43
  [2026-11-18 16:07:03] done prompts (56s)
44
+ [2026-11-18 16:13:46] done smoke_fp8 (425s)
runs/r2/logs/smoke_fp8.log CHANGED
@@ -9,3 +9,24 @@ PASS
9
  [16:11:17] vLLM ready after 266s
10
  [16:12:46] chat: 2 prompts done
11
  [16:12:55] code: 2 prompts done
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
9
  [16:11:17] vLLM ready after 266s
10
  [16:12:46] chat: 2 prompts done
11
  [16:12:55] code: 2 prompts done
12
+ [16:13:06] math: 2 prompts done
13
+ [16:13:45] longform: 2 prompts done
14
+
15
+ ================ emc2 eval summary ================
16
+ engine vLLM 0.31.0
17
+ method mtp (K = 3, context 16384)
18
+ prompts x repeats 8 x 1
19
+ acceptance length 2.876
20
+ chat 2.971
21
+ code 3.785
22
+ math 3.677
23
+ longform 2.390
24
+ TPOT (median) 15.28 ms per output token (this GPU)
25
+ TTFT (median, info) 107.5 ms (this GPU; not scored)
26
+ memory metric 0.607 GB (weights 0.540 + KV 0.067)
27
+ score 110.8 (baseline = 100)
28
+ components acceptance 1.041, memory 1.500, tpot 0.853
29
+ full results /workspace/emc/work/results/20261118-161345-results.json
30
+ ===================================================
31
+
32
+ FP8 probe acceptance 2.876 vs published baseline 2.802 (ratio 1.041)
runs/r2/logs/stats.log CHANGED
@@ -11,3 +11,4 @@
11
  [2026-11-18 16:06:42] ram 1.2/172 GB used (cache 0.0, shmem 0.0) | cgroup 1.2/172 GB | disk 66 GB used | gpu 0.0/95.6 GiB, 0%, 59 W | load 0.0 | largest python 1.7 GB, python 0.1 GB, python 0.1 GB
12
  [2026-11-18 16:07:03] ram 1.2/172 GB used (cache 0.1, shmem 0.0) | cgroup 1.2/172 GB | disk 66 GB used | gpu 0.0/95.6 GiB, 0%, 59 W | load 0.0 | largest vllm 1.2 GB, python 1.2 GB, python 0.1 GB
13
  [2026-11-18 16:12:55] ram 5.1/172 GB used (cache 0.3, shmem 0.0) | cgroup 5.1/172 GB | disk 66 GB used | gpu 83.1/95.6 GiB, 100%, 291 W | load 0.0 | largest VLLM::EngineCor 4.8 GB, vllm 2.5 GB, python 1.2 GB
 
 
11
  [2026-11-18 16:06:42] ram 1.2/172 GB used (cache 0.0, shmem 0.0) | cgroup 1.2/172 GB | disk 66 GB used | gpu 0.0/95.6 GiB, 0%, 59 W | load 0.0 | largest python 1.7 GB, python 0.1 GB, python 0.1 GB
12
  [2026-11-18 16:07:03] ram 1.2/172 GB used (cache 0.1, shmem 0.0) | cgroup 1.2/172 GB | disk 66 GB used | gpu 0.0/95.6 GiB, 0%, 59 W | load 0.0 | largest vllm 1.2 GB, python 1.2 GB, python 0.1 GB
13
  [2026-11-18 16:12:55] ram 5.1/172 GB used (cache 0.3, shmem 0.0) | cgroup 5.1/172 GB | disk 66 GB used | gpu 83.1/95.6 GiB, 100%, 291 W | load 0.0 | largest VLLM::EngineCor 4.8 GB, vllm 2.5 GB, python 1.2 GB
14
+ [2026-11-18 16:13:46] ram 0.2/172 GB used (cache 0.0, shmem 0.0) | cgroup 0.2/172 GB | disk 66 GB used | gpu 0.0/95.6 GiB, 0%, 146 W | load 0.0 | largest python 0.1 GB, python 0.1 GB, python 0.0 GB
runs/r2/logs/sync.log CHANGED
@@ -4,3 +4,4 @@ uploaded logs/results to mudassiruddin21/EMC:runs/r2
4
  uploaded logs/results to mudassiruddin21/EMC:runs/r2
5
  uploaded logs/results to mudassiruddin21/EMC:runs/r2
6
  uploaded logs/results to mudassiruddin21/EMC:runs/r2
 
 
4
  uploaded logs/results to mudassiruddin21/EMC:runs/r2
5
  uploaded logs/results to mudassiruddin21/EMC:runs/r2
6
  uploaded logs/results to mudassiruddin21/EMC:runs/r2
7
+ uploaded logs/results to mudassiruddin21/EMC:runs/r2
runs/r2/results/smoke_fp8.json ADDED
@@ -0,0 +1,99 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "method": "mtp",
3
+ "prompts": 8,
4
+ "repeats": 1,
5
+ "acceptance_length": 2.8757894736842102,
6
+ "acceptance_by_domain": {
7
+ "chat": 2.9710982658959537,
8
+ "code": 3.7848837209302326,
9
+ "math": 3.6769230769230767,
10
+ "longform": 2.390449438202247
11
+ },
12
+ "tpot_median_s": 0.015276186888992942,
13
+ "ttft_median_s": 0.10750844149970362,
14
+ "memory_bytes": 607177728,
15
+ "weight_bytes": 540068864,
16
+ "kv_cache_bytes": 67108864,
17
+ "decode_total_s": 77.94211177399984,
18
+ "num_speculative_tokens": 3,
19
+ "max_model_len": 16384,
20
+ "engine": {
21
+ "vllm_version": "0.31.0",
22
+ "vllm_version_pinned": "0.31.0",
23
+ "image": "pip:vllm==0.31.0",
24
+ "matches_pin": true
25
+ },
26
+ "per_prompt": [
27
+ {
28
+ "id": "chat-0000",
29
+ "domain": "chat",
30
+ "tokens": 512,
31
+ "ttft_s": 0.11134,
32
+ "total_s": 10.87529,
33
+ "tpot_s": 0.021064
34
+ },
35
+ {
36
+ "id": "chat-0001",
37
+ "domain": "chat",
38
+ "tokens": 512,
39
+ "ttft_s": 0.05799,
40
+ "total_s": 8.01027,
41
+ "tpot_s": 0.015562
42
+ },
43
+ {
44
+ "id": "code-0000",
45
+ "domain": "code",
46
+ "tokens": 139,
47
+ "ttft_s": 0.06155,
48
+ "total_s": 1.95345,
49
+ "tpot_s": 0.013709
50
+ },
51
+ {
52
+ "id": "code-0001",
53
+ "domain": "code",
54
+ "tokens": 512,
55
+ "ttft_s": 0.12115,
56
+ "total_s": 7.53887,
57
+ "tpot_s": 0.014516
58
+ },
59
+ {
60
+ "id": "math-0000",
61
+ "domain": "math",
62
+ "tokens": 384,
63
+ "ttft_s": 0.10368,
64
+ "total_s": 5.67368,
65
+ "tpot_s": 0.014543
66
+ },
67
+ {
68
+ "id": "math-0001",
69
+ "domain": "math",
70
+ "tokens": 333,
71
+ "ttft_s": 0.11574,
72
+ "total_s": 5.09248,
73
+ "tpot_s": 0.01499
74
+ },
75
+ {
76
+ "id": "longform-0000",
77
+ "domain": "longform",
78
+ "tokens": 805,
79
+ "ttft_s": 0.10182,
80
+ "total_s": 18.79314,
81
+ "tpot_s": 0.023248
82
+ },
83
+ {
84
+ "id": "longform-0001",
85
+ "domain": "longform",
86
+ "tokens": 895,
87
+ "ttft_s": 0.11159,
88
+ "total_s": 20.00494,
89
+ "tpot_s": 0.022252
90
+ }
91
+ ],
92
+ "score": {
93
+ "score": 110.85,
94
+ "ranked": true,
95
+ "s_acceptance": 1.0407,
96
+ "s_memory": 1.5,
97
+ "s_tpot": 0.8526
98
+ }
99
+ }
runs/r2/state/fp8_ok ADDED
@@ -0,0 +1 @@
 
 
1
+ 1.0407
runs/r2/state/smoke_fp8.done ADDED
@@ -0,0 +1 @@
 
 
1
+ 425