Remove run logs for QPS values outside paper's evaluated range

Crawler: removed 16 run log dirs for arrival=0.0625 (16 QPS) and
arrival=0.125 (8 QPS). Paper evaluates QPS 0.5–4.

ANNS: removed 24 run log dirs for arrival=0.0625 (16 QPS),
arrival=0.125 (8 QPS), and arrival=0.25 (4 QPS). Paper evaluates
QPS 0.25–2.

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

run_log/anns/H100_enhanced_schedulers_v1_full/default_vllm/10-16-2025_17-52-04/config_10-16-2025_17-52-04.yaml

@@ -1,35 +0,0 @@
-config_for_run: driver/anns/run_log/H100_enhanced_schedulers_v1_full/default_vllm
-data_dir: data/anns
-experiments:
-- concurrency: true
-  stream: true
-- concurrency: true
-  stream: false
-log_stats: true
-logging:
-  suppress_vllm_logs: true
-  verbose_console: false
-model:
-  gpu_memory_utilization: 0.8
-  max_token_budget: 8192
-  name: meta-llama/Meta-Llama-3.1-8B-Instruct
-  seed: 42
-  tensor_parallel_size: 2
-prefix_caching_flag: false
-query_range: 500
-query_trace_map: data/anns/query_trace_map_5k.json
-replay:
-  delay_multiplier: 1.0
-  end_template: '
-
-
-    Question:'
-  max_output_tokens: 1
-  poisson_avg_arrival_time: 0.0625
-  template: 'Answer the question based only on the provided context. If the answer
-    is not in the context, say "I don''t know."
-
-    Context:'
-scheduler: default_vllm
-trace_dir: data/anns/res
-use_saved_poisson_delays: driver/anns/configs/H100_enhanced_schedulers_v1_full/poisson_delays_0_0625.txt

run_log/anns/H100_enhanced_schedulers_v1_full/default_vllm/10-16-2025_17-52-04/run_metrics.csv

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:32441371598abda908010d5ec13cc8a75aa72a4581b4eb8583c7f10119b44e69
-size 996128

run_log/anns/H100_enhanced_schedulers_v1_full/default_vllm/10-16-2025_18-18-50/config_10-16-2025_18-18-50.yaml

@@ -1,35 +0,0 @@
-config_for_run: driver/anns/run_log/H100_enhanced_schedulers_v1_full/default_vllm
-data_dir: data/anns
-experiments:
-- concurrency: true
-  stream: true
-- concurrency: true
-  stream: false
-log_stats: true
-logging:
-  suppress_vllm_logs: true
-  verbose_console: false
-model:
-  gpu_memory_utilization: 0.8
-  max_token_budget: 8192
-  name: meta-llama/Meta-Llama-3.1-8B-Instruct
-  seed: 42
-  tensor_parallel_size: 2
-prefix_caching_flag: false
-query_range: 500
-query_trace_map: data/anns/query_trace_map_5k.json
-replay:
-  delay_multiplier: 1.0
-  end_template: '
-
-
-    Question:'
-  max_output_tokens: 1
-  poisson_avg_arrival_time: 0.125
-  template: 'Answer the question based only on the provided context. If the answer
-    is not in the context, say "I don''t know."
-
-    Context:'
-scheduler: default_vllm
-trace_dir: data/anns/res
-use_saved_poisson_delays: driver/anns/configs/H100_enhanced_schedulers_v1_full/poisson_delays_0_125.txt

run_log/anns/H100_enhanced_schedulers_v1_full/default_vllm/10-16-2025_18-18-50/run_metrics.csv

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:a9d7efec7fd4b2c951f27cfb78bf5aed2b5cdea63061c95d6138437dd02c26b8
-size 997844

run_log/anns/H100_enhanced_schedulers_v1_full/default_vllm/10-16-2025_18-44-01/config_10-16-2025_18-44-01.yaml

@@ -1,35 +0,0 @@
-config_for_run: driver/anns/run_log/H100_enhanced_schedulers_v1_full/default_vllm
-data_dir: data/anns
-experiments:
-- concurrency: true
-  stream: true
-- concurrency: true
-  stream: false
-log_stats: true
-logging:
-  suppress_vllm_logs: true
-  verbose_console: false
-model:
-  gpu_memory_utilization: 0.8
-  max_token_budget: 8192
-  name: meta-llama/Meta-Llama-3.1-8B-Instruct
-  seed: 42
-  tensor_parallel_size: 2
-prefix_caching_flag: false
-query_range: 500
-query_trace_map: data/anns/query_trace_map_5k.json
-replay:
-  delay_multiplier: 1.0
-  end_template: '
-
-
-    Question:'
-  max_output_tokens: 1
-  poisson_avg_arrival_time: 0.25
-  template: 'Answer the question based only on the provided context. If the answer
-    is not in the context, say "I don''t know."
-
-    Context:'
-scheduler: default_vllm
-trace_dir: data/anns/res
-use_saved_poisson_delays: driver/anns/configs/H100_enhanced_schedulers_v1_full/poisson_delays_0_25.txt

run_log/anns/H100_enhanced_schedulers_v1_full/default_vllm/10-16-2025_18-44-01/run_metrics.csv

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:02c6683da62bad389455314e63697a89c1f15f65c4f812b720ff4df164caa8cc
-size 1007311

run_log/anns/H100_enhanced_schedulers_v1_full/fcfs/10-16-2025_17-58-51/config_10-16-2025_17-58-51.yaml

@@ -1,38 +0,0 @@
-config_for_run: driver/anns/run_log/H100_enhanced_schedulers_v1_full/fcfs
-data_dir: data/anns
-experiments:
-- concurrency: true
-  stream: true
-log_stats: true
-logging:
-  suppress_vllm_logs: true
-  verbose_console: false
-model:
-  gpu_memory_utilization: 0.8
-  max_token_budget: 8192
-  name: meta-llama/Meta-Llama-3.1-8B-Instruct
-  seed: 42
-  tensor_parallel_size: 2
-prefix_caching_flag: false
-query_range: 500
-query_trace_map: data/anns/query_trace_map_5k.json
-replay:
-  delay_multiplier: 1.0
-  end_template: '
-
-
-    Question:'
-  max_output_tokens: 1
-  poisson_avg_arrival_time: 0.0625
-  template: 'Answer the question based only on the provided context. If the answer
-    is not in the context, say "I don''t know."
-
-    Context:'
-scheduler: fcfs
-trace_dir: data/anns/res
-use_recomputation_latency_predictor: 1
-use_saved_poisson_delays: driver/anns/configs/H100_enhanced_schedulers_v1_full/poisson_delays_0_0625.txt
-use_saved_recomputation_data_path: driver/recomputation_perf_model/H200_tp2_recomputation_latency.json
-use_saved_swap_latency_data_path: driver/swap_perf_model/H200_tp2_swap_kernel_latency.json
-use_swap_kernel: 1
-use_swap_latency_predictor: 1

run_log/anns/H100_enhanced_schedulers_v1_full/fcfs/10-16-2025_17-58-51/run_metrics.csv

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:f1aecd06f68ba42481922ce7544645e99bebb7a23c5d3c945af597ece88ba0bd
-size 622276

run_log/anns/H100_enhanced_schedulers_v1_full/fcfs/10-16-2025_18-26-14/config_10-16-2025_18-26-14.yaml

@@ -1,38 +0,0 @@
-config_for_run: driver/anns/run_log/H100_enhanced_schedulers_v1_full/fcfs
-data_dir: data/anns
-experiments:
-- concurrency: true
-  stream: true
-log_stats: true
-logging:
-  suppress_vllm_logs: true
-  verbose_console: false
-model:
-  gpu_memory_utilization: 0.8
-  max_token_budget: 8192
-  name: meta-llama/Meta-Llama-3.1-8B-Instruct
-  seed: 42
-  tensor_parallel_size: 2
-prefix_caching_flag: false
-query_range: 500
-query_trace_map: data/anns/query_trace_map_5k.json
-replay:
-  delay_multiplier: 1.0
-  end_template: '
-
-
-    Question:'
-  max_output_tokens: 1
-  poisson_avg_arrival_time: 0.125
-  template: 'Answer the question based only on the provided context. If the answer
-    is not in the context, say "I don''t know."
-
-    Context:'
-scheduler: fcfs
-trace_dir: data/anns/res
-use_recomputation_latency_predictor: 1
-use_saved_poisson_delays: driver/anns/configs/H100_enhanced_schedulers_v1_full/poisson_delays_0_125.txt
-use_saved_recomputation_data_path: driver/recomputation_perf_model/H200_tp2_recomputation_latency.json
-use_saved_swap_latency_data_path: driver/swap_perf_model/H200_tp2_swap_kernel_latency.json
-use_swap_kernel: 1
-use_swap_latency_predictor: 1

run_log/anns/H100_enhanced_schedulers_v1_full/fcfs/10-16-2025_18-26-14/run_metrics.csv

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:674f86d4dc2f0a9f924071ebcf8c54fa69d93f504df97834731b8c32162c7289
-size 626606

run_log/anns/H100_enhanced_schedulers_v1_full/fcfs/10-16-2025_18-55-07/config_10-16-2025_18-55-07.yaml

@@ -1,38 +0,0 @@
-config_for_run: driver/anns/run_log/H100_enhanced_schedulers_v1_full/fcfs
-data_dir: data/anns
-experiments:
-- concurrency: true
-  stream: true
-log_stats: true
-logging:
-  suppress_vllm_logs: true
-  verbose_console: false
-model:
-  gpu_memory_utilization: 0.8
-  max_token_budget: 8192
-  name: meta-llama/Meta-Llama-3.1-8B-Instruct
-  seed: 42
-  tensor_parallel_size: 2
-prefix_caching_flag: false
-query_range: 500
-query_trace_map: data/anns/query_trace_map_5k.json
-replay:
-  delay_multiplier: 1.0
-  end_template: '
-
-
-    Question:'
-  max_output_tokens: 1
-  poisson_avg_arrival_time: 0.25
-  template: 'Answer the question based only on the provided context. If the answer
-    is not in the context, say "I don''t know."
-
-    Context:'
-scheduler: fcfs
-trace_dir: data/anns/res
-use_recomputation_latency_predictor: 1
-use_saved_poisson_delays: driver/anns/configs/H100_enhanced_schedulers_v1_full/poisson_delays_0_25.txt
-use_saved_recomputation_data_path: driver/recomputation_perf_model/H200_tp2_recomputation_latency.json
-use_saved_swap_latency_data_path: driver/swap_perf_model/H200_tp2_swap_kernel_latency.json
-use_swap_kernel: 1
-use_swap_latency_predictor: 1

run_log/anns/H100_enhanced_schedulers_v1_full/fcfs/10-16-2025_18-55-07/run_metrics.csv

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:fa5b350889ca0f28f9ceb1c1ccd1731cd5aba495b0aa81b5c8f7edc19f8dd762
-size 632368

run_log/anns/H100_enhanced_schedulers_v1_full/lcas/10-16-2025_18-06-32/config_10-16-2025_18-06-32.yaml

@@ -1,38 +0,0 @@
-config_for_run: driver/anns/run_log/H100_enhanced_schedulers_v1_full/lcas
-data_dir: data/anns
-experiments:
-- concurrency: true
-  stream: true
-log_stats: true
-logging:
-  suppress_vllm_logs: true
-  verbose_console: false
-model:
-  gpu_memory_utilization: 0.8
-  max_token_budget: 8192
-  name: meta-llama/Meta-Llama-3.1-8B-Instruct
-  seed: 42
-  tensor_parallel_size: 2
-prefix_caching_flag: false
-query_range: 500
-query_trace_map: data/anns/query_trace_map_5k.json
-replay:
-  delay_multiplier: 1.0
-  end_template: '
-
-
-    Question:'
-  max_output_tokens: 1
-  poisson_avg_arrival_time: 0.0625
-  template: 'Answer the question based only on the provided context. If the answer
-    is not in the context, say "I don''t know."
-
-    Context:'
-scheduler: lcas
-trace_dir: data/anns/res
-use_recomputation_latency_predictor: 1
-use_saved_poisson_delays: driver/anns/configs/H100_enhanced_schedulers_v1_full/poisson_delays_0_0625.txt
-use_saved_recomputation_data_path: driver/recomputation_perf_model/H200_tp2_recomputation_latency.json
-use_saved_swap_latency_data_path: driver/swap_perf_model/H200_tp2_swap_kernel_latency.json
-use_swap_kernel: 1
-use_swap_latency_predictor: 1

run_log/anns/H100_enhanced_schedulers_v1_full/lcas/10-16-2025_18-06-32/run_metrics.csv

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:96e4af78015686bd18a606710347152b5a10fd54cf6cbd6b3d18db5bb9290377
-size 621123

run_log/anns/H100_enhanced_schedulers_v1_full/lcas/10-16-2025_18-31-54/config_10-16-2025_18-31-54.yaml

@@ -1,38 +0,0 @@
-config_for_run: driver/anns/run_log/H100_enhanced_schedulers_v1_full/lcas
-data_dir: data/anns
-experiments:
-- concurrency: true
-  stream: true
-log_stats: true
-logging:
-  suppress_vllm_logs: true
-  verbose_console: false
-model:
-  gpu_memory_utilization: 0.8
-  max_token_budget: 8192
-  name: meta-llama/Meta-Llama-3.1-8B-Instruct
-  seed: 42
-  tensor_parallel_size: 2
-prefix_caching_flag: false
-query_range: 500
-query_trace_map: data/anns/query_trace_map_5k.json
-replay:
-  delay_multiplier: 1.0
-  end_template: '
-
-
-    Question:'
-  max_output_tokens: 1
-  poisson_avg_arrival_time: 0.125
-  template: 'Answer the question based only on the provided context. If the answer
-    is not in the context, say "I don''t know."
-
-    Context:'
-scheduler: lcas
-trace_dir: data/anns/res
-use_recomputation_latency_predictor: 1
-use_saved_poisson_delays: driver/anns/configs/H100_enhanced_schedulers_v1_full/poisson_delays_0_125.txt
-use_saved_recomputation_data_path: driver/recomputation_perf_model/H200_tp2_recomputation_latency.json
-use_saved_swap_latency_data_path: driver/swap_perf_model/H200_tp2_swap_kernel_latency.json
-use_swap_kernel: 1
-use_swap_latency_predictor: 1

run_log/anns/H100_enhanced_schedulers_v1_full/lcas/10-16-2025_18-31-54/run_metrics.csv

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:71707a914914ad8e51aa1ae1d319cb81d21cdee5962de4e2592f6ace847e5953
-size 625948

run_log/anns/H100_enhanced_schedulers_v1_full/lcas/10-16-2025_19-07-43/config_10-16-2025_19-07-43.yaml

@@ -1,38 +0,0 @@
-config_for_run: driver/anns/run_log/H100_enhanced_schedulers_v1_full/lcas
-data_dir: data/anns
-experiments:
-- concurrency: true
-  stream: true
-log_stats: true
-logging:
-  suppress_vllm_logs: true
-  verbose_console: false
-model:
-  gpu_memory_utilization: 0.8
-  max_token_budget: 8192
-  name: meta-llama/Meta-Llama-3.1-8B-Instruct
-  seed: 42
-  tensor_parallel_size: 2
-prefix_caching_flag: false
-query_range: 500
-query_trace_map: data/anns/query_trace_map_5k.json
-replay:
-  delay_multiplier: 1.0
-  end_template: '
-
-
-    Question:'
-  max_output_tokens: 1
-  poisson_avg_arrival_time: 0.25
-  template: 'Answer the question based only on the provided context. If the answer
-    is not in the context, say "I don''t know."
-
-    Context:'
-scheduler: lcas
-trace_dir: data/anns/res
-use_recomputation_latency_predictor: 1
-use_saved_poisson_delays: driver/anns/configs/H100_enhanced_schedulers_v1_full/poisson_delays_0_25.txt
-use_saved_recomputation_data_path: driver/recomputation_perf_model/H200_tp2_recomputation_latency.json
-use_saved_swap_latency_data_path: driver/swap_perf_model/H200_tp2_swap_kernel_latency.json
-use_swap_kernel: 1
-use_swap_latency_predictor: 1

run_log/anns/H100_enhanced_schedulers_v1_full/lcas/10-16-2025_19-07-43/run_metrics.csv

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:aca50fe504b4673036ac7e591b35c8f07da42e1466fc818209128b421a625d78
-size 628145

run_log/anns/H100_enhanced_schedulers_v1_full/mcps/10-16-2025_18-13-01/config_10-16-2025_18-13-01.yaml

@@ -1,38 +0,0 @@
-config_for_run: driver/anns/run_log/H100_enhanced_schedulers_v1_full/mcps
-data_dir: data/anns
-experiments:
-- concurrency: true
-  stream: true
-log_stats: true
-logging:
-  suppress_vllm_logs: true
-  verbose_console: false
-model:
-  gpu_memory_utilization: 0.8
-  max_token_budget: 8192
-  name: meta-llama/Meta-Llama-3.1-8B-Instruct
-  seed: 42
-  tensor_parallel_size: 2
-prefix_caching_flag: false
-query_range: 500
-query_trace_map: data/anns/query_trace_map_5k.json
-replay:
-  delay_multiplier: 1.0
-  end_template: '
-
-
-    Question:'
-  max_output_tokens: 1
-  poisson_avg_arrival_time: 0.0625
-  template: 'Answer the question based only on the provided context. If the answer
-    is not in the context, say "I don''t know."
-
-    Context:'
-scheduler: mcps
-trace_dir: data/anns/res
-use_recomputation_latency_predictor: 1
-use_saved_poisson_delays: driver/anns/configs/H100_enhanced_schedulers_v1_full/poisson_delays_0_0625.txt
-use_saved_recomputation_data_path: driver/recomputation_perf_model/H200_tp2_recomputation_latency.json
-use_saved_swap_latency_data_path: driver/swap_perf_model/H200_tp2_swap_kernel_latency.json
-use_swap_kernel: 1
-use_swap_latency_predictor: 1

run_log/anns/H100_enhanced_schedulers_v1_full/mcps/10-16-2025_18-13-01/run_metrics.csv

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:6532a0c321e85ab59666991380b91f2e2dc734bcfb85d39081cd6aaaf88ad1cf
-size 621441

run_log/anns/H100_enhanced_schedulers_v1_full/mcps/10-16-2025_18-37-45/config_10-16-2025_18-37-45.yaml

@@ -1,38 +0,0 @@
-config_for_run: driver/anns/run_log/H100_enhanced_schedulers_v1_full/mcps
-data_dir: data/anns
-experiments:
-- concurrency: true
-  stream: true
-log_stats: true
-logging:
-  suppress_vllm_logs: true
-  verbose_console: false
-model:
-  gpu_memory_utilization: 0.8
-  max_token_budget: 8192
-  name: meta-llama/Meta-Llama-3.1-8B-Instruct
-  seed: 42
-  tensor_parallel_size: 2
-prefix_caching_flag: false
-query_range: 500
-query_trace_map: data/anns/query_trace_map_5k.json
-replay:
-  delay_multiplier: 1.0
-  end_template: '
-
-
-    Question:'
-  max_output_tokens: 1
-  poisson_avg_arrival_time: 0.125
-  template: 'Answer the question based only on the provided context. If the answer
-    is not in the context, say "I don''t know."
-
-    Context:'
-scheduler: mcps
-trace_dir: data/anns/res
-use_recomputation_latency_predictor: 1
-use_saved_poisson_delays: driver/anns/configs/H100_enhanced_schedulers_v1_full/poisson_delays_0_125.txt
-use_saved_recomputation_data_path: driver/recomputation_perf_model/H200_tp2_recomputation_latency.json
-use_saved_swap_latency_data_path: driver/swap_perf_model/H200_tp2_swap_kernel_latency.json
-use_swap_kernel: 1
-use_swap_latency_predictor: 1

run_log/anns/H100_enhanced_schedulers_v1_full/mcps/10-16-2025_18-37-45/run_metrics.csv

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:168f373a2803b00f7759e4051aa9eefdd8041578d645f15684e9aa2293bcb139
-size 629799

run_log/anns/H100_enhanced_schedulers_v1_full/mcps/10-16-2025_19-13-38/config_10-16-2025_19-13-38.yaml

@@ -1,38 +0,0 @@
-config_for_run: driver/anns/run_log/H100_enhanced_schedulers_v1_full/mcps
-data_dir: data/anns
-experiments:
-- concurrency: true
-  stream: true
-log_stats: true
-logging:
-  suppress_vllm_logs: true
-  verbose_console: false
-model:
-  gpu_memory_utilization: 0.8
-  max_token_budget: 8192
-  name: meta-llama/Meta-Llama-3.1-8B-Instruct
-  seed: 42
-  tensor_parallel_size: 2
-prefix_caching_flag: false
-query_range: 500
-query_trace_map: data/anns/query_trace_map_5k.json
-replay:
-  delay_multiplier: 1.0
-  end_template: '
-
-
-    Question:'
-  max_output_tokens: 1
-  poisson_avg_arrival_time: 0.25
-  template: 'Answer the question based only on the provided context. If the answer
-    is not in the context, say "I don''t know."
-
-    Context:'
-scheduler: mcps
-trace_dir: data/anns/res
-use_recomputation_latency_predictor: 1
-use_saved_poisson_delays: driver/anns/configs/H100_enhanced_schedulers_v1_full/poisson_delays_0_25.txt
-use_saved_recomputation_data_path: driver/recomputation_perf_model/H200_tp2_recomputation_latency.json
-use_saved_swap_latency_data_path: driver/swap_perf_model/H200_tp2_swap_kernel_latency.json
-use_swap_kernel: 1
-use_swap_latency_predictor: 1

run_log/anns/H100_enhanced_schedulers_v1_full/mcps/10-16-2025_19-13-38/run_metrics.csv

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:63309a5237d8843012236f237f32ca14db411bde2bd55705b2af21fe77b19d2d
-size 630461

run_log/anns/H200_enhanced_schedulers_v1_full/default_vllm/10-16-2025_18-09-21/collected_outputs_streaming.json

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:c27496e7fc8d3f3bf6fef6dbf7e9c6b9eb91f326c45de3bf8cc2ead7fb3ddcae
-size 76115256

run_log/anns/H200_enhanced_schedulers_v1_full/default_vllm/10-16-2025_18-09-21/config_10-16-2025_18-09-21.yaml

@@ -1,35 +0,0 @@
-config_for_run: driver/anns/run_log/H200_enhanced_schedulers_v1_full/default_vllm
-data_dir: data/anns
-experiments:
-- concurrency: true
-  stream: true
-- concurrency: true
-  stream: false
-log_stats: true
-logging:
-  suppress_vllm_logs: true
-  verbose_console: false
-model:
-  gpu_memory_utilization: 0.8
-  max_token_budget: 8192
-  name: meta-llama/Meta-Llama-3.1-8B-Instruct
-  seed: 42
-  tensor_parallel_size: 2
-prefix_caching_flag: false
-query_range: 500
-query_trace_map: data/anns/query_trace_map_5k.json
-replay:
-  delay_multiplier: 1.0
-  end_template: '
-
-
-    Question:'
-  max_output_tokens: 1
-  poisson_avg_arrival_time: 0.0625
-  template: 'Answer the question based only on the provided context. If the answer
-    is not in the context, say "I don''t know."
-
-    Context:'
-scheduler: default_vllm
-trace_dir: data/anns/res
-use_saved_poisson_delays: driver/anns/configs/H200_enhanced_schedulers_v1_full/poisson_delays_0_0625.txt

run_log/anns/H200_enhanced_schedulers_v1_full/default_vllm/10-16-2025_18-09-21/run_metrics.csv

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:22af42c7be4d52cf9884b4760e9f53c85eed66c5a724caef1fd18eab11174bb4
-size 996357

run_log/anns/H200_enhanced_schedulers_v1_full/default_vllm/10-16-2025_18-37-45/collected_outputs_streaming.json

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:0bb16e7f142859b08a5ce6ef373c2757e51f8571e024ac83349e55c8cf36fe1c
-size 76115252

run_log/anns/H200_enhanced_schedulers_v1_full/default_vllm/10-16-2025_18-37-45/config_10-16-2025_18-37-45.yaml

@@ -1,35 +0,0 @@
-config_for_run: driver/anns/run_log/H200_enhanced_schedulers_v1_full/default_vllm
-data_dir: data/anns
-experiments:
-- concurrency: true
-  stream: true
-- concurrency: true
-  stream: false
-log_stats: true
-logging:
-  suppress_vllm_logs: true
-  verbose_console: false
-model:
-  gpu_memory_utilization: 0.8
-  max_token_budget: 8192
-  name: meta-llama/Meta-Llama-3.1-8B-Instruct
-  seed: 42
-  tensor_parallel_size: 2
-prefix_caching_flag: false
-query_range: 500
-query_trace_map: data/anns/query_trace_map_5k.json
-replay:
-  delay_multiplier: 1.0
-  end_template: '
-
-
-    Question:'
-  max_output_tokens: 1
-  poisson_avg_arrival_time: 0.125
-  template: 'Answer the question based only on the provided context. If the answer
-    is not in the context, say "I don''t know."
-
-    Context:'
-scheduler: default_vllm
-trace_dir: data/anns/res
-use_saved_poisson_delays: driver/anns/configs/H200_enhanced_schedulers_v1_full/poisson_delays_0_125.txt

run_log/anns/H200_enhanced_schedulers_v1_full/default_vllm/10-16-2025_18-37-45/run_metrics.csv

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:041030ce63eff15fa98cf5a05c5a76587aa94fd617864aa5ca0d9a191a51a40f
-size 1004315

run_log/anns/H200_enhanced_schedulers_v1_full/default_vllm/10-16-2025_19-14-45/collected_outputs_streaming.json

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:44f4b615490d9f6bedb7205868468d36cc1c7eacfe8d80c0c1ebd645ae3d4dbf
-size 76115355

run_log/anns/H200_enhanced_schedulers_v1_full/default_vllm/10-16-2025_19-14-45/config_10-16-2025_19-14-45.yaml

@@ -1,35 +0,0 @@
-config_for_run: driver/anns/run_log/H200_enhanced_schedulers_v1_full/default_vllm
-data_dir: data/anns
-experiments:
-- concurrency: true
-  stream: true
-- concurrency: true
-  stream: false
-log_stats: true
-logging:
-  suppress_vllm_logs: true
-  verbose_console: false
-model:
-  gpu_memory_utilization: 0.8
-  max_token_budget: 8192
-  name: meta-llama/Meta-Llama-3.1-8B-Instruct
-  seed: 42
-  tensor_parallel_size: 2
-prefix_caching_flag: false
-query_range: 500
-query_trace_map: data/anns/query_trace_map_5k.json
-replay:
-  delay_multiplier: 1.0
-  end_template: '
-
-
-    Question:'
-  max_output_tokens: 1
-  poisson_avg_arrival_time: 0.25
-  template: 'Answer the question based only on the provided context. If the answer
-    is not in the context, say "I don''t know."
-
-    Context:'
-scheduler: default_vllm
-trace_dir: data/anns/res
-use_saved_poisson_delays: driver/anns/configs/H200_enhanced_schedulers_v1_full/poisson_delays_0_25.txt

run_log/anns/H200_enhanced_schedulers_v1_full/default_vllm/10-16-2025_19-14-45/run_metrics.csv

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:c3dc5ec855aa106838d79401ab8ab556433f10580b2ebaa72001fb9be469d5ca
-size 1006687

run_log/anns/H200_enhanced_schedulers_v1_full/fcfs/10-16-2025_18-16-56/collected_outputs_streaming.json

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:f62265866759f2980be81218f6520fd0208331cbde063ccb6368cf3da34f3ff5
-size 76115351

run_log/anns/H200_enhanced_schedulers_v1_full/fcfs/10-16-2025_18-16-56/config_10-16-2025_18-16-56.yaml

@@ -1,38 +0,0 @@
-config_for_run: driver/anns/run_log/H200_enhanced_schedulers_v1_full/fcfs
-data_dir: data/anns
-experiments:
-- concurrency: true
-  stream: true
-log_stats: true
-logging:
-  suppress_vllm_logs: true
-  verbose_console: false
-model:
-  gpu_memory_utilization: 0.8
-  max_token_budget: 8192
-  name: meta-llama/Meta-Llama-3.1-8B-Instruct
-  seed: 42
-  tensor_parallel_size: 2
-prefix_caching_flag: false
-query_range: 500
-query_trace_map: data/anns/query_trace_map_5k.json
-replay:
-  delay_multiplier: 1.0
-  end_template: '
-
-
-    Question:'
-  max_output_tokens: 1
-  poisson_avg_arrival_time: 0.0625
-  template: 'Answer the question based only on the provided context. If the answer
-    is not in the context, say "I don''t know."
-
-    Context:'
-scheduler: fcfs
-trace_dir: data/anns/res
-use_recomputation_latency_predictor: 1
-use_saved_poisson_delays: driver/anns/configs/H200_enhanced_schedulers_v1_full/poisson_delays_0_0625.txt
-use_saved_recomputation_data_path: driver/recomputation_perf_model/H200_tp2_recomputation_latency.json
-use_saved_swap_latency_data_path: driver/swap_perf_model/H200_tp2_swap_kernel_latency.json
-use_swap_kernel: 1
-use_swap_latency_predictor: 1

run_log/anns/H200_enhanced_schedulers_v1_full/fcfs/10-16-2025_18-16-56/run_metrics.csv

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:7bda09918d200a9d167c943f2e67363b44eb62aee5f51bc7270819685f590b01
-size 623281

run_log/anns/H200_enhanced_schedulers_v1_full/fcfs/10-16-2025_18-45-07/collected_outputs_streaming.json

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:292e4db0a42bb2ce59641f204c564940112c8aad3fecfca2b5a8956c064aaede
-size 76115311

run_log/anns/H200_enhanced_schedulers_v1_full/fcfs/10-16-2025_18-45-07/config_10-16-2025_18-45-07.yaml

@@ -1,38 +0,0 @@
-config_for_run: driver/anns/run_log/H200_enhanced_schedulers_v1_full/fcfs
-data_dir: data/anns
-experiments:
-- concurrency: true
-  stream: true
-log_stats: true
-logging:
-  suppress_vllm_logs: true
-  verbose_console: false
-model:
-  gpu_memory_utilization: 0.8
-  max_token_budget: 8192
-  name: meta-llama/Meta-Llama-3.1-8B-Instruct
-  seed: 42
-  tensor_parallel_size: 2
-prefix_caching_flag: false
-query_range: 500
-query_trace_map: data/anns/query_trace_map_5k.json
-replay:
-  delay_multiplier: 1.0
-  end_template: '
-
-
-    Question:'
-  max_output_tokens: 1
-  poisson_avg_arrival_time: 0.125
-  template: 'Answer the question based only on the provided context. If the answer
-    is not in the context, say "I don''t know."
-
-    Context:'
-scheduler: fcfs
-trace_dir: data/anns/res
-use_recomputation_latency_predictor: 1
-use_saved_poisson_delays: driver/anns/configs/H200_enhanced_schedulers_v1_full/poisson_delays_0_125.txt
-use_saved_recomputation_data_path: driver/recomputation_perf_model/H200_tp2_recomputation_latency.json
-use_saved_swap_latency_data_path: driver/swap_perf_model/H200_tp2_swap_kernel_latency.json
-use_swap_kernel: 1
-use_swap_latency_predictor: 1

run_log/anns/H200_enhanced_schedulers_v1_full/fcfs/10-16-2025_18-45-07/run_metrics.csv

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:8902523002297b4144c104ee706229fc8254594dc62f9b047c4fca65b2c30cd0
-size 627299

run_log/anns/H200_enhanced_schedulers_v1_full/fcfs/10-16-2025_19-21-15/collected_outputs_streaming.json

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:203928ba43905a334ca789e54c71ddaed1314631b838149d6952598f294dda76
-size 76115350

run_log/anns/H200_enhanced_schedulers_v1_full/fcfs/10-16-2025_19-21-15/config_10-16-2025_19-21-15.yaml

@@ -1,38 +0,0 @@
-config_for_run: driver/anns/run_log/H200_enhanced_schedulers_v1_full/fcfs
-data_dir: data/anns
-experiments:
-- concurrency: true
-  stream: true
-log_stats: true
-logging:
-  suppress_vllm_logs: true
-  verbose_console: false
-model:
-  gpu_memory_utilization: 0.8
-  max_token_budget: 8192
-  name: meta-llama/Meta-Llama-3.1-8B-Instruct
-  seed: 42
-  tensor_parallel_size: 2
-prefix_caching_flag: false
-query_range: 500
-query_trace_map: data/anns/query_trace_map_5k.json
-replay:
-  delay_multiplier: 1.0
-  end_template: '
-
-
-    Question:'
-  max_output_tokens: 1
-  poisson_avg_arrival_time: 0.25
-  template: 'Answer the question based only on the provided context. If the answer
-    is not in the context, say "I don''t know."
-
-    Context:'
-scheduler: fcfs
-trace_dir: data/anns/res
-use_recomputation_latency_predictor: 1
-use_saved_poisson_delays: driver/anns/configs/H200_enhanced_schedulers_v1_full/poisson_delays_0_25.txt
-use_saved_recomputation_data_path: driver/recomputation_perf_model/H200_tp2_recomputation_latency.json
-use_saved_swap_latency_data_path: driver/swap_perf_model/H200_tp2_swap_kernel_latency.json
-use_swap_kernel: 1
-use_swap_latency_predictor: 1

run_log/anns/H200_enhanced_schedulers_v1_full/fcfs/10-16-2025_19-21-15/run_metrics.csv

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:324bbaf183ba03bfaf75e385d96b61d19b24eaa09732305e13eb76c8013ca96c
-size 630349

run_log/anns/H200_enhanced_schedulers_v1_full/lcas/10-16-2025_18-23-52/collected_outputs_streaming.json

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:fba32e0d25b69fb5593e325e76bd86eb9d6b23c66e9b9c073644040eedd6c8f4
-size 76115364

run_log/anns/H200_enhanced_schedulers_v1_full/lcas/10-16-2025_18-23-52/config_10-16-2025_18-23-52.yaml

@@ -1,38 +0,0 @@
-config_for_run: driver/anns/run_log/H200_enhanced_schedulers_v1_full/lcas
-data_dir: data/anns
-experiments:
-- concurrency: true
-  stream: true
-log_stats: true
-logging:
-  suppress_vllm_logs: true
-  verbose_console: false
-model:
-  gpu_memory_utilization: 0.8
-  max_token_budget: 8192
-  name: meta-llama/Meta-Llama-3.1-8B-Instruct
-  seed: 42
-  tensor_parallel_size: 2
-prefix_caching_flag: false
-query_range: 500
-query_trace_map: data/anns/query_trace_map_5k.json
-replay:
-  delay_multiplier: 1.0
-  end_template: '
-
-
-    Question:'
-  max_output_tokens: 1
-  poisson_avg_arrival_time: 0.0625
-  template: 'Answer the question based only on the provided context. If the answer
-    is not in the context, say "I don''t know."
-
-    Context:'
-scheduler: lcas
-trace_dir: data/anns/res
-use_recomputation_latency_predictor: 1
-use_saved_poisson_delays: driver/anns/configs/H200_enhanced_schedulers_v1_full/poisson_delays_0_0625.txt
-use_saved_recomputation_data_path: driver/recomputation_perf_model/H200_tp2_recomputation_latency.json
-use_saved_swap_latency_data_path: driver/swap_perf_model/H200_tp2_swap_kernel_latency.json
-use_swap_kernel: 1
-use_swap_latency_predictor: 1

run_log/anns/H200_enhanced_schedulers_v1_full/lcas/10-16-2025_18-23-52/run_metrics.csv

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:26d18c3778f0dda363484903908bd4c208e2e46be2825b5172fb26d8373ab04e
-size 622330

run_log/anns/H200_enhanced_schedulers_v1_full/lcas/10-16-2025_18-55-07/collected_outputs_streaming.json

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:2a3cda1a7e1ea9c988bc5d790dd1b4e1daf6c1111cbfc5e7375e5811444eddfc
-size 76115301

run_log/anns/H200_enhanced_schedulers_v1_full/lcas/10-16-2025_18-55-07/config_10-16-2025_18-55-07.yaml

@@ -1,38 +0,0 @@
-config_for_run: driver/anns/run_log/H200_enhanced_schedulers_v1_full/lcas
-data_dir: data/anns
-experiments:
-- concurrency: true
-  stream: true
-log_stats: true
-logging:
-  suppress_vllm_logs: true
-  verbose_console: false
-model:
-  gpu_memory_utilization: 0.8
-  max_token_budget: 8192
-  name: meta-llama/Meta-Llama-3.1-8B-Instruct
-  seed: 42
-  tensor_parallel_size: 2
-prefix_caching_flag: false
-query_range: 500
-query_trace_map: data/anns/query_trace_map_5k.json
-replay:
-  delay_multiplier: 1.0
-  end_template: '
-
-
-    Question:'
-  max_output_tokens: 1
-  poisson_avg_arrival_time: 0.125
-  template: 'Answer the question based only on the provided context. If the answer
-    is not in the context, say "I don''t know."
-
-    Context:'
-scheduler: lcas
-trace_dir: data/anns/res
-use_recomputation_latency_predictor: 1
-use_saved_poisson_delays: driver/anns/configs/H200_enhanced_schedulers_v1_full/poisson_delays_0_125.txt
-use_saved_recomputation_data_path: driver/recomputation_perf_model/H200_tp2_recomputation_latency.json
-use_saved_swap_latency_data_path: driver/swap_perf_model/H200_tp2_swap_kernel_latency.json
-use_swap_kernel: 1
-use_swap_latency_predictor: 1

run_log/anns/H200_enhanced_schedulers_v1_full/lcas/10-16-2025_18-55-07/run_metrics.csv

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:0c90fb45c7b904cabcab51cade21ae6f1f807c6ac0d5de1f14e58a1297fdd58d
-size 625225

run_log/anns/H200_enhanced_schedulers_v1_full/lcas/10-16-2025_19-27-28/collected_outputs_streaming.json

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:514e0fa9255dc8c857503da15f735dd93d0e5b990ca5737f3a75e6bf6084e0f8
-size 76115302

run_log/anns/H200_enhanced_schedulers_v1_full/lcas/10-16-2025_19-27-28/config_10-16-2025_19-27-28.yaml

@@ -1,38 +0,0 @@
-config_for_run: driver/anns/run_log/H200_enhanced_schedulers_v1_full/lcas
-data_dir: data/anns
-experiments:
-- concurrency: true
-  stream: true
-log_stats: true
-logging:
-  suppress_vllm_logs: true
-  verbose_console: false
-model:
-  gpu_memory_utilization: 0.8
-  max_token_budget: 8192
-  name: meta-llama/Meta-Llama-3.1-8B-Instruct
-  seed: 42
-  tensor_parallel_size: 2
-prefix_caching_flag: false
-query_range: 500
-query_trace_map: data/anns/query_trace_map_5k.json
-replay:
-  delay_multiplier: 1.0
-  end_template: '
-
-
-    Question:'
-  max_output_tokens: 1
-  poisson_avg_arrival_time: 0.25
-  template: 'Answer the question based only on the provided context. If the answer
-    is not in the context, say "I don''t know."
-
-    Context:'
-scheduler: lcas
-trace_dir: data/anns/res
-use_recomputation_latency_predictor: 1
-use_saved_poisson_delays: driver/anns/configs/H200_enhanced_schedulers_v1_full/poisson_delays_0_25.txt
-use_saved_recomputation_data_path: driver/recomputation_perf_model/H200_tp2_recomputation_latency.json
-use_saved_swap_latency_data_path: driver/swap_perf_model/H200_tp2_swap_kernel_latency.json
-use_swap_kernel: 1
-use_swap_latency_predictor: 1