AgentPerfBench / README.md
booth-algo's picture
Keep MSE validation tabular in dataset repo (#8)
2b82ddc
|
Raw
History Blame
20.2 kB
metadata
license: apache-2.0
task_categories:
  - tabular-regression
language:
  - en
tags:
  - llm-inference
  - benchmarking
  - gpu-profiling
  - vllm
  - sglang
  - agentic-workloads
size_categories:
  - 100K<n<1M
pretty_name: AgentPerfBench
version: '1.0'
configs:
  - config_name: trace_replay
    data_files:
      - split: summary
        path: trace_replay/summary.parquet
  - config_name: synthetic_distributional
    data_files:
      - split: summary
        path: synthetic_distributional/summary.parquet
  - config_name: kernels_labeled
    data_files:
      - split: train
        path: kernel_profiles/kernels_labeled.parquet
  - config_name: mse_validation
    data_files:
      - split: summary
        path: mse_validation/summary.parquet
  - config_name: layer_roofline
    data_files:
      - split: summary
        path: layer_roofline/summary.parquet
dataset_info:
  - config_name: trace_replay
    features:
      - name: run_id
        dtype: string
      - name: model
        dtype: string
      - name: model_family
        dtype: string
      - name: hardware
        dtype: string
      - name: engine
        dtype: string
      - name: tensor_parallelism
        dtype: int64
      - name: profile
        dtype: string
      - name: concurrency
        dtype: int64
      - name: num_requests
        dtype: int64
      - name: duration_s
        dtype: float64
      - name: successful_requests
        dtype: int64
      - name: failed_requests
        dtype: int64
      - name: request_throughput
        dtype: float64
      - name: input_token_throughput
        dtype: float64
      - name: output_token_throughput
        dtype: float64
      - name: total_token_throughput
        dtype: float64
      - name: mean_ttft_ms
        dtype: float64
      - name: median_ttft_ms
        dtype: float64
      - name: p90_ttft_ms
        dtype: float64
      - name: p99_ttft_ms
        dtype: float64
      - name: mean_tpot_ms
        dtype: float64
      - name: median_tpot_ms
        dtype: float64
      - name: p90_tpot_ms
        dtype: float64
      - name: p99_tpot_ms
        dtype: float64
      - name: mean_itl_ms
        dtype: float64
      - name: median_itl_ms
        dtype: float64
      - name: p90_itl_ms
        dtype: float64
      - name: p99_itl_ms
        dtype: float64
      - name: mean_e2el_ms
        dtype: float64
      - name: median_e2el_ms
        dtype: float64
      - name: p90_e2el_ms
        dtype: float64
      - name: p99_e2el_ms
        dtype: float64
    splits:
      - name: summary
        num_examples: 3147
        num_bytes: 694254
  - config_name: synthetic_distributional
    features:
      - name: run_id
        dtype: string
      - name: model
        dtype: string
      - name: model_family
        dtype: string
      - name: hardware
        dtype: string
      - name: engine
        dtype: string
      - name: tensor_parallelism
        dtype: int64
      - name: profile
        dtype: string
      - name: concurrency
        dtype: int64
      - name: num_requests
        dtype: int64
      - name: duration_s
        dtype: float64
      - name: successful_requests
        dtype: int64
      - name: failed_requests
        dtype: int64
      - name: request_throughput
        dtype: float64
      - name: input_token_throughput
        dtype: float64
      - name: output_token_throughput
        dtype: float64
      - name: total_token_throughput
        dtype: float64
      - name: mean_ttft_ms
        dtype: float64
      - name: median_ttft_ms
        dtype: float64
      - name: p90_ttft_ms
        dtype: float64
      - name: p99_ttft_ms
        dtype: float64
      - name: mean_tpot_ms
        dtype: float64
      - name: median_tpot_ms
        dtype: float64
      - name: p90_tpot_ms
        dtype: float64
      - name: p99_tpot_ms
        dtype: float64
      - name: mean_itl_ms
        dtype: float64
      - name: median_itl_ms
        dtype: float64
      - name: p90_itl_ms
        dtype: float64
      - name: p99_itl_ms
        dtype: float64
      - name: mean_e2el_ms
        dtype: float64
      - name: median_e2el_ms
        dtype: float64
      - name: p90_e2el_ms
        dtype: float64
      - name: p99_e2el_ms
        dtype: float64
    splits:
      - name: summary
        num_examples: 245
        num_bytes: 70836
  - config_name: kernels_labeled
    features:
      - name: source
        dtype: string
      - name: gpu
        dtype: string
      - name: model
        dtype: string
      - name: kernel_family
        dtype: string
      - name: kernel_name
        dtype: string
      - name: dtype
        dtype: string
      - name: held_out
        dtype: bool
      - name: M
        dtype: float64
      - name: 'N'
        dtype: float64
      - name: K
        dtype: float64
      - name: bs
        dtype: float64
      - name: seq
        dtype: float64
      - name: n_heads
        dtype: float64
      - name: head_dim
        dtype: float64
      - name: kv_heads
        dtype: float64
      - name: numel
        dtype: float64
      - name: op_type
        dtype: string
      - name: gpu_time_duration_ms
        dtype: float64
      - name: launch_block_size
        dtype: float64
      - name: launch_grid_size
        dtype: float64
      - name: dram_bytes_sum
        dtype: float64
      - name: launch_registers_per_thread
        dtype: float64
    splits:
      - name: train
        num_examples: 148077
  - config_name: mse_validation
    features:
      - name: validation_id
        dtype: string
      - name: tier
        dtype: string
      - name: stage
        dtype: string
      - name: stage_label
        dtype: string
      - name: comparison_group
        dtype: string
      - name: raw_json_r2_uri
        dtype: string
      - name: per_turn_json_r2_uri
        dtype: string
      - name: replay_kind
        dtype: string
      - name: workload
        dtype: string
      - name: profile
        dtype: string
      - name: model
        dtype: string
      - name: hardware
        dtype: string
      - name: engine
        dtype: string
      - name: tensor_parallelism
        dtype: int64
      - name: concurrency
        dtype: int64
      - name: sessions
        dtype: int64
      - name: source_locked
        dtype: bool
      - name: prefix_caching_state
        dtype: string
      - name: chunked_prefill
        dtype: string
      - name: prefix_aware_synthetic
        dtype: bool
      - name: shared_prefix_tokens
        dtype: int64
      - name: shared_prefix_block_aligned
        dtype: bool
      - name: synthetic_filler_style
        dtype: string
      - name: synthetic_target_chars_per_token
        dtype: float64
      - name: max_model_len
        dtype: int64
      - name: success_rate
        dtype: float64
      - name: num_requests
        dtype: int64
      - name: duration_s
        dtype: float64
      - name: successful_requests
        dtype: int64
      - name: failed_requests
        dtype: int64
      - name: request_throughput
        dtype: float64
      - name: input_token_throughput
        dtype: float64
      - name: output_token_throughput
        dtype: float64
      - name: total_token_throughput
        dtype: float64
      - name: mean_ttft_ms
        dtype: float64
      - name: median_ttft_ms
        dtype: float64
      - name: p90_ttft_ms
        dtype: float64
      - name: p99_ttft_ms
        dtype: float64
      - name: mean_tpot_ms
        dtype: float64
      - name: median_tpot_ms
        dtype: float64
      - name: p90_tpot_ms
        dtype: float64
      - name: p99_tpot_ms
        dtype: float64
      - name: mean_itl_ms
        dtype: float64
      - name: median_itl_ms
        dtype: float64
      - name: p90_itl_ms
        dtype: float64
      - name: p99_itl_ms
        dtype: float64
      - name: mean_e2el_ms
        dtype: float64
      - name: median_e2el_ms
        dtype: float64
      - name: p90_e2el_ms
        dtype: float64
      - name: p99_e2el_ms
        dtype: float64
    splits:
      - name: summary
        num_examples: 28
        num_bytes: 40940
  - config_name: layer_roofline
    features:
      - name: record_type
        dtype: string
      - name: source_file
        dtype: string
      - name: model
        dtype: string
      - name: hardware
        dtype: string
      - name: engine
        dtype: string
      - name: dtype
        dtype: string
      - name: tensor_parallelism
        dtype: int64
      - name: phase
        dtype: string
      - name: batch_size
        dtype: int64
      - name: sequence_length
        dtype: int64
      - name: q_len
        dtype: int64
      - name: kv_len
        dtype: int64
      - name: component_name
        dtype: string
      - name: component_bound
        dtype: string
      - name: flops
        dtype: float64
      - name: bytes
        dtype: float64
      - name: operational_intensity_flop_per_byte
        dtype: float64
      - name: ridge_point_flop_per_byte
        dtype: float64
      - name: total_flops
        dtype: float64
      - name: total_bytes
        dtype: float64
      - name: overall_bound
        dtype: string
      - name: kernel_id
        dtype: int64
      - name: kernel_name
        dtype: string
      - name: block_size
        dtype: string
      - name: grid_size
        dtype: string
      - name: duration_us
        dtype: float64
      - name: compute_sm_throughput_pct
        dtype: float64
      - name: dram_throughput_pct
        dtype: float64
      - name: memory_throughput_pct
        dtype: float64
      - name: l1_tex_cache_throughput_pct
        dtype: float64
      - name: l2_cache_throughput_pct
        dtype: float64
      - name: sm_frequency_ghz
        dtype: float64
      - name: dram_frequency_ghz
        dtype: float64
      - name: artifact_path
        dtype: string
      - name: artifact_kind
        dtype: string
      - name: artifact_bytes
        dtype: int64
      - name: artifact_description
        dtype: string
      - name: notes
        dtype: string
    splits:
      - name: summary
        num_examples: 56
        num_bytes: 27430

AgentPerfBench

LLM inference benchmark: 3,392 main sweep rows measuring TTFT, TPOT, ITL, and throughput across 9 models, up to 14 GPU configurations, and 2 serving engines (vLLM 0.19.0, SGLang 0.5.9). All models served in BF16 except gpt-oss, which uses mxfp4 for projection weights. The dataset also includes 148,077 per-kernel NCU profiles, 28 curated tabular MSE validation rows for the distributional synthetic replay generator, and 56 layer-roofline validation rows.

Dataset configurations

The dataset provides five configurations. trace_replay replays exact input/output sequences from recorded agent sessions. synthetic_distributional samples from statistical distributions fitted to those same workloads, trading fidelity for faster sweeps across the hardware matrix. kernels_labeled contains per-kernel Nsight Compute labels. mse_validation contains tabular paired synthetic-vs-real validation rows and ablations for the final APC-aware synthetic generator, with raw JSON artifacts referenced in R2 rather than stored in the dataset repo. layer_roofline exposes the per-layer roofline validation artifacts as a loadable tabular subset.

trace_replay (3,147 rows)

Replays exact ISL/OSL sequences from recorded agent sessions (SWE-Bench, TerminalBench, OSWorld, ShareGPT). Covers 77 unique (model, hardware, engine) combinations across 17 profiles and 6 concurrency levels. The full 5-dimensional matrix is 12.2% filled; not all models run on all hardware.

17 profiles: chat-medium, chat-multiturn-long, chat-multiturn-medium, chat-multiturn-short, chat-short, chat-singleturn, coding-singleturn, decode-heavy, osworld-multiturn-long, osworld-multiturn-medium, osworld-multiturn-short, prefill-heavy, random-1k, swebench-multiturn-medium, swebench-multiturn-short, terminalbench-multiturn-medium, terminalbench-multiturn-short

synthetic_distributional (245 rows)

Samples ISL/OSL from lognormal distributions fitted to real workload statistics. Covers 42 unique (model, hardware, engine) combinations across 6 profiles and 7 concurrency levels (3.0% matrix fill). gpt-oss-120b, 3090x8, and A100-40GBx8 are excluded from this configuration.

6 profiles: chat-multiturn, chat-singleturn, coding-singleturn, osworld-multiturn, swebench-multiturn, terminalbench-multiturn

kernels_labeled (148,077 rows)

Per-kernel Nsight Compute (ncu) profiles across 4 GPUs (A100, H100, RTX 3090, RTX 2080 Ti) and 13 model/sweep sources.

mse_validation (28 rows)

Curated H100 / Llama-3.1-8B / vLLM validation table for the distributional synthetic replay generator. The main rows keep paired MSE distributional replay and real trace replay runs with success rate at least 75%; supplementary rows preserve no-replacement and high-concurrency debug runs. Raw aggregate/per-turn JSON artifacts are referenced through R2 URI columns, not stored as dataset files.

The headline SWE-bench C=5 source-locked cascade is:

Condition Turn 10-19 E2EL delta
APC on, English filler, source-locked +45.5%
APC on, code-morph filler, no shared prefix +31.1%
APC off, code-morph filler, no shared prefix +11.7%
APC on, code-morph filler, 1024-token shared prefix -3.2%

See mse_validation/README.md, mse_validation/manifest.csv, and mse_validation/fidelity_deltas.csv for the filtering rule and R2 raw-artifact URIs.

layer_roofline (56 rows)

Tabular view over the per-layer roofline evidence for Llama-3.1-8B on H100. This subset combines analytical component OI rows, selected NCU kernel summary rows, and an artifact manifest pointing to the raw per_layer_oi_cf/ evidence files.

Record types: analytical_total, analytical_component, ncu_kernel, artifact.

Concurrency filtering

The benchmark harness capped actual concurrent connections at the session pool size. Rows where declared concurrency exceeded the pool were excluded:

  • trace_replay: concurrency > 100 removed (session pool was 100). Remaining values: {1, 5, 10, 20, 40, 80}.
  • synthetic_distributional (pre-fix): concurrency > 10 removed (session pool was 10). Post-fix data has no cap. Remaining values: {1, 5, 10, 40, 80, 200, 320}.
Config Rows
trace_replay 3,147
synthetic_distributional 245
mse_validation 28
layer_roofline 56
kernels_labeled 148,077

Failed requests

Some runs produce request failures, typically at high concurrency where the engine hits memory or timeout limits. 30.8% of trace_replay rows and 42% of synthetic_distributional rows have failed_requests > 0. Summary metrics (TTFT, TPOT, throughput) are computed from successful requests only.

Coverage

Hardware

All benchmarks collected on PyTorch 2.10.0, CUDA 12.8.

GPU VRAM HBM bandwidth Peak half-precision TFLOPS
NVIDIA H100 SXM 80 GB 3.35 TB/s 989
NVIDIA A100 SXM4 40 GB 1.56 TB/s 312
NVIDIA RTX 3090 24 GB 936 GB/s 71
NVIDIA RTX 2080 Ti 11 GB 616 GB/s 27

Multi-GPU configurations: 1, 2, 4, or 8 GPUs with tensor parallelism. TP degree depends on model size and available GPUs.

Models

All models served in BF16 unless noted.

Model Family Parameters Architecture Notes
Llama-3.1-8B Llama 8B Dense
Llama-3.1-70B Llama 70B Dense
Llama-3.3-70B Llama 70B Dense
Qwen2.5-72B Qwen 72B Dense
Qwen3.5-9B Qwen 9B Dense
Qwen3.5-27B Qwen 27B Dense
Mixtral-8x7B Mixtral 46.7B (12.9B active) MoE
gpt-oss-20b GPT-OSS 21B (3.6B active) MoE mxfp4 projections
gpt-oss-120b GPT-OSS 117B (5.1B active) MoE mxfp4 projections

Model names in this table match the model column in the parquet files.

Engines

  • vLLM 0.19.0
  • SGLang 0.5.9

Schema

Each row in the serving summary.parquet configs:

Column Type Description
run_id string Deterministic hash of run parameters
model string Model short name
model_family string Model family (llama, qwen, gpt-oss, mixtral)
hardware string GPU configuration (e.g., H100x4)
engine string Serving engine (vllm, sglang)
tensor_parallelism int TP degree
profile string Workload profile name
concurrency int Concurrent request level
num_requests int Total requests in run
duration_s float Total run duration
successful_requests int Completed requests
failed_requests int Failed requests
request_throughput float Requests/second
input_token_throughput float Input tokens/second
output_token_throughput float Output tokens/second
total_token_throughput float Total tokens/second
mean/median/p90/p99_ttft_ms float Time to first token
mean/median/p90/p99_tpot_ms float Time per output token
mean/median/p90/p99_itl_ms float Inter-token latency
mean/median/p90/p99_e2el_ms float End-to-end latency

Loading

from datasets import load_dataset

ds = load_dataset("agent-perf-bench/AgentPerfBench", "trace_replay")
# or "synthetic_distributional", "kernels_labeled", "mse_validation", "layer_roofline"

Benchmark methodology

  • Closed-loop concurrency with semaphore control.
  • Concurrency levels: {1, 5, 10, 20, 40, 80} (trace_replay), {1, 5, 10, 40, 80, 200, 320} (synthetic_distributional).
  • 3-request warmup before each configuration.
  • Metrics: TTFT, TPOT, ITL, E2EL, request throughput, token throughput.
  • Summary statistics: mean, median, p90, p99.
  • Collection period: March 2026 onwards.
  • PyTorch 2.10.0, CUDA 12.8 on all machines. All models served in BF16 (gpt-oss: mxfp4 projection weights).

Future releases

  • Full per-request and multi-turn granularity data for the main sweep (pending raw JSON availability from collection infrastructure). Curated raw JSONs are referenced from R2 for mse_validation.
  • Additional per-kernel roofline profiles beyond the included kernels_labeled, layer_roofline, and raw per_layer_oi_cf artifacts.
  • This is version 1.0. Updates will be tagged with semantic versions.

Intended uses

  • Inference engine comparison under controlled conditions.
  • Capacity planning for LLM deployments.
  • TTFT scaling with context length in multi-turn sessions.

Limitations

  • Results are specific to tested hardware and software versions (vLLM 0.19.0, SGLang 0.5.9, PyTorch 2.10.0, CUDA 12.8).
  • Synthetic distributional profiles approximate but do not replicate production traffic patterns.
  • Consumer GPU coverage is limited to RTX 3090 and RTX 2080 Ti; no non-NVIDIA accelerators.
  • Closed-loop concurrency only; no open-loop (Poisson) arrivals.
  • The model-hardware-concurrency matrix is sparse (12.2% fill for trace_replay, 3.0% for synthetic_distributional). Not all model-hardware combinations are represented.
  • No model quality metrics. This is a systems benchmark.

Ethical considerations

No PII. Trace-replay profiles derive from open benchmarks (SWE-Bench MIT, TerminalBench, OSWorld). Synthetic profiles use random tokens.

License

Benchmark data released under Apache-2.0. Source datasets retain their original licenses.

Source datasets

Citation

@inproceedings{agentperfbench2026,
  title={AgentPerfBench: A Benchmarking and Evaluation Suite for Inference Performance of Agentic LLMs},
  author={Anonymous},
  booktitle={NeurIPS 2026 Evaluations and Datasets Track},
  year={2026}
}