isage-control-plane-benchmark 0.1.0.1__cp311-none-any.whl

isage_control_plane_benchmark-0.1.0.1.dist-info/METADATA

@@ -0,0 +1,596 @@
+Metadata-Version: 2.4
+Name: isage-control-plane-benchmark
+Version: 0.1.0.1
+Summary: Control Plane scheduling benchmark for the SAGE ecosystem
+Author-email: IntelliStream Team <shuhao_zhang@hust.edu.cn>
+License-Expression: MIT
+Project-URL: Homepage, https://github.com/intellistream/sage-control-plane-benchmark
+Project-URL: Documentation, https://github.com/intellistream/sage-control-plane-benchmark#readme
+Project-URL: Repository, https://github.com/intellistream/sage-control-plane-benchmark
+Project-URL: Issues, https://github.com/intellistream/sage-control-plane-benchmark/issues
+Keywords: sage,benchmark,control-plane,scheduling,evaluation,intellistream
+Classifier: Development Status :: 4 - Beta
+Classifier: Intended Audience :: Developers
+Classifier: Intended Audience :: Science/Research
+Classifier: Programming Language :: Python :: 3
+Classifier: Programming Language :: Python :: 3.11
+Classifier: Programming Language :: Python :: 3.12
+Classifier: Programming Language :: Python :: 3 :: Only
+Classifier: Topic :: Software Development :: Libraries :: Python Modules
+Classifier: Topic :: Scientific/Engineering :: Artificial Intelligence
+Requires-Python: >=3.11
+Description-Content-Type: text/markdown
+License-File: LICENSE
+Requires-Dist: isage-common
+Requires-Dist: isage-kernel
+Requires-Dist: isage-middleware>=0.2.4.0
+Requires-Dist: isage-libs
+Requires-Dist: aiohttp>=3.9.0
+Requires-Dist: numpy<2.3.0,>=1.26.0
+Requires-Dist: pandas>=2.0.0
+Requires-Dist: pyyaml>=6.0
+Requires-Dist: typer<1.0.0,>=0.15.0
+Requires-Dist: rich<14.0.0,>=13.0.0
+Requires-Dist: matplotlib>=3.7.0
+Requires-Dist: seaborn>=0.12.0
+Requires-Dist: jinja2>=3.1.0
+Provides-Extra: dev
+Requires-Dist: pytest>=7.0.0; extra == "dev"
+Requires-Dist: pytest-asyncio>=0.21.0; extra == "dev"
+Requires-Dist: pytest-cov>=4.0.0; extra == "dev"
+Requires-Dist: black>=23.0.0; extra == "dev"
+Requires-Dist: ruff==0.14.6; extra == "dev"
+Requires-Dist: pre-commit>=3.0.0; extra == "dev"
+Requires-Dist: mypy>=1.0.0; extra == "dev"
+Requires-Dist: types-PyYAML>=6.0.0; extra == "dev"
+Dynamic: license-file
+
+# sageLLM Control Plane Benchmark
+
+This module provides comprehensive benchmarking tools for evaluating different scheduling policies
+in sageLLM's Control Plane. It supports both **LLM-only** and **Hybrid (LLM + Embedding)**
+workloads.
+
+## Overview
+
+The benchmark measures key performance metrics across various scheduling strategies:
+
+- **Throughput**: Requests per second and tokens per second
+- **Latency**: End-to-end latency, Time to First Token (TTFT), Time Between Tokens (TBT)
+- **SLO Compliance**: Percentage of requests meeting their SLO deadlines
+- **Error Rates**: Failed requests and timeout rates
+- **Resource Utilization**: GPU memory and compute utilization (optional)
+
+## Architecture
+
+```
+                                 ┌─────────────────────────────────────────────┐
+                                 │           Control Plane                     │
+┌─────────────┐      HTTP        │  ┌─────────────────────────────────────┐   │
+│  Benchmark  │ ───────────────► │  │     Scheduler (Policy: X)           │   │
+│   Client    │                  │  │  ┌───────────┬───────────────────┐  │   │
+│             │                  │  │  │ LLM Queue │ Embedding Queue   │  │   │
+└─────────────┘                  │  │  └───────────┴───────────────────┘  │   │
+     │                           │  └─────────────────────────────────────┘   │
+     │                           └──────────────────┬──────────────────────────┘
+     │                                              │
+     │                           ┌──────────────────┴──────────────────────┐
+     │                           │                                         │
+     ▼                           ▼                                         ▼
+┌─────────────┐           ┌──────────────┐                         ┌──────────────┐
+│   Metrics   │           │ vLLM Inst 1  │                         │  Embedding   │
+│  Collector  │           │ (Qwen-7B)    │                         │   Server     │
+└─────────────┘           ├──────────────┤                         │  (BGE-M3)    │
+                          │ vLLM Inst 2  │                         └──────────────┘
+                          │ (Llama-13B)  │
+                          └──────────────┘
+```
+
+## Quick Start
+
+### Installation
+
+```bash
+# Install sage-benchmark package
+pip install isage-control-plane-benchmark
+
+# Or for development:
+pip install -e "packages/sage-benchmark[dev]"
+
+# CLI dependencies
+pip install typer aiohttp pyyaml
+
+# Visualization dependencies (optional)
+pip install matplotlib jinja2
+```
+
+### Running Your First Benchmark
+
+```bash
+# 1. Run a simple LLM benchmark
+sage-cp-bench run --mode llm --policy fifo --requests 100 --rate 10
+
+# 2. Run a hybrid (LLM + Embedding) benchmark
+sage-cp-bench run --mode hybrid --policy hybrid_slo --llm-ratio 0.7 --requests 100
+
+# 3. Compare multiple policies
+sage-cp-bench compare --mode llm --policies fifo,priority,slo_aware --requests 500
+
+# 4. Run a predefined experiment
+sage-cp-bench experiment --name throughput --policies fifo,priority
+```
+
+## CLI Reference
+
+### Commands Overview
+
+| Command      | Description                                  |
+| ------------ | -------------------------------------------- |
+| `run`        | Run benchmark for a single scheduling policy |
+| `compare`    | Compare multiple scheduling policies         |
+| `sweep`      | Sweep across multiple request rates          |
+| `experiment` | Run predefined experiments                   |
+| `visualize`  | Generate charts from existing results        |
+| `config`     | Show/save example configuration              |
+| `validate`   | Validate a configuration file                |
+
+### `run` Command
+
+```bash
+sage-cp-bench run [OPTIONS]
+
+Options:
+  --mode            -m   [llm|hybrid]  Benchmark mode (default: llm)
+  --control-plane   -c   TEXT          Control Plane URL (default: http://localhost:8080)
+  --policy          -p   TEXT          Scheduling policy (default: fifo)
+  --requests        -n   INTEGER       Number of requests (default: 100)
+  --rate            -r   FLOAT         Request rate req/s (default: 10.0)
+  --llm-ratio            FLOAT         LLM ratio for hybrid mode (default: 0.7)
+  --output          -o   TEXT          Output directory (default: ./benchmark_results)
+  --warmup          -w   INTEGER       Warmup requests (default: 10)
+  --timeout         -t   FLOAT         Request timeout seconds (default: 60.0)
+  --no-visualize                       Disable auto visualization
+  --config               TEXT          Load config from YAML/JSON file
+  --quiet           -q                 Suppress progress output
+```
+
+**Examples:**
+
+```bash
+# LLM-only benchmark
+sage-cp-bench run --mode llm --policy fifo --requests 100 --rate 10
+
+# Hybrid benchmark with 70% LLM, 30% Embedding
+sage-cp-bench run --mode hybrid --policy hybrid_slo --llm-ratio 0.7 --requests 100
+
+# Load configuration from file
+sage-cp-bench run --config benchmark_config.yaml
+```
+
+### `compare` Command
+
+```bash
+sage-cp-bench compare [OPTIONS]
+
+Options:
+  --mode            -m   [llm|hybrid]  Benchmark mode (default: llm)
+  --policies        -p   TEXT          Comma-separated policy list (default: fifo,priority,slo_aware)
+  --requests        -n   INTEGER       Requests per policy (default: 100)
+  --rate            -r   FLOAT         Request rate (default: 10.0)
+  --llm-ratio            FLOAT         LLM ratio for hybrid mode (default: 0.7)
+  --output          -o   TEXT          Output directory
+  --no-visualize                       Disable comparison charts
+```
+
+**Examples:**
+
+```bash
+# Compare LLM scheduling policies
+sage-cp-bench compare --mode llm --policies fifo,priority,slo_aware
+
+# Compare hybrid scheduling policies
+sage-cp-bench compare --mode hybrid --policies fifo,hybrid_slo --llm-ratio 0.7
+```
+
+### `sweep` Command
+
+```bash
+sage-cp-bench sweep [OPTIONS]
+
+Options:
+  --mode            -m   [llm|hybrid]  Benchmark mode (default: llm)
+  --policy          -p   TEXT          Policy to test (default: fifo)
+  --rates                TEXT          Comma-separated rates (default: 10,50,100,200)
+  --requests        -n   INTEGER       Requests per rate (default: 100)
+  --output          -o   TEXT          Output directory
+```
+
+**Examples:**
+
+```bash
+# Sweep request rates for LLM benchmark
+sage-cp-bench sweep --mode llm --policy fifo --rates 10,50,100,200
+
+# Sweep rates for hybrid benchmark
+sage-cp-bench sweep --mode hybrid --policy hybrid_slo --rates 10,50,100
+```
+
+### `experiment` Command
+
+```bash
+sage-cp-bench experiment [OPTIONS]
+
+Options:
+  --name            -e   TEXT          Experiment: throughput|latency|slo|mixed_ratio [required]
+  --control-plane   -c   TEXT          Control Plane URL
+  --requests        -n   INTEGER       Requests per test (default: 500)
+  --rate            -r   INTEGER       Request rate (default: 100)
+  --llm-ratio            FLOAT         LLM ratio (default: 0.5)
+  --policies        -p   TEXT          Policies to test (default: fifo,priority,slo_aware)
+  --output          -o   TEXT          Output directory
+  --no-visualize                       Skip visualization
+```
+
+**Available Experiments:**
+
+| Experiment    | Description                                   |
+| ------------- | --------------------------------------------- |
+| `throughput`  | Sweep request rates to find max throughput    |
+| `latency`     | Analyze latency distribution under fixed load |
+| `slo`         | Compare SLO compliance across policies        |
+| `mixed_ratio` | Test different LLM/Embedding ratios           |
+
+**Examples:**
+
+```bash
+# Run throughput experiment
+sage-cp-bench experiment --name throughput --policies fifo,priority
+
+# Run latency analysis
+sage-cp-bench experiment --name latency --rate 100 --requests 1000
+
+# Run SLO compliance comparison
+sage-cp-bench experiment --name slo --policies fifo,slo_aware
+
+# Run mixed ratio sweep (hybrid only)
+sage-cp-bench experiment --name mixed_ratio --rate 100
+```
+
+### `visualize` Command
+
+```bash
+sage-cp-bench visualize [OPTIONS]
+
+Options:
+  --input           -i   TEXT          Results JSON file [required]
+  --output          -o   TEXT          Output directory (default: ./visualizations)
+  --format          -f   TEXT          Output format: charts|html|markdown|all (default: all)
+```
+
+**Examples:**
+
+```bash
+# Generate all visualizations
+sage-cp-bench visualize --input results.json --output ./charts
+
+# Generate only HTML report
+sage-cp-bench visualize --input results.json --format html
+```
+
+### `config` and `validate` Commands
+
+```bash
+# Show example LLM configuration
+sage-cp-bench config --mode llm
+
+# Show and save hybrid configuration
+sage-cp-bench config --mode hybrid --output config.yaml
+
+# Validate configuration file
+sage-cp-bench validate config.json --mode llm
+sage-cp-bench validate config.yaml --mode hybrid
+```
+
+## Python API
+
+### LLM-only Benchmark
+
+```python
+import asyncio
+from sage.benchmark_control_plane import (
+    BenchmarkConfig,
+    BenchmarkRunner,
+    BenchmarkReporter,
+)
+
+# Configure benchmark
+config = BenchmarkConfig(
+    control_plane_url="http://localhost:8080",
+    policies=["fifo", "priority", "slo_aware"],
+    num_requests=1000,
+    request_rate=100.0,
+)
+
+# Run benchmark
+runner = BenchmarkRunner(config)
+result = asyncio.run(runner.run())
+
+# Generate report
+reporter = BenchmarkReporter(result)
+reporter.print_summary()
+reporter.save_all("./benchmark_results")
+```
+
+### Hybrid Benchmark (LLM + Embedding)
+
+```python
+import asyncio
+from sage.benchmark_control_plane.hybrid_scheduler import (
+    HybridBenchmarkConfig,
+    HybridBenchmarkRunner,
+    HybridBenchmarkReporter,
+)
+
+# Configure hybrid benchmark
+config = HybridBenchmarkConfig(
+    control_plane_url="http://localhost:8080",
+    num_requests=1000,
+    request_rate=100.0,
+    llm_ratio=0.7,              # 70% LLM, 30% Embedding
+    embedding_ratio=0.3,
+    policies=["fifo", "hybrid_slo"],
+)
+
+# Run benchmark
+runner = HybridBenchmarkRunner(config)
+result = asyncio.run(runner.run())
+
+# Generate report
+reporter = HybridBenchmarkReporter(result)
+reporter.print_summary()
+reporter.save_json("./results/hybrid_benchmark.json")
+```
+
+### Running Predefined Experiments
+
+```python
+import asyncio
+from sage.benchmark_control_plane.experiments import (
+    ThroughputExperiment,
+    LatencyExperiment,
+    SLOComplianceExperiment,
+    MixedRatioExperiment,
+)
+from sage.benchmark_control_plane.common.base_config import SchedulingPolicy
+
+# Throughput experiment
+exp = ThroughputExperiment(
+    name="throughput_sweep",
+    control_plane_url="http://localhost:8080",
+    policies=[SchedulingPolicy.FIFO, SchedulingPolicy.PRIORITY],
+    request_rates=[50, 100, 200, 500],
+)
+result = asyncio.run(exp.run_full())  # Includes visualization
+print(f"Best policy: {result.summary['best_policy']}")
+
+# Latency experiment
+exp = LatencyExperiment(
+    name="latency_analysis",
+    control_plane_url="http://localhost:8080",
+    request_rate=100,
+    num_requests=1000,
+)
+result = asyncio.run(exp.run_full())
+
+# Mixed ratio experiment (hybrid)
+exp = MixedRatioExperiment(
+    name="ratio_sweep",
+    control_plane_url="http://localhost:8080",
+    llm_ratios=[0.0, 0.25, 0.5, 0.75, 1.0],
+)
+result = asyncio.run(exp.run_full())
+```
+
+### Generating Visualizations
+
+```python
+from pathlib import Path
+from sage.benchmark_control_plane.visualization import (
+    BenchmarkCharts,
+    ReportGenerator,
+)
+
+# Generate charts
+charts = BenchmarkCharts(output_dir=Path("./charts"))
+charts.plot_throughput_comparison(policy_metrics)
+charts.plot_latency_distribution(latency_data)
+charts.plot_slo_compliance(slo_data)
+
+# Generate reports
+report_gen = ReportGenerator(result=benchmark_result, charts_dir=Path("./charts"))
+report_gen.generate_html_report(Path("./report.html"))
+report_gen.generate_markdown_report(Path("./report.md"))
+```
+
+## Supported Scheduling Policies
+
+| Policy           | Mode   | Description                                     |
+| ---------------- | ------ | ----------------------------------------------- |
+| `fifo`           | Both   | First-In-First-Out scheduling                   |
+| `priority`       | Both   | Priority-based scheduling                       |
+| `slo_aware`      | Both   | SLO-deadline aware scheduling                   |
+| `cost_optimized` | LLM    | Cost-optimized scheduling                       |
+| `adaptive`       | LLM    | Adaptive scheduling based on system state       |
+| `aegaeon`        | LLM    | Advanced scheduling with multiple optimizations |
+| `hybrid`         | Hybrid | Hybrid LLM/Embedding scheduling                 |
+| `hybrid_slo`     | Hybrid | Hybrid with SLO awareness                       |
+
+## Configuration Options
+
+### LLM Benchmark Configuration
+
+| Option                  | Description                        | Default                             |
+| ----------------------- | ---------------------------------- | ----------------------------------- |
+| `control_plane_url`     | Control Plane HTTP address         | `http://localhost:8080`             |
+| `policies`              | List of policies to benchmark      | `["fifo", "priority", "slo_aware"]` |
+| `num_requests`          | Total requests per policy          | `100`                               |
+| `request_rate`          | Target request rate (req/s)        | `10.0`                              |
+| `arrival_pattern`       | Request arrival pattern            | `poisson`                           |
+| `model_distribution`    | Request distribution across models | `{"default": 1.0}`                  |
+| `priority_distribution` | Request priority distribution      | `{"NORMAL": 1.0}`                   |
+| `timeout_seconds`       | Request timeout                    | `60.0`                              |
+| `warmup_requests`       | Warmup requests before measurement | `10`                                |
+
+### Hybrid Benchmark Configuration
+
+| Option                      | Description                       | Default       |
+| --------------------------- | --------------------------------- | ------------- |
+| `llm_ratio`                 | Ratio of LLM requests (0.0-1.0)   | `0.5`         |
+| `embedding_ratio`           | Ratio of Embedding requests       | `0.5`         |
+| `embedding_model`           | Embedding model name              | `BAAI/bge-m3` |
+| `embedding_batch_size`      | Batch size for embedding requests | `32`          |
+| `llm_slo_deadline_ms`       | SLO deadline for LLM requests     | `5000`        |
+| `embedding_slo_deadline_ms` | SLO deadline for embedding        | `500`         |
+
+## Output Formats
+
+### Terminal Output
+
+```
+============================================================
+       sageLLM Hybrid Scheduling Benchmark Report
+============================================================
+Config: 1000 requests @ 100 req/s | LLM: 70% | Embedding: 30%
+------------------------------------------------------------
+
+| Policy     | Throughput | LLM Avg | Emb Avg | LLM SLO | Emb SLO | Errors |
+|------------|------------|---------|---------|---------|---------|--------|
+| fifo       | 95.2 req/s | 156 ms  | 23 ms   | 71.2%   | 92.1%   | 0.3%   |
+| hybrid_slo | 98.5 req/s | 132 ms  | 18 ms   | 93.7%   | 98.2%   | 0.1%   |
+
+Best Throughput: hybrid_slo (98.5 req/s)
+Best LLM SLO: hybrid_slo (93.7%)
+Best Embedding SLO: hybrid_slo (98.2%)
+```
+
+### JSON Report
+
+Full results saved to `report_<timestamp>.json` including:
+
+- Configuration summary
+- Per-policy metrics
+- Raw request results
+- Summary statistics
+
+### HTML Report
+
+Interactive HTML report with embedded charts and tables.
+
+### Markdown Report
+
+Markdown format suitable for documentation and GitHub.
+
+## Module Structure
+
+```
+benchmark_control_plane/
+├── __init__.py              # Module exports (backward compatible)
+├── cli.py                   # CLI interface (sage-cp-bench)
+├── config.py                # Legacy config (→ llm_scheduler)
+├── workload.py              # Legacy workload (→ llm_scheduler)
+├── client.py                # Legacy client (→ llm_scheduler)
+├── metrics.py               # Legacy metrics (→ llm_scheduler)
+├── runner.py                # Legacy runner (→ llm_scheduler)
+├── reporter.py              # Legacy reporter (→ llm_scheduler)
+├── README.md                # This file
+│
+├── common/                  # Shared components
+│   ├── __init__.py
+│   ├── base_config.py       # Base configuration classes
+│   ├── base_metrics.py      # Base metrics classes
+│   ├── gpu_monitor.py       # GPU resource monitoring
+│   └── strategy_adapter.py  # Scheduling strategy adapter
+│
+├── llm_scheduler/           # LLM-only benchmark
+│   ├── __init__.py
+│   ├── config.py            # LLM benchmark config
+│   ├── workload.py          # LLM workload generation
+│   ├── client.py            # LLM HTTP client
+│   ├── metrics.py           # LLM metrics collection
+│   ├── runner.py            # LLM benchmark runner
+│   └── reporter.py          # LLM result reporting
+│
+├── hybrid_scheduler/        # Hybrid LLM+Embedding benchmark
+│   ├── __init__.py
+│   ├── config.py            # Hybrid benchmark config
+│   ├── workload.py          # Hybrid workload generation
+│   ├── client.py            # Hybrid HTTP client
+│   ├── metrics.py           # Hybrid metrics collection
+│   ├── runner.py            # Hybrid benchmark runner
+│   └── reporter.py          # Hybrid result reporting
+│
+├── visualization/           # Charts and reports
+│   ├── __init__.py
+│   ├── charts.py            # Matplotlib chart generation
+│   ├── report_generator.py  # HTML/Markdown reports
+│   └── templates/           # Report templates
+│       ├── benchmark_report.html
+│       └── comparison_report.html
+│
+└── experiments/             # Predefined experiments
+    ├── __init__.py
+    ├── base_experiment.py   # Experiment base class
+    ├── throughput_exp.py    # Throughput sweep
+    ├── latency_exp.py       # Latency analysis
+    ├── slo_compliance_exp.py # SLO compliance
+    └── mixed_ratio_exp.py   # LLM/Embedding ratio sweep
+```
+
+## Related Documentation
+
+- [DATA_PATHS.md](./DATA_PATHS.md) - Data directory structure and formats
+- [VISUALIZATION.md](./VISUALIZATION.md) - Chart types and report formats
+- [examples/run_llm_benchmark.py](../../../../examples/benchmark/run_llm_benchmark.py) - LLM
+  benchmark example
+- [examples/run_hybrid_benchmark.py](../../../../examples/benchmark/run_hybrid_benchmark.py) -
+  Hybrid benchmark example
+
+## Control Plane Integration
+
+### Required API Endpoints
+
+| Endpoint               | Method | Description                          |
+| ---------------------- | ------ | ------------------------------------ |
+| `/health`              | GET    | Health check                         |
+| `/v1/chat/completions` | POST   | OpenAI-compatible LLM endpoint       |
+| `/v1/embeddings`       | POST   | OpenAI-compatible embedding endpoint |
+| `/admin/set_policy`    | POST   | Switch scheduling policy             |
+| `/admin/metrics`       | GET    | Get Control Plane metrics            |
+
+### Request Headers
+
+- `X-Request-ID`: Unique request identifier
+- `X-Request-Priority`: Request priority (HIGH, NORMAL, LOW)
+- `X-SLO-Deadline-Ms`: SLO deadline in milliseconds
+- `X-Request-Type`: Request type (llm_chat, llm_generate, embedding)
+
+## Troubleshooting
+
+### Common Issues
+
+1. **Connection refused**: Ensure Control Plane is running at the specified URL
+1. **Timeout errors**: Increase `--timeout` or reduce `--rate`
+1. **No visualization**: Install matplotlib: `pip install matplotlib`
+1. **YAML config error**: Install pyyaml: `pip install pyyaml`
+
+### Debug Mode
+
+```bash
+# Enable verbose logging
+export SAGE_LOG_LEVEL=DEBUG
+sage-cp-bench run --mode llm --policy fifo --requests 10
+```
+
+______________________________________________________________________
+
+*Updated: 2025-11-28*

isage_control_plane_benchmark-0.1.0.1.dist-info/RECORD

@@ -0,0 +1,46 @@
+isage_control_plane_benchmark-0.1.0.1.dist-info/licenses/LICENSE,sha256=vBNVIGkYYZY0B8f0Ui1ITYwRu7WNtSwyxvIAVGYS6jU,1075
+sage/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
+sage/benchmark_control_plane/__init__.py,sha256=Zp1bFy7qeiIGaXOtPL_uImkKoy7MFtuj4Df4UKL5uK8,5350
+sage/benchmark_control_plane/_version.py,sha256=Z1QHTVqqrdwKiMdqUj-e4F0qu5MMUIw8jhHVJSYgGT8,157
+sage/benchmark_control_plane/cli.py,sha256=ITXtS1UyPrWz7DHbM8iKSqggjHcxJ4hMkLRoxRBAYaQ,42855
+sage/benchmark_control_plane/client.py,sha256=tFbafZOGj7I4ouPqoflRL3WAJ_ggiFRXVbEdbxPQQGE,15128
+sage/benchmark_control_plane/config.py,sha256=_Li77HAEMGNuk6EU8IcNaTQPHWJNn3dRLyS3I0gG7WM,9295
+sage/benchmark_control_plane/metrics.py,sha256=s8Vn6dCCz4h1YM52aMGLZh4xHSxIXq43W5PFL3GhGlY,11991
+sage/benchmark_control_plane/reporter.py,sha256=qSkCHhQHQkEpKqDJNhfAnMoxc0r34KuPoDkQ8SEypPQ,11367
+sage/benchmark_control_plane/runner.py,sha256=JH3a4prTxlgs7I3aLv2Y6shGtn95rH54QgxCCn-rHLk,11873
+sage/benchmark_control_plane/workload.py,sha256=gNMCWLgXrzTZ-PE275m6aIqcHG4v95ssSuVHHLLY9-c,12653
+sage/benchmark_control_plane/common/__init__.py,sha256=1xrrrYTOVAPNuNHLxzrSf4VQKcHH4LOZWvpYKb-1GcY,1158
+sage/benchmark_control_plane/common/base_config.py,sha256=lmSWFiB5m-oy2kugsed989kxiU4msY_-b0NWsgMFxAw,7185
+sage/benchmark_control_plane/common/base_metrics.py,sha256=XFlVsX4SP0noDBdeGp7gqmyOvvZMZARq_080YBlU9xk,12357
+sage/benchmark_control_plane/common/gpu_monitor.py,sha256=smU8bDUmHq6jQCy-jeFIKoELNKlsjdF2aON-MP-o_1I,18144
+sage/benchmark_control_plane/common/strategy_adapter.py,sha256=QPXf3MLBTJS7yGUGZ3UwfSZGT_-QIBJl1O5ShU_nPQc,11655
+sage/benchmark_control_plane/experiments/__init__.py,sha256=GzH5YZ-evPy3TELxNAYPnn2oUt4MvjnQ7ms6kNOjkcY,1816
+sage/benchmark_control_plane/experiments/base_experiment.py,sha256=wa3HwRFtUVEv9pEgzQfV0b8Z3TxIdPK52Zvz5d87zbw,9505
+sage/benchmark_control_plane/experiments/latency_exp.py,sha256=8dbq-p9O0qJ26LjrAfc_wjPSQlmx4cfV3NXw9qJlRnY,12299
+sage/benchmark_control_plane/experiments/mixed_ratio_exp.py,sha256=KrYxj3Kln7nt4umONkFwoApO2sqPO5Zbaq14_-7b5PA,14286
+sage/benchmark_control_plane/experiments/slo_compliance_exp.py,sha256=kvzZrok5mIYH6xu0XQJpiR41lk2YhwLED0wo_T_nEDQ,14513
+sage/benchmark_control_plane/experiments/throughput_exp.py,sha256=eQ1Sc02IXBGErwM-4KRYtEhqa136iPCY8fFzahGVXjI,9828
+sage/benchmark_control_plane/hybrid_scheduler/__init__.py,sha256=2rvc376TWkGy5NQoHmb_aTjJXMRGRC7f8IKQK_mFSEY,2037
+sage/benchmark_control_plane/hybrid_scheduler/client.py,sha256=cSHPtd5PA8nTn1WATUqzy1xmojF-RsQC22Gtj2tQAzs,19803
+sage/benchmark_control_plane/hybrid_scheduler/config.py,sha256=Ms-HgV8cU3jXntox0kHsr0Z3QamRJVx_V2WWXLWcmnk,13444
+sage/benchmark_control_plane/hybrid_scheduler/metrics.py,sha256=B87ivvSPdHsL0dRhJpVKa0zaff9QKf47WxnRDlnKYgU,18296
+sage/benchmark_control_plane/hybrid_scheduler/reporter.py,sha256=p4GE8v9gszsrLwM5_DIlqwjCqcnNsru0KelKol8iCLw,18102
+sage/benchmark_control_plane/hybrid_scheduler/runner.py,sha256=zDfNYnLP70ls9Unws2q2B5PKns0FlVmXWWObldiwdnQ,20050
+sage/benchmark_control_plane/hybrid_scheduler/workload.py,sha256=-irkMjJAXzR3KaEullcZVdyCpmKtlCJmW5r-jechBqU,20607
+sage/benchmark_control_plane/llm_scheduler/__init__.py,sha256=OWYYnsil5dIoiciRWW9xUZrSYR0cGz52yBzaYszFuqM,1754
+sage/benchmark_control_plane/llm_scheduler/client.py,sha256=gXl096l2O99Mq5U-pO69upMdwScovUC_h8Cye4YvXlY,13777
+sage/benchmark_control_plane/llm_scheduler/config.py,sha256=Yum_cB3kqhuqwLd4YRUSNktIkt-m33_r7tUJJ_lgC58,6245
+sage/benchmark_control_plane/llm_scheduler/metrics.py,sha256=lls_-fKtdIHiHBwQAFjrJVasIKpDbHA55Q8cmbYgbVE,6950
+sage/benchmark_control_plane/llm_scheduler/reporter.py,sha256=DTP7uI-YBzC1SMT0y3mstk53MV5UWeH8j1ZTz7Bk5nw,12434
+sage/benchmark_control_plane/llm_scheduler/runner.py,sha256=ZIyTQweZQBT8vxmilOmfAzSmMGbqh8dAkU0yNWqcPxo,15834
+sage/benchmark_control_plane/llm_scheduler/workload.py,sha256=ioqV5p0vx9KiMz66wQqVZAuGcM4KeTOEDltzfGfg5x8,13058
+sage/benchmark_control_plane/visualization/__init__.py,sha256=7qE5ghCkceOH8wIT4c8076TFboXJeCoZN0U7yZHGsoA,1496
+sage/benchmark_control_plane/visualization/charts.py,sha256=dbD20L1omtpPv9XZJh-obBQXrZa4d55Hr0mqM50phhE,36132
+sage/benchmark_control_plane/visualization/report_generator.py,sha256=cO712ZgbVER8Cc4bdgNOX3w3DoQ3jigyEvXC20c_MO8,25078
+sage/benchmark_control_plane/visualization/templates/benchmark_report.html,sha256=rxjE9lSePfSaCR_qsTSPlNWh9JpC6dqwMvnD3EUz9d8,16616
+sage/benchmark_control_plane/visualization/templates/comparison_report.html,sha256=pmsC1HiT1MIotody_Hy5P40y6tr6w_rVVJjNS2vhncE,15217
+isage_control_plane_benchmark-0.1.0.1.dist-info/METADATA,sha256=hhm-ypZm__VFDpsBYgBa-5M6uQBp_RUFkta13cQ37ho,23128
+isage_control_plane_benchmark-0.1.0.1.dist-info/WHEEL,sha256=yk-B4c9kYsinhQ_MzhPAVcDm9mhkAVmdo0rg0jgFCmo,94
+isage_control_plane_benchmark-0.1.0.1.dist-info/entry_points.txt,sha256=Vz3Grx2gvseJda-U9vCB-vRNJGfGtUnw4uYXWlm2xqQ,72
+isage_control_plane_benchmark-0.1.0.1.dist-info/top_level.txt,sha256=hibFyzQHiLOMK68qL1OWsNKaXOmSXqZjeLTBem6Yy7I,5
+isage_control_plane_benchmark-0.1.0.1.dist-info/RECORD,,

isage_control_plane_benchmark-0.1.0.1.dist-info/WHEEL

@@ -0,0 +1,5 @@
+Wheel-Version: 1.0
+Generator: setuptools (80.10.2)
+Root-Is-Purelib: true
+Tag: cp311-none-any
+

isage_control_plane_benchmark-0.1.0.1.dist-info/entry_points.txt

@@ -0,0 +1,2 @@
+[console_scripts]
+sage-cp-bench = sage.benchmark_control_plane.cli:main

isage_control_plane_benchmark-0.1.0.1.dist-info/licenses/LICENSE

@@ -0,0 +1,21 @@
+MIT License
+
+Copyright (c) 2026 IntelliStream Team
+
+Permission is hereby granted, free of charge, to any person obtaining a copy
+of this software and associated documentation files (the "Software"), to deal
+in the Software without restriction, including without limitation the rights
+to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+copies of the Software, and to permit persons to whom the Software is
+furnished to do so, subject to the following conditions:
+
+The above copyright notice and this permission notice shall be included in all
+copies or substantial portions of the Software.
+
+THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+SOFTWARE.

isage_control_plane_benchmark-0.1.0.1.dist-info/top_level.txt

@@ -0,0 +1 @@
+sage