slowai 0.3.0

One command to find why your PyTorch model is slow — and fix it.

slowai

One command to find why your PyTorch model is slow — and fix it.

slowai diagnoses which performance regime your workload is stuck in (compute-bound, memory-bound, or overhead-bound), prescribes the right fix, auto-applies it, and proves the speedup with before/after measurements. No guesswork. No manual profiler interpretation.

$ slowai fix model.py

==============================================================
  BASELINE: model.py: COMPUTE_BOUND (confidence: 0.85)
  wall time: 7.523s
==============================================================

  Tried 4 remedies:

  1. [10.00x] bf16_autocast  ** BEST **
     Run under bfloat16 automatic mixed precision
     7.523s >>> 0.752s
     regime: compute (confidence: 0.85)

  2. [6.32x] tf32_tensor_cores
     Enable TF32 tensor cores (~2x matmul throughput on Ampere+)
     7.523s >>> 1.191s

  3. [6.22x] high_matmul_precision
     Set float32 matmul precision to 'high'
     7.523s >>> 1.210s

  4. [1.31x] cudnn_benchmark
     Enable cuDNN auto-tuner for conv kernels
     7.523s >>> 5.719s

--------------------------------------------------------------
  WINNER: bf16_autocast
  7.523s >>> 0.752s  (10.00x, +900% faster)
  How: Run under bfloat16 automatic mixed precision
--------------------------------------------------------------

Why this exists

Every deep learning workload is stuck in one of three performance regimes (Horace He, 2022):

Regime	What's happening	Wrong fix = no speedup
Compute-bound	GPU is saturated doing math (matmuls, convolutions)	Fusing ops won't help — the math itself is the bottleneck
Memory-bound	GPU is waiting for data (pointwise ops, activations)	Smaller model won't help — you need less data movement
Overhead-bound	GPU is idle waiting for Python/dispatcher (tiny ops)	Lower precision won't help — you need fewer, bigger ops

The fix for each regime is different. Applying a compute-bound fix to a memory-bound workload does nothing. Engineers waste hours in profiler UIs figuring this out manually.

slowai does it in one command.

How it works

slowai diagnose model.py    # Classify the regime + prescribe fixes
slowai fix model.py          # ^ plus auto-apply fixes and measure speedup

Under the hood:

1. Profile — Runs your workload under torch.profiler with CUDA timing, warmup pass, and op-level statistics
2. Classify — A heuristic classifier analyzes op shares (matmul, normalization, pointwise, tiny-op fraction) to determine the dominant regime
3. Prescribe — Returns a ranked list of fixes for that regime, cheapest first
4. Remediate — Auto-applies each applicable fix (TF32 tensor cores, bf16/fp16 autocast, cuDNN benchmark, matmul precision), re-profiles, and ranks by measured speedup

No code changes required. Remedies are environment-level transforms — they modify PyTorch's runtime settings, not your model code.

Benchmarks

Tested on NVIDIA Jetson Orin Nano Super (Ampere GPU, 1024 CUDA cores, JetPack 6.2, PyTorch 2.8.0):

Workload	Regime	Baseline	Best remedy	After	Speedup
Dense GEMM (4096x4096)	Compute	7.523s	bf16_autocast	0.752s	10.00x
Pointwise chain (8192x8192)	Memory	2.400s	tf32_tensor_cores	0.575s	4.17x
Tiny ops (5000 micro-ops)	Overhead	3.281s	tf32_tensor_cores	1.141s	2.88x
MobileNetV2 inference	Compute	1.778s	cudnn_benchmark	1.681s	1.06x
ResNet-50 inference	Compute	2.105s	bf16_autocast	1.969s	1.07x

Production models (MobileNet, ResNet) show modest gains because PyTorch already optimizes common architectures well. The real value is that slowai finds the right fix automatically — cuDNN benchmark wins for convolution-heavy MobileNet, bf16 autocast wins for matmul-heavy ResNet. Different models, different winners, zero guesswork.

Installation

git clone https://github.com/ricojallen37-sketch/slowai.git
cd slowai
pip install -e .

Requires Python 3.10+ and PyTorch >= 2.1 with CUDA support.

Writing a workload

slowai profiles any Python script that exposes a main() function:

# my_model.py
import torch
from torchvision.models import resnet50

model = resnet50().cuda().eval()
data = torch.randn(8, 3, 224, 224, device="cuda")

def main():
    with torch.no_grad():
        for _ in range(30):
            model(data)

slowai fix my_model.py

Architecture

slowai/
  schema.py      # Regime enum, Diagnosis dataclass — the product thesis in types
  profiler.py    # torch.profiler wrapper → ProfileResult (op stats + wall time)
  diagnose.py    # Heuristic classifier → Diagnosis (regime + confidence + prescriptions)
  remediate.py   # Auto-fix engine → FixReport (before/after speedup per remedy)
  cli.py         # CLI entry points: diagnose, fix

The classifier is a pure function of profiler output — no torch dependency, fully unit-testable. The remediate engine applies fixes as environment transforms (global flags, autocast context managers) so it never modifies user code.

What's different

Other tools in this space are profiler UIs that show you data and leave interpretation to you. slowai is the only tool that goes from raw workload to regime classification to ranked prescriptions to auto-applied fixes to measured speedup in a single CLI command.

Tool	Profiles	Classifies regime	Prescribes fixes	Auto-applies	Measures speedup
PyTorch Profiler	Yes	No	No	No	No
NVIDIA Nsight	Yes	No	No	No	No
torch.utils.bottleneck	Yes	No	No	No	No
DeepSpeed Flops Profiler	Yes	No	No	No	No
slowai	Yes	Yes	Yes	Yes	Yes

Roadmap

• V1 (shipped) — Profile + classify regime for synthetic workloads
• V2 (shipped) — Noise filtering (sync ops, init ops), normalization-aware classification, real model support
• V3 (shipped) — Auto-remediate: apply fixes and measure before/after speedup
• V4 (next) — torch.compile integration, channels_last auto-transform, batch size search, export optimized config

Built by

Rico Allen — @ricojallen37-sketch

Built and tested on NVIDIA Jetson Orin Nano Super Developer Kit.