Artificial IntelligenceTechnical Deep Dive

PyTorch 2.15 Supercharges Apple Silicon Mac Performance

Published
EElectricBuzz Editorial Team
PyTorch 2.15 Supercharges Apple Silicon Mac Performance
3 min read592 wordsElectricBuzz Editorial Team

The Gist

“The latest PyTorch update introduces native Metal-accelerated kernels, drastic memory optimizations for video workloads, and massive speed gains for local LLM inference.”

Revolutionizing Mac-Based AI Development

PyTorch 2.15 has arrived, marking a significant milestone for developers utilizing Apple silicon hardware. This release, driven by extensive contributions from the Hugging Face team and the PyTorch community, shifts the MPS (Metal Performance Shaders) backend away from generic graph-based execution toward highly optimized, custom Metal kernels. This architectural pivot directly addresses performance bottlenecks, memory management issues, and feature parity gaps that previously hindered Mac-based machine learning workflows.

For practitioners working with local Large Language Models (LLMs), video generation, or complex scientific computing on macOS 15 and above, PyTorch 2.15 offers a transformative experience. By replacing indirect execution paths with native Metal operations, the framework now maximizes the throughput of M-series chips, turning Apple hardware into an even more formidable platform for cutting-edge AI development.

Breakthroughs in LLM Inference

The update brings substantial speed improvements for the decoding phase of transformer models. By optimizing linear layers and integrating native support for grouped matrix multiplications, 2.15 dramatically boosts token throughput. Specifically, Mixture-of-Experts (MoE) models—which are notoriously difficult to run efficiently—see decoding speeds increase between 5 and 6.5 times compared to version 2.14. A model like Qwen3.6-35B-A3B jumps from roughly 9 tokens per second to over 60 tokens per second on an M3 Ultra, making real-time local interaction a reality for much larger models than previously possible.

Complementing these gains, the framework now includes optimized softmax and log_softmax kernels. These are critical for the final stages of token sampling, with tests showing an 8x speedup for large vocabulary sizes (e.g., [1, 128256] shapes). This optimization ensures that the sampling stage no longer acts as a throttle during high-speed text generation.

Massive Memory Efficiency for Creative Workflows

Perhaps the most significant technical achievement in this release is the drastic reduction in memory footprint for video generation models. In earlier versions, processes like the tiled Variational Autoencoder (VAE) decoding—central to models such as Wan2.2—suffered from inefficient padding operations that caused driver memory usage to balloon to upwards of 270 GiB. By moving constant padding to Metal-native kernels, PyTorch 2.15 has reduced the peak driver memory usage to just 8 GiB for the same workload.

This memory efficiency is a game-changer for content creators and researchers using tools like ComfyUI, as it prevents system crashes and enables the execution of heavy generative workflows on consumer-grade unified memory configurations that previously lacked the overhead to handle such massive memory spikes.

Expanding Capabilities with FP8 and Linear Algebra

PyTorch 2.15 introduces native support for FP8 (float8_e4m3fn) tensors on MPS. This format halves the memory storage requirements compared to traditional BF16 or FP16, enabling more aggressive model quantization and larger context windows within the same memory constraints. The inclusion of scaled matrix multiplication (F.scaled_mm) ensures these FP8 tensors can be processed directly on the GPU, bridging a critical feature gap with CUDA-based workflows.

Furthermore, the linear algebra suite has been expanded, with complex-valued LU factorization and triangular solves now running natively on the GPU. Small matrix operations, essential for computer vision and camera geometry, have also been streamlined through dedicated kernels, ensuring that specialized scientific tasks run with minimal CPU-to-GPU synchronization latency.

Why it Matters

  • Performance: Massive reductions in latency for LLM token decoding and complex linear algebra operations.
  • Efficiency: Video generation tasks that once required massive memory overhead now fit comfortably on standard Apple silicon configurations.
  • Compatibility: Expanded operator support ensures that modern generative AI tools and libraries run reliably without the need for manual workarounds.
  • Control: Introduction of explicit MPS stream management allows developers to fine-tune execution and memory allocation for complex, parallelized pipelines.
SPONSORED
The 5 Best Over-Ear ANC Headphones of 2026, Tested & Ranked
Editor's Pick Guide
92/100
Tech & Gadgets•12 min read

The 5 Best Over-Ear ANC Headphones of 2026, Tested & Ranked

We locked five over-ear ANC picks for 2026 — Sony WH-1000XM6, Bose QuietComfort Ultra 2, Soundcore Space One, Sennheiser Momentum 5, and Apple AirPods Max 2 — then stress-tested them on lab metrics, long-term owner truth, and live street prices.

Related Stories

Semantically matched articles, ranked by topic overlap and freshness.

Six Intelligent Gadgets to Supercharge Your Home Office
Artificial Intelligence

Six Intelligent Gadgets to Supercharge Your Home Office

From AI-powered writing tools to smart standing desks, these six hardware devices are designed to streamline your daily workflow and enhance productivity.

Hebbia Leverages GPT-6 Astra to Revolutionize Financial Diligence
Artificial Intelligence

Hebbia Leverages GPT-6 Astra to Revolutionize Financial Diligence

Financial research platform Hebbia reports significant efficiency gains and improved output quality after integrating OpenAI’s latest model, GPT-6 Astra, into its Matrix system.

OpenAI Unveils MedChemBench to Put AI Reasoning to the Medicinal Test
Artificial Intelligence

OpenAI Unveils MedChemBench to Put AI Reasoning to the Medicinal Test

A new benchmark designed to evaluate how well AI models handle the complex, multi-variable world of pharmaceutical research and molecular chemistry.

OpenAI Bolsters Frontier Models Against 'Adversarial Distillation' Threats
Artificial Intelligence

OpenAI Bolsters Frontier Models Against 'Adversarial Distillation' Threats

OpenAI is implementing advanced safeguards to prevent bad actors from extracting 'protected reasoning' traces to train unaligned, unauthorized AI models.

Thinking Machines Unveils Inkling 2 Mini: A New Standard for Edge AI
Artificial Intelligence

Thinking Machines Unveils Inkling 2 Mini: A New Standard for Edge AI

The arrival of Inkling 2 Mini marks a significant milestone in localized multimodal AI, offering a high-performance 26B parameter model designed for specialized setups.

Anthropic’s Massive IPO Signals a Shift in the AI Financial Ecosystem
Artificial Intelligence

Anthropic’s Massive IPO Signals a Shift in the AI Financial Ecosystem

As Anthropic enters the public markets, its IPO marks a pivotal transition from circular venture funding to a self-sustaining model backed by public capital.

Europa Fusion: The AI-Driven Approach to Solving the Energy Crisis
Artificial Intelligence

Europa Fusion: The AI-Driven Approach to Solving the Energy Crisis

Spotify founder Daniel Ek is betting on a new fusion startup, Europa Fusion, which aims to crack the fusion code through data-heavy simulation rather than traditional hardware-first methods.

A Mathematical Revolution: The Duel to Solve the Riemann Hypothesis
Artificial Intelligence

A Mathematical Revolution: The Duel to Solve the Riemann Hypothesis

In an unprecedented week for mathematics, two competing proofs for the 167-year-old Riemann hypothesis have emerged—one assisted by AI and one crafted entirely by hand.