Machine Learning

Artificial Intelligence

Google Cloud Platform

Tpu

Explore recommended reads, summarized by AI

TPU Implementation and Optimization

Technical guides and analysis of deploying and accelerating machine learning workloads on TPU hardware.

Language Model Serving on TPUs

TPU Architecture and Comparisons

Popular this week

Harshit Sinha

Beyond the Hype: The Fundamental Architecture Difference Between CPUs and GPUs

In

Google Cloud - Community

by

Glen Yu

Google Colab, but in your favourite terminal

Ashutoshkumarsingh

PyTorch, Keras, and Flax in 2026: Architecture, Performance, and Real-World Adoption

xbill

TPU Deployments with Gemma 31B, v6e-4, and Antigravity CLI

Ayaka

Adding Direct CMEM Loads to Pallas’s Closed-Source Backend

Latest stories

Youngbin Lee

Under the Hood of JAX on TPU: From Asynchronous Dispatch to Custom Kernels with Pallas

Youngbin Lee

Benchmarking JAX on Cloud TPU v5e vs. PyTorch on NVIDIA A100: Why Peak TFLOPs/s Lies

Youngbin Lee

How to Think About TPUs vs. GPUs: TensorCore, MXU, VMEM, and Compiler-Driven Execution

Youngbin Lee

Tracing Y = ReLU(XW + b) from Python to Silicon: JAX (OpenXLA) vs. PyTorch 2

Fahim ul Haq

How to reason about GPU vs TPU when you design AI systems

See more stories

Who to follow

Popular this week

Harshit Sinha

Beyond the Hype: The Fundamental Architecture Difference Between CPUs and GPUs

In

Google Cloud - Community

by

Glen Yu

Google Colab, but in your favourite terminal

Ashutoshkumarsingh

PyTorch, Keras, and Flax in 2026: Architecture, Performance, and Real-World Adoption

xbill

TPU Deployments with Gemma 31B, v6e-4, and Antigravity CLI

Ayaka

Adding Direct CMEM Loads to Pallas’s Closed-Source Backend

Latest stories

Youngbin Lee

Under the Hood of JAX on TPU: From Asynchronous Dispatch to Custom Kernels with Pallas

Youngbin Lee

Benchmarking JAX on Cloud TPU v5e vs. PyTorch on NVIDIA A100: Why Peak TFLOPs/s Lies

Youngbin Lee

How to Think About TPUs vs. GPUs: TensorCore, MXU, VMEM, and Compiler-Driven Execution

Youngbin Lee

Tracing Y = ReLU(XW + b) from Python to Silicon: JAX (OpenXLA) vs. PyTorch 2

Fahim ul Haq

How to reason about GPU vs TPU when you design AI systems

See more stories

Who to follow