Pallas
PulseAugur coverage of Pallas — every cluster mentioning Pallas across labs, papers, and developer communities, ranked by signal.
-
LLM research explores faster inference, efficient training, and novel adaptation techniques
Multiple research papers explore methods for improving the efficiency and performance of large language models (LLMs). One paper introduces DSpark, a technique that significantly speeds up LLM inference by using a light…
-
Grumpy Pallas kittens debut at Japan zoo
Five young Pallas's cats, known for their perpetually grumpy expressions, have begun public display at the Kobe Animal Kingdom in Japan. Born in May, these kittens are now over a kilogram and are transitioning to solid …
-
JAXBench launches to optimize AI kernels on Google TPUs
A new benchmark suite called JAXBench has been developed to specifically address the optimization of AI kernel performance on Google Cloud TPUs. This suite includes 50 JAX workloads derived from prominent AI models like…
-
Google optimizes Qwen 3.5-397B MoE on Ironwood TPUs for 4.7x speedup
Google has optimized the Qwen 3.5-397B Mixture-of-Experts (MoE) model to run on its Ironwood Tensor Processing Units (TPUs). This optimization, achieved using JAX and Pallas, resulted in a 4.7x speedup for prefill workl…
-
Pallas DSL Enables Custom Kernel Optimization on Google TPUs
This guide introduces Pallas, a Python-based domain-specific language (DSL) designed for writing custom kernels that can be compiled and optimized for Google's Tensor Processing Units (TPUs). Pallas aims to provide deve…