English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

Artificial intelligence (AI) methods have become critical in scientificapplications to help accelerate scientific discovery. Large language models(LLMs) are being considered as a promising approach to address some of thechallenging problems because of their superior generalization capabilitiesacross domains. The effectiveness of the models and the accuracy of theapplications is contingent upon their efficient execution on the underlyinghardware infrastructure. Specialized AI accelerator hardware systems haverecently become available for accelerating AI applications. However, thecomparative performance of these AI accelerators on large language models hasnot been previously studied. In this paper, we systematically study LLMs onmultiple AI accelerators and GPUs and evaluate their performancecharacteristics for these models. We evaluate these systems with (i) amicro-benchmark using a core transformer block, (ii) a GPT- 2 model, and (iii)an LLM-driven science use case, GenSLM. We present our findings and analyses ofthe models' performance to better understand the intrinsic capabilities of AIaccelerators. Furthermore, our analysis takes into account key factors such assequence lengths, scaling behavior, sparsity, and sensitivity to gradientaccumulation steps.

A Comprehensive Performance Study of Large Language Models on Novel AI Accelerators