# A Hardware Evaluation Framework for Large Language Model Inference

arXiv: https://arxiv.org/abs/2312.03134 · HTML: https://mlsanity.com/paper/2312.03134/

## Key facts

- Cited by 14 papers in the map (data through 14 Oct 2026).
- Builds on 27 of its 72 references that are in the map.
- First posted on arXiv 5 Dec 2023.

## Builds on (most cited first, 27 of 27)

- [Attention Is All You Need](https://mlsanity.com/paper/1706.03762/) (2017)
- [BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding](https://mlsanity.com/paper/1810.04805/) (2018)
- [Language Models are Few-Shot Learners](https://mlsanity.com/paper/2005.14165/) (2020)
- [LLaMA: Open and Efficient Foundation Language Models](https://mlsanity.com/paper/2302.13971/) (2023)
- [Scaling Laws for Neural Language Models](https://mlsanity.com/paper/2001.08361/) (2020)
- [PaLM: Scaling Language Modeling with Pathways](https://mlsanity.com/paper/2204.02311/) (2022)
- [FlashAttention: Fast and Memory-Efficient Exact Attention with IO-Awareness](https://mlsanity.com/paper/2205.14135/) (2022)
- [Gaussian Error Linear Units (GELUs)](https://mlsanity.com/paper/1606.08415/) (2016)
- [Switch Transformers: Scaling to Trillion Parameter Models with Simple and Efficient Sparsity](https://mlsanity.com/paper/2101.03961/) (2021)
- [Training Compute-Optimal Large Language Models](https://mlsanity.com/paper/2203.15556/) (2022)
- [FlashAttention-2: Faster Attention with Better Parallelism and Work Partitioning](https://mlsanity.com/paper/2307.08691/) (2023)
- [Emergent Abilities of Large Language Models](https://mlsanity.com/paper/2206.07682/) (2022)
- [Megatron-LM: Training Multi-Billion Parameter Language Models Using Model Parallelism](https://mlsanity.com/paper/1909.08053/) (2019)
- [FlexGen: High-Throughput Generative Inference of Large Language Models with a Single GPU](https://mlsanity.com/paper/2303.06865/) (2023)
- [Efficiently Scaling Transformer Inference](https://mlsanity.com/paper/2211.05102/) (2022)
- [Recent Advances in Natural Language Processing via Large Pre-Trained Language Models: A Survey](https://mlsanity.com/paper/2111.01243/) (2021)
- [DeepSpeed Inference: Enabling Efficient Inference of Transformer Models at Unprecedented Scale](https://mlsanity.com/paper/2207.00032/) (2022)
- [SpAtten: Efficient Sparse Attention Architecture with Cascade Token and Head Pruning](https://mlsanity.com/paper/2012.09852/) (2020)
- [Can Foundation Models Wrangle Your Data?](https://mlsanity.com/paper/2205.09911/) (2022)
- [Online normalizer calculation for softmax](https://mlsanity.com/paper/1805.02867/) (2018)
- [EdgeBERT: Sentence-Level Energy Optimizations for Latency-Aware Multi-Task NLP Inference](https://mlsanity.com/paper/2011.14203/) (2020)
- [Gemmini: Enabling Systematic Deep-Learning Architecture Evaluation via Full-Stack Integration](https://mlsanity.com/paper/1911.09925/) (2019)
- [SpreadsheetCoder: Formula Prediction from Semi-structured Context](https://mlsanity.com/paper/2106.15339/) (2021)
- [Interstellar: Using Halide's Scheduling Language to Analyze DNN Accelerators](https://mlsanity.com/paper/1809.04070/) (2018)
- [Mind Mappings: Enabling Efficient Algorithm-Accelerator Mapping Space Search](https://mlsanity.com/paper/2103.01489/) (2021)
- [A Full-Stack Search Technique for Domain Optimized Deep Learning Accelerators](https://mlsanity.com/paper/2105.12842/) (2021)
- [The Cost of Application-Class Processing: Energy and Performance Analysis of a Linux-ready 1.7GHz 64bit RISC-V Core in 22nm FDSOI Technology](https://mlsanity.com/paper/1904.05442/) (2019)

## Built on by (most cited first, 14 of 14)

- [LLM in a flash: Efficient Large Language Model Inference with Limited Memory](https://mlsanity.com/paper/2312.11514/) (2023)
- [Towards Greener LLMs: Bringing Energy-Efficiency to the Forefront of LLM Inference](https://mlsanity.com/paper/2403.20306/) (2024)
- [Scaling Down to Scale Up: A Cost-Benefit Analysis of Replacing OpenAI's LLM with Open Source SLMs in Production](https://mlsanity.com/paper/2312.14972/) (2023)
- [Memory Is All You Need: An Overview of Compute-in-Memory Architectures for Accelerating Large Language Model Inference](https://mlsanity.com/paper/2406.08413/) (2024)
- [NoMAD-Attention: Efficient LLM Inference on CPUs Through Multiply-add-free Attention](https://mlsanity.com/paper/2403.01273/) (2024)
- [Demystifying AI Platform Design for Distributed Inference of Next-Generation LLM models](https://mlsanity.com/paper/2406.01698/) (2024)
- [Performance Modeling and Workload Analysis of Distributed Large Language Model Training and Inference](https://mlsanity.com/paper/2407.14645/) (2024)
- [TokenSim: Enabling Hardware and Software Exploration for Large Language Model Inference Systems](https://mlsanity.com/paper/2503.08415/) (2025)
- [Position: Enough of Scaling LLMs! Lets Focus on Downscaling](https://mlsanity.com/paper/2505.00985/) (2025)
- [Achieving Peak Performance for Large Language Models: A Systematic Review](https://mlsanity.com/paper/2409.04833/) (2024)
- [Hysteresis Activation Function for Efficient Inference](https://mlsanity.com/paper/2411.10573/) (2024)
- [MIST: A Co-Design Framework for Heterogeneous, Multi-Stage LLM Inference](https://mlsanity.com/paper/2504.09775/) (2025)
- [Can AI automatically analyze public opinion? A LLM agents-based agentic pipeline for timely public opinion analysis](https://mlsanity.com/paper/2505.11401/) (2025)
- [Think Before You Grid-Search: Floor-First Triage for LLM Serving](https://mlsanity.com/paper/2607.05876/) (2026)

## Abstract (arXiv)

The past year has witnessed the increasing popularity of Large Language Models (LLMs). Their unprecedented scale and associated high hardware cost have impeded their broader adoption, calling for efficient hardware designs. With the large hardware needed to simply run LLM inference, evaluating different hardware designs becomes a new bottleneck. This work introduces LLMCompass, a hardware evaluation framework for LLM inference workloads. LLMCompass is fast, accurate, versatile, and able to describe and evaluate different hardware designs. LLMCompass includes a mapper to automatically find performance-optimal mapping and scheduling. It also incorporates an area-based cost model to help architects reason about their design choices. Compared to real-world hardware, LLMCompass' estimated latency achieves an average 10.4% error rate across various operators with various input sizes and an average 4.1% error rate for LLM inference. With LLMCompass, simulating a 4-NVIDIA A100 GPU node running GPT-3 175B inference can be done within 16 minutes on commodity hardware, including 26,400 rounds of the mapper's parameter search. With the aid of LLMCompass, this work draws architectural implications and explores new cost-effective hardware designs. By reducing the compute capability or replacing High Bandwidth Memory (HBM) with traditional DRAM, these new designs can achieve as much as 3.41x improvement in performance/cost compared to an NVIDIA A100, making them promising choices for democratizing LLMs. LLMCompass is planned to be fully open-source.