# An Inquiry into Datacenter TCO for LLM Inference with FP8

arXiv: https://arxiv.org/abs/2502.01070 · HTML: https://mlsanity.com/paper/2502.01070/

## Key facts

- Cited by 5 papers in the map (data through 14 Oct 2026).
- Builds on 30 of its 56 references that are in the map.
- First posted on arXiv 3 Feb 2025.

## Builds on (most cited first, 30 of 30)

- [DeepSeek-R1: Incentivizing Reasoning Capability in LLMs via Reinforcement Learning](https://mlsanity.com/paper/2501.12948/) (2025)
- [Efficient Memory Management for Large Language Model Serving with PagedAttention](https://mlsanity.com/paper/2309.06180/) (2023)
- [PaLM: Scaling Language Modeling with Pathways](https://mlsanity.com/paper/2204.02311/) (2022)
- [RoFormer: Enhanced Transformer with Rotary Position Embedding](https://mlsanity.com/paper/2104.09864/) (2021)
- [FlashAttention-2: Faster Attention with Better Parallelism and Work Partitioning](https://mlsanity.com/paper/2307.08691/) (2023)
- [SmoothQuant: Accurate and Efficient Post-Training Quantization for Large Language Models](https://mlsanity.com/paper/2211.10438/) (2022)
- [GQA: Training Generalized Multi-Query Transformer Models from Multi-Head Checkpoints](https://mlsanity.com/paper/2305.13245/) (2023)
- [DeepSeek-V2: A Strong, Economical, and Efficient Mixture-of-Experts Language Model](https://mlsanity.com/paper/2405.04434/) (2024)
- [SparseGPT: Massive Language Models Can Be Accurately Pruned in One-Shot](https://mlsanity.com/paper/2301.00774/) (2023)
- [Enhancing Chat Language Models by Scaling High-quality Instructional Conversations](https://mlsanity.com/paper/2305.14233/) (2023)
- [A Simple and Effective Pruning Approach for Large Language Models](https://mlsanity.com/paper/2306.11695/) (2023)
- [QuaRot: Outlier-Free 4-Bit Inference in Rotated LLMs](https://mlsanity.com/paper/2404.00456/) (2024)
- [Efficient Large-Scale Language Model Training on GPU Clusters Using Megatron-LM](https://mlsanity.com/paper/2104.04473/) (2021)
- [In-Datacenter Performance Analysis of a Tensor Processing Unit](https://mlsanity.com/paper/1704.04760/) (2017)
- [DistServe: Disaggregating Prefill and Decoding for Goodput-optimized Large Language Model Serving](https://mlsanity.com/paper/2401.09670/) (2024)
- [Splitwise: Efficient generative LLM inference using phase splitting](https://mlsanity.com/paper/2311.18677/) (2023)
- [SpinQuant: LLM quantization with learned rotations](https://mlsanity.com/paper/2405.16406/) (2024)
- [Massive Activations in Large Language Models](https://mlsanity.com/paper/2402.17762/) (2024)
- [Atom: Low-bit Quantization for Efficient and Accurate LLM Serving](https://mlsanity.com/paper/2310.19102/) (2023)
- [A Study of BFLOAT16 for Deep Learning Training](https://mlsanity.com/paper/1905.12322/) (2019)
- [QServe: W4A8KV4 Quantization and System Co-design for Efficient LLM Serving](https://mlsanity.com/paper/2405.04532/) (2024)
- [No Token Left Behind: Reliable KV Cache Compression via Importance-Aware Mixed Precision Quantization](https://mlsanity.com/paper/2402.18096/) (2024)
- [Breaking the Computation and Communication Abstraction Barrier in Distributed Machine Learning Workloads](https://mlsanity.com/paper/2105.05720/) (2021)
- [M'elange: Cost Efficient Large Language Model Serving by Exploiting GPU Heterogeneity](https://mlsanity.com/paper/2404.14527/) (2024)
- [Efficient Post-training Quantization with FP8 Formats](https://mlsanity.com/paper/2309.14592/) (2023)
- [When Precision Meets Position: BFloat16 Breaks Down RoPE in Long-Context Training](https://mlsanity.com/paper/2411.13476/) (2024)
- [Sustainable Supercomputing for AI: GPU Power Capping at HPC Scale](https://mlsanity.com/paper/2402.18593/) (2024)
- [SCOOT: SLO-Oriented Performance Tuning for LLM Inference Engines](https://mlsanity.com/paper/2408.04323/) (2024)
- [HLAT: High-quality Large Language Model Pre-trained on AWS Trainium](https://mlsanity.com/paper/2404.10630/) (2024)
- [Debunking the CUDA Myth Towards GPU-based AI Systems](https://mlsanity.com/paper/2501.00210/) (2024)

## Built on by (most cited first, 5 of 5)

- [A Cost-Benefit Analysis of On-Premise Large Language Model Deployment: Breaking Even with Commercial LLM Services](https://mlsanity.com/paper/2509.18101/) (2025)
- [Faster Inference of LLMs using FP8 on the Intel Gaudi](https://mlsanity.com/paper/2503.09975/) (2025)
- [Position: LLM Serving Needs Mathematical Optimization and Algorithmic Foundations, Not Just Heuristics](https://mlsanity.com/paper/2605.01280/) (2026)
- [Adaptive Data Flywheel: Applying MAPE Control Loops to AI Agent Improvement](https://mlsanity.com/paper/2510.27051/) (2025)
- [DPS: Dual-Mode Precision LLM Serving with Semi-Unified Memory](https://mlsanity.com/paper/2609.34380/) (2026)

## Abstract (arXiv)

As large language models (LLMs) continue to scale, the high power consumption of AI accelerators in datacenters presents significant challenges, substantially increasing the total cost of ownership (TCO) for cloud service providers (CSPs) that provide LLM inference. In this work, we analyze the computational characteristics of LLM inference from a TCO perspective and present a generalizable framework to compare AI accelerators across diverse operational requirements. Using this model, we investigate key workload characteristics influencing TCO for AI accelerators from Intel (Gaudi 2 & 3) and NVIDIA (H100 & H200), especially thin GEMM utilization and FP8 quantization. In particular, as FP8 emerges as the baseline precision for next-generation LLMs, understanding how different architectures implement and benefit from low-precision computation is increasingly critical. Throughput on thin GEMMs has a greater impact on TCO than theoretical hardware peak throughput because the memory-bound decode phase is dominated by GEMV-like computations. We find that Gaudi HPUs achieve superior utilization on thin GEMMs compared to their counterparts, especially in FP8-quantized models. Our result underscores the importance of empirical, workload-level analysis in evaluating accelerator performance, rather than relying solely on theoretical hardware specifications. By studying the interaction between power consumption, quantization strategies, and hardware architecture, we provide insights to support informed deployment decisions and guide future accelerator designs aimed at improving the TCO of LLM inference workloads.