Sukru Burc Eryilmaz

S. Burc Eryilmaz is a senior GPU architect and GPU architecture manager at NVIDIA. His team focuses on optimizing the end-to-end performance and energy efficiency of neural network training both at single-node scale and supercomputer scale through hardware and software improvements. He received his PhD from Stanford University, and bachelor's degree from Bilkent University.

Posts by Sukru Burc Eryilmaz

Data Center / Cloud Jun 04, 2025

Reproducing NVIDIA MLPerf v5.0 Training Scores for LLM Benchmarks

The previous post, NVIDIA Blackwell Delivers up to 2.6x Higher Performance in MLPerf Training v5.0, explains how the NVIDIA platform delivered the fastest time... 11 MIN READ

Data Center / Cloud Jun 04, 2025

NVIDIA Blackwell Delivers up to 2.6x Higher Performance in MLPerf Training v5.0

The journey to create a state-of-the-art large language model (LLM) begins with a process called pretraining. Pretraining a state-of-the-art model is... 12 MIN READ

Data Center / Cloud Nov 13, 2024

NVIDIA Blackwell Doubles LLM Training Performance in MLPerf Training v4.1

As models grow larger and are trained on more data, they become more capable, making them more useful. To train these models quickly, more performance,... 8 MIN READ

Data Center / Cloud Jun 12, 2024

NVIDIA Sets New Generative AI Performance and Scale Records in MLPerf Training v4.0

Generative AI models have a variety of uses, such as helping write computer code, crafting stories, composing music, generating images, producing videos, and... 11 MIN READ

Data Center / Cloud Nov 08, 2023

Setting New Records at Data Center Scale Using NVIDIA H100 GPUs and NVIDIA Quantum-2 InfiniBand

Generative AI is rapidly transforming computing, unlocking new use cases and turbocharging existing ones. Large language models (LLMs), such as OpenAI’s GPT... 19 MIN READ

Data Center / Cloud Jun 27, 2023

Breaking MLPerf Training Records with NVIDIA H100 GPUs

At the heart of the rapidly expanding set of AI-powered applications are powerful AI models. Before these models can be deployed, they must be trained through a... 15 MIN READ