本中心與「台大醫院醫學研究部次世代核酸定序 (NGS) 服務核心實驗室」於 2024 年 3 月 7 日 合辦一場 NVIDIA 軟體教學工作坊，主題為：
Workshop on Leveraging NVIDIA Technologies Across the Genomics Workflow
This workshop focuses on three essential tools necessary for comprehensive genomics analysis. Firstly, NVIDIA® TensorRT™: an SDK designed for high-performance deep learning inference. It features a deep learning inference optimizer and runtime, ensuring low latency and high throughput for inference applications (compatible with both TensorFlow and PyTorch). During the demonstration, we will compare the speed of TensorRT-based inference with native PyTorch. Next, RAPIDS cuML: a suite of fast, GPU-accelerated machine learning algorithms tailored for data science and analytical tasks (mirroring sklearn's API). As datasets grow larger, traditional CPU-based algorithms become slow and cumbersome. RAPIDS offers users a streamlined approach, where data is initially loaded onto the GPU, allowing compute tasks to be performed directly on it. In this demo, we will compare the speed of cuML with scikit-learn. Lastly, NVIDIA® Parabricks®: a scalable genomics analysis software suite that harnesses full-stack accelerated computing to process data in minutes. During this demonstration, participants will learn how to utilize the tool for alignment and variant calling.
Speakers from NVIDIA:
- Warren Tseng, Solution Architect
- Ying-Ja Chen, Solutions Architect
- Chi-Cheng Fu, Senior Developer Relations Manager
【工作坊資訊】
日期：2024 年 3 月 7 日 (星期四)
時間：15:30-16:30
地點：台大醫學院 103 教室
報名連結：https://forms.gle/sHzSCi3MnKsK3jES8 (報名截止日期：2024/03/05)
機會難得，歡迎大家踴躍參與！
