{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2025,10,1]],"date-time":"2025-10-01T15:59:37Z","timestamp":1759334377355,"version":"build-2065373602"},"reference-count":14,"publisher":"IEEE","license":[{"start":{"date-parts":[[2025,4,28]],"date-time":"2025-04-28T00:00:00Z","timestamp":1745798400000},"content-version":"stm-asf","delay-in-days":0,"URL":"https:\/\/doi.org\/10.15223\/policy-029"},{"start":{"date-parts":[[2025,4,28]],"date-time":"2025-04-28T00:00:00Z","timestamp":1745798400000},"content-version":"stm-asf","delay-in-days":0,"URL":"https:\/\/doi.org\/10.15223\/policy-037"}],"content-domain":{"domain":[],"crossmark-restriction":false},"short-container-title":[],"published-print":{"date-parts":[[2025,4,28]]},"DOI":"10.1109\/aicas64808.2025.11173089","type":"proceedings-article","created":{"date-parts":[[2025,9,25]],"date-time":"2025-09-25T17:52:35Z","timestamp":1758822755000},"page":"1-5","source":"Crossref","is-referenced-by-count":0,"title":["Enhancing Training Efficiency: A Novel Approach to Handling GPU Failures in Large-Scale Distributed System for LLM Training"],"prefix":"10.1109","author":[{"given":"Kiran","family":"Gunnam","sequence":"first","affiliation":[{"name":"Micron Technology"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Pranshu Tapan","family":"Mandal","sequence":"additional","affiliation":[{"name":"Micron Technology"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Shivam","family":"Khandelwal","sequence":"additional","affiliation":[{"name":"Micron Technology"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Rajesh","family":"Bhagwat","sequence":"additional","affiliation":[{"name":"Micron Technology"}],"role":[{"role":"author","vocabulary":"crossref"}]}],"member":"263","reference":[{"key":"ref1","first-page":"68","volume":"2","author":"Dubey","year":"2024","journal-title":"The llama 3 herd of models"},{"key":"ref2","first-page":"4403","article-title":"Polynomial Codes: an Optimal Design for High-Dimensional Coded Matrix Multiplication","author":"Yu","year":"2017","journal-title":"Advances In Neural Information Processing Systems (NIPS)"},{"key":"ref3","doi-asserted-by":"publisher","DOI":"10.1109\/ISIT.2018.8437563"},{"key":"ref4","first-page":"11117","article-title":"Cramming: Training a Language Model on a single GPU in one day","volume-title":"International Conference on Machine Learning","author":"Geiping"},{"key":"ref5","doi-asserted-by":"publisher","DOI":"10.1145\/3579370.3594777"},{"journal-title":"Prediction of GPU Failures Under Deep Learning Workloads","year":"2022","author":"Liu","key":"ref6"},{"key":"ref7","doi-asserted-by":"publisher","DOI":"10.3390\/s22072665"},{"key":"ref8","doi-asserted-by":"publisher","DOI":"10.1145\/1327452.1327492"},{"key":"ref9","doi-asserted-by":"publisher","DOI":"10.5555\/1863103.1863113"},{"journal-title":"Training Deep Nets with Sublinear Memory Cost","year":"2016","author":"Chen","key":"ref10"},{"key":"ref11","first-page":"693","article-title":"Hogwild!: A Lock-Free Approach to Parallelizing Stochastic Gradient Descent","volume-title":"Proc. 24th Int. Conf. Neural Inf. Process. Syst. (NIPS \u201911)","author":"Recht"},{"key":"ref12","first-page":"1223","article-title":"More Effective Distributed ML via a Stale Synchronous Parallel Parameter Server","volume-title":"Proc. 26th Int. Conf. Neural Inf. Process. Syst. (NIPS \u201913)","author":"Ho"},{"key":"ref13","doi-asserted-by":"publisher","DOI":"10.1109\/HPCA.2015.7056044"},{"key":"ref14","first-page":"1","article-title":"Characterizing and Mitigating the Impact of GPU Failures on Deep Learning Workloads","volume-title":"Proc. 22nd IEEE Int. Symp. High Perform. Comput. Archit. (HPCA \u201916)","author":"Nie"}],"event":{"name":"2025 IEEE 7th International Conference on Artificial Intelligence Circuits and Systems (AICAS)","start":{"date-parts":[[2025,4,28]]},"location":"Bordeaux, France","end":{"date-parts":[[2025,4,30]]}},"container-title":["2025 IEEE 7th International Conference on Artificial Intelligence Circuits and Systems (AICAS)"],"original-title":[],"link":[{"URL":"http:\/\/xplorestaging.ieee.org\/ielx8\/11172731\/11173086\/11173089.pdf?arnumber=11173089","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2025,9,30]],"date-time":"2025-09-30T15:23:27Z","timestamp":1759245807000},"score":1,"resource":{"primary":{"URL":"https:\/\/ieeexplore.ieee.org\/document\/11173089\/"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2025,4,28]]},"references-count":14,"URL":"https:\/\/doi.org\/10.1109\/aicas64808.2025.11173089","relation":{},"subject":[],"published":{"date-parts":[[2025,4,28]]}}}