Saved in:
| Main Author: | |
|---|---|
| Format: | Recurso digital |
| Language: | English |
| Published: |
Zenodo
2025
|
| Subjects: | |
| Online Access: | https://doi.org/10.5281/zenodo.17283177 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866901626867941376 |
|---|---|
| author | Alex, Anish |
| author_facet | Alex, Anish |
| contents | <p>This article presents a comprehensive overview of specialized cloud hardware for artificial intelligence workloads, addressing the shift from general-purpose computing to purpose-built architectures. As AI applications grow in complexity and scale, traditional computing infrastructures struggle to meet the demanding computational requirements of modern deep learning models. The emergence of dedicated hardware accelerators including Graphics Processing Units, Tensor Processing Units, and Field-Programmable Gate Arrays has revolutionized AI computation, offering substantial performance and efficiency advantages. The integration of these specialized hardware solutions with optimized software frameworks, advanced storage systems, and high-performance networking infrastructure creates a synergistic ecosystem that enables training and deployment of increasingly sophisticated AI models. Additionally, the article examines emerging technologies such as neuromorphic computing, photonic computing, quantum machine learning, and processing-in-memory architectures that promise to further transform AI hardware capabilities in the coming years </p> |
| format | Recurso digital |
| id | zenodo_https___doi_org_10_5281_zenodo_17283177 |
| institution | Zenodo |
| language | eng |
| publishDate | 2025 |
| publisher | Zenodo |
| record_format | zenodo |
| spellingShingle | Specialized cloud hardware for AI workloads: Current state and future directions Alex, Anish Hardware Acceleration Neuromorphic Computing AI Infrastructure Distributed Training Photonic Computing <p>This article presents a comprehensive overview of specialized cloud hardware for artificial intelligence workloads, addressing the shift from general-purpose computing to purpose-built architectures. As AI applications grow in complexity and scale, traditional computing infrastructures struggle to meet the demanding computational requirements of modern deep learning models. The emergence of dedicated hardware accelerators including Graphics Processing Units, Tensor Processing Units, and Field-Programmable Gate Arrays has revolutionized AI computation, offering substantial performance and efficiency advantages. The integration of these specialized hardware solutions with optimized software frameworks, advanced storage systems, and high-performance networking infrastructure creates a synergistic ecosystem that enables training and deployment of increasingly sophisticated AI models. Additionally, the article examines emerging technologies such as neuromorphic computing, photonic computing, quantum machine learning, and processing-in-memory architectures that promise to further transform AI hardware capabilities in the coming years </p> |
| title | Specialized cloud hardware for AI workloads: Current state and future directions |
| topic | Hardware Acceleration Neuromorphic Computing AI Infrastructure Distributed Training Photonic Computing |
| url | https://doi.org/10.5281/zenodo.17283177 |