Abstract
Recent advancements in deep learning techniques facilitate intelligentquery support in diverse applications, such as content-based image retrieval and audio texturing. Unlike conventional key-based queries, these intelligent queries lack efficient indexing and require complex compute operations for feature matching. To achieve highperformance intelligent querying against massive datasets, modern computing systems employ GPUs in-conjunction with solid-state drives (SSDs) for fast data access and parallel data processing. However, our characterization with various intelligent-query workloads developed with deep neural networks (DNNs), shows that the storage I/O bandwidth is still the major bottleneck that contributes 56%-90% of the query execution time. To this end, we present DeepStore, an in-storage accelerator architecture for intelligent queries. It consists of (1) energy-efficient in-storage accelerators designed specifically for supporting DNNbased intelligent queries, under the resource constraints in modern SSD controllers; (2) a similarity-based in-storage query cache to exploit the temporal locality of user queries for further performance improvement; and (3) a lightweight in-storage runtime system working as the query engine, which provides a simple software abstraction to support different types of intelligent queries. DeepStore exploits SSD parallelisms with design space exploration for achieving the maximal energy efficiency for in-storage accelerators. We validate DeepStore design with an SSD simulator, and evaluate it with a variety of vision, text, and audio based intelligent queries. Compared with the state-of-the-art GPU+SSD approach, DeepStore improves the query performance by up to 17.7, and energy-efficiency by up to 78.6.
| Original language | English (US) |
|---|---|
| Title of host publication | MICRO 2019 - 52nd Annual IEEE/ACM International Symposium on Microarchitecture, Proceedings |
| Publisher | IEEE Computer Society |
| Pages | 224-238 |
| Number of pages | 15 |
| ISBN (Electronic) | 9781450369381 |
| DOIs | |
| State | Published - Oct 12 2019 |
| Event | 52nd Annual IEEE/ACM International Symposium on Microarchitecture, MICRO 2019 - Columbus, United States Duration: Oct 12 2019 → Oct 16 2019 |
Publication series
| Name | Proceedings of the Annual International Symposium on Microarchitecture, MICRO |
|---|---|
| ISSN (Print) | 1072-4451 |
Conference
| Conference | 52nd Annual IEEE/ACM International Symposium on Microarchitecture, MICRO 2019 |
|---|---|
| Country | United States |
| City | Columbus |
| Period | 10/12/19 → 10/16/19 |
Fingerprint
Keywords
- Hardware accelerators
- In-storage computing
- Information retrieval
- Intelligent query
- Solid-state drive
ASJC Scopus subject areas
- Hardware and Architecture
Cite this
DeepStore : In-storage acceleration for intelligent queries. / Mailthody, Vikram Sharma; Qureshi, Zaid; Liang, Weixin; Feng, Ziyan; Gonzalo, Simon Garcia De; Li, Youjie; Franke, Hubertus; Xiong, Jinjun; Huang, Jian; Hwu, Wen Mei.
MICRO 2019 - 52nd Annual IEEE/ACM International Symposium on Microarchitecture, Proceedings. IEEE Computer Society, 2019. p. 224-238 (Proceedings of the Annual International Symposium on Microarchitecture, MICRO).Research output: Chapter in Book/Report/Conference proceeding › Conference contribution
}
TY - GEN
T1 - DeepStore
T2 - In-storage acceleration for intelligent queries
AU - Mailthody, Vikram Sharma
AU - Qureshi, Zaid
AU - Liang, Weixin
AU - Feng, Ziyan
AU - Gonzalo, Simon Garcia De
AU - Li, Youjie
AU - Franke, Hubertus
AU - Xiong, Jinjun
AU - Huang, Jian
AU - Hwu, Wen Mei
PY - 2019/10/12
Y1 - 2019/10/12
N2 - Recent advancements in deep learning techniques facilitate intelligentquery support in diverse applications, such as content-based image retrieval and audio texturing. Unlike conventional key-based queries, these intelligent queries lack efficient indexing and require complex compute operations for feature matching. To achieve highperformance intelligent querying against massive datasets, modern computing systems employ GPUs in-conjunction with solid-state drives (SSDs) for fast data access and parallel data processing. However, our characterization with various intelligent-query workloads developed with deep neural networks (DNNs), shows that the storage I/O bandwidth is still the major bottleneck that contributes 56%-90% of the query execution time. To this end, we present DeepStore, an in-storage accelerator architecture for intelligent queries. It consists of (1) energy-efficient in-storage accelerators designed specifically for supporting DNNbased intelligent queries, under the resource constraints in modern SSD controllers; (2) a similarity-based in-storage query cache to exploit the temporal locality of user queries for further performance improvement; and (3) a lightweight in-storage runtime system working as the query engine, which provides a simple software abstraction to support different types of intelligent queries. DeepStore exploits SSD parallelisms with design space exploration for achieving the maximal energy efficiency for in-storage accelerators. We validate DeepStore design with an SSD simulator, and evaluate it with a variety of vision, text, and audio based intelligent queries. Compared with the state-of-the-art GPU+SSD approach, DeepStore improves the query performance by up to 17.7, and energy-efficiency by up to 78.6.
AB - Recent advancements in deep learning techniques facilitate intelligentquery support in diverse applications, such as content-based image retrieval and audio texturing. Unlike conventional key-based queries, these intelligent queries lack efficient indexing and require complex compute operations for feature matching. To achieve highperformance intelligent querying against massive datasets, modern computing systems employ GPUs in-conjunction with solid-state drives (SSDs) for fast data access and parallel data processing. However, our characterization with various intelligent-query workloads developed with deep neural networks (DNNs), shows that the storage I/O bandwidth is still the major bottleneck that contributes 56%-90% of the query execution time. To this end, we present DeepStore, an in-storage accelerator architecture for intelligent queries. It consists of (1) energy-efficient in-storage accelerators designed specifically for supporting DNNbased intelligent queries, under the resource constraints in modern SSD controllers; (2) a similarity-based in-storage query cache to exploit the temporal locality of user queries for further performance improvement; and (3) a lightweight in-storage runtime system working as the query engine, which provides a simple software abstraction to support different types of intelligent queries. DeepStore exploits SSD parallelisms with design space exploration for achieving the maximal energy efficiency for in-storage accelerators. We validate DeepStore design with an SSD simulator, and evaluate it with a variety of vision, text, and audio based intelligent queries. Compared with the state-of-the-art GPU+SSD approach, DeepStore improves the query performance by up to 17.7, and energy-efficiency by up to 78.6.
KW - Hardware accelerators
KW - In-storage computing
KW - Information retrieval
KW - Intelligent query
KW - Solid-state drive
UR - http://www.scopus.com/inward/record.url?scp=85074449092&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85074449092&partnerID=8YFLogxK
U2 - 10.1145/3352460.3358320
DO - 10.1145/3352460.3358320
M3 - Conference contribution
AN - SCOPUS:85074449092
T3 - Proceedings of the Annual International Symposium on Microarchitecture, MICRO
SP - 224
EP - 238
BT - MICRO 2019 - 52nd Annual IEEE/ACM International Symposium on Microarchitecture, Proceedings
PB - IEEE Computer Society
ER -