Download presentation
Presentation is loading. Please wait.
Published byArleen Carson Modified over 9 years ago
1
Yi-Lin, Tu 2013 IEE5011 –Fall 2013 Memory Systems Wide I/O High Bandwidth DRAM Yi-Lin, Tu Department of Electronics Engineering National Chiao Tung University b22531423@hotmail.com
2
NCTU IEE5011 Memory Systems 2013Yi-Lin, Tu Outline Introduction Proximity Communication A Wide I/O DRAM Architecture Conclusion Reference 2
3
NCTU IEE5011 Memory Systems 2013Yi-Lin, Tu Introduction Memory gap How to solve this problem? Other techniques? 3
4
NCTU IEE5011 Memory Systems 2013Yi-Lin, Tu Proximity Communication A wireless chip-to-chip communication technology. Two chips are placed face to face and their bonding pads are allowed to come within close proximity of each other without touching. 4
5
NCTU IEE5011 Memory Systems 2013Yi-Lin, Tu Proximity Communication 5
6
NCTU IEE5011 Memory Systems 2013Yi-Lin, Tu Proximity Communication Advantages Increase I/O density Remove the on/off chip wires Remove the on-die termination Ease of testability Remove the ESD structures 6
7
NCTU IEE5011 Memory Systems 2013Yi-Lin, Tu Proximity Communication 7
8
NCTU IEE5011 Memory Systems 2013Yi-Lin, Tu Proximity Communication Challenges Mechanical misalignment Supplying power to chips Thermal removal 8
9
NCTU IEE5011 Memory Systems 2013Yi-Lin, Tu Proximity Communication Mechanical misalignment Six axis Multiple source 9
10
NCTU IEE5011 Memory Systems 2013Yi-Lin, Tu Proximity Communication Electronic sensor Chip to chip separation sensor Vernier scale(translation) 10
11
NCTU IEE5011 Memory Systems 2013Yi-Lin, Tu Proximity Communication Electronic re-alignment Use receiver and transmitter array. This array has the ability to electrically reposition the transmitter pads to align the transmitter and receiver pads. 11
12
NCTU IEE5011 Memory Systems 2013Yi-Lin, Tu A Wide I/O DRAM Architecture Using 4Gb DRAM as the starting point for developing a wide I/O DRAM architecture. 12
13
NCTU IEE5011 Memory Systems 2013Yi-Lin, Tu A Wide I/O DRAM Architecture Pad moving Moving the I/O channel to the edge. Data and comment signals will need to be buffered at the center. Allows the local column circuitry to be moved to the center. 13
14
NCTU IEE5011 Memory Systems 2013Yi-Lin, Tu A Wide I/O DRAM Architecture Centralizing Limit the bandwidth of the column and row path. It’s possible by using proximity communication. Increase the array efficiency. 14
15
NCTU IEE5011 Memory Systems 2013Yi-Lin, Tu A Wide I/O DRAM Architecture Conventional DRAM chips operate with eight internal memory bank. Enable eight wordlines to be active at once, one of each bank. Possible to perform sequential column access to each bank. Remove the large row access latency. 15
16
NCTU IEE5011 Memory Systems 2013Yi-Lin, Tu A Wide I/O DRAM Architecture 512Mb bank structure 4 possible arrangements for creating a 512Mb memory bank. Keep the global I/O metal lines short allows for a higher bandwidth on an open page. C and D are preferred. 16
17
NCTU IEE5011 Memory Systems 2013Yi-Lin, Tu A Wide I/O DRAM Architecture 17
18
NCTU IEE5011 Memory Systems 2013Yi-Lin, Tu A Wide I/O DRAM Architecture Challenges Number of metal layers Global I/O routing Local I/O routing 18
19
NCTU IEE5011 Memory Systems 2013Yi-Lin, Tu A Wide I/O DRAM Architecture Number of metal layers and global I/O routing A wide I/O architecture with 64 data pins operating with burst length of eight, and therefore a pre-fetch of 8n, requires 512 bits to be accessed in parallel. The highest level of metal is used for global I/O routing and metal one is for global wordlines. Increase the parasitic of each wordline. 19
20
NCTU IEE5011 Memory Systems 2013Yi-Lin, Tu A Wide I/O DRAM Architecture Divide the 8k page 256 bits per half-bank This enables a possibility of increasing the number of global I/O tracks from 512 to 1024 or higher. 20
21
NCTU IEE5011 Memory Systems 2013Yi-Lin, Tu A Wide I/O DRAM Architecture Local I/O routing The large number of global I/O tracks requires 32 data signals from each 256 kb memory array. Moving 32 data signals from the bitline sense amplifiers to the global I/O track is a major challenge due to the limited routing space above the bitline sense amplifiers. Signals can be routed to the top and bottom of each 256kb memory segment. 21
22
NCTU IEE5011 Memory Systems 2013Yi-Lin, Tu A Wide I/O DRAM Architecture Summary Developing a wide I/O DRAM architecture that is suitable for Proximity Communication requires the communication channel to be moved to the side of the DRAM chip. A distributed page and bank structure was developed to enable the possibility of using Proximity Communication with 32 data pins. Reaching the use of 64 data pins required architectural changes that would not increase the manufacturing cost compared to current DRAM architectures. 22
23
NCTU IEE5011 Memory Systems 2013Yi-Lin, Tu Conclusion 23 A DRAM architecture that uses proximity communication to increase the off-chip bandwidth while scaling the number of data pins. Proximity communication allows for an increase of I/O density, ease of testability, removal of ESD structures and resistive termination. Electrical sensors and electrical re-alignment techniques has enabled proximity communication to become a viable I/O technology. The challenges of creating a wide I/O architecture were found to be in the global and local I/O routing.
24
NCTU IEE5011 Memory Systems 2013Yi-Lin, Tu Reference Q. Harvard, “Wide I/O DRAM architecture utilizing proximity communication,” Master’s thesis, Boise State University, December 2009. Q. Harvard, R. J. Baker, and R. Drost, “Main memory with proximity communication: A wide I/O DRAM architecture,” in Proc. IEEE Workshop Microelectron. Electron Devices, Apr. 2010, pp. 1–4. Harvard, Q., Baker, R.J., “A scalable I/O architecture for wide I/O DRAM,” 54th International Midwest Symposium on Circuits and Systems (MWSCAS), 7-10 Aug. 2011, Seoul, 2011 24
25
Yi-Lin, Tu 2013 Thank you 25
Similar presentations
© 2025 SlidePlayer.com. Inc.
All rights reserved.