Unleashing the Power of Embedded DRAM

نویسنده

  • Peter Gillingham
چکیده

Embedded DRAM technology offers many advantages in System On Chip products. Computing applications demand memory with low latency and zero soft error rate. Graphics and networking need high bandwidth. Mobile applications require extremely low power. All applications benefit from the high density afforded by embedded DRAM technology. A single DRAM architecture cannot provide an optimal solution across the full range of applications. This paper describes in detail the advantages of embedded DRAM technology over external memory and embedded SRAM, and presents three 90nm embedded DRAM architectures optimized for specific applications. Embedded DRAM Value Proposition DRAM technology offers the highest density random access memory due to a simple 1T1C structure consisting of a single access transistor and a single storage capacitor. Typical 90nm embedded DRAM processes offer cell sizes in the range of 0.2?m2. In contrast, 90nm SRAM cells typically occupy around 1.0?m2 due to a more complex 6T structure comprised of cross-coupled latch and dual access devices. Taking into account the more extensive peripheral circuitry associated with DRAM, the overall density of a completed embedded DRAM marcocell is about four times that of the SRAM alternative. In DRAM a read operation is destructive since the charge on the storage capacitor is shared with the larger bitline capacitance. This data must be sensed and fully restored to the cell capacitors during each memory cycle. In a typical DRAM row operation, many thousands of bits are sensed and are available for use on-chip. In commodity DRAM, only a small fraction of the bits which are sensed in each row cycle are made available off-chip. In a 256M DDR2 SDRAM with a x16 400Mb/s/pin interface for example, 8192 High Bandwidth High Speed Low Power Capacity 16Mb 2Mb 8Mb Interface 2048bit synchronous 128bit synchronous 64bit asynchronous Array Size 256r x 1024c 128r x 256c 256r x 512c Page Size 8192bit 512bit 512bit # of Banks 2 banks 16 banks 1 bank Macro Size 6.0mm2 1.4mm2 3.3mm2 Density 2.69Mb/mm2 1.32Mb/mm2 2.45Mb/mm2 Data Rate 500MHz 1GHz 166MHz Cycle Time 16ns 4ns 6ns Bandwidth 128GB/s 16GB/s 1.33GB/s Supply 1.2v 1.2v 1.0v (0.7v in standby) Active Power 1.2W 220mW 25mW Standby Power 1mW 66mW 35μW Table 1. Summary of Optimized 90nm Embedded DRAM Macrocells Copyright © 2005 Design And Reuse S.A. All rights reserved.

برای دانلود متن کامل این مقاله و بیش از 32 میلیون مقاله دیگر ابتدا ثبت نام کنید

ثبت نام

اگر عضو سایت هستید لطفا وارد حساب کاربری خود شوید

منابع مشابه

Destructive-read in embedded DRAM, impact on power consumption

This paper explores power consumption for destructive-read embedded DRAM. Destructive-read DRAM is based on conventional DRAM design, but with sense amplifiers optimized for lower latency. This speed increase is achieved by not conserving the content of the DRAM cell after a read operation. Random access time to DRAM was reduced from 6 ns to 3 ns in a prototype made by Hwang et. al. A write-bac...

متن کامل

Embedded DRAM Technologies: Comparisons and Design Tradeoffs

The use of embedded DRAM technology has become widespread, especially in higher-end system designs, because of its superior performance, silicon area savings, and low power compared to discrete memory solutions. Traditionally, in cost-sensitive consumer applications, large memory arrays of 64 megabits and above were usually better suited to discrete commodity memory implementations. But as the ...

متن کامل

VIRAM1: A Media-Oriented Vector Processor with Embedded DRAM

Processors for mobile multimedia devices must be low power while having excellent performance on media applications. Our processor, VIRAM1, accomplishes this by combining vector processing with embedded DRAM. VIRAM1 includes a scalar core, 13 megabytes (104 megabits) of DRAM, and four vector datapaths. It consumes 2 watts at 200 MHz and executes up to 9.6 giga-ops (16 bit) per second.

متن کامل

Reducing On-Chip DRAM Energy via Data Transfer Size Optimization

This paper proposes a software-controllable variable linesize (SC-VLS) cache architecture for low power embedded systems. High bandwidth between logic and a DRAM is realized by means of advanced integrated technology. System-in-Silicon is one of the architectural frameworks to realize the high bandwidth. An ASIC and a specific SRAM are mounted onto a silicon interposer. Each chip is connected t...

متن کامل

A 120-mW 3-D Rendering Engine With 6-Mb Embedded DRAM and 3.2-GB/s Runtime Reconfigurable Bus for PDA Chip

A low-power three-dimensional (3-D) rendering engine is implemented as part of a mobile personal digital assistant (PDA) chip. Six-megabit embedded DRAM macros attached to 8-pixel-parallel rendering logic are logically localized with a 3.2-GB/s runtime reconfigurable bus, reducing the area by 25% compared with conventional local frame-buffer architectures. The low power consumption is achieved ...

متن کامل

ذخیره در منابع من


  با ذخیره ی این منبع در منابع من، دسترسی به آن را برای استفاده های بعدی آسان تر کنید

برای دانلود متن کامل این مقاله و بیش از 32 میلیون مقاله دیگر ابتدا ثبت نام کنید

ثبت نام

اگر عضو سایت هستید لطفا وارد حساب کاربری خود شوید

عنوان ژورنال:

دوره   شماره 

صفحات  -

تاریخ انتشار 2005