scispace - formally typeset
Proceedings ArticleDOI

A 16Gb 18Gb/S/pin GDDR6 DRAM with per-bit trainable single-ended DFE and PLL-less clocking

Reads0
Chats0
TLDR
This paper presents a 16Gb 18Gb/s/pin GDDR6 DRAM with a die architecture and high-speed circuit techniques on 1.35V DRAM process and introduces a dual channel for a data granularity of 32B with a BL16, per-bit training of l/REF, and an equalizer with PLL-less clocking.
Abstract
Starting at 512Mb 6Gb/s/pin [1], GDDR5's speed and density have been steadily developing for about 10 years; recently achieving 8Gb 9Gb/s/pin [2] with per-pin timing training. Although 8Gb GDDR5X can operate at 12Gb/s [3] by increasing the burst length (BL) from 8 to 16, a degradation in system performance at a data granularity of 64B is seen. The I/O specification, using PLL clocking that additionally causes PLL jitter, has not changed much compared with GDDR5. To overcome these issues, GDDR6 introduced a dual channel for a data granularity of 32B with a BL16, per-bit training of l/ REF , and an equalizer with PLL-less clocking. This paper presents a 16Gb 18Gb/s/pin GDDR6 DRAM with a die architecture and high-speed circuit techniques on 1.35V DRAM process.

read more

Citations
More filters
Proceedings ArticleDOI

Siena: exploring the design space of heterogeneous memory systems

TL;DR: This paper systematically explore the organization of heterogeneous memory systems on a framework called Siena, which facilitates quick exploration of memory architectures with flexible configurations of memory systems and realistic memory workloads.
Journal ArticleDOI

Recent Evolution in the DRAM Interface: Mile-Markers Along Memory Lane

TL;DR: As semiconductor technologies have matured and computing applications have proliferated, memory technology has experienced a corresponding evolution and diversification to address application-specific requirements and topological targets.
References
More filters
Proceedings ArticleDOI

18.1 A 20nm 9Gb/s/pin 8Gb GDDR5 DRAM with an NBTI monitor, jitter reduction techniques and improved power distribution

TL;DR: To cover operation up to 9Gb/s, which is the highest data-rate among implemented GDDR5 DRAMs, this work includes an NBTI monitor, a WCK clock receiver with equalizing and duty-cycle correction modes, CML-to-CMOS converters with wide range operation, active resonant loads at the end of WCK lane, and an on-chip de-emphasis circuit at a 4- to-1 multiplexer output.
Related Papers (5)