Column
The Connected Evolution of Memories and Microprocessors
Rochester’s legacy support from yesterday to today
SRAMs for high-speed applications and niche FIFO/Dual-Port memories to buff er the large quantities of data passing throughout the systems. As we moved into the 2000s, microprocessors continued their
performance advancements, with multi-core CPUs becoming common. However, performance was not the only metric receiving attention. Initially, laptops, then smartphones and tablets showed that power was another critical parameter. T e microprocessor and memory suppliers respond accordingly with dedicated products to address the application diversities. Microprocessors continue to progress, adding more cores and
E
lectronics has grown and developed rapidly since its early days. Nowhere is this more visible than in memory and microprocessor products. T ese products have existed like siblings, continuously
competing and pushing each other to new heights. T eir technological developments and successes have been intertwined for decades. In early microprocessor designs, memory requirements were simple
and fulfi lled by SRAM for operating memory and EPROM for non- volatile storage requirements. In the early 1980s, the extent of the relationship between memory and microprocessors became evident. Products like Motorola’s MC68000 family and similar CPUs drove demand for higher-density memories. At the same time, the standard parallel bus connecting to the memory started to show performance limitations. With its unique addressing scheme and cost-eff ectiveness, DRAM
became the leading memory choice. To advance their product lines, memory manufacturers used reductions in process geometries to improve speed and lower power. Introducing a synchronous interface to memory enabled further performance improvements and set the standard to be further developed over future DRAM generations. T e PC industry boomed in the late 1980s and early 1990s, and
Intel x86 microprocessors greatly infl uenced the memory market. Synchronous DRAM (SDRAM) became the dominant solution. Increased soſt ware requirements also increased the demand for higher- density memory, and standardised modules were developed to handle capacity requirements. For non-volatile storage, developing Parallel NOR Flash memory from
suppliers such as AMD/Spansion replaced EEPROM by providing re- programmability and in-system programming, which benefi ted design fl exibility and automated manufacturing. Growth was not limited to the PC Industry; other graphics,
networking, storage and gaming applications were, and still are being developed. Architectures like PowerPC, SPARC, a nascent ARM CPU and dedicated graphic processors captured designs with their memory requirements. T ese applications utilise SDRAM and Synchronous
14 Dec 2024/Jan 2025
www.electronicsworld.co.uk
increasing clock speed. SDRAM evolved into DDR SDRAM, which uses both clock edges to transfer data. T is evolution continued with the development of DDR2, DDR3, DDR4, and DDR5 versions, along with several low-power iterations (Mobile DDR, LPDDR). T e Flash memory market is active as well. T e development of
NAND Flash now competes with NOR Flash to provide a cost- eff ective option for high-density, non-volatile storage. Higher-density NAND solutions such as eMMC, UFS and SSD have begun to replace some mechanical storage solutions such as HDDs. Flash also sees a migration to a serial interface, reducing pin count and allowing smaller, more cost-eff ective packaging.
So, what does all this mean? T e evolution is represented by systems that are still in production today. Medical, automotive, industrial, military and avionics systems require long design cycles and extensive qualifi cation testing, which requires government and international agency approvals. Redesigns due to component obsolescence can consume excessive time and resources, making them cost-prohibitive. Since its founding, Rochester Electronics has been helping to avoid
redesign decisions by providing a source for critical microprocessors and memories, thereby allowing systems to remain in production. Working with our suppliers and customers and staying abreast of industry changes, Rochester continuously evaluates its inventory to provide the best possible coverage for long-lifecycle applications. T is includes fi nished goods inventory on active and obsolete lifecycle devices, investment in wafer and die stock, along with test capabilities for continued manufacturing, and die replication for devices where the original die source is no longer available. When faced with a redesign due to an obsolete memory,
microprocessor, or other device, Rochester may have a solution. Work with Rochester before obsolescence to proactively assist you with your lifecycle plans. For more information visit:
www.rocelec.com
Page 1 |
Page 2 |
Page 3 |
Page 4 |
Page 5 |
Page 6 |
Page 7 |
Page 8 |
Page 9 |
Page 10 |
Page 11 |
Page 12 |
Page 13 |
Page 14 |
Page 15 |
Page 16 |
Page 17 |
Page 18 |
Page 19 |
Page 20 |
Page 21 |
Page 22 |
Page 23 |
Page 24 |
Page 25 |
Page 26 |
Page 27 |
Page 28 |
Page 29 |
Page 30 |
Page 31 |
Page 32 |
Page 33 |
Page 34 |
Page 35 |
Page 36 |
Page 37 |
Page 38 |
Page 39 |
Page 40 |
Page 41 |
Page 42 |
Page 43 |
Page 44 |
Page 45 |
Page 46 |
Page 47 |
Page 48 |
Page 49 |
Page 50 |
Page 51 |
Page 52 |
Page 53 |
Page 54
