SDRAM
Memory chips synchronised to a clock signal for higher bandwidth
SDRAM is a type of dynamic memory that uses an external clock to coordinate data transfers. Samsung released the first commercial chip in 1992, leading to widespread adoption in computers by 2000 due to improved concurrency and speed over older asynchronous designs.
- First commercial chip released by Samsung in 1992
- Replaced asynchronous DRAM in most computers by 2000
- Uses pipelining to accept new commands during ongoing tasks
- Divided into independent banks for simultaneous access
- JEDEC adopted the first standard for it in 1993
What it is
Synchronous dynamic random-access memory, known as SDRAM, is a form of computer storage where operations are timed by an external clock signal. Unlike earlier asynchronous memory, which responded directly to control inputs, SDRAM waits for a clock edge before acting. This coordination allows the chip to pipeline commands, meaning it can accept new instructions before finishing previous ones. The result is higher data transfer rates and better concurrency. By the year 2000, this technology had replaced almost all other dynamic memory types in modern computing systems due to its superior performance characteristics.
How it came about
Early dynamic memory from the 1970s was often synchronised with processor clocks, but the industry shifted to asynchronous designs by the mid-1970s. Synchronous operation returned in the 1990s. IBM demonstrated dual-edge clocking features in the late 1980s, though these were not standard SDRAM. Samsung Electronics produced the first commercial SDRAM chip, the KM48SL2000, in 1992 using a CMOS process. Mass production followed in 1993. JEDEC adopted the first formal standard for this memory in 1993. Samsung later demonstrated Double Data Rate (DDR) SDRAM in 1997, releasing the first commercial DDR chip in June 1998.
Inside the chip
The chip divides its storage into independent sections called banks, allowing simultaneous access to different areas. An internal finite-state machine steps through commands based on the clock signal. Control inputs such as row address strobe, column address strobe, and write enable select specific operations. A clock enable signal can pause the chip, while a chip select line ignores inputs when inactive. Data mask lines suppress input or output as needed. Pipelining lets the device start new tasks while old ones complete, improving throughput without reducing inherent latency. CAS latency defines the delay between a column address request and data availability.
How it was made
Samsung manufactured the initial commercial units using complementary metal–oxide–semiconductor fabrication. Today, major producers include SK Hynix, Micron Technology, ChangXin Memory Technologies, and Nanya Technology. Most modern modules comply with open standards set by JEDEC to ensure components work together. These standards cover various generations, including DDR, DDR2, and DDR3. The manufacturing process has evolved to support higher densities and faster interface speeds, although fundamental timing limits like read cycle time have remained relatively stable across recent generations.
Where it went
This memory type is found in virtually all modern personal computers and servers. Registered varieties are used in workstations and data centres where scalability is critical. Intel published the PC100 standard to ensure modules could reliably operate at 100 MHz, a guideline that influenced naming conventions like PC133. The technology supports high-bandwidth applications by interleaving operations across multiple banks. It serves as the primary system memory for processors that require fast, synchronous data access for both reading and writing tasks.
Versions and family
Double Data Rate SDRAM emerged as a successor, first shown by Samsung in 1997. Subsequent generations include DDR2 and DDR3, which use wider mode registers and additional address bits. DDR2 removed the burst terminate command, while DDR3 repurposed it for calibration. DDR4 introduced a new activate signal encoding. Earlier single-data-rate versions are now largely obsolete. The family also includes registered modules for enterprise systems and unbuffered types for consumer devices, all sharing the same basic command structure but differing in timing and voltage requirements.
Why it matters
SDRAM became the dominant memory technology by 2000, displacing asynchronous DRAM and burst EDO types. Its ability to interleave bank operations allowed for significant gains in effective bandwidth, even if raw latency did not drop immediately. The establishment of JEDEC standards facilitated interoperability among components from different manufacturers. This uniformity helped drive down costs and accelerate adoption in the personal computer market. The transition to synchronous design laid the groundwork for future high-speed memory interfaces used in today’s computing infrastructure.
Problems and reception
Early implementations were sometimes slower than contemporaneous burst EDO DRAM due to added logic complexity. Some manufacturers sold modules labelled as 100 MHz that could not sustain that speed reliably. This led to the creation of strict certification standards like PC100. Despite initial teething issues, the industry embraced the technology for its scalability and pipelining benefits. The shift to synchronous operation was widely accepted as necessary for keeping pace with increasing processor speeds and bus frequencies.




Questions people ask about SDRAM
How does SDRAM differ from older asynchronous memory?
Asynchronous memory responds directly to control signals with only pathway delays. SDRAM uses an external clock signal to coordinate actions. Changes on control inputs are recognised only after a rising clock edge, allowing for pipelined operations and better concurrency across multiple memory banks.
What is the benefit of having multiple banks in SDRAM?
Banks are independent sections of memory. The chip can process a command in one bank while another bank is busy or refreshing. This interleaving allows for higher data transfer rates and greater concurrency compared to accessing a single continuous block of memory.
Who sets the standards for SDRAM compatibility?
JEDEC, an electronics industry association, establishes open standards for these components. They adopted the first SDRAM standard in 1993 and have since defined specifications for DDR, DDR2, and DDR3 generations to ensure interoperability between chips from different manufacturers.
Good reading and sources
- Synchronous dynamic random-access memory on WikipediaEncyclopedia
- More photos on Wikimedia CommonsPhotos
- Wikidata recordOpen data
The facts on this page come from the Wikipedia article "Synchronous dynamic random-access memory", written by its contributors and used under the CC BY-SA 4.0 licence, and from its Wikidata record (CC0). EveryChip rewrote the text in its own words, with AI assistance, and compares every page with the article to make sure no passage is copied. It may be reused under the same CC BY-SA 4.0 licence with a link back.
It may contain mistakes, and figures change over time. Nothing here is investment or buying advice. If something is wrong, tell us. Last refreshed 2026-11-14.