Encyclopedia of Computer Science and Technology

memory 301about prospective drugs and treatments. (See personalhealth information management.)Further ReadingChen, Hsinchun, et al., eds. Medical Informatics: Knowledge Managementand Data Mining in Biomedicine. New York: Springer,2005.LinuxMedNews [open-source medical software]. Available online.URL: http://www.linuxmednews.com/. Accessed August 14,2007.Medical Software (Yahoo! Directory). Available online. URL:http://dir.yahoo.com/Business_and_Economy/Business_to_Business/Health_Care/Software/Me dical/. Accessed August14, 2007.Shortliffe, Edward H., and James J. Cimino, eds. Biomedical Informatics:Computer Applications in Health Care and Biomedicine.New York: Springer, 2006.Sullivan, Frank, and Jeremy Wyatt. ABC of Health Informatics. Malden,Mass.: Blackwell, 2006.Wooton, Richard. Introduction to Telemedicine. 2nd ed. London:Royal Society of Medicine Press, 2006.memoryGenerally speaking, memory is a facility for temporarilystoring program instructions or data during the courseof processing. In modern computers the main memory israndom access memory (RAM) consisting of silicon chips.Today’s personal computers typically have from between64MB (megabytes) and 512MB of main memory.Development of the TechnologyIn early calculators “memory” was stored as the positionsof various dials. Charles Babbage conceived of a “store” ofsuch dials that could hold constants or other values neededduring processing by his Analytical Engine (see Babbage,Charles).A number of forms of memory were used in early electronicdigital computers. For example, a circuit with aninherent delay could be used to store a series of pulses thatcould be “refreshed” every fraction of a second to maintainthe data values. The Univac I, for example, used a mercurydelay line memory. Researchers also experimented withcathode ray tubes (CRTs) to store data patterns.The most practical early form of memory was the ferritecore, which consisted of an array of tiny donut-shaped magnets,crisscrossed by electrical lines so that any elementcan be addressed by row and column number. By convertingdata into appropriate voltage levels, the magnetic stateof the individual elements can be switched on and off torepresent 1 or 0. In turn, a current can be passed throughany element to read its current state—although the elementmust then be remagnetized. Ferrite cores were relativelyfast but expensive, and “core” became programmers’ shorthandfor the amount of precious memory available.By the 1960s, the use of transistors and integrated circuitsmade electronic solid-state memory systems possible. Sincethen, the MOSFET (Metal Oxide Semiconductor Field EffectTransistor) using CMOS (Complementary Metal Oxide) fabricationhas been the dominant way to implement DRAM(dynamic random access memory). Here “dynamic” meansthat the memory must be “refreshed” by applying currentafter data is read in each cycle, and “random access” meansthat any desired memory location can be accessed directlyrather than requiring locations to be read sequentially.Static RAM is used in some computer components wheremaximum memory speed is desirable. Static memory is fasterbecause it does not need to be refreshed after each readingcycle. However, it is also considerably more expensive.Memory performance is also dependent on how quicklylocations in the memory can be addressed. The earliestforms of DRAM required that the row and column of thedesired memory location be sent in separate cycles. EDO(Extended Data Out) and more recent technologies allowthe row to be requested one time, and then just the columngiven for adjacent or nearby locations. Timing and pipeliningtechniques can also be used to start a new request whilethe previous one is still being processed.For SDRAM (synchronous DRAM), memory speed islimited by the inherent response time of the memory chip,but also by the number of clock cycles per second initiatedby the data bus (see bus). Double data rate (DDR) SDRAMis able to use both the “rising” and “falling” part of theclock cycle to transfer data, doubling throughput. It is beingsuperseded by DDR2, which achieves another doublingbecause it can run the data-transfer bus at twice the systemclock speed. However, this does increase latency (the timeneeded to begin an access). On the horizon is DDR3, whichcan run the bus at four times clock speed—yet anotherdoubling. Possible future memory technologies include“spintronics,” or the use of the quantum state or “spin” ofelectrons to hold data. The speed, compactness, and reliabilityof this technology could exceed current devices by afactor of hundreds to thousands.As memory gets faster, it continues to get cheaper (atleast for all but the latest technology). At the same time,memory demands continue to increase. Today’s PCs generallycome with 1 GB (billion bytes) of RAM, and 2 GB ormore is often recommended, particularly for memory-hungryoperating systems such as Microsoft Windows Vista andapplications such as Adobe PhotoShop and video editing.Another popular kind of memory is “flash” (nonvolatile)memory that does not require power to maintain itscontents. This kind of memory is used in a wide variety ofdevices, including digital cameras, PDAs, and media players(see also flash drive).In actual systems, a small amount of faster memory (seecache) is used to hold the data that is most likely to beimmediately needed. A proper balance between primaryand secondary cache and main memory in the system chipsetmakes it less necessary to use the fastest, most expensiveform of main memory.Many computers also have ROM (Read-Only Memory) orPROM (Programmable Read-Only Memory). This memoryholds permanent system settings and data (see bios) that areneeded during the startup process (see boot sequence).Further ReadingJacob, Bruce, Spencer Ng, and David Wang. Memory Systems:Cache, DRAM, Disk. San Francisco: Morgan Kaufmann, 2007.