Why DIMM has 64 bit data width?

Discussion in 'Computing and Networks' started by spartai, Feb 2, 2015.

  1. spartai

    Thread Starter New Member

    Dec 4, 2014
    8
    0
    Wikipedia’s definition of DIMM says:

    "Most DIMMs are built using "×4" ("by four") or "×8" ("by eight") memory chips with nine chips per side; "×4" and "×8" refer to the data width of the DRAM chips in bits."

    So if DRAM chips have 8 bit data width why DIMM has 64 bit data width? I mean doesn’t memory controller issue read or write request to one DRAM chip at a time? If it does what is the purpose of other 56 bits?
     
  2. MrChips

    Moderator

    Oct 2, 2009
    12,440
    3,361
    Computers use buses to connect the various functions and modules. The first microchips used a 4-bit data bus. The DEC PDP-8/S used a 1-bit data path.

    An 8-bit data bus is very common because 8-bit data is a convenient size for processing text information.

    When working with numbers, 8-bit data is not sufficient and 16, 32 and 64 bits are commonly used. Using an 8-bit data bus to fetch a 64-bit number will require four fetches. A single 64-bit data bus will allow 64-bit access in a single operation, four times faster than an 8-bit bus.
     
  3. spartai

    Thread Starter New Member

    Dec 4, 2014
    8
    0
    If u look at the image below all dram chips have seperate data bits,let's suppose that cpu is executing "mov dword ptr [ebp-40h],0Ch" instruction,it will write 32 bit data to specific address,so these 32 bit data will be separated to 4 dram chips in DIMM module or the data will be written to one dram chip by sending 4x8 byte sequentially?

    [​IMG]
     
  4. MrChips

    Moderator

    Oct 2, 2009
    12,440
    3,361
    The memory chips themselves have 8-bit data paths. Four chips are arranged in parallel to create a single 32-bit data path. Hence a 32-bit access can be performed in a single operation.
     
Loading...