These technical notes are provided as a service to our customers.
Above is a 30 PIN SIMM. It has 30 gold pins on the front and the back. It comes in either a one MEG or four Megabyte Strip. Most computers use this in a set of four. Length is 3.5 inches long and the height varies.
Animated GIF by Donnie Heath.
Above is a 72 PIN SIMM. It has 72 gold or tin pins on the front and the back. Most mother board manufacturers such as Intel are requesting that you use tin leads to avoid plating. It comes in either a Four 1X, Eight 2X, Sixteen 4X or thirty-two 8X Megabyte Strip. Most 486Based computers can use these one at a time and can use either parity X36 or non parity X32 RAM. Most Pentium based computers use this in a set of two. Pentiums use these in either EDO or non parity X32 arrangements. Length is 4.25 inches long and the height varies.
SOJ Video RAM. Known as
a 256K X 16 SOJ Can either be EDO or Non EDO check your video card manual to be sure. If
you don't know then use non EDO. Length is 1 inch by 7/16ths wide.
SDRAM DIMM SyncDRAM, which is the latest type of PC memory
architecture. This provides a significant performance improvement over conventional Fast
Page Mode (FPM) or Extended Data Output (EDO). A key difference found in SyncDRAM
is that it is synchronized to the system clock signal, which means that all operations run
at the same speed as the processor bus, i.e. 60 or 66MHz. This is compared to conventional
memory architectures which run asynchronously to the processor bus. Due to this
synchronization, data bits are retrieved faster and thereby increase system performance.
PC applications commonly access DRAM data in 4-quad word (Qword) burst lengths. Optimum
system performance is achieved when data is fed to the processor on par with the system
clock. Data access from an L2 cache comes close to meeting this requirement as typical
only one or two wait states are inserted while accessing the first data. However, the
remaining three Qwords are delivered on a par with the processor speed, denoted as a
2/1/1/1 burst rate. When, however, the requested data must come directly from DRAM, the
burst rate is degraded. Assuming the burst is from a DRAM page that is already open (page
hit), the achievable burst rate at 66MHz using 60ns FPM DRAMs is 5/3/3/3 cycles; for EDO
DRAMs it is 6/2/2/2. SyncDRAMs can reduce burst rates to 7/1/1/1 for L2
cache misses. While SyncDRAM takes more time to locate the first data
bit, it then is faster at retrieving the next three Qwords. Furthermore, it really excels
at back to back reads which assumes a page hit:
SDRAM
Synchronous DRAM (SDRAM) is “in synch” or synchronized to the system clock that
controls the CPU. The clock that controls the microprocessor also controls the SDRAM, thus
eliminating wait states and reducing data retrieval times. This synchronization allows the
memory controller to know on which clock cycle data requests will be available. Data is
thus input to the rising edge of the clock instead of with every two clock cycles (like
EDO) or every three clock cycles (like FPM). SDRAM also utilizes multiple memory banks
that function simultaneously, in addition to a burst mode feature that addresses an entire
block rather than just one piece of data. SDRAM is currently in production.
2.5V, 168-pin (64/128 Mbit Non-ECC or 72/144 Mbit ECC), 400 MHz, 356 MHz, and
300 MHz RDRAM RIMM* modules. The devices are known as PC800, PC700, and PC600
respectively.
Note: There can be no empty RIMM connectors. RIMM connectors without a RIMM installed must be populated with Continuity RIMMs. The Continuity RIMMs are used to provide signal paths for the memory modules.
Buffered vs. Non-Buffered: what is the difference?
Most of the requirements related to memory design
depend on the computer system board design. This is also the case with Buffered and
Non-Buffered DIMMs. A buffer is a driver. Some system boards include drivers for the
memory control signals, while others rely on the memory module to drive its own memory
control signals. A buffered module has drivers on it; a non-buffered module does not. A
system that requires buffered modules will not function with a non-buffered module and
vise versa. If you have a clone system and must order a generic DIMM, make sure you
know which kind is needed.
What is SPD?
When a computer system boots up, it needs to “detect” the configuration of the
memory modules in order to run properly. SPD (Serial Presence Detect), uses an EPROM
to store information about the module. The SPD then tells the computer what kind of
RAM is installed in the system
The difference between 2-clock and 4-clock RAM.
SDRAM requires clock lines running from the system clock to the memory module. Two clock
means there are two clock lines running to the module, and four clock means there are four
clock lines running to the module. Four clock designs are faster because they allow less
chips per clock line, enabling a quicker interface with data.
Intel has introduced the PC 100 main memory bus in their system designs. PC 100 is a faster main memory bus. At 100MHz, PC 100 is roughly 33% faster than the earlier 66MHz main memory bus. To match this faster 100MHz bus speed, 100MHz SDRAM modules will be the required memory technology for PC 100.
Systems with CPU's from 233 to 333 have a 66 MHz bus. CPU's 350 MHz and above use a 100
MHz bus.
Definition: Just what is a Bus?
The bus is the main communication avenue in a PC. The bus is the primary data traffic lane
from the CPU to all key subsystems. It consists of various parallel connections to
which the CPU, memory, and all input-output devices are connected. The bus can send data
in either direction between any two system devices. For example, a 16-bit bus transfers
two bytes at a time over 16 wires, a 32-bit bus transfers four bytes at a time over 32
wires. Much like a highway, more lanes equals greater traffic flow.