DIMM, SIMM, DDR & SDRAM

DIMM

The DIMM (Dual In-line Memory Module) slots were made for multitasking in computing devices. It was made to replace SIMM (Single in Line Memory Module) as it had high-speed data flow and gave users more options. DIMM contains one or more RAM chips on a small PCB with pins that connects to the computer motherboard. The DIMM stores each bit of data in a separate memory cell. DIMMs are used in personal computers, laptops, mobile phones and etc. DIMM has many DRAM integrated circuits. The modules are attached to the PCB, with a lot RAM chips on a single circuit board, which is in the motherboard. With DMA (Direct memory access), a Desktop processor can access any sector of the memory directly without having to see in chronological order. With DRAM, RAM directly accesses all parts of the memory.

TYPES OF DIMM-

SO-DIMM

SO-DIMM (Small outline Dual Input Memory Module) slots are used since there is a lack of space in smaller and compact devices. SO-DIMM slots take half the space in comparison to DIMM slots, but it draws more power than DIMM.

MicroDIMM

MicroDIMM is a type of memory module used in small form factor computers. It is the smallest RAM form factor. MicroDIMM is about 45.5 millimeters long and 30 millimeters wide in physical size. It is over 50 percent smaller than a SO-DIMM.

Error correction(ECC) 

ECC DIMMs have extra data bits which can be used by the memory controller to detect and correct errors. They most of the time carry a multiple of 9 instead of a multiple of 8 chips. There are many ECC schemes, the most common is Single Error Correct, Double Error Detect (SECDED) uses an extra byte per 64-bit word.

SIMM

SIMM (Single In-Line Memory Module) is a PCB with a lot of memory chip on it. It is an older type of computer memory. It was made by Wang laboratories in 1983. It holds 6 to 9 memory chips per board and the ninth chip was most of the time is used for parity error checking. They are calculated in bytes instead of bits. SIMM to the actual memory chips is 32 bits wide from the bus unlike DIMM 64 bit wide.

DDR & SDRAM

SDRAM

SDRAM (synchronous DRAM) is just DRAM with synchronized cloak speeds which the microprocessor is made for. It increases the number of instructions in a given time the processor can perform. Hence, it is more efficient. Samsung made the first commercial SDRAM in 1993. The internal finite state machine (Pipelines incoming instructions) enables SDRAM to work in more complicated instructions than simple DRAM, so it functions at way higher speeds. Cause of this SDRAM is able to keep two sets of memory addresses open simultaneously. By transferring data alternately from one set of addresses, then the other. Term - Pipelining is a process in which the SDRAM can accept a new instruction before it has finished processing the previous one.

SDRAM types: DDR versions and QDR

SDRAM technology underwent a huge amount of development. As a result, several successive families of the memory was introduced, each with improved performance over the previous generation.

• SDR SDRAM: First SDRAM that was introduced. It has now been demolished by the newer ones. It is known as single data rate SDRAM, or simply SDRAM.

• DDR SDRAM: DDR SDRAM or DDR1 SDRAM has its name because it is Double Data Rate SDRAM. These types of SDRAM provide data transfer double the speed of the original SDRAM memory. This is done by transferring data twice per cycle.

• DDR2 SDRAM: DDR2 SDRAM came in 2003 can use the external bus double the speed of DDR SDRAM.

• DDR3 SDRAM: DDR3 SDRAM is a more developed version of the double-data rate type of SDRAM. It provides better overall performance and speeds then the last-gen.

• DDR4 SDRAM: In 2014 DDR4 SDRAM was the new generation of DDR SDRAM. It provided better performance for day-to-day needs.

• DDR5 SDRAM:

The next generation of SDRAM incoming, labeled as DDR5 is at the moment under development. The specification was shown in 2016 with expected first production starting in 2020. DDR5 is going to reduce power consumption while having double the bandwidth and capacity from the last-gen.

• QDR SDRAM: It can transfer up to four words of data in each clock cycle. It is not double the speed of DDR SRAM but is 100% efficient when reads and writes are inserted.

DDR

DDR (Double Data Rate) is an advanced version of SDRAM. It is a method in calculating with whic.
h a computer bus transfers data at double the rate sending data at up and down edges of a clock cycle. This process allows for the ram to send 2 signals per clock cycle which is double the speed of SDR (Single Data Rate). In August 2000, the first retail PC motherboard using DDR SDRAM was released. There are many forms of DDR Ram, DDR2 in 2003, DDR3 2007, DDR4 2014. Every clock cycle has unidirectional data flowing. It has the same frequency as the clock cycle. The method is used to incorporate unstable storage in Graphic Cards and is known as GDDR (graphics double data rate). It is made to work with video cards only. Are used in analog-to-digital converters.

This method was an improved over its predecessor SDR (single data rate) and has been succeeded by an enhanced version named QDR (quad data rate). 

GDDR

Graphics DDR is a type of  SDRAM which is made for GPUs, it has similarities with the DDR SDRAM which is used as a system ram, but it is much faster, has higher clock speeds, higher bandwidths, and power management. 

Some major differences between GDDR and DDR are- 

• GDDR has a larger bit bus, which lets it have greater bandwidth.
• GDDR is made especially to handle Graphics workloads whereas DDR is made especially for handling assorted computer workloads.
• GDDR can send data and receive data on the same clock cycle and DDR handles data packets in small bits with low latency.

Both are made for what they are best at.

As with DDR, GDDR also has its Generations, currently the leading generation is GDDR6 and Nvidias new 3000 series lineup introduces GDDR6X, which increases data transfer data rate by using PAM4 (Pulse Amplitude Modulation 4), which implies that instead of two power levels which indicate 1 and 0, it is able to communicate with 4 power levels, allowing it to transfer four signals, being, 00, 01, 10, 11, and it doubles the data transfer rate than traditional methods, and each GDDR6X chip on an RTX 3080 can transfer 84GB/s and a total memory bandwidth of 840GB/s. But it comes with a lower memory clock in order to make the power consumption not go beyond System's capabilities, otherwise, it's a fair way for your PSU.

But GDDR is not the only technology used for making video memory modules, another technology used to make VRAM modules is HBM, which is High Bandwidth Memory, the leading lineup of HBM is HBM2E used in  Nvidia Titan V 12GB, AMD Radeon VII 16GB, AMD Radeon RX Vega 

HBM2 comes with a maximum bandwidth of 256Gb/s whereas GDDR6 can reach 768GB/s at its maximum bandwidth.

In a Graphics card GDDR VRAM' s single chip will b replaced alongside the GPU, the HBM chips are stacked on one top of each other and Through-Silicon Vias (TSVs) connects the chips, it saves space and form factor.

               GDDR6                                                                          HBM2

Advanages over GDDR:

• Much Higger bandwidth compared to GDDR standards, GDDR5 can provide 10.6GB/s per watt, whereas HBM2 can do 30+GB/s per watt.
  
• Smaller form factor and saves space.
• Lower power consumption and heat production.

But HBM can be more expensive than GDDR, by a lot.

HBM2 is a better option if you looking for performance and low power consumption, but we can only find it in Founder's GPUs and are very expensive. The best HBM2 Graphics card in the market is AMD Radeon VII 16GB so far, with a cheaper price tag than other HBM2 Graphics cards, and also with a better price to performance ratio.

Enable GingerCannot connect to Ginger Check your internet connection
or reload the browserDisable in this text fieldRephraseRephrase current sentenceEdit in Ginger×