Memory Bus Interface Width: Every DDR

Memory bandwidth is the speed at which information could be learn from or stored into a semiconductor memory by a processor. Memory bandwidth is usually expressed in items of bytes/second, although this will differ for programs with pure information sizes that are not a multiple of the generally used 8-bit bytes. Memory bandwidth that is marketed for a given memory or system is often the utmost theoretical bandwidth. In practice the noticed memory bandwidth will be lower than (and is assured to not exceed) the advertised bandwidth. Quite a lot of laptop benchmarks exist to measure sustained memory bandwidth utilizing a variety of access patterns. These are meant to supply perception into the memory bandwidth that a system should sustain on varied classes of actual purposes. 1. The bcopy convention: counts the quantity of knowledge copied from one location in memory to another location per unit time. For example, copying 1 million bytes from one location in memory to a different location in memory in a single second could be counted as 1 million bytes per second.



The bcopy convention is self-consistent, however will not be simply prolonged to cowl circumstances with extra complex access patterns, for instance three reads and one write. 2. The Stream convention: sums the amount of information that the application code explicitly reads plus the quantity of knowledge that the applying code explicitly writes. Utilizing the previous 1 million byte copy example, the STREAM bandwidth could be counted as 1 million bytes read plus 1 million bytes written in one second, for a complete of 2 million bytes per second. The STREAM convention is most directly tied to the user code, however may not count all the data site visitors that the hardware is definitely required to carry out. 3. The hardware convention: counts the actual quantity of information learn or written by the hardware, whether the data movement was explicitly requested by the person code or not. Utilizing the same 1 million byte copy example, the hardware bandwidth on pc techniques with a write allocate cache policy would come with an extra 1 million bytes of traffic as a result of the hardware reads the goal array from Memory Wave into cache earlier than performing the shops.



This provides a total of 3 million bytes per second actually transferred by the hardware. The hardware convention is most immediately tied to the hardware, but could not represent the minimal amount of information traffic required to implement the person's code. Quantity of information transfers per clock: Two, in the case of "double data price" (DDR, DDR2, DDR3, DDR4) Memory Wave. Memory bus (interface) width: Every DDR, DDR2, or DDR3 memory interface is sixty four bits extensive. Number of interfaces: Modern private computer systems sometimes use two memory interfaces (dual-channel mode) for an effective 128-bit bus width. This theoretical most memory bandwidth is referred to as the "burst rate," which will not be sustainable. The naming convention for DDR, DDR2 and DDR3 modules specifies either a maximum speed (e.g., DDR2-800) or a most bandwidth (e.g., PC2-6400). The pace ranking (800) isn't the utmost clock velocity, but twice that (because of the doubled information price).



The specified bandwidth (6400) is the maximum megabytes transferred per second using a 64-bit width. In a twin-channel mode configuration, that is effectively a 128-bit width. Thus, the memory configuration in the example may be simplified as: two DDR2-800 modules running in twin-channel mode. Two Memory Wave Audio interfaces per module is a typical configuration for Laptop system memory, but single-channel configurations are widespread in older, low-finish, or low-power units. Some private computers and most fashionable graphics cards use more than two memory interfaces (e.g., four for Intel's LGA 2011 platform and the NVIDIA GeForce GTX 980). High-efficiency graphics cards working many interfaces in parallel can attain very excessive whole memory bus width (e.g., 384 bits within the NVIDIA GeForce GTX TITAN and 512 bits within the AMD Radeon R9 290X utilizing six and eight 64-bit interfaces respectively). In systems with error-correcting memory (ECC), the extra width of the interfaces (sometimes 72 relatively than 64 bits) is just not counted in bandwidth specifications as a result of the additional bits are unavailable to retailer consumer data. ECC bits are better considered part of the memory hardware fairly than as information stored in that hardware.
reference.com