Conversion from Mebibits per second to Teraoctets per second

Information about the Mebibit per second (Mibit/s)

The mebibit per second is a unit of data transfer rate equal to 1,048,576 bits per second. Its symbol is Mibit/s. It is the binary counterpart of the megabit per second, and the two differ by 4.9 per cent — small, but large enough to change a headline figure.

That five per cent is where the unit starts to matter. A connection advertised at 100 megabits per second and a connection at 100 mebibits per second differ by about five megabits, which is a whole high-definition video stream. In a specification, a contract or a benchmark report, quoting one and delivering the other is a real discrepancy rather than a rounding difference.

In practice the confusion is rare in networking, because network equipment is always specified in decimal. What produces mebibit figures is measurement software: a tool that counts a transfer in mebioctets and divides by elapsed seconds is reporting a binary rate, and multiplying by eight makes it mebibits per second. A person comparing that reading with an advertised rate must convert twice, once for the base and once for the factor of eight.

Where a genuine binary rate does arise is inside a machine. A memory bus transfers a fixed number of bits per clock cycle, and the width is a power of two — sixty-four bits at a time, for instance — so the quantity moved per cycle is binary even though the clock frequency is not. Rates derived from such a structure are naturally expressed with binary prefixes.

For scale, a mebibit per second is 131,072 octets per second, or about 128 kibioctets per second. That is roughly a photograph every two seconds, or a plain-text novel every four. It is a rate at which the modern web is slow but usable, which puts it in the range that mobile networks fall to when congested.

The correct symbol has the lowercase i, and its presence is the only reliable way to tell the two conventions apart. A document writing Mbit/s in a context where the underlying figure came from a binary computation has misstated its own measurement by five per cent.

One mebibit per second equals 1,048,576 bits per second, 131,072 octets per second, or about 1.049 megabits per second.

Information about the Teraoctet per second (To/s)

The teraoctet per second is a unit of data transfer rate equal to a thousand gigaoctets per second, or eight terabits per second. Its symbol is To/s. It describes memory bandwidth inside the fastest processors and the aggregate throughput of large computing systems, rather than any link between separate machines.

Graphics and accelerator chips are the clearest example. A modern accelerator uses stacked memory placed alongside the processor on the same package, and the bandwidth between them reaches several teraoctets per second. That figure is what allows thousands of arithmetic units to be kept busy at once, and it is now the property that most often determines how fast a machine learning workload runs.

The reason bandwidth rather than arithmetic has become the limiting factor is worth stating. Processors have grown far faster at computing than memory has at supplying data, so a modern chip can perform tens of operations for every octet it reads. Any calculation that touches memory more often than that is limited by the memory, and most real calculations are.

Supercomputer interconnects reach this range in aggregate. The network joining thousands of nodes carries teraoctets per second across the whole machine, though no single link does. The design problem is to arrange the topology so that any node can reach any other quickly, which is why these networks are built as multi-dimensional meshes and trees rather than as simple stars.

To make the number concrete, one teraoctet per second would fill a large consumer hard drive in twenty seconds and transfer the entire text of every book in a national library within a minute. Nothing in ordinary use approaches it, and no external cable of any kind carries it.

The unit also appears in descriptions of storage arrays. A large parallel filesystem, spread across thousands of drives, can deliver a few teraoctets per second in aggregate to a supercomputer, and that figure is quoted as a headline specification because it determines how quickly a simulation can be written out and read back.

One teraoctet per second equals 1,000 gigaoctets per second, 8 terabits per second, or about 0.9095 tebioctets per second.