FLOPS  What are floating point operations per second? ▷ 2020

Floating point operations per second, also known as FLOPS, are a way to measure the performance of a computer or graphics card.. This measurement mainly indicates the number of operations of this type that the hardware processor can solve in one second by mixing small, large and even fractional numbers.

The FLOPS is a very small measure, so developers currently refer to TeraFLOPS to give the user the data. Thus, the higher this value, the better the computational power of the devices.

FLOPS have become a new standard that can be crucial when buying a computing device or game console. That is why in this post we are going to explain what they are, in what units they are measured and how we can know how many our device has.

What are the FLOPS of my processor and GPU?

As we said at the beginning, FLOPS are the floating point operations that a CPU or GPU can perform per second. They represent very large or very small numbers and account for the number of realizable operations. Current teams already use the unit Teraflops to be able to do trillions of calculations per second.

This unit, in its beginnings, was used to design the operations that supercomputers could do. Currently, this type of device already uses the pentaflops that correspond to 1000 trillion operations per second.

In any case, although this measure is important, When purchasing a computer we must also take into account the frequencies of the processor or the amount of RAM and ROM it has. A good set of all these elements will be the one that results in a powerful team, but if it exceeds any of the values ​​too much without being in balance with the rest, it will be useless.

This concept also affects video game consoles, so that the more teraflops they have, the better virtual universes they can offer to video game consumers. But, of course, and as with computers, this is not entirely true, since in the case of game consoles, drivers and graphics optimization are also important.

Flops types, measures and units

Of course, we come up with different types and units to refer to this concept, so that we can always speak properly, being more accurate and using concepts that adapt to the new capabilities that technology, always in constant evolution, is capable of. offer us.

Types

There are currently two types of FLOPS:

  • Sustained FLOPS: Used to calculate the performance of a device based on the number of flops that contains. In any case, it must be remembered that it is not a determining characteristic to acquire or not a team.
  • Current FLOPS: It is about getting a measurement of the number of operations performed by a team at the moment in which it is being used.

Measurements and Units

flop units

The most used measures and that correspond to units of the flops are the following:

  • Kiloflops = Thousand flops
  • Megaflops = A million flops
  • Gigaflops (GFlops) = One billion flops
  • Teraflops (TFlops) = One trillion flops = One thousand gigaflops
  • Petaflops (PFlops) = One thousand trillion flops = One thousand teraflops = One million gigaflops
  • Exaflops = A trillion flops
  • Zetaflops = A thousand trillion flops
  • Yotaflops = A quadrillion flops

As graphic examples, it can be said that a card Nvidia GeForce GTX 1080 it has a throughput of 9 teraflops or an i7-6700K processor has roughly 1.2 teraflops. In the case of petaflops, an example of a device that uses them is the Tianhe-2 supercomputer. This has 33 petaflops, although the most powerful computer in the world currently has 125 petaflops.

The concept used to measure capabilities in video games in teraflops, where, for example, the Tegra X1 processor on the Nintendo switch has a performance of 1 teraflop. The XBox One has 1,3115 Teraflops, the PlayStation 4 has 1,840 Teraflops, and the PlayStation 4 Pro has a total of 4.2 Teraflops.

Thus, we see that it is teraflop that is used as a benchmark for game developers, so that they can know how much computational power they can count on for each platform.

How can I know and calculate how many TeraFlops my computer has?

In order to locate how many teraflops or gigaflops has the device that interests us, we would only have to access Google and search for its name followed by the word flops. Although it is true that also There are several websites that you can access in order to know the data.

Some are so complete that they show you the information even in the form of a graph about the performance of the device you are interested in. It is likely that if you browse the manufacturer’s website (or even if you send them an email to ask them) you can locate this completely updated information because, as we have said, it is a value that the user is increasingly interested in and that, therefore, the manufacturer will be happy to make known.

The information provided by the brands will consist of the Gigaflops in reference to hardware with stock speed. For this reason, we are going to be left without knowing the value when it is performing (current flops), which is called overlock.

This is logical, because they should indicate the value in reference to the combination of the other components with which it works and this is not very operational at the same time that it is totally changing.

It is possible that you are not interested in doing this search for each device and what you want is to know if there is any simple operation with which you can calculate these values. Indeed, yes it exists and it is a mathematical formula that will give result in Mflops.

In order to use it we have to have two data for each device that we want to know. On the one hand, there would be the shaders and on the other we would need the data of its clock speed. The formula is very easy. We must multiply both quantities with each other and once we have that result, we multiply it again by 2.

MFLOPS = Shaders * Clock * 2
formula know teraflops

This operation will give us the Megaflops that the device has, but, as the data that indicates the most reality is the Teraflop, we would have to divide the resulting quantity by a thousand and thus we would change the unit.

If you have any questions, leave them in the comments, we will answer you as soon as possible, and it will surely be of great help to more members of the community. Thank you! 😉

You may be interested:

Rate this post

Leave a Comment