What are timings and how do they affect the speed of RAM

The choice of RAM in a game assembly can turn into a nightmare if you start to understand the intricacies of its work. The requirements of modern gaming and work tasks dictate their own conditions, so memory is now almost the most important and difficult part in assembling a computer. Among the many models, you need to choose the only suitable option and this is scary. And the most difficult thing in this is why the memory with a lower frequency works faster and shows more frames in games than the one with a higher frequency. To do this, you need to figure out how the memory speed is measured and what parameters affect it.

The power of a computer is measured by FLOPS, which indicates the number of computational operations per second. Due to the fact that computers can simultaneously perform millions of operations, the prefix “giga” is added to the flops.

In a familiar environment, we can confuse power and frequency, so we consider the performance of computers not as gigaflops, but as the maximum operating frequency. This is easier in ordinary situations, when the speakers know the topic well and automatically correlate power with hertz in their heads.

At the same time, such linguistic simplification makes adjustments to the understanding of the practical part of the issue. Taking the context out of the forums, the average user really thinks that the power of memory can be expressed in hertz. Simply because the race for frequency has become a trend among hobbyists and enthusiasts alike. This prevents an inexperienced person from understanding why his high-frequency processor can lose to one with a few hundred hertz less. It’s simple – one has two cores and four threads, while the other has four real ones. And that’s a big difference.

RAM and its speed
Random access memory consists of thousands of elements interconnected in chips-microcircuits. They are called banks, which store rows and columns with an electric charge. The electric charge itself is information (pictures, programs, text in the clipboard, and much more). As soon as the system needs data, the bank gives a charge and waits for a command to fill in with new data. This process is directed by the memory controller.

1Z0-1083-21 Dumps PDF

1Z0-1080-21 pdf Dumps

1Z0-1080-21 Exam Dumps

1Z0-1080-21 Practice Questions

CDPSE

NSE4_FGT-7.0

NSE4_FGT-7.0

NSE4_FGT-7.0

NSE4_FGT-7.0

HPE6-A66 exam dumps

HPE6-A66 questions

2V0-51.21 dumps

1Z0-1104-21 exam

1Z0-1104-21 practice

CCDAK practice dumps

For an analogy, let’s compare the work of RAM and the work of a cafe. Chips can be represented in the form of decanters with tomato juice. Each is filled with juice and pulp of ripe tomatoes (electrical charge, information). A client (computer user) comes to the cafe and orders juice (starts the game ). The bartender (the controller, the one who manages the cans) accepts the order, goes to the kitchen (asks for information from the cans), pours the juice (takes the game files) and brings it to the guest, and then returns and fills the decanter with new juice (new information about what he launched user). So ad infinitum.

Timings – quality
The work of memory, contrary to the stereotype, is measured not only by hertz. Memory speed is usually measured in nanoseconds. All memory elements work in nanoseconds. The more often they are discharged and charged, the faster the user receives information. The time for which banks must complete the tasks was called in one word – timing (timing – calculation of time, terms). The fewer ticks (seconds) in timing, the faster the banks work.

So you. If we need to climb a ladder with 100 steps to the top, we will take 100 steps. If we need to get to the top faster, we can go through a step. This is already twice as fast. Or you can take two steps. It will be three times faster. Each person has their own speed limit. As well as for chips – some allow to reduce timings, some do not.

Frequency – quantity
Now for the memory frequency. In RAM operation, the frequency does not affect the time, but the amount of information that the controller can steal in one go. For example, a customer comes to a cafe again and demands tomato juice, as well as whiskey on ice and a milkshake. The bartender can bring first one drink, then a second, then a third. The client does not want to wait. Then the bartender will have to carry everything at once in one go. If he has no coordination problems, he will put all three drinks on the tray and fulfill the demand of the whimsical client.

The memory frequency works in a similar way: it increases the width of the data channel and allows you to receive or send a larger amount of information in one go.

Timings plus frequency – speed
Accordingly, the frequency and timings are interconnected and set the overall speed of the RAM. In order not to get confused in complex formulas, we will present the work of the frequency / timings tandem as a graphical example:

Let’s analyze the scheme. The shopping center has two technical departments. One sells video cards, the other sells game consoles. The lack of gaming technology has driven customers crazy, and they are ready to buy a video card or console just to play the new Assassin’s Creed. The trading conditions are as follows: the waiting area in the department of the first seller allows serving only one client at a time, and the second can accommodate two at once. But the first has a warehouse with video cards two times closer than the second one with consoles. Therefore, he brings the goods faster than the second. However, the second seller will serve two customers at once, although he will have to go twice as far to get the goods. In this case, the speed of work of both will be the same. Now let’s imagine that the warehouse with attachments is at the same distance as the first one with video cards. Now the console seller will start working twice as fast as the first one and take most of the profits for himself. And, the closer the warehouse and the more customers in the department, the faster it makes money.

So, we understand how the frequency interacts with the timings in the memory speed.

A queue is a user who requests information from RAM.
The salesperson is the memory controller (which delivers information).
Stock technology is information for the user. Counter is the memory bandwidth in hertz (frequency).
Distance to the warehouse – timings (the time it takes for the controller to find information upon request).
Accordingly, the fewer meters the controller travels to the cans with an electric charge, the faster the user receives information. If the memory frequency allows more information to be delivered at the same distance, then the memory speed increases. If the memory frequency drags on the increase in the distance to the banks (high timings), then the overall memory speed will drop.

You can compare the speed of different RAM modules in nanoseconds using the formula: timing * 2000 / memory frequency. So, RAM with a frequency of 3600 and CL14 timings will operate at a speed of 14 * 2000/3600 = 7.8 ns. A 4000 on CL16 will show exactly 8 ns. It turns out that both options are approximately the same in speed, but the second is preferable due to the higher bandwidth. At the same time, if we take memory with a frequency of 4000 at CL14, then it will already be 7 ns. In this case, the throughput will become even higher, and the information delivery time will decrease by 1 ns.

Memory chip structure and timings
In theory, RAM has speeds in nanoseconds and megabytes per second. However, in practice, there are more than a dozen timings, and each sets the time for a certain work in the microcircuit.

They are divided into primary, secondary and tertiary. Basically, a group of primary timings is used for marketing purposes. They can be found in the characteristics of the modules. For instance: