RAM Speed – What You Need to Know about – Does it matter?
A more comprehensive look at RAM speeds and how they affect gaming, workstation, and every day
For years now, PC enthusiasts have been obsessed with getting the fastest available hardware components for their new PC builds. When it comes to RAM, however, this might not actually be the best way to go.
When buying RAM, you’ll see a large number followed by MHz (Megahertz). This, as you probably already know, represents the speed of the memory you’re looking to purchase. But how does RAM speed affect your computer and the tasks it’s trying to perform?
Is higher speed RAM good for gaming? Is higher RAM frequency (MHz) better in general? And what’s the best RAM speed overall?
These are the sort of questions that gamers and keen PC builders ask themselves regularly and the very reason why we’re going to try to explain what RAM speed does and whether it’s worth the extra money.
VIDEO: Does RAM Speed Matter for Gaming?
RAM Speed Breakdown
To make your life a bit easier, we’ve compiled the essential pieces of information from this article and have laid them out in bullet point format. These are staples that you should consider when purchasing new RAM, especially if you’re looking to increase in-game performance:
- Tight timings > Speeds: The clear pattern in all of our tests were found within the tighter times. By tightening the times, you improve FPS across most of the games exponentially. Well, I say exponentially, you’d probably still be better off putting the extra cash towards a better GPU. But for those that already have a top-of-the-line card, this is certainly a way of squeezing extra performance out of your build.
- The Sweet Spot: This might spark some speculation, but from our results, the best way to maximize in-game performance (from RAM) is to purchase a 3200MHz kit and tighten the timings right down. You’ll save a decent amount by opting for 3200MHz over 3600MHz, and once you’ve tuned in the timings, the difference isn’t that wide.
- Not messing with the timings?: If you plan on getting RAM and using it straight out of the box, we recommend getting the cheapest 3000MHz+ kit you can find. All the tests suggested that out-the-box speeds made very little difference (less than 1% in most cases) to in-game performance.
What Is RAM?
Random Access Memory (RAM) is used as a temporary, super-fast data storage for your processor (CPU). When you launch a program, the hard drive sends the relevant data for that program to the RAM, where the processor can access it much quicker than going directly to the hard drive.
How Does RAM Work?
Random Access Memory is an integrated circuit chip made from millions of transistors and capacitors. Each pair of transistors and capacitors make up a cell, and these cells are where the data is stored.
These cells hold and release electrical charges to write, rewrite, and erase data. This allows data to regularly change much faster than your traditional mechanical drive (HDD) that uses platters and actuator arms. Even SSD’s can’t compete with the speeds RAM provides.
RAM is also volatile, meaning that any data that is held within its cells will be lost when it loses power. This is why we don’t permanently write data to our RAM modules.
The Random Access part of Random Access Memory comes from the fact that data can be written to any cell in any order, and can be read from any cell as long as the physical location of the cell containing the data is known.
What Does RAM Speed Do?
So, what does RAM speed do? The speed of RAM is a tricky subject and one that actually comes down to several different factors, not just the MHz clock speed. To get a real sense of how fast our RAM is, we need to take into consideration the CAS latency and clock speed simultaneously.
RAM Frequency (MHz)
RAM frequency works off of clock cycles (people often call this the RAM speed even though it is only part of the speed equation). Each read and write is done on a cycle. RAM is measured by how many cycles per second it can perform. For example, if RAM is rated at 3200 MHz, it performs 3.2 billion cycles per second.
The more cycles your RAM can perform per second translates to how much data can be stored and read – making for smoother user experiences. There’s usually a direct correlation between higher DDR (Double Data Rate) RAM and the clock cycles those modules accommodate. This can easily be seen when comparing DDR3, DDR4, and now DDR5 (Due to release in 2021).
Column Access Strobe (CAS) latency, or CL, is the delay time of your RAM receiving a command and then being able to issue it. The numbers for the timings will look something like this; 15-17-17-35. Those numbers indicate how many clock cycles it takes for the RAM to respond to the command. However, faster MHz RAM with slower CAS timings might be slower than lower MHz RAM with faster CAS timings.
This is where it can be a little confusing, but let’s try and simplify the data a little more.
Finding Your Real RAM Speed
Finding the real speed of your RAM comes down to a couple of factors, and has been embodied in a handy little formula. However, before we show you the formula, let’s understand the different factors to consider when finding your RAM speed.
The first factor is the RAM clock speed; let’s take 3200Mhz as an example.
The next factor relates to what RAM you have. For example, DDR (Double Data Rate) transfers data on both sides of the clock cycle. Meaning, it transfers data twice in one cycle.
And lastly, your RAM CL Timing, which is the first of the four numbers that make up the RAM timing. We’ll use (14-15-15-39) for this example.
Those are the factors, so let’s put them into an equation to work out the exact RAM speed.
You can use this formula to work out exactly how fast any RAM kit is – a great little tool to have in your locker.
By default, the maximum stock clock speed for DDR4 RAM is 2400 MHz. When you see RAM with speeds rated over this, it means the module has been overclocked to that speed by the manufacturer.
To access the RAM’s boosted clock speed, you must first locate and enable the RAM’s XMP/DOCP profile.
An XMP, or extreme memory profile, is a separate module that must be enabled within the motherboard’s BIOS to boost the RAM to its advertised speed. If you do not select the RAM’s XMP/DOCP in the BIOS, your RAM will be running at a slower speed of 2400MHz.
It’s always a good idea to make sure all of your hardware is compatible with each other as well.
Do I Need The Fastest RAM Speed?
For the majority of us, the answer is no. Alternatively, there are sweet spots to consider that largely depend on the CPU in your build. We found that RAM kits with tighter timings can improve FPS in some games and while this improvement isn’t quite as good as a GPU upgrade, it is still worth noting. Regardless, if you’ve gone for AMD or Intel, RAM kits in the 3200MHz to 3600MHz will serve you well.
In regards to the hugely popular Ryzen 5000 series processors, it appears that 4000MHz RAM may well be worth your investment. That being said, you can expect RAM of that speed to add significant costs to your overall budget, especially if you plan on grabbing a kit with tight timings.
Does RAM speed matter? Well, we are going to explore that theory in a bit more detail later on – making us of some in-house benchmarking.
More RAM Or Faster RAM?
If you are in a position where you are unsure to go for 8GB of faster RAM or 16GB, you should always opt for the 16GB – especially if you are a price-conscious gamer. Sure, if you can afford it then grabbing a 16GB+ kit with fast speeds will have obvious benefits but RAM speed is only really relevant at the high-end, where small performance gains are important.
The fact is, for most of us gamers, we are barely going to notice a difference – and 16GB of RAM is going to give you a much better gaming experience in the long run.
Can I Mix RAM Speeds?
The short answer is yes, but only sometimes.
RAM itself is compatible with other RAM of different timings and speeds. This being said, the greater the gap between the two speeds, the greater the stress your motherboard is going to incur trying to run them simultaneously.
For example, two modules of RAM with the same speed (let’s say 2400 MHz) and slightly different CAS timings probably wouldn’t be an issue. Your motherboard would pick the slowest one and run them both at those speeds.
The further away you get in speeds and timings, the harder your motherboard is going to have to work to run them simultaneously. Furthermore, it may require manual control over the speeds and timings in the form of over or underclocking the modules to ensure stability.
In other words, you could probably do it, but it would definitely be a bigger headache than it is worth.
Bandwidth & Capacity
RAM speed doesn’t exist in isolation; you also need to consider bandwidth and capacity when making a purchasing decision. Because of the number of factors, RAM bandwidth (and efficacy as a whole) can be a somewhat complex topic that’s deserving of its own article – something we might do in the near future.
As a concept, it’s relatively simple: bandwidth affects the amount of data that can get through at any given time. And of course, memory is the space.
Some like to explain RAM as being like a highway. The bandwidth is like the number of lanes on the road; more lanes mean more cars that can go through. The speed limit represents the literal speed, and memory is like a big parking garage that all the cars are going in and out of – except in this scenario, the cars entering and leaving help you play video games.
Putting it simply, if you don’t have enough parking spaces, if the speed limit is too low, or if there aren’t enough lanes on the highway then everything gets bogged down – and if you’re following my riveting civil engineering metaphor, you understand that this is not great news for the performance of your system.
AMD APUs And Fast RAM
As many will know, the new batch of AMD’s processors tend to perform better under higher speed RAM. With that in mind, we thought it would be suitable to touch upon how fast RAM boosts the performance of the new line of processors that have become such a huge hit amongst both gamers and productivity users.
Let’s start with their APU range.
An APU, or accelerated processing unit, is AMD’s answer to a processor with a powerful, built-in graphics card. AMD & Intel have both been designing some form of APU for years now, allowing customers to get a slice of both pies at a hugely reduced price. But how does an APU utilize faster RAM?
First, let’s understand how an APU works. Unlike a PC that comes with a freestanding GPU, the APU has the graphics processor integrated into the CPU.
One of the benefits of having a standalone GPU is that it comes with its own super quick onboard VRAM to use when processing graphics. The APU, on the other hand, has to utilize your computer’s system RAM – RAM that is already being utilized for other processes.
So, what’s our point? Well, faster RAM, and more of it, is extremely beneficial to your APU system – faster RAM like the kind you would find onboard a GPU.
With AMD’s CPU range, the story doesn’t really change.
Since Ryzen was released, AMD has been making use of its impressive multi-threaded nature.
Multi-threading is when a CPU can double its cores by making use of virtual cores. If your CPU has six cores, it would have 12 threads, and so on. This being said, each thread will try to access your RAM, meaning quicker RAM will be more suitable to accommodate a multi-threaded CPU over say, an Intel CPU that doesn’t have this facility.
We’re currently in the process of putting together some benchmarking results of AMD vs. Intel and how faster RAM affects them in real-world situations. Stay tuned to see those results soon.
Does RAM Speed Matter?
Alright, so we’ve figured out what RAM speeds are, how they work, and even how they affect your system. But how does RAM speed impact gaming? Is it worth buying RAM that’s rated a little higher to try and get those higher FPS counts?
Does RAM speed matter is a question the WePC team are asked all the time (and one that I’ve asked myself numerous times as well). For that reason, we decided to do some independent research.
Here’s the system I’ll be using for the tests:
- ASUS TUF X570 Gaming PLUS WIFI
- AMD Ryzen 9 3900X
- EVGA GTX 1080Ti SC Black Edition
- 16 G.Skill Trident Z RGB (CL16)
- Corsair Vengeance LPX 3200MHz (CL16)
- Corsair Vengeance LPX 3000MHz (CL15)
- Samsung 970 EVO 512GB M.2 SSD
Using this setup, I’ll be running six separate tests using eight different speeds/timing combinations:
- The G.Skill will be used to test “2666MHz CL11” “3600MHz CL16” “3600MHz CL14” “3800MHz CL16“
- We will use Corsair’s Vengeance LPX 3200MHz module for “3200MHz CL16” “3200MHz CL14“
- And the Vengeance LPX 3000MHz for “2666MHz CL16” “3000MHz CL15“
Testing a range of RAM capacities, timings, and speeds will help give us a larger pool of information for both high-end and budget builds alike.
For benchmarking, I’ll be using the Time Spy demo from 3DMark to test both CPU and GPU and track FPS. I’ll be running the same test for each RAM configuration.
We’ll also be accumulating in-game FPS from a number of different games which range from:
- Counter-Strike: Global Offensive
- Far Cry 5
- The Witcher 3
- Total War: Warhammer II [Battle]
- Forza Horizon 4
Counter-Strike: Global Offensive
The first game we decided to run was CS:GO, and not because it’s my favourite competitive esports title either.
We went for CS as we knew it was a CPU-intensive game and thought we’d see a variance in FPS across the different speeds and timings. However, as you’ll soon see from the results, there was almost no difference across the entire range of tests we ran.
Stay tuned for further tests in this particular game
Far Cry 5
The next game we ran was Far Cry 5. We were much happier with the results of this test due to the fact we actually saw a difference.
We decided to do a second test in ultra settings to see if the difference was greater than our original in-game settings.
The Witcher 3
Like Far Cry 5, we decided to run a second test in ultra settings for The Witcher 3. We used the same two RAM configurations as Far Cry, and to no one’s surprise, the results were as expected.
Total War: Warhammer II [Battle]
Forza Horizon 4
3DMark Time Spy
We finished by running 3DMark Time Spy to see if different configurations had any difference here.
What Speed RAM Should I Get?
So, on to the big question surrounding RAM speeds – which speed should I get?
Well, I suppose that depends on a number of different factors. As far as gaming is concerned, the most important component that affects frame rates is the graphics card. Ultimately, this will determine whether or not you’re going to get 140+ FPS in AAA game titles. It’ll also determine what resolution and settings you’ll be able to push as well. Having said that, as you can see from the above data, getting the right RAM kit can make a huge difference.
|2666MHz CL16||2666MHz CL11|
|Far Cry 5||118||87||159||125||94||84|
|The Witcher 3||136||96||78||157||111||92|
|Total War: Warhammer II [Battle]||134||87||76||140||94||82|
|Forza Horizon 4||153||-||-||156||-||-|
|3DMark Time Spy||9759||10138|
|3000MHz CL15||3200MHz CL16|
|Far Cry 5||125||94||85||127||98||89|
|The Witcher 3||149||105||86||150||104||85|
|Total War: Warhammer II [Battle]||140||92||82||141||93||85|
|Forza Horizon 4||158||-||-||158||-||-|
|3DMark Time Spy||10033||9955|
|3200MHz CL14||3600MHz CL16|
|Far Cry 5||136||104||96||126||95||87|
|The Witcher 3||168||118||93||153||109||85|
|Total War: Warhammer II [Battle]||146||101||90||143||96||86|
|Forza Horizon 4||162||-||-||157||-||-|
|3DMark Time Spy||10156||10026|
|3600MHz CL14||3800MHz CL16|
|Far Cry 5||141||111||99||135||105||94|
|The Witcher 3||178||124||97||170||121||94|
|Total War: Warhammer II [Battle]||149||105||95||147||103||90|
|Forza Horizon 4||162||-||-||160||-||-|
|3DMark Time Spy||9901||10138|
|3600MHz CL14||3200MHz CL16|
|Far Cry 5 [ULTRA]||130||102||92||117||90||84|
|The Witcher 3 [ULTRA]||162||108||77||145||101||72|
Before we talk about the differences, let’s quickly brush over where RAM doesn’t seem to make a difference in performance. The 3DMark Time Spy test seemed to showcase very little to no effect when under different RAM speeds or timings. Furthermore, CS:GO was also unaffected by differing configurations.
Where you will see RAM speeds benefit you, however, is in the newer titles from the last few years. Every other game we tested showcased the same increase in performance across the board. Interestingly, we saw a dip in performance when reaching 3800MHz. We’ll discuss this in more detail shortly, but the bottom line is: tighter timings are better than faster speeds.
It’s also worth mentioning that better speed and timing configurations are going to make multi-tasking scenarios much more accessible – think streaming and rendering.
Check out our buying guides on Best DDR3 & DDR4 RAM.
What Is a Good RAM Speed?
What’s considered a good RAM speed is determined by the nature of your build and what you want your computer to be capable of. If you’ve put together a high spec system with a top of the range CPU, GPU, and monitor, you’ll need to match them with high performance RAM to ensure you’re using your build to its full potential.
For a high-performance, dual-channel setup, we’d recommend two DDR4 sticks with a minimum combined capacity of 32GB, each with speeds of 3000MHz. You can buy even better RAM than that if you want, but this is all you really need for epic gaming and seamless streaming and content creation.
If you’re running hardware that you didn’t have to remortgage your house and sell your car to afford, you don’t need such impressive RAM. For mid-spec modern gaming, you can settle for an 8GB DDR4 stick or two 4GB DDR4 sticks with speeds between 2400 and 2666Mhz.
Does RAM Speed Matter?
RAM speed does have an impact on frames per second, especially in 1080p, but it’s not as essential as you may think. RAM capacity is often considered more important on the whole.
The reason RAM matters less the higher up the resolution ladder you go is that the workload is shifted over to the integrated memory in your graphics card. RAM still supports a smooth gaming experience, but the onus is largely removed.
Let’s take Metro Exodus, for example. Average frames in 4K on Ultra settings using a DDR4 2400 stick are 28fps with lows of 18fps. In the same scenario only with a 3600MHz stick, the average remains 28fps, but the lows go up by 1 frame to 19fps – Not exactly earth-shattering, is it? Now let’s discuss the same game, the same sticks, but in 1080p on low settings. The 2400MHz stick musters averages of 182fps with lows of 104, while the 3600MHz stick storms the show with averages of 207fps and lows of 120. That’s over 20fps boosted averages for the faster card.
Now, if you’re looking to match frame rates up with a particularly powerful monitor with a high refresh rate, that’s when super speedy RAM can really help you out.
Can I Mix RAM Speeds?
Yes, you can absolutely mix RAM speeds in a dual-channel setup. Hell, you can mix it up a bit with RAM capacity or even bring different brands into the fold. That said, there are a few things to consider before you try and spice up those expansion slots with some variety.
RAM is at its finest when the sticks match up, and it’s not like, say, a food or cosmetics company printing copy on their products that reads something along the lines of ‘goes best with our jerk sauce’ or ‘ for best results, use with our face cream’. Identical RAM actually does work better.
Another thing you should bear in mind is that, in a dual-memory system, your computer will run both RAM sticks to the spec of the weaker one. So, if you have one 4GB 2400MHz stick and one 8GB 3200MHz stick, performance will be equivalent to two 4GB 2400MHz sticks.
Is Higher MHz RAM Better?
Generally speaking, yes, the higher the MHz of your Ram, the better your computer will perform, but only if it blends well alongside all your other hardware. It’s all about balance. Hooking up an entry-level build with 32GB, 4000 MHz DDR4 RAM doesn’t make much sense, besides, lower spec motherboards will be capped at a certain speed and capacity of RAM anyway. Choosing the best RAM that your system can support, on the other hand, will indeed make for a faster machine.
Is it Bad to Overclock Your RAM?
Overclocking RAM isn’t bad at all. In fact, as most motherboards won’t pass POST if your RAM has been calibrated in a dangerous way, it’s one of the safest bits of hardware to tinker with. You can then just tweak the settings until your system passes the self-test.
Overclocking your RAM can lead to bolstered memory speeds and an overall more responsive performance, but as with any overclocking, quality thermal solutions and suitable power supplies are essential.
So we’ve learned how RAM works, what the speeds tell us, and we’ve even looked at some benchmarks to find out exactly how it affects your system when it comes to graphical tasks vs. processor tasks.
I’m currently running 16GB of 3600MHz RAM for my personal gaming computer, and it never misses a beat. Whether I’m playing one of the latest AAA game titles or doing some fairly intensive video editing, 16GB is more than enough to keep my system running at optimal levels.
Lets us know what RAM speed you’re currently running and if this article helped in any way. Better yet, head on over to our Community Hub, where you can start a thread and discuss everything RAM related.
WePC is reader-supported. When you buy through links on our site, we may earn an affiliate commission. Learn more
15 thoughts on “RAM Speed – What You Need to Know about – Does it matter?”
I already tried to post this but it didn’t work. If that’s regular time spy your scores are way low. Like 50% of what I get with the same CPU, GPU and 3600 CL16 ram. Both the GPU and CPU are capable of scoring 10k in time spy and that’s thousands away from record high for that hardware.
What time spy are you running to test this? Is it time spy extreme or something? The reason I ask is because I have a ryzen 7 3800x and a 5700xt with 2 x 8 3600 CL16 ram and I’m scoring 9800 without even trying or just over 10000 with a little effort. Both my CPU and GPU are capable of scoring 10k on their own and I’m definitely not alone there. Just comparing the results online I see people doing way better on the same hardware too. I have a better motherboard but it shouldn’t make double the performance. Just saying it was a great article but if that’s all you’re getting out of that hardware then you should probably be putting in an RMA. You can look on the 3DMark sore and search results by hardware to check it out yourself.
I think RAM speed does matter more than the article states. Your testing method could be improved a bit. The fact of the matter is that we need a bottleneck hierarchy to be stated. You are correct that the GPU is the one doing most of the work in these graphically intensive games, and this is the point I want to bring up. When the GPU is able to run the settings you are trying to play with without breaking a sweat anymore, then it can go ahead and use the rest of the GPU resources to try and pump in a few more FPS. (I think in your test, the RX 580’s resources were already being fully utilized and it makes sense that having faster RAM or more ram wouldn’t make a difference.) In order to do this, it seeks help from the CPU. Since there are GPU resources available, the CPU (which is faster than the RAM) assists in the process of pumping in more FPS. The thing is that RAM needs to send this data to the CPU. If the CPU and GPU are “waiting” too long and spending several clock cycles not doing anything productive to increase the FPS, it is because they are “waiting” for the RAM to bring the resources to the CPU. In your case, the GPU’s resources were already being fully taken advantage of. So no matter what, the different RAM kits you used did not make any impact on the FPS because there was no “juice” available left to draw from the GPU. Had you used a more superior GPU and CPU (I say CPU because the single core / single thread score is not very good on Ryzen’s which is required for a higher FPS output), then yes, faster memory would have allowed for a higher FPS output. However, I will mention that your system is more or less a perfect case of no bottleneck. The only disadvantage is that in terms of gaming, upgrading anyone of the components to increase your gaming FPS will provide diminishing returns. If you upgrade the GPU, your CPU will be too slow to take advantage of it and will barely give better FPS and actually lead to stuttering. If you upgrade the CPU, you won’t get better FPS because your RX 580 was already being used to its full potential. If you upgrade the memory, same scenario, you won’t get better FPS because your RX 580 is already being used to its full potential.
RX 580? Blind or some?
WePc, I’m building my own pc right now this is a really well define article love it.
Thanks for that info hope to find the best ram suitable for my needs.
Do some testing with ACTUAL GAMES. 3dmark is basically perfectly optimized, something that you’ll never encounter in the real world. In some games you’ll see up to a 10% increase in avg fps going from 2133MHz to 4000MHz and even in games that don’t gain much avg fps the frametime stability improves.
Thanks very much. Valuable information. I made a mistake buying a double processor motherboard a few years back. I need faster but fewer cores I’ve come to experience. Wanted to check what the effect and importance is of faster memory. It seems that it’s not that crazily important. That’s then saving my a few pop. Let me know where to donate 🙂
Thanks for explaining what memory bandwidth means better than any other article I could find. I’m an artist/filmmaker building a workstation for animation in both 3D (Maya) and 2D (After Effects) as well as editing (4k or less footage).
Today I was trying to decide whether or not to buy one of the latest gen Ryzen processors vs. the 2nd gen version of the same processor, and it concerns me a lot that they seem to have decreased the memory bandwidth for writes by half on the new 3rd gen Ryzens. (source: https://www.techspot.com/review/1869-amd-ryzen-3900x-ryzen-3700x/#add-new-comment) Do you have any thoughts on how this might impact an animation workflow, where new versions of clip are being constantly cached/reloaded into RAM as the artist works? It’s possible I’m not completely understanding how animation utilizes RAM, as all this is really over my head as an artist trying to build a PC. I’m just trying to figure out whether these new processors would actually slow down my workflow as opposed to speed it up?
You have a typo in your “Clock cycles” section: “if RAM is rated at 3200 MHz, it performs 3200 cycles per second.”. You’re leaving out the “MHz” of the second part. 3200 cycles per second is 3.2 kHz, or 0.0032 MHz.
Please take a look not only at FPS but 1% and 0.1% frametime lows. Please use more than Time Spy, such as a CPU-heavy game title (heck, even OG Crysis can batter a modern CPU, although GPU not so much).
I would love to see a memory bandwidth/latency comparison with regards to stutters and framepacing in CPU heavy games.
Isn’t 1 Hertz (1 – 1 millionth of 1000 Mhz) defined as 1 clock per second and by that default making 3200MHz (effective speed of 1700Mhz) 1.7 million cycles per second not 3200?
DDR stands for double data rate, meaning you clock data on the rising edge and falling edge of the clock signal. I haven’t looked into your question much but if you are only thinking that 1 clock cycle is 1 data out then you need to double your results to account for DDR mode
There is sometimes big difference betwen for example 1×8 vs 2×4 gb because second option uses dual channel
Nice article, thanx for the read.
You should test the various RAM configs, going from 2Ghz-3.6Ghz with games. IE BF, GTA5, Doom, Witcher, Mordor etc. This will help those with newer systems make a good choice.
Also older systems, IE 2nd-4th gen i7’s befit HUGELY from 2133-2400Mhz RAM VS 1600-1866Mhz RAM in games, esp when a 1060-1080Ti is used. As GPU efficiency increases, so the system needs faster RAM to be able to middleman the FPS to the CPU.
Comments are closed.