In almost any location on the internet, whether forums, blogs or tech review sites, you can find someone asking the question, “Is my PC good enough to do this?”, and the biggest downside to this question, is that there isn’t just one simple answer.
Throughout this article, I will use the terms “cores”, “modules” and “threads”. For this article’s purpose, “cores” and “modules” will refer to the same thing, “a physical compute unit of a CPU”, whereas “threads” will mean “a stream of instructions”. CPU is an acronym for “Central Processing Unit” and is effectively the brain of your PC; also referred to as “the processor”. Using the term “CPU” does not mean the same as using the term “the computer”. Prices mentioned may vary from locale to locale.
To start off, consider two lower-mid end, current generation, CPUs from Intel and AMD: (Similarly priced, within ~$20 of each other)
Intel i3-4130 @ 3.4GHz / AMD Fx-6350 @ 3.9GHz
Generally speaking, when it comes to the battle between Intel or AMD, you have to consider a few things: single- vs. multi-core performance, the price/performance ratio and the overclocking ability of the CPU.
Single- vs. Multi-core
Most games nowadays utilize at most, 2 cores of the CPU. When comparing Intel vs AMD with this info, it would make more sense to use an Intel CPU over an AMD CPU, would it not? Not necessarily.
With just performance while gaming, the i3-4130 would likely be sufficient, but if you wish to perform numerous other tasks while gaming, one being streaming the game, then this is where the FX-6350 could be very handy to have. The AMD FX-6350 has 6 physical cores for computation on the unit, whereas the Intel i3-4130 only has 2 physical cores.
When a CPU has more cores, you would think that it would clearly have the advantage, because more cores means more available resources, and so it should be able to perform more tasks at once, or split singular tasks between cores to execute them quicker. Sadly, this isn’t up to the CPU to determine how many cores or threads are used by the programs, but by the developers who write the programs themselves.
When you think of how a CPU functions, it receives a set of instructions, executes them, and opens itself for the next set to continue the cycle, over and over again; the faster a CPU can execute these instructions, the more overall power the CPU is said to have. The biggest difference in how CPUs execute these instructions when comparing Intel to AMD CPUs, is that Intel utilizes threads in a slightly different way than AMD, with something they call “hyperthreading”.
Hyperthreading vs. Non-hyperthreading
Traditionally, a single core on a CPU could execute a single thread or stream of instructions at a time, and you would need more cores to be able to perform more tasks simultaneously; this is the route AMD is on with their currently manufactured CPUs. Intel on the other hand has implemented hyperthreading with almost all of their current gen (Haswell) CPUs, which allows a single core on the CPU to execute 2 threads simultaneously, effectively allowing a single core act as 2 cores. Note though, that not all CPUs previous to Intel’s Haswell generation have hyperthreading enabled, this was generally reserved for the i7 models only.
Hypertheading is essentially virtualization of a secondary core within each physical core of Intel’s CPUs, meaning the dual-core i3-4130 can act similarly to a quad-core CPU, but it does not match the full performance of a dedicated quad-core CPU (looking upwards to the i5 and i7 series).
What AMD does to counter Intel’s hyperthreading, is use more physical computations units instead; the FX-6350 utilizes 6 physical cores. These cores tend to work in tandem when performing many tasks, which is also why many people refer to them as modules and not explicitly as cores.
Due to the way Intel CPUs utilize hypertheading, this generally gives them greater single-core performance compared to AMD CPUs, but because AMD CPUs have more overall instruction execution lanes, AMD CPUs are generally more successful at multi-core applications. In most everyday applications, like web browsing, instant messaging through Skype/IRC/etc., or simply watching a video, the performance of these two CPU brands will be nearly the same.
When you look up a review of a specific CPU on tech sites, such as Arstechnica, Anandtech, Tom’s Hardware, bit-tech or xbit labs, they generally give a price/performance comparison of that CPU compared to other CPUs, and purchasing either a pre-built PC or purchasing parts for your own build, price/performance is always something to take into account, unless money happens to be no issue for you.
The price/performance ratio generally happens to be better for AMD CPUs over Intel CPUs due to better multi-core performance and being cheaper in the larger scope.
Let’s again look at the i3-4130 vs the FX-6350.
(Benchmarks courtesy of xbit labs, click here for link to full documentation)
CPU: i3-4130 FX-6350
Price: $125 $140
SYSmark 2012 is an application that simulates general everyday tasks, with most of these being single-core performance benchmarks.
As you can easily see, the i3-4130 scores better than the FX-6350.
Next they tested the same CPUs on a few different games, while choosing games that rely more on the CPU than most games do. (In case readers don’t wish to read the full Xbit labs article, the GPU in the test rig was a GTX 780.)
As you can see here as well, the i3-4130 wins almost every time, until we run into a game like Metro: Last Light, which was coded to utilize up to 4 cores of the CPU. While playing games like Batman: Arkham Origin and Civ 5: Brave New World, that only utilize at most 2 cores, the i3-4130 will win every time and the FX-6350 will show some FPS bottlenecking, whereas the opposite is true with games like Metro: Last Light. Games nowadays are moving towards utilizing more cores available to them from the CPU, which gives AMD a good outlook with the future for gaming, but for now Intel has a good grasp on performance in gaming.
Next they ran two very popular synthetic benchmarking applications, 3DMark Cloud Gate (meant more for multi-threading usage) and 3DMark Fire Strike (meant more for single-threading usage).
As you can see again, in the Cloud Gate test, barring the FX-4350 and A10, the AMD CPUs are on top again, but in Fire Strike, nearly all of the Haswell i3s beat out the FX-6350.
The last battery of tests were running these same CPUs through Cinebench R15, dPoweramp Music Converter, Lightroom 5.2, Photoshop CC, Truecrypt 7.1a, WinRAR 5.0, x264 (r2358), and Freemake Video Converter 4.1.0. Each of these applications tests various things from video encoding/converting, audio encoding/converting, encryption, decompression, or dealing with audio/visual manipulation. Across the board the FX-6350 wins but mainly due to the fact that these types of programs can easily utilize more cores to perform the tasks faster. (Pictures not shown to limit length of article. Readers are encouraged to read full article here.)
All of these tests have consistently shown that Intel will win more often than not in single-threaded/single-core applications, whereas AMD can put up a good fight in multi-threading/multi-core applications.
After everything is said and done between these two CPUs, they would receive very similar scores in the price/performance category; although when it comes to streaming, the more powerful the CPU the better. If the games you are playing, aren’t very CPU intensive, you could get away with using a processor in line with the i3-4130 or FX-6350; although if you can manage it, using something in the i5-4570/FX-8320 range or higher would be best.
Overclocking (OC) Ability
When it comes to the battle for overclocking ability between Intel and AMD CPUs, AMD is generally the better bet for one main reason: every single one of their CPUs has the ability to be overclocked.
Overclocking is increasing the base clock speed of your CPU to gain extra performance out of it, and it can’t be done to all CPUs, including most of Intel’s; anything not denoted with a “k” after it’s model number can not be manually overclocked. Recently Intel has implemented the ability for the operating system to dynamically boost the CPUs base clock if near 100% power is required by the operating system and is referred to as “turbo boost”; beyond that, the base clock stays put. On the other hand, all of AMD’s CPUs are unlocked and can be manually overclocked to increase performance across the board for the CPU. For this reason alone, many people people prefer to go with AMD CPUs because they get a little bit more freedom with their hardware.
Many wonder why overclocking is something to even be bothered with, but if you’re a real enthusiast about pushing your system to the absolute limits and squeezing every bit of power out of your CPU, overclocking can sometimes give you up to a 10% performance boost. Keep in mind, overclocking can be potentially harmful to your system if you don’t know what you are doing; please read up on guides and tutorials about how to overclock before attempting to do anything yourself.
Higher end CPUs
When comparing low end CPUs, it can pretty much be a wash while comparing benchmarks and real world performances, as was shown in the battery of tests performed by Xbit labs. All of the previous findings are amplified though, when you start looking at higher end CPUs.
i5-4570/4670/4670k vs FX-8320/8350
Taking a few steps above the i3-4130 and the FX-6350 will get you to the i5-4570 and the FX-8320. Each of these new CPUs are higher end, i.e. they have more cores and yield a power gain over the previously mentioned. As with all the previous sections, these 2 CPUs still carry on most of the same characteristics from their lower end counterparts.
The i5-4570 has 2 more physical cores over the i3-4130, likewise for the FX-8320 over the FX-6350, it has 8 cores versus 6 cores; each higher end CPU also has a higher base clock, further increasing the overall power.
According to various tech sites, the upper end of the i5 series is considered to be better overall. It has a much better single-threaded performance, consumes much less power and uses a newer manufacturing process (allowing for more “perks”). The counter to this is that the FX-8320/8350 can most times outperform the i5 series in multi-threaded/multi-cored applications.
It is also worth it to note, that when looking at benchmarks, synthetic or real time, most FX-8350 scores can be achieved by overclocking the FX-8320. In the line of these i5s, another good note, is that the non-k versions are not hyperthreaded; only the i5-4670k has hyperthreading (and overclocking ability), making it worth the $20-30 jump if you’re already willing to go over $200.
The prices for these CPUs are as follows:
i5-4670k = $240-250
i5-4670 = $220
i5-4570 = $200
FX-8350 = $200
FX-8320 = $160
Price/performance ranking: i5-4670k > i5-4670 > i5-4570 > FX-8320 > FX-8350
Even higher, i7+
If money is of no concern, or you just want the best you can get, the i7 range is most likely where you’ll be hangning out. This range consists of mainly the i7-4770, i7-4771 and the unlocked i7-4770k.
The i7-4770 has a base clock of 3.4GHz while the other two are at 3.5GHz. The i7-4770 and i7-4771 both have a turbo boost up to 3.9Ghz, while the i7-4770k is unlocked so it will go as far as you can push it. All 3 of these are hyperthreaded, making them the highest performing CPUs currently on the market (in the more user friendly, 1150 socket category).
The prices for these CPUs are as follows:
i7-4770 = $310
i7-4771 = $320
i7-4770k = $340-350
Price/performance ranking: i7-4770k > i7-4771/4770
The only real choice difference between the i7-4771 and i7-4770 is whether or not you want to pay $10 more as they are near identical performance.
After reading all of this has the question of “Which CPU is best for me?” been answered? Let’s hope not. No one CPU is right for everyone, due to prices and personal preference, among other things.
Depending on what you wish to do with your PC, I will place two sets of recommendations here.
For streaming plus gaming purposes, you’ll need a CPU that is fairly good at multitasking, and in this case, the bigger the better. Of course anything above the FX-8320 would work, but these are the best choices in their respective price ranges.
If it can be helped, don’t stream with anything under the FX-8320. Other, lower end CPUs will work, but either your game, your stream or both will suffer from the lowered power, and will cause neither you nor your audience to have an enjoyable time.
For gaming only purposes, you’ll need a CPU that is powerful, but not the most powerful. Anything over the FX-6300 will be perfectly fine, but these are the best choices in their respective price ranges. A quick thing to note as well, is that most of the time, the performance of the FX-6350 can be achieved by overclocking the FX-6300.
Anandtech has a good bench comparison tool built right into their site, containing data from all of the benchmarks they have ever performed. They don’t have every CPU on there, but they do have a lot; including the most popular ones.
Cpuboss is another good comparison site. They have their own benchmarking scores compiled into one easy to see locataion.
Don’t take any one tech site as fact, as not all benchmarks will be exactly the same everytime. It’s a good rule of thumb to check multiple sites before making your final choice, or if you have the option, talk to others on forums who have had personal experience with these hardware components.