With a team of dedicated hardware experts behind Tech4Gamers, we have created a specific methodology dedicated to testing and reviewing processors. This process helps us comb through every nook and cranny of a CPU and test its real-world performance based on a variety of scenarios. Furthermore, following the same set of testing methods for each processor also helps us compare them on similar grounds, which is very useful in creating the best CPU roundups.
So in this guide, we will be lifting the curtains and giving you a behind-the-scenes of how Tech4Gamers tests its processors. The goal of this guide is to tell you about everything that goes behind reviewing and testing a CPU. In addition to that, this guide will also help you buy the right processor because after reading it, you will know the importance of each benchmark and how it affects a processor’s performance and usage.
- First off, we perform a 5-step performance analysis. It contains general tests, rendering tests, transcoding, web-based benchmarks, and gaming benchmarks. After the 5-step testing process is over, we have a complete overview of a processor’s performance.
- During the testing, we keep an eye on the frequency behavior of the CPU. It lets us understand how high a processor can clock under intensive workloads.
- We also monitor the temperature and power consumption of the processors we test. This way, we can determine the best CPU cooler and power supply for it.
- Lastly, we overclock the processors step-by-step to determine exactly how much more performance we can get out of them without running into stability issues.
How We Test Processors
Being the brains of your gaming PC, the CPU is practically the most important part of your build. It impacts not only your PC’s gaming capabilities but also its productivity performance and overall experience. As such, there is a lot to test to gauge a processor’s performance accurately.
In addition to that, a processor’s success (or lack of it) does not end at just sheer performance alone. There are several other key factors to keep in mind as well. The most important of them is the CPU’s thermal and power efficiency, as it ultimately dictates the overall cost of the build and the tier of components required to handle the processor.
Then comes another aspect that we always test in processors: its overclocking potential. While this factor is mostly aimed towards enthusiasts who love to push their components to their limits, just knowing how much performance you can get out of your processors helps us understand how much power and thermal headroom it has.
With that said, let’s go through each step of our testing methodology.
Obviously, first and foremost, we dive into the processor’s performance. But, there are several different performance tests and benchmarks that we check to evaluate a CPU over various use cases. We start off with general tests to check the overall performance and move our way towards finally testing the gaming performance with several AAA titles running at different resolutions.
We begin our testing with several general benchmarks to see how the CPU performs in day-to-day tasks.
7-Zip is an open-source test benchmark that tests a processor’s file compression and decompression capabilities. It uses different algorithms to compress files in various formats and then decompresses them. The output result is in MIPS (million instructions per second), which is the higher, the better.
AIDA64 offers a suite of benchmarking tests to gauge a processor over a series of different tests. We first perform the CPU AES (Advanced Encryption Standard) test, which measures a CPU’s performance using data encryption. After that, we go for CPU Queen, which measures a CPU’s integer performance by solving the famous “Queens problem” on a 10×10-sized chessboard. Lastly, we go for the Memory benchmark, which tests the memory’s write, read, and copy speeds.
PCMark 10 tests a CPU over a series of tasks to gauge its performance based on web browsing, office productivity, digital content creation like photo/video editing, and other day-to-day tasks. The result is an overall score that is based on the CPU’s performance on all of these tasks.
PassMark PerformanceTest offers benchmarks to test different hardware. To test our processors, we use its Advanced CPU Benchmark Test, which is comprised of integer math, floating point math, prime numbers, extended instructions (SSE, AVX, and FMA), encryption, compression, sorting, and physics.
Super-PI is used to test a CPU’s single-threaded performance by calculating the value of PI after the decimal point to a specific number of digits; it can calculate up to 32 million digits.
To test the multi-core and multi-threaded benchmarks, we use wPrime, which uses Newton’s method to calculate square roots for estimating functions.
By using the Deep Fritz 12 engine to analyze as many chessboard positions as possible, Fritz Chess gauges the CPU performance of your machine. The outcome is given in kilo nodes per second or the number of chess positions your computer can analyze in a split second. The better the score, the faster your computer’s CPU performs.
3DMark CPU Profile
3DMark CPU profile is another way of testing a CPU’s multi-threaded performance. It features six different profile tests to test a processor’s capabilities. You can go for max threads, 16 threads, 8 threads, 4 threads, 2 threads, and 1 thread. Similarly, you can also do a custom run where you can remove or add a specific profile test into your benchmark run.
The Vera-Crypt disc encryption software includes a benchmarking tool called the Vera-Crypt CPU Test. It performs disc encryption and decryption of large files to evaluate a processor’s performance.
Once we have gone through the general tests to test the processor’s overall and basic performance, we move towards rendering tests that judge a processor based on how fast it renders a given scene, which is a very CPU-intensive task.
The Cycles render engine in the Blender Benchmark assesses CPU performance by rendering a number of scenes. The benchmark calculates hardware rendering speed by counting the number of samples it can render each minute. The faster the CPU is at rendering, the higher the score. We mostly use the Classroom and BMW27 rendering scenes.
Another similar rendering tool we use is Corona. It starts the test the moment the software is opened. The result is displayed in two ways: rendering time per second and rays per second.
Frybench, a multi-core CPU test, is developed utilizing the underlying framework of Fryrender, which is a photo-realistic rendering engine. Every aspect involved in producing the final render, including materials, lights, and cameras, is constructed based on meticulously accurate physical models.
V-Ray Benchmark is a standalone tool available for free, designed specifically to evaluate the rendering speed of your hardware. Depending on the type of processor you wish to test, the V-Ray provides two separate test scenes: one dedicated to GPUs and another for CPUs.
POV-Ray, short for Persistence of Vision Ray Tracer, is a ray tracing program that produces images using a text-based scene description. It can be used on various computer platforms and has its roots in DKBTrace, a program created by David Kirk Buck and Aaron A. Collins for Amiga computers.
Indigo Renderer is a cutting-edge GPU and CPU renderer renowned for its unbiased and photorealistic approach, aiming to achieve the highest possible image quality through accurate simulation of light physics. It simplifies rendering performance, materials, and camera models with an interactive photographic methodology that minimizes the need for abstract settings. This enables users to concentrate on the crucial aspects of lighting and the creation of their own visually stunning imagery.
Cinebench is a robust real-world test suite developed to assess computer performance across multiple platforms. Cinebench is an excellent tool for measuring and comparing system performance on both Windows and OS X systems, thanks to its ability to compare CPU and graphics performance. The most recent iteration, R23, includes the AVX load and numerous improvements.
Geekbench 5, a benchmark tool developed by Primate Labs, is frequently used for evaluating CPU and GPU performance on multiple platforms. It employs a variety of workloads to evaluate components’ capabilities comprehensively. These workloads are painstakingly designed to mimic real-world operations such as gaming, video editing, and 3D rendering. The Geekbench 5 benchmark result is displayed as a score, which is an overall representation of the CPU’s single-core and multi-core performance.
CPU transcoding performance, also known as the CPU’s capacity to encode or decode media files using different codecs, plays a crucial role in efficiently and swiftly processing and converting media content, including videos and audio.
The significance of CPU transcoding performance extends to numerous scenarios where media files require transformation or compression. In the realm of video editing, for instance, transcoding becomes essential for converting videos into alternative formats or compressing them to ensure optimal storage or seamless streaming.
To test a CPU’s ability to quickly and efficiently transcode, we perform two benchmarks: x264 HD and x265 HD Benchmarks.
x264 HD Benchmark
The x264 HD Benchmark is well-known as a benchmarking tool built primarily to analyze CPU performance when it comes to encoding a 1080p video with the x264 codec. Notably, the x264 codec is famous for being an open-source version of the H.264 video compression standard, recognized for its ability to produce outstanding video quality while being efficient with encoding.
x265 HD Benchmark
Similarly, we also use the x65 HD Benchmark, which encodes 1080p videos using the HEVC x265 video standard. Specifically designed for the benchmark, it utilizes the advanced and efficient x265 codec, known for its superior performance and recent advancements compared to the older x264 codec.
All of the above tests were related to productivity workload. However, when you build a PC, you don’t use it just for work. You also browse the internet and perform general browsing tasks. When we review processors, we test them for all of the real-life scenarios to stay as thorough as possible, and that includes testing their web-based performance.
It might seem obvious that if a CPU nails productivity tasks, it can easily handle web-based tasks as well. But, we believe in showing our readers the complete picture, and that includes telling them about each and every aspect of a processor’s performance.
Lastly, we move our heads towards the most important CPU benchmark, the gaming performance. It wouldn’t be a complete CPU review if it didn’t involve gaming benchmarks, and we always save the best for last. Anyways, to test a processor’s gaming performance, we run two software tests: Fire Strike and Time Spy, and after that, we move towards gaming benchmarks @ 1080p and 1440p resolutions.
Fire Strike is an exceptional benchmark developed exclusively for contemporary gaming PCs, utilizing DirectX 11 technology. It sets a new standard in real-time visuals, surpassing other benchmarks and games with its advanced complexity and sophistication. The benchmark comprises two test cases. One is for the physics, and the other is a combined test that rigorously tests the performance of both the CPU and GPU components.
Time Spy is one of the first programs designed exclusively for DirectX 12 to take advantage of the performance benefits offered by this new programming interface. Due to its use of a pure DirectX 12 engine, it is an appropriate benchmark for measuring the capabilities of the most contemporary graphics cards. This engine supports the API’s advanced capabilities, such as explicit multi-adapter, asynchronous computing, and multithreading.
When it comes to gaming, 1080p resolution is the most processor-bound. As you increase the resolution, the performance relies on GPU more and more. Therefore, 1080p gaming is more CPU-intensive than 1440p gaming, and 1440p gaming is more CPU-intensive than 4K.
We run a variety of AAA titles on 1080p resolution with ultra settings to record their framerates.
Similarly, we also test 1440p gaming to see how the performance shifts as we crank up the resolution.
During our performance testing, we also keep an eye on the clock speeds of the processors we are testing. We review how the frequencies behave under different workloads and how the processor reaches its boost clock speed as we ramp up the loads. It should be noted that there are a ton of factors that have an impact on a CPU’s ability to reach its turbo frequency. However, considering the fact that our test build and workload remain the same throughout different processors, these factors are minimized to some extent.
Nevertheless, analyzing the frequency behavior helps us check whether a given processor is performing at its rated speed or not. Furthermore, many times we have noticed that our units actually boost at clock speeds higher than their rated ones, which is always a good sight.
Thermal And Power Efficiency
It is integral to test the power consumption and heat output of processors. The new generation of CPUs from both Intel and AMD consumes lots of power and, in turn, generates quite some heat. Thermal and power efficiency are two fundamentals that must always be reviewed thoroughly when testing processors.
These two factors not only impact performance and overclocking but also dictate the tier of PSU and CPU cooler you will need in your PC build.
First, to set the base, we leave the test bench idle for exactly 30 minutes and then note down the temperature and power consumption. After that, we run a 30-minute Cinebench R23 test to test the power consumption and thermal efficiency of the CPU.
We perform this test for both single-core load and multi-core load.
Once we have complete data about the CPU’s performance, power consumption, and temperatures, we move towards the final phase of our thorough testing, and that is overclocking. It wouldn’t be a true processor review if it didn’t involve overclocking. This test puts into perspective exactly how much performance you should expect from your processor and how much more juice you can squeeze out of it.
With that said, we gradually overclock the silicon die and check its stability by running some of the above-mentioned benchmarking tests along with a few games. If the CPU maintains its overclocked state without running into any errors or crashes, we take things another step further by pushing it even more.
This way, we determine exactly how much performance you can get out of a given processor. Of course, at each step of the overclocking phase, we monitor the performance gains, power consumption, and temperatures. Doing so helps us figure out the sweet spot between running at the stock configuration and overclocking to the extreme.
The Tech4Gamers hardware reviewers follow a specific guideline to test and review processors. This guideline is created to keep the readers as informed as possible while also maintaining authenticity and reliability. As such, we test our processors over a series of benchmarking tools along with games to judge them based on their real-world performance.
From day-to-day tasks to CPU-intensive rendering, transcoding, and web-based benchmarks, we leave no stone unturned in testing our processors. We also benchmark high-end games at 1080p and 1440p resolutions to review the gaming performance of our processors.
To ensure accurate performance representation, our testing rig is equipped with state-of-the-art equipment, including the GIGABYTE RTX 4090 Gaming OC 24G. Rest assured that the figures you see reflect the actual performance you can expect from these processors.
Apart from processors, we also test a variety of gaming hardware. If you want to learn more about our testing process, be sure to check out the following guides: