One of the key factors that determine a laptop purchase is CPUs. Both Intel and AMD make very good CPUs. In this mammoth of a page, you will get a massive list of AMD and Intel CPUs, sorted according to the performance. Also, read through the following guide to know how to pick a CPU and how not to pick a CPU. Also, do not miss our “Recommended Smartphones” list that contains best smartphones for every budget.
CPU
Intel and AMD make the majority of CPUs for laptops and only recently we have started seeing CPUs from ARM. Both are based on AMD’s x86-64 architecture, which is an extension of Intel’s x86 architecture. At this point in time, AMD’s new Zen 3 chips have a clear upper hand over Intel’s 10th generation and 11th generation Core chips.
Cores
Unlike desktop processors where you can get up to 64 core/128 thread CPU, laptop processors max out at 8 cores and 16 threads, unless you find a laptop with desktop-class CPU (highly recommend not to buy such laptops as they will have a major problem in controlling heat and struggle to keep the CPU cool). Stay away from dual-core CPUs as you will end up with a laptop that is very slow to use. Also, keep an eye on the number of threads. A fraction of these CPUs come without hyperthreading and the CPU with 4 core / 4 threads will be slower than a CPU with 4 cores/8 threads.
Cache
A higher cache definitely helps. All CPUs come with three levels of cache (L1/L2/L3). Cache is used to store tiny amount of data and L1 is the volatile storage that is closest to the CPU core(s). A L1 cache typically 100 times faster than RAM. L2 and L3 caches are usually on a separate module on same CPU and is typically 25 times faster than RAM and slower than L1. It costs lot more to add L1 cache than to add L2 /L3/RAM. This is the reason why we see L1 cache in kilobytes and L2/L3 in Megabyte, RAM in Gigabytes.
In short, when you compare two very similar CPUs, one with higher L1-L3 cache tends to be faster given that everything else is similar.
IPC (Instructions per Clock/cycle)
IPC is one of the most important factors to assess performance of a CPU. IPC is calculated as the average number of instructions executed for each clock cycle (Source: Wikipedia). In simple terms, the CPU that has higher IPC. As an example, take Ryzen 5 3600X and Ryzen 5 5600X. The newer 5600X (3.7GHz) has 100MHz lower clock speed when compared to 3600X (3.8GHz), and has same number of cores/threads, same amount cache and made on same 7nm fab. Yet, the 5600X is 20%-30% faster than 3600x and in fact, in some cases, is faster than a Ryzen 7 3900X which has 12 cores/16 threads. The key reason is that AMD has made massive imporvements to IPC. According to AMD, the new Zen 3 chips (5xxx processors) have up to 25% higher IPC compared to previous (3xxx) chips.
In short, a CPU with higher IPC does very well overall when compared to a CPU with lower IPC, at similar clock speed.
Clock Speed
While IPC is all about number of instructions CPU can execute per cycel, clock is the number of cycles per second. For example, a CPU running at 3.2GHz can execute 3.2 billion cycles per second. Intel CPUs generally tend to have higher clock speeds but thanks to superior IPC, current Zen 3 AMD chips are able to easily beat Intel chips even though Intel chips run at higher boost clock speed. On a CPU, you see two clock speeds. One is called base clock and the other is called boost clock. Base clock is what the actual CPU clock speed is and boost clock is overclocking the CPU for small amount of time. When you buy a CPU, you have to remember that boost clock is not the speed at which CPU runs, CPU hits it for few seconds and go back to lower speed. This is to protect the CPU from overheating.
TDP and heat
Thermal Design Power (TDP) is a crucial factor for CPU. TDP, measured in watts, measures power consumption of a CPU under maximum load. This is usually directly linked to Fab process. The lower the size of transistor (7nm vs 14nm), the lower the power consumption is and lower the heat generation is. Just like a car engine, a CPU can run at its optimum best when the heat is under control. If too much heat is generated, CPU will throttle (lower the speed) or system will shutdown/restart. AMD’s Zen 2 and Zen 3 chips are on 7nm fab process and Intel is on 10nm and 14nm. As a result, we see AMD chips consuming less power and generate less amoutn heat. This has in turn increased efficiency.
Fab Process (5nm, 7nm, 8nm, 14nm etc)
Semiconductor device fabrication is the process used to create the integrated circuits that are present in everyday electrical and electronic devices. It is a multiple-step sequence of photo lithographic and chemical processing steps during which electronic circuits are gradually created on a wafer made of pure semiconducting material. Silicon is almost always used, but various compound semiconductors are used for specialized applications. -Wikipedia
CPUs and GPUs are filled with millions and millions of transistors. In chips, these function more like gates switching on and off, to allow electrons pass. In manufacturing plant, multiple chips are made out of single wafer of silicon. The nm figure that you see in specifications (14nm, 28nm etc) is the size of this transistor. When a company says that their chip is designed on 14nm fabrication process, in layman’s terms, it means that the size of the transistor/gate is 14nm. When the transistor size is reduced, there are various advantages. Electrons travel less distance and this improves efficiency and speed. Transistors require less power to turn on and this reduces amount of power required. As a result, there will be noticeable increase in processing power. Also, with smaller transistor size, the size of cpu decreases allowing phone makers to have more space for other components. For example, the battery can be bigger, niche features can be added, phone can be thinner.
Yield rate increases as manufacturing plants move to lower fabrication process. Take a scenario where the transistors are so big that you can only create two processors from a wafer. If one of these two processor is faulty, the yield rate for the wafer is just 50%. Shrink the size of transistor and now you can create 10 processor from the same wafer. If one is faulty ,the yield rate will be 90%. Now, create 100 processors from same wafer and if 1 is faulty, you have a yield rate of 99%. A smaller fabrication processor not only improves performance and power efficiency, it also lets chip manufacturing companies have higher yield rate, faster production and higher revenue.
Here is the order in which we rate the laptop CPUs. Tap on the model number to see key specifications:
