ZisaLog: Beginner’s Guide to Building a Custom PC 
SMT is a technology that allows two threads to run on a single core, which enables more effective use of CPU resources and improves overall performance.
In recent years, this technology has become standard in modern CPUs.
This article will clearly explain how SMT works, how to identify CPUs that support SMT, and the advantages and disadvantages of SMT—important points for choosing a CPU for a custom-built PC.
- SMT is a technology that allows one CPU core to execute two threads simultaneously.
- It is called “Hyper-Threading” by Intel and “SMT” by AMD.
- SMT makes effective use of idle execution units with other threads.
- It increases CPU utilization and improves multi-thread performance.
- Performance does not improve for all tasks, and resource contention can occur.
- Check for SMT support by looking at the official core/thread count or supported features list.
- SMT was developed as a way to overcome the limits of single-thread performance.
- With the same number of threads, more physical cores give better performance, but tend to cost more.
Basic knowledge such as how to read CPU manufacturer and model numbers, performance benchmarks, and how to choose a CPU from the perspective of performance and compatibility are also explained.
≫ Related article: How to Choose a CPU for a Custom PC [Performance / Features / Compatibility]
Select PC parts and online stores to instantly generate an estimate, check compatibility, and calculate power requirements. You can save up to five different builds, making it easy to try out multiple configurations.
≫ Tool：PC Parts Estimation & Compatibility Check Tool
Table of Contents
What is SMT?
This section explains the basics of SMT and how to identify CPUs that support SMT when building a custom PC.
About SMT
SMT (Simultaneous Multithreading) is a technology that allows one physical CPU core to execute two threads at the same time.
The main purpose is to increase CPU resource utilization and improve processing power.
Normally, a CPU can only execute one thread per core, but with SMT, it can process two threads on one core at the same time.
This allows more effective use of CPU resources and improves overall performance.
AMD calls this technology “Simultaneous MultiThreading (SMT)”, while Intel calls it “Intel Hyper-Threading Technology”.
The names and details differ slightly by CPU manufacturer, but the basic idea of running two processes on one core at the same time is the same.
How SMT Works
The basic idea of SMT is to make use of unused CPU resources (such as execution units) with a second thread while the first thread is processing, increasing the overall efficiency of the core.
To achieve this, the execution units in each CPU core are shared by multiple threads.
Execution units are the parts of the CPU that actually perform calculations and processing, and there are several types:
- Integer Arithmetic Unit (ALU)
Performs addition, subtraction, and logical operations. - Floating Point Unit (FPU)
Performs calculations with decimals (e.g., 1.23 × 4.56). - Load/Store Unit
Handles data transfer between memory and the CPU. - Branch Unit
Handles decision-making for conditional statements like if statements. - Vector Arithmetic Unit (SIMD)
Performs parallel processing used in image processing and AI.
With SMT, while one thread is using a specific execution unit (for example, the integer arithmetic unit), other execution units (such as the floating point unit or branch unit) may be idle.
The basic mechanism of SMT is that the second thread uses these idle execution units (sharing them between multiple threads).
In addition to using different types of execution units, the same type of execution unit can be used at different times.
For example, while one thread is temporarily stopped waiting for memory or branch prediction results, another thread on the same core can use the idle execution units to continue processing.
This means that even if one thread is temporarily stopped, the core as a whole continues processing, increasing CPU efficiency.
In this way, SMT aims to improve performance not only by making effective use of different execution units but also by filling timing gaps in the use of the same execution unit.
How to Check if a CPU Supports SMT
When considering a CPU for a custom PC, sometimes it is necessary to check if the CPU supports SMT.
In that case, it can be determined from the official specifications for core and thread count or from the list of supported features.
As for recent CPUs, most of them support SMT, but lower-end models like Core i3 and Ryzen 3 do not support it.
Also, the latest Intel Core Ultra Series 2 desktop CPUs do not support SMT.
Mobile CPUs still support SMT, so it has not disappeared, but it may be unsupported in the future.
How to Check by Core and Thread Count
If a CPU supports SMT, there will be two threads per core, so if the thread count is double the core count, the CPU supports SMT.
However, from the 12th generation of Intel Core i Series, there are two types of cores: P-cores and E-cores, and only P-cores support SMT.
Therefore, the thread count is calculated by doubling the number of P-cores and adding the number of E-cores. If this matches the thread count, the CPU supports SMT.
≫ Related article: About Intel CPU P-cores and E-cores
How to Check by Supported Features List
Intel and AMD specifications list the support status for each feature.
For Intel, check the [Intel Hyper-Threading Technology] item under [Advanced Technologies].
If it says “Yes”, the CPU supports SMT. If not, this item will not appear.
For AMD, check the [Multithreading (SMT)] item under [General Usage].
Similarly, if the CPU does not support SMT, this item will not appear.
Advantages of SMT
This section explains the advantages of SMT.
Improved Multi-Thread Performance
The biggest advantage of SMT is improved multi-thread performance.
Multi-thread performance means how quickly and efficiently a CPU can handle many tasks at once.
Normally, when a CPU is running a single thread, there may be delays such as waiting for memory.
However, with SMT, the CPU can process multiple threads at the same time, making efficient use of CPU resources that a single thread cannot fully utilize.
As a result, CPU idle time (time spent waiting and doing nothing) is reduced, and CPU utilization increases, improving multi-thread performance.
For example, in systems like database servers or web servers that need to handle many requests at once, SMT increases throughput and shortens response times.
For general use, multi-thread performance is also important when opening multiple browser tabs to listen to music on YouTube, research something in another tab, or check webmail, or when watching videos while downloading files or running a virus scan in the background.
These multiple light tasks each run as independent threads, and SMT allows the CPU to process them efficiently in parallel.
Hiding Latency
SMT also has the advantage of hiding latency.
CPUs often experience “latency”—waiting times that cannot be controlled by the CPU itself, such as reading data from memory or input/output processing.
If a CPU does not support SMT, it cannot share resources with other threads, so during these waits, the CPU cannot use its full processing power and resources are wasted.
SMT helps reduce this wasted waiting time.
While one thread is waiting for data to be read, another thread’s instructions can be executed immediately, making effective use of idle execution units and maintaining overall CPU efficiency.
This does not make each thread much faster, but it improves the system’s overall responsiveness and smoothness.
Disadvantages of SMT
This section explains the disadvantages of SMT.
Performance Gains May Be Limited
SMT improves multi-thread performance, but it does not improve performance for every task.
SMT increases CPU efficiency by running multiple threads at the same time, but since threads share resources, resource contention can occur.
For example, if both threads need the same execution unit at the same time.
Especially when cache memory or memory bandwidth becomes a bottleneck, the benefits of SMT may not be fully realized.
In particular, performance gains may be limited in the following cases:
- When threads are highly dependent on each other (for example, when one thread must wait for the result of another)
- When threads are performing heavy calculations at the same time
For example, FPS games rely heavily on single-thread performance for high frame rates and low latency.
If SMT causes cache or execution units to be shared between two threads, resource contention may actually lower frame rates.
Some professional gamers and advanced users even disable SMT to achieve more stable performance.
However, just because a task is gaming does not mean it is necessary to avoid CPUs with SMT.
This is only noticeable in ultra-high frame rate environments (such as consistently above 240fps), and for most players, there is not a significant difference in gaming experience.
In these situations, the benefits of SMT may not be as great as expected, so keep in mind that having an SMT-enabled CPU does not guarantee improved performance for every task.
Increased Power Consumption and Heat
SMT improves CPU efficiency, but it also has the disadvantage of increased power consumption and heat generation.
Running multiple threads at the same time makes full use of CPU resources, but as a result, the CPU consumes more power and generates more heat.
This increases the burden on the cooling system and can affect overall system stability.
Especially in high-performance PCs and servers, managing heat is important and proper cooling is necessary.
≫ Related article: How to Choose a CPU Cooler for a Custom PC [Performance / Features / Compatibility]
The Background of SMT
This section explains the background of SMT, including the performance limits and market needs at the time.
The Limits of Single-Thread Performance
In the past, CPU performance improvements were mainly achieved by increasing clock speed.
The higher the clock speed, the more instructions can be executed per second, and the better the single-thread performance.
However, in recent years, this approach has reached clear limits.
As clock speed increases, power consumption and heat generation rise sharply, making cooling and stability more difficult.
In addition, as semiconductors become smaller, new problems such as leakage current, design complexity, and increased manufacturing costs have emerged.
As a result, instead of trying to force single-thread performance higher, the trend has shifted toward using all CPU resources more efficiently.
Ken
Wasted Resources
Modern CPUs are designed with highly parallel execution units and memory systems, which should allow for high processing power.
However, in actual operation, not all resources are always used efficiently.
For example, while a single thread is waiting for data from memory or when there are dependencies in instruction order, some execution units or data paths may be idle.
It is not uncommon for resources to be idle inside the CPU.
SMT was developed to reduce this “waste of resources” and use the CPU more efficiently.
By running multiple threads at the same time, SMT allows one thread to use resources while the other is stopped, making effective use of idle time.
As a result, CPU utilization increases, more processing can be done, and overall system efficiency and responsiveness improve.
Growing Need for Multi-Threading
Recently, the ability to process multiple tasks at the same time (parallel processing) has become more important than ever, and the need for multi-threading technology is rapidly increasing.
This is because it has become clear that simply improving single-thread performance has its limits for computer performance.
This is due to changes in everyday usage environments.
For example, it is now common for users to play videos while searching the web in another tab, with software updates or cloud file syncing running in the background.
In addition, web applications and smartphone apps usually process user input, communication, and rendering in multiple threads asynchronously.
Servers and cloud environments are also designed to handle many requests at the same time.
Web servers, databases, game servers, and virtual machine environments often require thousands or tens of thousands of parallel processes per second.
To meet these needs, CPUs must be able to efficiently process multiple threads.
With SMT, a CPU can process multiple threads on a single physical core, achieving high parallel performance with limited hardware resources.
Practical Use and Spread of SMT
SMT is now widely used as an important technology in modern CPU design to improve performance and optimize resource efficiency.
This is because traditional methods of improving CPU performance, such as increasing clock speed and transistor density, are reaching their physical and practical limits due to heat, power consumption, and other factors.
In the past, performance was simply increased by running at higher clock speeds, but now this is difficult, and architectural innovations are needed to improve performance.
SMT was introduced in this context, and its practical use began in the early 2000s. A well-known example is Intel’s Hyper-Threading Technology (HTT).
This has been used in a wide range of CPUs, from the early Pentium 4 to the current Core and Xeon series.
AMD adopted SMT much later, starting with the Zen architecture in 2017, providing two threads per core.
Frequently Asked Questions about SMT
This section answers common questions about SMT.
Which Should Be Prioritized: Core Count or Thread Count?
Sometimes it is hard to decide whether to choose a CPU without SMT but with more cores, or a CPU with SMT but fewer cores, when the thread count is the same.
For example:
- 8-core configuration without SMT (8 cores / 8 threads)
- 4-core configuration with SMT (4 cores / 8 threads)
Cores each have independent execution units and can process instructions in parallel at full speed.
Therefore, for tasks with heavy computational loads or when multiple threads compete for resources, more cores provide more stable performance.
On the other hand, SMT shares execution units in the core among multiple threads, so peak performance per thread is lower.
Therefore, if only performance is considered, more cores (without SMT) generally provide higher performance for the same number of threads.
This is especially true for rendering, computation, gaming, scientific calculations, and other tasks with heavy per-thread loads.
On the other hand, for handling many lightweight tasks at the same time, CPUs with SMT can also process efficiently.
However, having more cores means each core needs its own instruction decoder, execution units, and cache, making the physical circuit larger and increasing manufacturing costs.
So, while more cores provide better performance, considering price, CPUs with SMT usually offer better cost performance.
SMT is an efficient way to increase processing power with fewer hardware resources, so CPU makers actively use it to balance cost and performance.
As a result, most modern CPUs support SMT by default.
Summary: SMT Is a Technology for Efficient Resource Utilization
This article explained the basics, advantages, and disadvantages of SMT (Simultaneous Multithreading).
Here is a summary of the key points:
- SMT is a technology that allows one CPU core to execute two threads simultaneously.
- It is called “Hyper-Threading” by Intel and “SMT” by AMD.
- SMT makes effective use of idle execution units with other threads.
- It increases CPU utilization and improves multi-thread performance.
- Performance does not improve for all tasks, and resource contention can occur.
- Check for SMT support by looking at the official core/thread count or supported features list.
- SMT was developed as a way to overcome the limits of single-thread performance.
- With the same number of threads, more physical cores give better performance, but tend to cost more.
SMT is a technology that increases processing efficiency by making full use of limited CPU resources.
In today’s world where multitasking is the norm, CPUs with SMT support are a reliable way to boost performance.
However, SMT does not always have a dramatic effect in every situation, and in some cases, having more physical cores is better depending on the workload.
When choosing a CPU, check the number of cores, threads, and whether SMT is supported, and find the right balance for your needs.
Basic knowledge such as how to read CPU manufacturer and model numbers, performance benchmarks, and how to choose a CPU from the perspective of performance and compatibility are also explained.
≫ Related article: How to Choose a CPU for a Custom PC [Performance / Features / Compatibility]
Select PC parts and online stores to instantly generate an estimate, check compatibility, and calculate power requirements. You can save up to five different builds, making it easy to try out multiple configurations.
≫ Tool：PC Parts Estimation & Compatibility Check Tool
