One of the stories around AMD’s initial generations of Zen processors was the effect of Simultaneous Multi-Threading (SMT) on performance. By running with this mode enabled, as is default in most situations, users saw significant performance rises in situations that could take advantage. The reasons for this performance increase rely on two competing factors: first, why is the core designed to be so underutilized by one thread, or second, the construction of an efficient SMT strategy in order to increase performance. In this review, we take a look at AMD’s latest Zen 3 architecture to observe the benefits of SMT.

What is Simultaneous Multi-Threading (SMT)?

We often consider each CPU core as being able to process one stream of serial instructions for whatever program is being run. Simultaneous Multi-Threading, or SMT, enables a processor to run two concurrent streams of instructions on the same processor core, sharing resources and optimizing potential downtime on one set of instructions by having a secondary set to come in and take advantage of the underutilization. Two of the limiting factors in most computing models are either compute or memory latency, and SMT is designed to interleave sets of instructions to optimize compute throughput while hiding memory latency. 

An old slide from Intel, which has its own marketing term for SMT: Hyper-Threading

When SMT is enabled, depending on the processor, it will allow two, four, or eight threads to run on that core (we have seen some esoteric compute-in-memory solutions with 24 threads per core). Instructions from any thread are rearranged to be processed in the same cycle and keep utilization of the core resources high. Because multiple threads are used, this is known as extracting thread-level parallelism (TLP) from a workload, whereas a single thread with instructions that can run concurrently is instruction-level parallelism (ILP).

Is SMT A Good Thing?

It depends on who you ask.

SMT2 (two threads per core) involves creating core structures sufficient to hold and manage two instruction streams, as well as managing how those core structures share resources. For example, if one particular buffer in your core design is meant to handle up to 64 instructions in a queue, if the average is lower than that (such as 40), then the buffer is underutilized, and an SMT design will enable the buffer is fed on average to the top. That buffer might be increased to 96 instructions in the design to account for this, ensuring that if both instruction streams are running at an ‘average’, then both will have sufficient headroom. This means two threads worth of use, for only 1.5 times the buffer size. If all else works out, then it is double the performance for less than double the core design in design area. But in ST mode, where most of that 96-wide buffer is less than 40% filled, because the whole buffer has to be powered on all the time, it might be wasting power.

But, if a core design benefits from SMT, then perhaps the core hasn’t been designed optimally for a single thread of performance in the first place. If enabling SMT gives a user exact double performance and perfect scaling across the board, as if there were two cores, then perhaps there is a direct issue with how the core is designed, from execution units to buffers to cache hierarchy. It has been known for users to complain that they only get a 5-10% gain in performance with SMT enabled, stating it doesn't work properly - this could just be because the core is designed better for ST. Similarly, stating that a +70% performance gain means that SMT is working well could be more of a signal to an unbalanced core design that wastes power.

This is the dichotomy of Simultaneous Multi-Threading. If it works well, then a user gets extra performance. But if it works too well, perhaps this is indicative of a core not suited to a particular workload. The answer to the question ‘Is SMT a good thing?’ is more complicated than it appears at first glance.

We can split up the systems that use SMT:

  • High-performance x86 from Intel
  • High-performance x86 from AMD
  • High-performance POWER/z from IBM
  • Some High-Performance Arm-based designs
  • High-Performance Compute-In-Memory Designs
  • High-Performance AI Hardware

Comparing to those that do not:

  • High-efficiency x86 from Intel
  • All smartphone-class Arm processors
  • Successful High-Performance Arm-based designs
  • Highly focused HPC workloads on x86 with compute bottlenecks

(Note that Intel calls its SMT implementation ‘HyperThreading’, which is a marketing term specifically for Intel).

At this point, we've only been discussing SMT where we have two threads per core, known as SMT2. Some of the more esoteric hardware designs go beyond two threads-per-core based SMT, and use up to eight. You will see this stylized in documentation as SMT8, compared to SMT2 or SMT4. This is how IBM approaches some of its designs. Some compute-in-memory applications go as far as SMT24!!

There is a clear trend between SMT-enabled systems and no-SMT systems, and that seems to be the marker of high-performance. The one exception to that is the recent Apple M1 processor and the Firestorm cores.

It should be noted that for systems that do support SMT, it can be disabled to force it down to one thread per core, to run in SMT1 mode. This has a few major benefits:

It enables each thread to have access to a full core worth of resources. In some workload situations, having two threads on the same core will mean sharing of resources, and cause additional unintended latency, which may be important for latency critical workloads where deterministic (the same) performance is required. It also reduces the number of threads competing for L3 capacity, should that be a limiting factor. Also should any software be required to probe every other workflow for data, for a 16-core processor like the 5950X that means only reaching out to 15 other threads rather than 31 other threads, reducing potential crosstalk limited by core-to-core connectivity.

The other aspect is power. With a single thread on a core and no other thread to jump in if resources are underutilized, when there is a delay caused by pulling something from main memory, then the power of the core would be lower, providing budget for other cores to ramp up in frequency. This is a bit of a double-edged sword if the core is still at a high voltage while waiting for data in an SMT disabled mode. SMT in this way can help improve performance per Watt, assuming that enabling SMT doesn’t cause competition for resources and arguably longer stalls waiting for data.

Mission critical enterprise workloads that require deterministic performance, and some HPC codes that require large amounts of memory per thread often disable SMT on their deployed systems. Consumer workloads are often not as critical (at least in terms of scale and $$$), and so the topic isn’t often covered in detail.

Most modern processors, when in SMT-enabled mode, if they are running a single instruction stream, will operate as if in SMT-off mode and have full access to resources. Some software takes advantage of this, spawning only one thread for each physical core on the system. Because core structures can be dynamically partitioned (adjusts resources for each thread while threads are in progress) or statically shared (adjusts before a workload starts), situations where the two threads on a core are creating their own bottleneck would benefit having only a single thread per core active. Knowing how a workload uses a core can help when designing software designed to make use of multiple cores.

Here is an example of a Zen3 core, showing all the structures. One of the progress points with every new generation of hardware is to reduce the number of statically allocated structures within a core, as dynamic structures often give the best flexibility and peak performance. In the case of Zen3, only three structures are still statically partitioned: the store queue, the retire queue, and the micro-op queue. This is the same as Zen2.


SMT on AMD Zen3 and Ryzen 5000

So much like AMD’s previous Zen-based processors, the Ryzen 5000 series that uses Zen3 cores also have an SMT2 design. By default this is enabled in every consumer BIOS, however users can choose to disable it through the firmware options.

For this article, we have run our AMD Ryzen 5950X processor, a 16-core high-performance Zen3 processor, in both SMT Off and SMT On modes through our test suite and through some industry standard benchmarks. The goals of these tests are to ascertain the answers to the following questions:

  1. Is there a single-thread benefit to disabling SMT?
  2. How much performance increase does enabling SMT provide?
  3. Is there a change in performance per watt in enabling SMT?
  4. Does having SMT enabled result in a higher workload latency?*

*more important for enterprise/database/AI workloads

The best argument for enabling SMT would be a No-Lots-Yes-No result. Conversely the best argument against SMT would be a Yes-None-No-Yes. But because the core structures were built with having SMT enabled in mind, the answers are rarely that clear.

Test System

For our test suite, due to obtaining new 32 GB DDR4-3200 memory modules for Ryzen testing, we re-ran our standard test suite on the Ryzen 9 5950X with SMT On and SMT Off. As per our usual testing methodology, we test memory at official rated JEDEC specifications for each processor at hand.

Test Setup
AMD AM4 Ryzen 9 5950X MSI X570
AGESA 1100
4x32 GB
GPU Sapphire RX 460 2GB (CPU Tests)
PSU OCZ 1250W Gold
SSD Crucial MX500 2TB
OS Windows 10 x64 1909
Spectre and Meltdown Patched
VRM Supplimented with Silversone SST-FHP141-VF 173 CFM fans

Also many thanks to the companies that have donated hardware for our test systems, including the following:

Hardware Providers for CPU and Motherboard Reviews
RX 460 Nitro
RTX 2080 Ti
Crucial SSDs Corsair PSUs
DDR4-3200 32GB

CPU Performance
Comments Locked


View All Comments

  • abufrejoval - Thursday, December 3, 2020 - link

    It's hard to imagine a transistor defect that would break *only* SMT. As you say all non-SMT chips are really SMT chips internally and the decision to disable SMT doesn't really result in huge chunks of transistors going dark (the potential target area for physical defects).

    I'd say most of the SMT vs. no-SMT decisions on individual CPUs are binning related: SMT can create significantly more heat because there is less idle which allows the chip to cool. So if you have a chip with higher resistance in critical vias and require higher voltage to function, you need to sacrifice clocks, TDP or utilization (and permutations).
  • leexgx - Saturday, December 5, 2020 - link

    With HT off I have definitely noticed less smoothness windows, as with HT it can keep the cpu active when a thread is slightly stuck
  • iranterres - Thursday, December 3, 2020 - link

    Why are people still testing SMT in 2020? Cache coherency and hierarchy design is mature enough to offset the possible instruction bottleneck issues. I don't even know the purpose of this article at all... Anyways, perhaps fallng back to 2008? Come on...
  • quadibloc - Friday, December 4, 2020 - link

    Well, instead of testing the concept of SMT, which has been around for a while, perhaps one could think of it as testing the implementation of SMT found on the chips we can get in 2020.
  • eastcoast_pete - Friday, December 4, 2020 - link

    Thanks Ian! I always thought of SMT as a way of using whatever compute capacity a core has, but isn't being used in the moment. Hence it's efficient if many tasks need doing that each don't take a full core most of the time. However, that hits a snag if the cores get really busy. Hence (for desktop or laptop), 6 or 8 real cores are usually better than 4 cores that pretend to be 8.
  • AntonErtl - Friday, December 4, 2020 - link

    I found the "Is SMT an good thing" discussion (and later discussion of the same topics) strange, because it seemed to take the POV of someone who wants to optimize some efficiency or utilization metric of someone who can choose the number of resources in the core. If you are in that situation, then the take of the EV8 designers was: we build a wide machine so that single-threaded applications can run fast, even though we know the wideness leads to low utilization; we also add SMT so that multi-threaded applications can increase utilization. Now, 20 years later, such wide cores become reality, although interestingly Apple and ARM do not add SMT.

    Anyway, buyers and users of off-the-shelf CPUs are not in that situation, and for them the questions are: For buyers: How much benefit does the SMT capabilty provide, and is it worth the extra money? For users: Does disabling SMT on this SMT-capable CPU increase the performance or the efficiency?

    The article shows that the answers to these questions depend on the application (although for the Zen3 CPUs available now the buyer's question does not pose itself).

    It would be interesting to see whether the wider Zen3 design gives significantly better SMT performance than Zen or Zen2 (and maybe also a comparison with Intel), but that would require also testing these CPUs.

    I did not find it surprising that the 5950X runs into the power limit with and without SMT. The resulting clock rates are mentioned in the text, but might be more interesting graphically than the temperature. What might also be interesting is the power consumed at the same clock frequency (maybe with fewer active cores and/or the clock locked at some lower clock rate).

    If SMT is so efficient (+91%) for 3DPMavx, why does the graphics only show a small difference?
  • Bensam123 - Friday, December 4, 2020 - link

    Anand, while I value your in depth articles you guys really need to drop the 95th percentile frame times and get on board with 1% and .1% lows. What disrupts gaming the most is the hiccups, not looking at a statistically smooth chart. SMT/HT effects these THE most, especially in heavily single threaded games. If you aren't testing what it influences, why test it at all? Youtube reviews are also having problems with tests that don't reflect real world scenarios as well. Sometimes it's a lot more disagreeable then others.

    Completely invalid testing methodology at this point.

    My advice based on my own testing. You turn off SMT/HT except in scenarios in which you become CPU bound, across all cores, not one. This improved .1 and 1% frame time... IE stutters. You turn it on when you reach a point of 90%+ utilization as it helps and a lot when your CPU is maxed out. Generally speaking <6 and soon to be 8 cores should always have it on.

    You didn't even test where this helps the most and that's low end CPUs vs high end CPUs where you find the Windows scheduler messes things up.

    Also if you're testing this on your own, always turn it off in the bios. If you use something like process lasso or manually change affinity, windows will still put protected services and process onto those extra virtual cores causing contention issues that lead to the stuttering.

    Most obvious games that get a benefit from SMT/HT off are heavily single threaded games, such as MOBAS.
  • Gloryholle - Friday, December 4, 2020 - link

    Testing Zen3 with 3200CL16?
  • peevee - Friday, December 4, 2020 - link

    "Most modern processors, when in SMT-enabled mode, if they are running a single instruction stream, will operate as if in SMT-off mode and have full access to resources."

    Which would have access to the whole microinstruction cache (L0I) in SMT mode?
  • Arbie - Friday, December 4, 2020 - link

    Another excellent AT article, which happens to hit my level of knowledge and interest; thanks!

Log in

Don't have an account? Sign up now