Cores vs Speed (I9 vs Threadripper)

by | Jul 19, 2022

Today’s article will be about one of the hottest topics for PC enthusiasts: when choosing a CPU for my computer is having multiple cores better or is having higher clock speeds better? Someone from outside looking in would say why not both? In truth, that would be the best case scenario, everyone running 64 core CPUs at 10GHz or higher. However, in reality we usually have to pick between one and the other. We’ll go over a couple of cool topics like cores, threads, clock speeds, gaming and productivity software. At the end of the article we will then put our new gained knowledge to the test with the industry’s top dogs: the Intel I9 9900K and the AMD Threadripper 2950x. Both of those CPUs have the highest core counts in their respective lineups along with some of the highest clock speeds.

Who remembers the first computer wars of the 90s and early 2000s? Both CPU manufacturers at the time (AMD and Intel) were on a race to get their respective lineups to hit the 1GHz processor speed range. In an unbelievable outcome both companies released their first 1GHz CPUs literally days from one another back in 1999. From that point on the two companies would keep ramping up the processing speed of their CPUs as fast as they could and we quickly saw clock speeds reach 1.5, 2 and even 3GHz. However, as time went on we started noticing that the improvements in clock speed were slowing down; manufacturers were reaching the limits of what was possible with current transistor technology and fabrication. Then, in an extremely similar situation with the release of the first 1GHz CPU, both Intel and AMD released their first dual core computers in May of 2005 also just days from one another (with AMD claiming the win). Many enthusiasts were skeptical at first because these new processors had lower clock speeds than older CPUs, so what gives? Well, all our doubts were put to rest when we experienced how much faster dual core processors with slower clock speeds were over single clock processors and their much faster (at first) clock speeds.

So that concludes the article, multiple cores win over single core speeds; sometimes, under certain conditions. We will have to take a step back and talk about a few related topics first however, starting with clock speeds. When we talk about the speed of a CPU we are often referring to its clock speed and this number is almost always measured in gigahertz (GHz) and it represents the clock rate at which the CPU operates; for example, a 1.5GHz CPU operates at 1.5 billion clock cycles per second. Each instruction that the CPU performs will take a few clock cycles to complete so even CPUs with equal clock speeds may perform differently depending on their architecture and instruction set. Next, let us move on to cores. Up until 2005 all CPUs that normal users could get their hands on were single core meaning that there was a single physical device that was handling all computations. With dual core CPUs (and quad core, etc) a single die (piece of silicon) will contain multiple physical devices that can each perform computations. The big idea with multiple cores is as follows: in a single core computer the CPU is running tasks one at a time, if you have multiple programs open the CPU will take turns giving each program attention and it does this so fast that to a person it looks like the computer is multitasking. Only in a multiple core CPU however, do you have real multitasking. If the scheduler sees that you are running multiple programs at once then it will divide the workload between the two (or more) cores; you could have your game running on one core and your 3-d modeling software running on another core thus allowing both programs to run faster. Lastly, let us talk about something you might have heard advertised on more expensive CPUs: multithreading. Whereas a core is a physical object a thread (this is not to be mistaken with threads which are the smallest piece of instructions that a CPU must perform in order to keep a certain program running) is a software created entity. With Intel’s Hyperthreading and AMD’s SMT each CPU core is divided into two threads each and to the rest of the computer they act as if they were multiple devices. In other words, if you have a dual core CPU with multithreading you can have up to 4 threads. Intel’s I9 9900KS for example has 8 cores and 16 threads while AMDs Threadripper 2950x has 16 cores and 32 threads and if you were to check for logical processors in your task manager you would indeed see that the computer sees the number of threads as if they were physical cores.

Putting all this together we can begin to see how CPU speed and threads work in conjunction: a thread will run programs and the faster the clock speed the faster the thread can work. Having multiple threads will allow multiple programs to have a thread of their own which means they will be able to perform better. Unfortunately, like with some good things, there is a catch. This catch is what makes the i9-9900K perform better than the Threadripper 2950x in some areas and what makes the Threadripper 2950x perform better than its rival in others. It all boils down to the people who are designing the software and whether they take the time to make their programs multi-core compatible. Simply because a CPU has 4 cores and 8 threads does not mean that a piece of software will use more than one; effectively rendering the rest of the cores useless wasting power for nothing. There’s also the fact that making software (or games) multi-core compatible takes far more time and work and therefore multiple core count is not always the deciding factor on how fast a piece of software runs. In cases like these, where a program can only utilize a single core, high clock speeds are the most important factor on the speed at which the program runs.

So now this brings us to the main showdown between the two consumer CPUs with the most cores and some of the highest clock speeds: Intel’s I9 9900K vs AMD’s Threadripper 2950X. Taking into consideration all the information we’ve given above you probably have begun making your guesses on which of the two is the better choice; we have the 9900K sporting 8 physical cores and 16 threads while running them at an average 3.6GHz with a max turbo clock speed of 5GHz. With the red team we have the 2950X equipped with a whopping 16 physical cores and 32 threads running at a base clock of 3.5GHz and a max boost clock of up to 4.4GHz. While we won’t run you through the benchmarks here (there are thousands of YouTube videos covering those) we will summarize what those videos would show you. When it comes to real world applications (software that serves some purpose) there are very few software examples that actually benefit from having multiple cores; a few that do would include video editing software (Adobe Premiere), 3D modeling programs (AutoCAD), scientific simulation software (MATLAB) or demanding productivity applications (Excel). In software like those mentioned above the 2950X does beat out the 9900K but not by margins as large as we would imagine from a CPU with twice as many physical cores. The reason this happens is that even multi-core compatible software will not scale infinitely with the number of cores; in other words, Adobe Premiere for example will use 4-8 cores to do its work and both of the CPUs we are comparing have the necessary hardware to supply that. When it comes to synthetic benchmarks (software that has no real world application whatsoever other than stressing the computer’s components to their limits) we do see results that are a little more like what we expect. Synthetic benchmarks are programmed in a way that they can stress any number of cores and therefore Threadripper with more cores than the 9900K will tend to beat its rival. Things might be a little bleak for the 9900K but things definitely change when we compare the two CPUs with software that the everyday person might use. Basically any kind of software that operates with at most 8 cores (meaning it cannot utilize any more whatsoever) will almost always crown the 9900K the victor. The reason for that goes back to our explanation of core clocks. Essentially, if a software can only use 4 cores then the determining factor between the two CPUs is their maximum core clock in which the 9900K beats out the Threadripper by a considerable margin. PC games are a very good example of this phenomenon. PC games are programmed in such a way that they will use 4 or perhaps 6 cores at most and hence the 9900K having sufficient cores but higher clock speeds will beat out the Threadripper in gaming performance on basically every game in the market. The same goes for most software nowadays as well, very few applications are coded to properly take advantage of more than 2 or  4 cores, many program are still even only single core compatible making all those extra cores in these CPUs useless. So, what is the verdict? Well, if you are the type of person that does some serious video editing or 3D designing then the Threadripper would be a solid choice. If you are a gamer or a normal user then the 9900K wins out ten times out of ten. Of course, if you are eyeing either of these processors you might not be a normal user at all. Both processors are in the 500+ dollar range.

There are a few other things to consider as well before making a choice between the two. One is hidden costs. The I9-9900K requires an 1151 type socket which is actually the socket Intel has used in the past three generations of processors which is good because motherboard prices can be bought at affordable prices. The Threadripper 2950X on the other hand uses a different socket than the current or past AMD Ryzen processors which means that you would not be able to upgrade your current CPU since the motherboard would not fit the new Threadripper CPU and the cheapest Threadripper compatible motherboard start out at $300. Another issue to consider is their power efficiency. The 9900K draws about 95W while the Threadripper draws almost twice that amount at 180W; while this may not seem like a big deal it does mean that you have to get better cooling systems to keep the Threadripper cool and a beefier power supply to give it enough power and that means more money to have to spend.

So the I9-9900K might be the obvious choice for most (if money is no issue) but that doesn’t mean you should completely disregard the Threadripper either. The software industry has been designing their programs to take advantage of more and more cores at a time which means that in the future the Threadripper might start gaining an edge and surpass the 9900K if software begins to use more than 8 cores or 16 threads. A Threadripper would also perform better as a server computer than a 9900K due to its larger amount of physical cores. At the end of the day you should pick a CPU that better aligns with the type of work that you plan to perform (there is absolutely no reason why you would get either of these if all you plan to do is browse YouTube and play Minesweeper). And while there is no such thing as future proofing a computer, either of these CPUs will definitely provide more than enough horsepower for many years to come. Thank you for joining us in today’s article, hopefully you learned a thing or two along the way. Do you have any topics or questions that you would like answered and talked about? Let us know. Until next time, have a good day!