The Evolution of CPU Technology: A Comparison of Xeon and Core i9 Processors

When it comes to high-performance computing, two processors that often come up are the Xeon and Core i9. While both share some similarities, they differ significantly in terms of their design, functionality, and purpose. In this article, we'll delve into the world of these processors and explore how they compare to each other.

At first glance, it may seem like Xeon and Core i9 are interchangeable terms, but that's not entirely accurate. While both processors belong to Intel's lineup, they cater to different markets and offer distinct features. Xeon is designed specifically for servers and data centers, whereas Core i9 is aimed at gaming laptops and high-performance desktops.

The Xeon processor is a server-grade CPU that offers scalability, reliability, and performance. It's built using Intel's 14nm process technology and features up to 28 cores and 56 threads on a single chip. This makes it an ideal choice for applications that require intense parallel processing, such as video editing, 3D modeling, and scientific simulations.

On the other hand, the Core i9 processor is a high-performance desktop CPU that's designed to deliver exceptional processing power and efficiency. It's built using Intel's 14nm process technology and features up to 8 cores and 16 threads on a single chip. While it's not as powerful as Xeon, it's still an excellent choice for gaming laptops and high-performance desktops.

However, in recent years, the lines between Xeon and Core i9 have blurred. With advancements in processor design and manufacturing, it's become possible to build servers that can rival the performance of workstations using a single CPU. This has led to a shift in the market, with more consumers looking for high-performance desktops that can handle demanding tasks.

So, do you still need multiple CPUs in a multi-socketed motherboard to handle intensive multi-threaded workloads? The answer is no. With advancements in processor design and manufacturing, it's now possible to build servers that can handle these workloads using a single CPU. In fact, a single Xeon Platinum 8180 offers 28 cores and 56 threads on a single chip, making it an ideal choice for applications that require intense parallel processing.

To demonstrate this, let's take a look at the Asus C621E Sage, a dual-socket motherboard that's designed to support multiple CPUs. This motherboard features two LGA 3647 sockets, allowing you to install two Xeon processors or other compatible CPUs. The idea behind this design is to provide a platform for enthusiasts who want to build high-performance servers or workstations using multiple CPUs.

But how does this setup affect performance? To answer that question, we'll need to look at some benchmarks and testing data. In our tests, we pitted the Asus C621E Sage against the fastest single CPU we've tested to date: the Core i9 Extreme Edition. We ran a variety of workloads, including Cinebench, 7-zip, Y-Cruncher, and Blender.

The results were surprising. While the dual-socket motherboard was faster than expected in some tests, it didn't always deliver the performance we expected. In fact, 7-zip showed a smaller gain over the Core i9 Extreme Edition, while Y-Cruncher found itself losing ground. However, in other tests, such as Cinebench and Blender, the dual-socket motherboard emerged victorious.

So what's going on here? The answer lies in the design of the dual-socket configuration. With two separate CPUs, each with its own set of cache and memory, there are latency issues that need to be addressed. This is where virtualization comes into play. By running multiple operating systems on top of a hypervisor like Red Hat KVM, we can allocate devices and memory to each CPU, reducing latency and improving performance.

In fact, our tests showed that by using virtualization, we could turn the dual-socket motherboard into two independent computing machines. This is exactly what's happening in the industry, where companies are using virtualization to optimize their systems and improve performance.

The implications of this are significant. With advancements in processor design and manufacturing, it's now possible to build servers that can rival the performance of workstations using a single CPU. This has led to a shift in the market, with more consumers looking for high-performance desktops that can handle demanding tasks.

However, there's a catch. The cost of these processors is astronomical, with the Xeon Platinum 8180 costing around $10,000 per CPU. While it may be tempting to buy one of these processors for your personal rig, it's unlikely you'll ever need the level of performance that an Xeon processor offers.

In conclusion, while Xeon and Core i9 processors share some similarities, they differ significantly in terms of their design, functionality, and purpose. The dual-socket motherboard is a testament to the evolution of CPU technology, where advancements in processor design and manufacturing are leading to new and innovative ways of building high-performance systems. Whether you're a gamer or a content creator, there's never been a better time to invest in a high-performance desktop or server.

"WEBVTTKind: captionsLanguage: enXeon WXeon scalableYou knowThey're both called XeonBut these things are really differentOne of them is basically a Core i9 with ECC memory supportAnd the other one is a server CPU that I fangirl all overBecause I love super high-end expensive tech toysNow, in the pastYou needed multiple CPU'sin one multi socketed motherboard in order to handle intensive multi-threaded workloadsBut, is that still the case today?Do you still need 2 of these, given that a single Xeon Platinum 8180is 28 cores and 56 threads on a single chip.I don't knowWhat is the purpose today ofA dual socket machine like this one and how much have single high core count CPUs?Eroded the market that they used to enjoyLet's find out. Shall we?Alright there's a lot more room down here and we are gonna need itfor this honking, not to mention heavy test bench.On this test bench, You will find the Asus C621E SageThis is a dual socket motherboard, rocking 2 LGA 3647sockets for Intel's Xeon scalable lineup of CPU'sAnd setups like this have actually been around as far back as the 486 in 1989With the resulting second-hand hardware giving enthusiasts the ability to get multiple physical coresin their homes over the years with the peak being somewhere in the mid 2000s or soBut, that was then and this is nowNow you can get multiple processing cores in a single chipTo see how far things have come what we're gonna do is pit this machine against the fastestSingle CPU that we've tested to dateWe're gonna try to keep the number of variables to a minimum in order to gauge the impactThat these extra CPU cores will have on our setupThough it should be noted that there aren't many options when it comes to aftermarket LGA 3647 coolersbecause most of the folks selling these kinds of systems would figure out their own solutionSo that means that our dual socket workstation will run a little bit toastyBut we didn't observe any thermal throttling so it shouldn't affect our performanceLet's start off then with good old-fashionedCinebenchI mean, we've seen this run before but it's always fun to see it finish that quicklySo, in a surprise to no one the dual socket machine is fasterbut considering it's 56 processing coresnot all of our workloads scale in the way that we might expect7-zip for example, shows a smaller than expected gain over our Core i9 Extreme EditionY-Cruncher even finds itself losing groundAsus real bench demonstrates thisthough with that said the encoding benchmark eats out a lead over our Core i9 7980XEand then Blender, well here we actually get a victory for our dual CPU system againShowing this platform's potential for expanding render farmsbutWhat is really going on hereWell, something you guys have to realize isthat there is more to a dual socket configuration than just more coresDo you remember when AMD managed a three percent improvement inIPC with 2nd Gen Ryzen just by improving cache latency?So on this motherboard, we've got two separate CPU's with two separate sets of cache and memorySee, these 6 banks go to this one and these 6 banks are wired into this oneand that means a lot of latency for compute tasks that require the same data setsThis latency is a necessary evil in the design of multiprocessor systemsbecause of the need for Non-Uniform Memory Access (NUMA) for shortThat allows these two processors to efficiently share resources or as efficiently as they canSo the short version of this is that it works by transparently allocating devices and memory to each CPUwhich means they can more easily avoid interrupting each other while accessing those resourcesThis in turn reduces the amount of waiting around that they have to do for those resources to become availableSo, that's what we're seeing during our testing like in Y-Cruncher for example...Where both CPU's are working on the same data, but...*babbling*What if we could use different data setsThen we should be able to find this kind of setups true callingand how better to do thatThan to effectively turn this system into two independent computing machines using virtualizationSo let's fire up unRAID, which uses Red Hat KVM as a hypervisor to see what kind of results we getsplitting these resources into multiple independent machinesImmediately, we see worse results from our VM's than our original 56 Core testingBut, look closely at how much lower it isIt's not a whole lotIn every test, it's basically the same story here and we are stillWwwaay out ahead of the Core i9 Extreme Edition, particularly when it comes to BlenderNow if we consider the fact that we are getting simultaneous work doneThat gives us a good look at what an optimized workload might look like I mean or heckVirtualization itself is a legitimate task to, I mean this thingCould be so many gamers in one PC ( ͡° ͜ʖ ͡°)but I digress I mean nobody is gonna buy something like this for their personal rig anytime soon given the$10,000 per CPU price tag which puts it squarely in the territory of\"big the size of the check doesn't matter\" businessWhat they care about is densityThe more processing power, a single computer can manage, the more processing power that can bephysically fit into a building and this is perhaps most important for data centers and render farms in particularThe less those guys have to spend on setting up the electricity and cooling management for a data centerVersus the amount of performance they can get the bettersooooooWill multiple sockets make a comeback in the prosumer space?Outside of you know oil and gas exploration wherethere are still workloads that can benefit from this kind of thingthe chances look pretty slim if you ask meBut I don't necessarily think that it's Intel's intent to sell these chips in the prosumer space and for that matterEven in the enterprise space. I don't think they knew a ton of themFor me, I look at a product like this as more of like a future crafting exercise where it is available today, butit's more of a representation of what actually might be attainable a generation or two from nowJust like the 22 core processors that we were playing around with a couple of years agoNowadays, those are much more affordable and businesses are using them to power your cloud computing servicesbut if it was awesome get subscribed hit that like button or check out the link to where to buy the stuff we featuredin the video descriptionAlso link down. There's our merch store which has cool shirts like this one and our community forum, which you should totally joinnow I'm off toFinally put this to use for the reason that I obtained itMany video editors, one CPU\n"