Techgage logo

Intel 12th-gen Core i9-12900K & i5-12600K Workstation Performance Review

Intel 12th-gen Core i9 Gaming PC

Date: November 30, 2021
Author(s): Rob Williams

Intel’s 12th-gen Core architecture represents a huge shift from previous designs, with ‘performance’ cores being teamed up with ‘efficient’ cores in launch models. For our first performance deep-dive, we’re going to put the i5-12600K and i9-12900K up against many encoding, rendering, and system benchmarks.



Introduction

With as exciting as the CPU market has been the past few years, it’s easy to forget that we had a long run leading up to 2017 where it felt like true advancement in CPUs just wasn’t happening. AMD’s chips were notably slower than Intel’s, and Intel’s seemed to only see modest performance improvements with each new generation. Compare this to GPUs, where a new generation is typically a lot faster than the previous one.

We all know the story of AMD’s Zen. When its first generation was released, it felt like we had true competition back, and even at that time, we had no idea what was coming. In the years that followed, core counts exploded, and with its most recent Zen 3 generation, AMD managed to match or best Intel from most angles.

Intel 12th-gen Core Alder Lake Review Kit
Intel 12th-gen Core Alder Lake Review Kit

Well, with Intel’s 12th-gen Core processors, it’s clear that while many were wondering when or how Intel would strike back at AMD, Intel was busy churning out a majorly different architecture that would get it back on track. It’s not often that our expectations are exceeded, but the 12th-gen Core series manages to do just that. We have plenty of benchmarks to explain why.

“It’s not often that our expectations are exceeded, but the 12th-gen Core series manages to do just that.”

We talked a length about what Intel’s new 12th-gen chips bring to the table a few weeks ago, but let’s reiterate some of the most important points here. In a huge departure from previous architecture designs, Intel has opted for one with Alder Lake that’s similar to Arm’s big.LITTLE, in that energy-efficient cores are saddled up with high-performance cores. On Intel’s new mainstream flagship, the Core i9-12900K, eight high-performance cores are paired up with eight energy-efficient cores.

This design confuses things a wee bit, because it’s strange to see a chip like the 12900K, which has 16 cores, but only 24 threads. We’re so used to seeing the thread count being double the core count, but that’s just not the case with these new designs. As you’ve likely surmised by now, the performance cores support SMT / HyperThreading, while the efficiency cores do not.

Intel 12th-gen Desktop Launch Lineup
Intel’s current 12th-gen Core CPUs, with more expected soon

With this design change, Intel is using a hardware solution called Thread Director to make sure that processes are allocated to the correct threads at the right time. Ideally, applications that actually require high-performance will use those respective cores, while background or simple tasks can use the efficiency cores. Ultimately, the goal is to deliver performance when it’s needed, and to sip power when it’s not.

Currently, Thread Director is only guaranteed to work ideally in Windows 11, as that OS features kernel upgrades that were performed in anticipation of this architecture. Intel is planning future improvements to Windows 10, and likewise Linux. If you’re devoted to Windows 10 or use Linux, now is probably not the time to upgrade to Intel’s 12th-gen.

While we didn’t test it for this review, this latest-gen of Core also sports Intel’s latest and greatest integrated GPU, called UHD Graphics 770. We’ll be giving this IGP a test in the future, probably first with VEGAS Pro, since that solution seems to jive well with Intel graphics.

As for other hardware, Intel’s latest chips require a brand-new chipset (Z690) as well as a brand-new socket (LGA1700), so you can’t plop a new chip into your preexisting motherboard. However, if you already have a CPU cooler for an Intel Core platform, you’re likely good on that front. That said, some of the top 12th-gen SKUs get toasty, so you won’t want to cheap out on your cooling.

Intel Z690 Chipset Block Diagram

One of the most significant changes with the Z690 chipset is that support for DDR5 is enabled, with DDR5-4800 being the officially supported spec. Faster kits exist, but it’s never guaranteed that you’ll enjoy complete stability (even though you will probably be fine if you stick to the XMP). Fortunately, DDR4 is also supported, so if you happen to have a beefy kit already, you may not need to buy another.

While we didn’t test DDR5 vs. DDR4 for this article, what we’ve seen so far has shown that the differences are generally not major, with strengths between either flip-flopping all over the place. To add to that, it’s really difficult to find DDR5 in stock right now, and as we wrote about in our news last week, price-gouging is intense, with some kits selling on eBay for five times as much as their SRP.

At this point in time, we’d suggest considering going the DDR4 route, since current DDR5 kits are going to be considered modest when even faster kits arrive later on. It’s currently cheaper to go with DDR4, allowing creators to opt for a healthy amount of memory (64~128GB) for a better price than DDR5. If you’re a gamer and can find a 32GB DDR5 kit at its expected price, then it’s not a bad way to go.

Intel 12th-gen Core i9-12900K CPU and ASUS' Z690-PLUS WiFi D4 Motherboard
Intel’s Core i9-12900K installed in ASUS’ Z690-PLUS WiFi D4

On that note, we do have plans to conduct some DDR4 vs. DDR5 performance testing soon, as we’re eager to see which workloads may actually benefit from that added bandwidth.

In addition to DDR5 support, another major update is to PCIe, version 5.0. It seems like only yesterday AMD introduced the first PCIe 4.0 platform, but here we are with 5.0 already. While PCIe 5.0 is going to have virtually no use over 4.0 at the moment, future graphics cards and SSDs will naturally update to it, so it’s nice that this platform will prepare you for them.

With that all covered, this feels like a good time to move on. On the next page, we’re going to discuss our tests, methodologies, and the PCs used for testing. If you don’t care about that, feel free to hop on over to page three.

Test Methodology & Systems

Benchmarking a CPU may sound like a simple enough task, but in order to deliver accurate, repeatable results, strict guidelines need to be adhered to. This makes for rigorous, time-consuming testing, but we feel that the effort is worth it.

This page exists so that we can be open about how we test, and give those who care about testing procedures an opportunity to review our methodology before flaming us in the comments. Here, you can see a breakdown of all of our test machines, specifics about the tests themselves, and other general information that might be useful.

For testing AMD’s Ryzen AM4-based CPUs, we’re using ASRock’s X570 TAICHI motherboard. For Intel’s 11th-gen Core processors, we use ASUS’ ROG MAXIMUS XIII HERO. For Intel’s 12th-gen, we’re using ASUS’ Z690-PLUS WiFi D4. Note that while Intel’s 12th-gen supports DDR5, we’re kicking off our coverage with a DDR4 platform, as our DDR5 memory kit didn’t arrive in time for our testing to get underway. That of course will be rectified soon, as we plan to get down and dirty with DDR4 vs. DDR5 testing shortly.

The same DDR4 64GB kit of Corsair DOMINATOR Platinum memory is used on each of the test platforms. After resetting each EFI to its default values, we enable the primary XMP option, and then manually downgrade the DDR speed to 3200MHz. We do this because DDR4-3600 has given us enough headaches in the past that we just don’t want to set ourselves up to having to waste a bunch of time because one platform decides to be fussy (and yes, we’ve had platforms suddenly become fussy.)

In a similar note, while in the past, we obsessed over changing motherboard settings to replicate “stock” or “reference” values, for this article, we simply enabled the memory profile, and left it at that. That means that in some cases, these CPUs may run a bit faster than they will on your own rig, if you disable things like ASUS’ MultiCore Enhancement. Similar to how DDR4-3600 has been fussy for us, we obsessed too much in the past about “apples to apples”, which increasingly seems like a pipe dream due to how complex motherboards have become. Just bear in mind that your platform, when all built, could be either a bit slower or faster based on your settings.

As of the time we got down to testing, Intel’s 12th-gen CPUs were primarily encouraged for use with Windows 11, whereas launch issues for that OS made things a bit iffy for other architectures. For this article, we stuck to Windows 10 for both Intel’s 11th-gen chips, as well as AMD’s AM4. For Intel’s 12th-gen, we used a fully up-to-date Windows 11. As always, an updated chipset driver was used for each platform.

Most of the tests in this article were run with the help of automated test scripts to help ensure that the same settings and actions are applied each and every time. This generally means that we don’t need to run a test too many times over to generate an accurate and repeatable result – but generally speaking, most tests are run at least three times over.

Here’s the full breakdown of the test platforms:

Techgage’s CPU Testing Platforms

AMD AM4 Test Platform
Processors AMD Ryzen 9 5950X (3.4GHz, 16C/32T)
AMD Ryzen 5 5600X (3.7GHz, 6C/12T)
Motherboard ASRock X570 TAICHI
CPUs tested with BIOS P4.60 (August 3, 2021)
Memory Corsair VENGEANCE (CMT64GX4M4Z3600C16) 16GB x4
Operates at DDR4-3200 16-18-18 (1.35V)
Graphics NVIDIA RTX 3070 (8GB; GeForce 496.49)
Storage WD Blue 3D NAND 1TB (SATA 6Gbps)
Power Supply Corsair RM850X (850W)
Chassis Fractal Design Define C
Cooling Corsair Hydro H100i PRO RGB (240mm)
Et cetera Windows 10 Pro (21H1, Build 19043)

Intel LGA1700 Test Platform
Processors Intel Core i9-12900K (3.2GHz, 16C/24T)
Intel Core i5-12600K (3.7GHz, 10C/16T)
Motherboard ASUS Z690-PLUS WiFi D4
CPUs tested with BIOS 0503 (September 10, 2021)
Memory Corsair VENGEANCE (CMT64GX4M4Z3600C16) 16GB x4
Operates at DDR4-3200 16-18-18 (1.35V)
Graphics NVIDIA RTX 3070 (8GB; GeForce 496.49)
Storage WD Blue 3D NAND 1TB (SATA 6Gbps)
Power Supply EVGA Bronze 600B1 (600W)
Chassis Corsair Crystal X570 RGB
Cooling Corsair Hydro H115i PRO RGB (280mm)
Et cetera Windows 11 Pro (21H2, Build 22000)

Intel LGA1200 Test Platform
Processors Intel Core i9-11900K (3.7GHz, 10C/20T)
Intel Core i5-11600K (4.1GHz, 6C/12T)
Motherboard ASUS ROG MAXIMUS XIII HERO
CPUs tested with BIOS 1007 (July 13, 2021)
Memory Corsair VENGEANCE (CMT64GX4M4Z3600C16) 16GB x4
Operates at DDR4-3200 16-18-18 (1.35V)
Graphics NVIDIA RTX 3070 (8GB; GeForce 496.49)
Storage WD Blue 3D NAND 1TB (SATA 6Gbps)
Power Supply EVGA Bronze 600B1 (600W)
Chassis Corsair Crystal X570 RGB
Cooling Corsair Hydro H115i PRO RGB (280mm)
Et cetera Windows 10 Pro (21H1, Build 19043)

Testing Considerations

As covered earlier, we use an up-to-date Windows for our testing, as well as the latest chipset driver for each respective platform. In the pursuit of accurate, repeatable benchmarks, here are some basic guidelines we follow:


Encoding Tests

Adobe Lightroom Classic
Adobe Premiere Pro
Agisoft Metashape
Blackmagic RAW Speed Test
LameXP
VEGAS Pro

Go straight to test:


Rendering Tests

Autodesk Maya with Arnold Autodesk Maya with Arnold
Autodesk Maya with Arnold
Autodesk Maya with Arnold
Blender Blender
Blender
Blender
Maxon Cinebench Maxon Cinebench
Maxon Cinebench
Maxon Cinebench
Autodesk 3ds Max with Corona Renderer Autodesk 3ds Max with Corona Renderer
Autodesk 3ds Max with Corona Renderer
Autodesk 3ds Max with Corona Renderer
Luxion KeyShot Luxion KeyShot
Luxion KeyShot
Luxion KeyShot
LuxMark LuxMark
LuxMark
LuxMark
POV-Ray POV-Ray
POV-Ray
POV-Ray
Autodesk 3ds Max with V-Ray Autodesk 3ds Max with V-Ray
Autodesk 3ds Max with V-Ray
Autodesk 3ds Max with V-Ray

Go straight to test:


Synthetic Tests

UL Procyon UL Procyon
UL Procyon
UL Procyon
SiSoftware Sandra SiSoftware Sandra
SiSoftware Sandra
SiSoftware Sandra

Go straight to test:


If you think there’s some information lacking on this page, or you simply want clarification on anything in particular, don’t hesitate to leave a comment.

Encoding: Premiere Pro, VEGAS Pro & Metashape

Premiere Pro: Project Encodes

Adobe Premiere Pro - 1080p YouTube CPU Encoding (AVC) Performance (Intel 12th-gen Core)
Adobe Premiere Pro - 4K YouTube CPU Encoding (AVC) Performance (Intel 12th-gen Core)

Adobe’s Premiere Pro helps kick-off our look at 12th-gen performance, and right out of the gate, we’re impressed. AMD’s Zen 3, and especially the 16-core Ryzen 9 5950X, seemed like it’d be a hard to beat for a while with a test like Premiere Pro, but Intel proves that it still has what it takes to release a new architecture that can effectively “wow” us. In both encodes here, the 8+8 design of the i9-12900K bests AMD’s 16-core 5950X.

As for the i5-12600K, it too impresses, using its 6+4 core design to leap far ahead of not just the 6-core Ryzen 5 5600X, but the last-gen 8-core i9-11900K.

As with many creator workloads nowadays, using only the CPU isn’t always the best way to go about getting an encode process done quickly. If we enable CPU+GPU encoding, scaling becomes much tighter:

Adobe Premiere Pro - 4K YouTube CPU Encoding (CUDA, AVC) Performance (Intel 12th-gen Core)

Even with the benefit of the GPU helping out so much in these encodes, both of Intel’s new 12th-gen chips tested here still stay glued to the top.

The above results represent real project encodes, but let’s get a little more granular and see how things shake up when we test specific codec transcodes instead:

Premiere Pro: Transcoding

Adobe Premiere Pro - 4K AVC to AVC CPU Encoding Performance (Intel 12th-gen Core)
Adobe Premiere Pro - 4K AVC to HEVC CPU Encoding Performance (Intel 12th-gen Core)
Adobe Premiere Pro - 8K RED to AVC CPU Encoding Performance (Intel 12th-gen Core)
Adobe Premiere Pro - 8K RED to HEVC CPU Encoding Performance (Intel 12th-gen Core)
Adobe Premiere Pro - 8K ProRes to AVC CPU Encoding Performance (Intel 12th-gen Core)
Adobe Premiere Pro - 8K ProRes to HEVC CPU Encoding Performance (Intel 12th-gen Core)

This is quickly turning into a rather exciting performance review. We’ve been so used to seeing Intel release a new generation of chips that didn’t feel that revolutionary, and that’s making Alder Lake feel like the most refreshing Intel launch we’ve seen in a while.

In all of the codec tests above, Intel’s Core i9-12900K beats out AMD’s Ryzen 9 5950X, and in the RED tests specifically, the i5-12600K manages to pull that off, too. It’s clear that with some encodes, it’s not always additional cores that can make a big difference, but more efficient and higher-performing cores instead.

Let’s reintroduce the GPU into our testing:

Adobe Premiere Pro - 4K AVC to AVC (CUDA) CPU Encoding Performance (Intel 12th-gen Core)
Adobe Premiere Pro - 4K AVC to HEVC (CUDA) CPU Encoding Performance (Intel 12th-gen Core)
Adobe Premiere Pro - 8K RED to AVC (CUDA) CPU Encoding Performance (Intel 12th-gen Core)
Adobe Premiere Pro - 8K RED to HEVC (CUDA) CPU Encoding Performance (Intel 12th-gen Core)
Adobe Premiere Pro - 8K ProRes to AVC (CUDA) CPU Encoding Performance (Intel 12th-gen Core)
Adobe Premiere Pro - 8K ProRes to HEVC (CUDA) CPU Encoding Performance (Intel 12th-gen Core)

As we saw with our CPU+GPU project encodes earlier, Intel still manages to keep glued to the top of most CPU+GPU codec tests, with the exception of the 5950X proving better in the RED test, and on par in the ProRes test.

Premiere Pro is just one video editor / encoding tool, so let’s see how Intel’s latest handles the also-popular VEGAS Pro next:

VEGAS Pro

VEGAS Pro - Median FX CPU Encoding Performance - (Intel 12th-gen Core)
VEGAS Pro - Colorize FX CPU Encoding Performance - (Intel 12th-gen Core)
VEGAS Pro - Style Transfer Encoding Performance - (Intel 12th-gen Core)

We’re not including CPU+GPU encode tests here because we encountered sporadic results overall, and couldn’t put much faith into all of them. Of the tests we did feel confident in, it seems CPU+GPU is even more important in VEGAS than in Premiere Pro – the deltas between all of them were modest.

Nonetheless, when engaging only the CPU, Intel’s and AMD’s current top-end mainstream chips match each other – with the exception of the Median FX test, which saw Intel gain a notable advantage. Both the Colorize and Style Transfer tests use CPU-bound AI, so it’s really interesting to see both the 12900K and 5950X perform identically in both.

We didn’t have time to update our tests for VEGAS Pro 19 in time for this article, but we’ll be getting down to exploring performance there soon, including taking a look at any new FX that may exist that could take good advantage of the CPU (or GPU, for that matter.)

Up next, a look at something a little different: photogrammetry.

Metashape

Agisoft Metashape Photogrammetry Performance - Build Depth Maps (Intel 12th-gen Core)
Agisoft Metashape Photogrammetry Performance - Build Dense Cloud (Intel 12th-gen Core)
Agisoft Metashape Photogrammetry Performance - Build Mesh (Intel 12th-gen Core)

Unlike most other workloads featured in this article, photogrammetry is one that makes it difficult to choose which CPU is the “best”, as each process of the entire photogrammetry task will use the processors differently. Aligning Photos uses both the CPU and GPU, as does Build Depth Maps. Meanwhile, Build Dense Cloud and Build Mesh uses only the CPU.

Overall, Intel’s top-end i9-12900K continues to prove that it means business, ranking at the top of most tests here. The Build Mesh result is the most interesting, since Intel claims the first three spots, while the 16-core 5950X drops to the bottom. This is behavior we’ve seen before, where the 5600X could manage to beat out the 5950X, and that really comes down to what we said above about why it’s difficult to find a “best” CPU for these tasks. Some CPU designs, their core counts, and their per-core performance can dramatically change behavior.

Ultimately, both of these new Intel chips perform very well here. It’s important to note that the GPU matters quite a bit in photogrammetry, as well. You can see recent performance in the Build Depth Maps test across 19 GPUs on the same CPU in our recent Radeon Pro W6600 and W6800 review.

Encoding: Lightroom, BRAW, HandBrake & LameXP

Lightroom Classic

Adobe Lightroom Classic - RAW to JPEG Export Performance (Intel 12th-gen Core)
Adobe Lightroom Classic - RAW to DNG Export Performance (Intel 12th-gen Core)

On the previous page, we saw the Core i9-12900K prove to be a fierce competitor to the Ryzen 9 5950X, beating it out in the majority of tests. In Lightroom, we see the 5950X claw back in its fight, keeping pegged to the top of both export tests. This is one workload that seems to favor not just efficient architectures, but additional cores, as well.

What’s perhaps most impressive about the results above is how much Intel has improved in this workload from one generation to the next. It used to be that even modest Ryzens could outperform Intel’s higher-end CPUs, just as we see here with the 6-core 5600X outperforming the 8-core i9-11900K.

Before going into this test, we wondered it the unique design of the new Alder Lake-derived chips would result in strange Lightroom results, but that’s not the case at all. There’s no other way to say it: Intel dramatically improved Lightroom performance between its 11th- and 12th-gen Core CPUs.

The above results also make us daydream about how prospective Core X-series CPUs infused with the Alder Lake architecture would fare in this workload. From previous testing, we’ve seen AMD’s Ryzen Threadripper series perform great in the JPG export, but not so much in the DNG. If Intel could dominate both with Core X, that’d be pretty cool to see.

BRAW Speed Test

Blackmagic RAW Speed Test (Intel 12th-gen Core)

We’ve long seen AMD’s higher core count CPUs perform great in BRAW Speed Test, but this becomes yet another example of how efficient Intel’s new architecture is, as both the 12600K and 12900K perform seriously well against the competition they’re going up against. Just don’t forget to also equip your rig with a powerful GPU, since DaVinci Resolve will take good advantage of that, too.

HandBrake

HandBrake AVC Encoding Performance - (Intel 12th-gen Core)
HandBrake HEVC Encoding Performance - (Intel 12th-gen Core)

We need to be honest: when we first began benchmarking Intel’s 12th-gen chips, we were ignorant of just how much better the architecture was over 11th-gen. Naturally, companies always talk positively about architecture upgrades with any new generation, but the 12th-gen release has been downright refreshing due of the stark improvements we’re seeing.

HandBrake is the latest example: the 8+8 design of the 12900K has no problem being efficient with either AVC or HEVC encoding here. While the 16-core 5950X ultimately wins, it’s impressive that Intel’s 8 high-performance cores tied to 8 energy-efficient cores are working this well together.

LameXP

LameXP - FLAC to MP3 Encoding Performance - (Intel 12th-gen Core)

We wrap up our encoding results with one of our favorite scenarios: music encoding. This test couldn’t be simpler: just toss 500 FLAC files at LameXP, and encode them to MP3. This is a process that used to be inefficient on many-core CPUs, but that’s sure not the case nowadays. Yet again, we see Intel performing extremely well here, with the 6+4 core design of the i5-12600K being enough to beat out last-gen’s 8-core i9-11900K.

On the next two pages, we’ll get into a performance scenario that always leads to taking great advantage of an efficient architecture and as many cores as you can throw at it: rendering.

Rendering: Arnold, Blender, KeyShot, V-Ray

Arnold

Autodesk Arnold CPU Rendering Performance - Jaguar E-Type (Intel 12th-gen Core)
Autodesk Arnold CPU Rendering Performance - Sophie (Intel 12th-gen Core)

With its unique architectural design, we really weren’t sure what to expect out of Intel’s 12th-gen Alder Lake chips in workloads that scale extremely well with additional cores. Could we possibly see the 8P+8E design of the 12900K beat out a traditional 16-core like the 5950X? Well, we have our answer thanks to the results above: “yes”.

It really is impressive just how fast these 12th-gen chips are in all of the workloads we’ve tested so far. With the original Zen, AMD helped make CPU-bound rendering tests interesting again, and it only became more interesting as higher core count chips were released. Now, Intel helps make it interesting by pairing performance and energy cores together to beat out AMD’s same-count native cores.

Let’s see if this trend continues:

Blender

Blender Cycles CPU Rendering Performance - BMW (Intel 12th-gen Core)
Blender Cycles CPU Rendering Performance - Classroom (Intel 12th-gen Core)
Blender Cycles CPU Rendering Performance - Controller (Intel 12th-gen Core)

With Blender, we’re seeing similar performance out of these chips as we did with Arnold, with Intel’s i9-12900K edging out the 5950X ever-so-slightly in the majority of tests. Notably, AMD did manage to win the Classroom render test, so overall, both chips perform pretty similarly, we must say.

What could change once a GPU is introduced?

Blender Cycles CPU + GPU Rendering Performance - BMW (Intel 12th-gen Core)
Blender Cycles CPU + GPU Rendering Performance - Classroom (Intel 12th-gen Core)
Blender Cycles CPU + GPU Rendering Performance - Controller (Intel 12th-gen Core)

If you ever needed proof that a graphics card can dramatically improve render times, then the above results should prove sufficient. Once again, Intel’s 12900K takes the top spot in the same projects as we saw in our CPU-only testing, but really – most of the end result is being helped by the GPU more than it is the CPU.

One thing the above results leave out is what things would look like if we used NVIDIA’s OptiX API instead of CPU or CUDA. As we’ve seen in the past, there are indeed times when CPU+GPU can beat out OptiX with a fast GPU, but it’s likely going to require many cores. You can peruse this yourself in our Blender 2.93 performance deep-dive.

We’re currently in the middle of benchmarking for the upcoming Blender 3.0 launch, so we’ll find out soon enough if the new Cycles X render engine boosts CPU performance as it does GPU performance.

KeyShot

Luxion KeyShot - Character Rendering Performance (Intel 12th-gen Core)
Luxion KeyShot - Room Rendering Performance (Intel 12th-gen Core)

KeyShot becomes the first rendering result that safely gives AMD’s Ryzen 9 5950X the lead, and we’d suspect that will be the case with most KeyShot projects, since we’ve generally seen nearly identical scaling with any project we’ve tested with.

While the 5950X beats out the 12900K here, the 12600K easily beats out the 5600X. Both chips have six high-performing cores, but those four additional cores with the 12600K clearly help out.

As with most of the results on this page so far, KeyShot does a good job highlighting the massive improvements Intel CPUs have seen going from 11th- to 12th-gen.

V-Ray

Chaos Group V-Ray - Flowers CPU Rendering Performance (Intel 12th-gen Core)

The above result looks similar to others we’ve seen on this page, but don’t worry, it’s about to get more interesting in a second. When identical core counts are involved, Intel has historically seemed to edge out AMD a bit in V-Ray, although in our CPU-only test here, the 5950X does manage to place ahead of the 12900K. But, with 16 high-performing cores, you’d expect that. What you might not expect is just how tight the delta is between that 16-core, and Intel’s 8P+8E design. Yet again, we can see enormous gains in performance from one generation of Intel CPU to the next.

Alright – let’s add a GPU to the mix:

Chaos Group V-Ray - Flowers CPU+GPU Rendering Performance (Intel 12th-gen Core)

We’ve seen countless examples over the years of how important a GPU can be to rendering performance, and V-Ray yet again backs that up. What’s interesting, though, is that while the CPU only render put the 5950X in the lead, the heterogeneous render gives an edge to both of the 12th-gen CPUs tested.

So far, we’ve seen Intel’s new chips perform extremely competitively in rendering, with the 12900K performing better against the 5950X than we were expecting. Because we like to be thorough, we have four more rendering tests on the following page, so let’s see if Intel can continue its seriously strong showing.

Rendering: Cinebench, Corona, LuxMark, POV-Ray

Cinebench

Maxon Cinebench - Multi-threaded Score (Intel 12th-gen Core)
Maxon Cinebench - Single-threaded Score (Intel 12th-gen Core)

When Cinebench R20 released a few years ago, the performance scaling we saw was a bit different than the previous version, as the new one made sure modern CPU architectures were being better taken advantage of. That being the case, AMD’s Zen CPUs were really impressing all over, especially as core counts grew. With Cinebench R23, Zen CPUs have typically won over Intel in each core-for-core test we’ve done.

Well, Alder Lake has sure thrown a wrench into AMD’s gears. In yet another surprising (to us) result, Intel’s new 8P+8E CPU design with the 12900K is strong enough to have it outperform AMD’s 16-core 5950X. We’re simply blown away by this. Those strong advantages carry over to the single-threaded test, as well.

We were without a Cinema 4D license when we got down to testing Alder Lake, so that’s your answer for why those additional tests are not here. We’re planning on taking a deeper look at C4D R25 soon, so we’ll take care of that before long.

Corona Renderer

Chaos Czech Corona Rendering Performance - Livingroom Scene (Intel 12th-gen Core)
Chaos Czech Corona Rendering Performance - Sales Gallery Scene (Intel 12th-gen Core)

With Corona Renderer, Intel’s top-end i9-12900K slightly beats out the 5950X in the living room scene, while AMD’s mainstream champion pulls further ahead in the sales gallery scene. All in all, Intel is still proving very impressive considering half of the 12900K’s cores are “energy-efficient”, while four of the 12600K’s are – and yet they are still performing really well.

LuxMark

LuxMark Food CPU Rendering Performance (Intel 12th-gen Core)
LuxMark Hall Bench CPU Rendering Performance (Intel 12th-gen Core)

Some renderers and scenes let the 16-core 5950X strut its stuff well, but LuxMark becomes one of the others that really seems to like Intel’s new chips. In both scenes, the 12900K places ahead of the 5950X. Most significant is still the fact that gen-over-gen results for Intel look really good – the 12900K soars past the 11900K in both tests.

If we saw these results out-of-context, we’d assume that LuxCoreRender using Intel’s Embree ray tracing kernels would be the reason for the lead, but other results here have proven it’s not only with Embree that Intel can leap ahead in rendering.

POV-Ray

POV-Ray 3.8 Multi-threaded Score (Intel 12th-gen Core)
POV-Ray 3.8 Single-threaded Score (Intel 12th-gen Core)

Wrapping up our rendering results, we’re seeing more of what we’ve largely seen throughout most of them: Intel’s 12th-gen chips are a force to be reckoned with. In both the single- and multi-threaded tests, the 12900K places in the top spot. In the multi-thread test, the 12900K pulls ahead of the 5950X further than in other rendering tests we’ve looked at, and the single-thread performance from all of the tested Intel chips put themselves on a different level entirely than Zen 3.

Synthetic: Procyon, Sandra


Procyon

As the previous handful of pages hopefully prove, we favor real-world benchmarks over synthetic ones, but the latter can still provide useful information at times. Since we’ve never given UL’s Procyon suite a go, we decided to this go-around, and are fairly happy with the results, and how it works.

For those unaware (as we were until recently), Procyon is in effect a replacement for PCMark. Its goal is to use real applications to run the same tests over and over, and generate a score based on the overall performance. The Video Editing test requires Adobe’s Premiere Pro, while the Photo Editing test requires both Lightroom and Photoshop. The Office Productivity test requires Microsoft Office (surprised?).

UL Procyon - Office Productivity Performance (Intel 12th-gen Core)

Intel has long pushed Office as a good benchmarking tool, and business-wise, it’s for good reason. Intel’s CPUs have largely been unbeatable in single-threaded operations, which are more common in office scenarios than multi-threaded ones. So, it’s not really a surprise that Intel’s latest-gen Core chips manage to stick to the top – but it’s an especially good look for a chip like the 12600K to outperform not just the previous-gen 11900K, but also the AMD chips here.

UL Procyon - Photo Editing Performance (Intel 12th-gen Core)

In our real-world Lightroom test, we saw AMD’s Ryzen 9 5950X lead the pack. We don’t see that reflected here, which could be because Photoshop was involved in this Procyon test in addition to Lightroom – so Intel’s single-threaded performance could have made up the slack. The gap between 5950X and 12900K was decently wide in the office productivity test, but the deltas are much tighter here.

UL Procyon - Video Editing Performance (Intel 12th-gen Core)

We saw Intel’s new chips perform exceptionally well in our real-world Premiere Pro tests a few pages back, and Procyon seems to agree that they are indeed pretty, pretty, pretttttty good.

Multimedia

SiSoftware Sandra 2020 - Multi-media Performance (Intel 12th-gen Core)

Most of Sandra’s CPU benchmarks will take great advantage of any architecture they’re given, and as many cores as they’re given. With the multimedia test, AMD really proves what the 5950X is capable of, with this being the first test we’ve seen where that chip scores quite a bit better than the 12900K. Similarly, we’re not seeing gen-over-gen Intel increases that are quite as stark as what we saw in our rendering tests, but it could be that future Sandra versions will utilize this unique Alder Lake architecture better.

Arithmetic

SiSoftware Sandra 2020 - Arithmetic Performance (Intel 12th-gen Core)

The multimedia test above showed the 5950X to place quite a good deal ahead of the 12900K, but when we math it up, both the 12900K and 5950X perform virtually the same. Again, just check out those generational improvements on the Intel side.

Cryptography

SiSoftware Sandra 2020 - Cryptography (High) Performance (Intel 12th-gen Core)
SiSoftware Sandra 2020 - Cryptography (Higher) Performance (Intel 12th-gen Core)

AMD has done well with CPU-based crypto for a while, and nothing has changed here. It’s interesting to see the 12900K and 11900K both perform identically in the AES256 + SHA256 test, another hint that Sandra may not be using the new architecture to its full potential. More interesting is the secondary result, for AES256 + SHA512 – here, the 11900K wins thanks to its inclusion of AVX-512. That’s something that also allows the 6-core 11600K to match AMD’s 16-core 5950X.

Memory Bandwidth

SiSoftware Sandra 2020 - Memory Bandwidth (Intel 12th-gen Core)

As we covered before, Intel’s new 12th-gen platform supports either DDR4 or DDR5 memory – entirely dependent on whichever one is equipped on the motherboard you buy. Our DDR5 memory kit didn’t arrive in time for testing, so all of the results above reflect our DDR4-3200 CL16 configuration.

Oddly, this is another case where Intel’s last-gen CPUs performed better than the current-gen, although based on our real-world performance, it doesn’t seem like the degraded memory performance has hurt the 12th-gen chips much. We’re eager to see how this chart changes up as soon as we get our DDR5 testing completed.

Power Consumption & Final Thoughts

Power Consumption (Intel 12th-gen Core)

Throughout all of our performance results, the overall impression is that Intel’s 12th-gen Core i9-12900K can either match or beat out AMD’s Ryzen 9 5950X more often than not. In the above power chart, we can begin to understand one thing that’s helping Intel: increased power usage.

Looking at the 12th- and 11th-gen parts first, we see that the 12900K uses only a bit more power than the 11900K, despite having eight additional cores. The 12600K result is a little more baffling, as it’s managing to use less power than the 11600K. We admittedly didn’t sanity check this one, since we would have had to reinstall the old motherboard and gear into the same chassis, but our testing procedure is so stringent, we believe we’d see the same thing if we did go through with sanity checking.

Nonetheless, gen-over-gen, Intel’s new chips seem to have decent power usage traits, but when you compare to AMD’s results, perceptions can really change. While the 12900K can regularly beat out the 5950X in many tests, it draws 125W more from the wall. The 12600K draws 50W more than the 5600X, but that almost seems modest in comparison!

Final Thoughts

Since Techgage launched over fifteen-years-ago, we’ve used and thoroughly benchmarked every Intel CPU generation to come along. After a while, things admittedly became a little boring, as we were simply not seeing the kind of gains from gen-to-gen that we were hoping to – and this all happened during a time when AMD’s chips weren’t wowing many people, either. Intel’s 12th-gen, aka Alder Lake, is downright exciting in comparison to other recent Intel Core launches.

When AMD first launched Zen, there was a lot to love. We finally saw the company deliver a product that deserved lots of attention. Vast improvements over previous generation AMD CPUs were made, and thankfully, the company iterated nicely on those improvements over time. When Zen 3 launched, AMD finally managed to catch up to Intel’s single-threaded performance, and it seemed like Intel would have a real mountain to climb in order to properly strike back.

Intel 12th-gen Core Alder Lake Review Kit

Well, Intel’s recent troubles are no secret. The company seemed to get into a rut where CPU development was concerned, and some wondered if AMD’s sudden resurgence was too much to handle in a swift manner. Alder Lake proves that we really had little reason to ever doubt the company. As with the original Zen release, we see vast improvements with Alder Lake, with performance hikes that are likely higher than most people would have expected. That seems like a successful launch to us.

As hinted at earlier, we didn’t expect the Core i9-12900K to perform strongly enough to take on AMD’s Ryzen 9 5950X. After all, we’re talking about an 8+8 CPU design, vs. a native 16-core. Well, as it happens, Intel’s architectural gains combined with those new energy-efficient cores means the company doesn’t currently need the same number of “high-performance” cores as its competitor to deliver extremely competitive performance.

While we didn’t touch on gaming for this article, we’ve seen some results from around the web that show Intel again proves strong there, giving Intel some validation for when it calls the 12th-gen the best gaming CPU series. For our first look at performance, we wanted to focus squarely on workstation and creator use, but will tackle gaming once we get down to our DDR4 vs. DDR5 testing.

As is probably obvious by now, there’s a lot to like about Intel’s 12th-gen chips, but there are some caveats to bear in mind. As it stands today, DDR5 is horrifically hard to acquire, at least if you don’t want to pay exorbitant prices to scalpers. At this point in time, we’d suggest any creator looking at these 12th-gen CPUs to consider DDR4 instead.

Intel 12th-gen Core i9-12900K CPU and ASUS' Z690-PLUS WiFi D4 Motherboard

From what we’ve seen, DDR4 vs. DDR5 performance isn’t so different that DDR4 should be ignored. You can currently get more DDR4 memory for less – at SRP, you could save at least $200 on a 64GB kit, but that actually assumes you could buy DDR5 in the first place. Launch DDR kits are always spec’d more modestly than later-gen kits, and are generally more expensive, so we’d expect that in the months and year ahead, we’ll see DDR5 kits create a bigger gap between it and DDR4.

We’d be remiss to talk power here, because as the results above have shown, Intel’s 12th-gen chips are not exactly modest when it comes to satiating their appetite for power. The 12900K did in fact beat out the 5950X in most of our tests, but it drew so much more power to attain it. In addition, because of all of that extra power use, a great CPU cooler will be needed for a chip like the 12900K. Think triple-rad AIO at least.

With this new platform, pricing is going to be a big consideration all over. The i9-12900K is priced at $589-599 USD by Intel, which is a bit of a hike over the i9-11900K, which was priced at $539-549 (all prices based on per 1,000 units, so you could expect to pay more at retail.)

In even more price talk, Intel Z690 motherboards are currently more expensive than last-gen boards overall, but you can still get some that look pretty decent for $250 and less. The cheapest Z690 board we could find so far was priced at $220 – MSI’s PRO Z690-A. If you want as feature-rich a motherboard as possible, you’ll be spending at least almost double that.

Ahead of receiving the 12th-gen chips to test, we felt like the insistence on using Windows 11, and also the need of a hardware-based Thread Director, would make things complicated while testing. We were happy to see that wasn’t the case at all – every one of our tests seemed to scale well, and for general OS use, we never noticed anything that felt unusual. That said – you will absolutely want to use Windows 11 for the best possible performance; Windows 10 and Linux will be OK for the most part, but will have degradation in some areas for the time-being.

One last thing: it might surprise you that the Core i9-12900K outperformed the Ryzen 9 5950X more often than not in our testing… while overheating. Unfortunately, a dual-sized radiator AIO isn’t quite enough to sufficiently cool a chip like the 12900K at max load, and with stress tests, we could easily hit 100C. That’s not ideal, so we’re in the process of getting some new gear that will do a better job of cooling for our next round of tests. Still, it’s quite something when a sweaty Intel chip still manages to outpace a properly cooled AMD competitor.

Pros

Cons

Copyright © 2005-2022 Techgage Networks Inc. - All Rights Reserved.