Intel’s high-end desktop dominance has been under siege ever since the name Threadripper was first uttered. The blue team’s first round of Core X CPUs, topping out with the i9-7980XE’s 18 cores and 36 threads, held a front against the first wave of AMD’s high-end Ryzens, but PC builders groused nonetheless—and rightfully so. Where did the solder that had joined heat spreaders to dies of past high-end Intel CPUs go, especially on a $2000 chip? Why did X299 motherboards have problems keeping their VRMs cool enough under extreme loads with the platform’s highest-core-count CPUs? Why did quad-core parts exist for the X299 platform at all?
Next to the unrestricted, segmentation-free approach of Threadripper CPUs and the X399 motherboards, the X299 platform and the breadth of the CPUs that could light it up looked by turns stingy and scattered. The Core i7-7800X and i7-7820X offered only 28 PCIe lanes from the CPU, compared to the 44 from the Core i9-7900X and better CPUs in the lineup. The entry-level Core i7-7800X didn’t even benefit from Turbo Boost Max 3.0, one of the headlining innovations of the Core X lineup. Worse, the 16 CPU-powered PCIe lanes from the Kaby Lake-powered Core i5-7640X and Core i7-7740X required motherboard makers to employ complex lane-switching schemes even on high-end mobos that seemed unlikely to ever play host to their four cores.
Intel may have had good intentions in providing builders a wide range of choices and an upgrade path in putting together high-end systems, but the initial headaches of X299 suggested that strategy had stretched the platform a bit too far.
AMD didn’t stand still with its high-end desktop CPUs in the intervening time, either. The Threadripper 2990WX didn’t just challenge the i9-7980XE in some workloads—it actually beat Intel’s highest-end desktop chip in some tasks for less money (though not in every test). Glancing though that blow may have been, the fact that AMD was even able to lay a finger on Intel’s high-end desktop performance crown was an indignity unimaginable just a couple of years ago. The ball has been in Intel’s court since, and that brings us to the new range of Core X CPUs launching today.
Base clock speed (GHz) |
Peak single-core boost speed (GHz) |
Turbo Boost Max 3.0 peak speed (GHz) |
Cores/ threads |
TDP (W) |
L3 cache (MB) |
CPU PCIe lanes |
Memory support |
Price | |
i9-9980XE | 3.0 | 4.4 | 4.5 | 18/36 | 165 | 24.75 | 44 | Four channels DDR4-2666 |
$1979 |
i9-9960X | 3.1 | 16/32 | 22 | $1684 | |||||
i9-9940X | 3.3 | 14/28 | 19.25 | $1387 | |||||
i9-9920X | 3.5 | 12/24 | $1189 | ||||||
i9-9900X | 10/20 | $989 | |||||||
i9-9820X | 3.3 | 4.1 | 4.2 | 16.5 | $889 | ||||
i7-9800X | 3.8 | 4.4 | 4.5 | 8/16 | $589 |
Intel isn’t classifying these chips as anything other than members of the Skylake family in its official materials, but that nonchalant code-naming scheme hides a range of under-the-hood improvements in the i9-9980XE and its stablemates. These chips benefit from some of the improvements in both eighth-gen and ninth-gen Coffee Lake mainstream CPUs.
First off, these new high-end chips are fabricated on Intel’s 14-nm++ process. 14-nm++ allows Intel’s engineers to lay down transistors that can be driven harder for better performance in exchange for only a slight increase in leakage current. In short, we can expect a better performance foundation for these CPUs without drastic increases in power consumption, and that reinforcement comes out in some minor clock-speed adjustments from top to bottom. Intel now specifies a Turbo Boost Max 3.0 speed of 4.5 GHz across the board for these chips (save for the Core i9-9820X and its victim-of-segmentation 4.2-GHz TBM 3.0 speed). Depending on the chip in question, peak Turbo Boost speeds have also increased anywhere from 100 MHz to 200 MHz (again excluding the odd-man-out i9-9820X).
Indeed, the existence of the Core i9-9820X suggests the product managers in charge of revitalizing the Core X lineup couldn’t keep the segmentation goblins entirely at bay. Those mischief-makers managed to get one weird chip into the new lineup. The i9-9820X has the 10 cores and 20 threads of its immediate superior, the i9-9900X, but in exchange for a $100 lower suggested price, it loses 2.25 MB of L3 cache, 300 MHz of peak Turbo Boost speed from any given core, and 300 MHz of Turbo Boost Max 3.0 speed from the two best cores on the chip. Perhaps this CPU is meant for overclockers trying to get ahold of 10 Skylake cores for as little cash as possible. For most high-end builders, though, we’d guess the extra $100 for the much-better-on-paper i9-9900X isn’t going to be a major obstacle.
For overclockers who do want to try and push those 10 cores to their limit, Intel has come to its senses about the material it uses to conduct heat from chip to cooler. Following in the footsteps of the Core i9-9900K and friends, refreshed Core X CPUs enjoy the return of solder thermal interface material (TIM). In tandem with the large dies that naturally arise from putting as many as 18 cores on a CPU, that metallic TIM could let overclockers cool these chips without resorting to the risks of delidding and repasting with more thermally conductive materials than Intel’s factory goop.
The benefits of big chips for heat transfer and cooling could apply to the entire Core X refresh lineup, too. You’ll note that many chips in this new lineup sport more L3 cache than the 1.375 MB per Skylake Server core would naturally add up to. That’s because Intel can disable cores on these chips without turning off the associated slice of L3 cache on the mesh that joins cores and shared caches together, and that fact offers a tantalizing clue as to the silicon being used to make these chips.
As Ian Cutress at Anandtech has pointed out, the fact that refreshed Core X CPUs boast more L3 cache than active cores would normally offer—especially at the low end—suggests that Intel is using its high-core-count (HCC) Skylake Server die as the starting point for all of the chips in this lineup. Another bit of backup for that suggestion comes from the fact that all Core X CPUs now come with a 165-W TDP, a figure previously reserved for only the four highest-core-count CPUs in the Core X lineup. In tandem with solder TIM and the process improvements of 14-nm++, the use of a uniformly large die across the refreshed Core X lineup could offer better overclocking potential across the board, thanks to the fact that there’s more surface area that can be joined to the heat spreader above by way of that solder.
One final segmentation demon that’s been banished from refreshed Core X CPUs is the 28-PCIe-lane switch that used to get flipped on Intel’s entry-level high-end parts. Every refreshed Core X part offers 44 CPU-powered PCIe lanes for motherboard makers to distribute as they please. While that figure still doesn’t match the 60 PCIe lanes AMD fans can enjoy from every Threadripper CPU, across-the-board consistency from the blue team is a welcome olive branch for I/O- or peripheral-hungry builders who might not have wanted to spend extra for cores, threads, and consequent cooling hardware that might not have been needed on the road to expansion bliss.
Getting to know the Core i9-9980XE
Intel only sent us one chip to test today: the highest-end Core i9-9980XE. At $1979, this chip is the latest Extreme Edition standard-bearer. If you want the very best of refreshed Core X, this chip is it. Rather than the vague spec table that Intel provides, let’s dig into the i9-9980XE’s per-core Turbo table and see just what the combination of process tech improvements and solder buys us versus the outgoing Core i9-7980XE.
Number of cores active | 1 | 2 | 3 | 4 | 5 | 6 | 7 | 8 | 9 | 10 | 11 | 12 | 13 | 14 | 15 | 16 | 17 | 18 |
i9-9980XE Turbo Boost speeds (GHz) | 4.5 | 4.5 | 4.2 | 4.2 | 4.1 | 4.1 | 4.1 | 4.1 | 4.1 | 4.1 | 4.1 | 4.1 | 3.9 | 3.9 | 3.9 | 3.9 | 3.8 | 3.8 |
i9-7980XE Turbo Boost speeds (GHz) | 4.2 | 4.2 | 4.0 | 4.0 | 3.9 | 3.9 | 3.9 | 3.9 | 3.9 | 3.9 | 3.9 | 3.9 | 3.5 | 3.5 | 3.5 | 3.5 | 3.4 | 3.4 |
Our test motherboard suggests the i9-9980XE incorporates its two Turbo Boost Max 3.0-capable cores into the up-front Turbo table of the chip, rather than leaving the operating system to work with an Intel driver to identify and pin workloads to those cores. Around the time the first round of Skylake-X CPUs launched, Intel said it was working with Microsoft to expose favored cores like those for Turbo Boost Max 3.0 to the operating system directly. Perhaps the incorporation of those 4.5-GHz bins into the Turbo table (rather than the official 4.4-GHz peak Turbo Boost 2.0 speed of the chip) is one puzzle piece in that larger effort.
From three to 12 active cores, the i9-9980XE boasts only a 200-MHz clock-speed boost over its predecessor. Once we reach 15 to 18 active cores, however, Intel seems to have taken advantage of some of the headroom the 14-nm++ process and solder TIM afford to keep Turbo Boost clocks as much as 400 MHz higher than those of the i9-7980XE. We’ll have to check just how much power that move burns later on, but a 400-MHz bump across 18 cores and 36 threads is a juicy improvement. Let’s see just how that improvement plays out in our test suite.
Our testing methods
As always, we did our best to deliver clean benchmarking numbers. We ran each benchmark at least three times and took the median of those results. Our test systems were configured as follows:
Processor | Intel Core i7-8700K | Intel Core i5-8400 | Intel Core i7-9700K | Intel Core i9-9900K | |
CPU cooler | Corsair H100i Pro 240-mm closed-loop liquid cooler | ||||
Motherboard | Gigabyte Z390 Aorus Master | ||||
Chipset | Intel Z390 | ||||
Memory size | 16 GB | ||||
Memory type | G.Skill Flare X 16 GB (2x 8 GB) DDR4 SDRAM | ||||
Memory speed | 3200 MT/s (actual) | ||||
Memory timings | 14-14-14-34 2T | ||||
System drive | Samsung 960 Pro 512 GB NVMe SSD |
Processor | AMD Ryzen 7 2700X | AMD Ryzen 5 2600X | ||
CPU cooler | EK Predator 240-mm closed-loop liquid cooler | |||
Motherboard | Gigabyte X470 Aorus Gaming 7 Wifi | |||
Chipset | AMD X470 | |||
Memory size | 16 GB | |||
Memory type | G.Skill Flare X 16 GB (2x 8 GB) DDR4 SDRAM | |||
Memory speed | 3200 MT/s (actual) | |||
Memory timings | 14-14-14-34 2T | |||
System drive | Samsung 960 EVO 500 GB NVMe SSD |
Processor | Threadripper 2950X | Threadripper 1920X | Threadripper 2920X | Threadripper 2970WX | Threadripper 2990WX | |
CPU cooler | Enermax Liqtech TR4 240-mm closed-loop liquid cooler | |||||
Motherboard | Gigabyte X399 Aorus Xtreme | |||||
Chipset | AMD X399 | |||||
Memory size | 32 GB | |||||
Memory type | G.Skill Flare X 32 GB (4x 8 GB) DDR4 SDRAM | |||||
Memory speed | 3200 MT/s (actual) | |||||
Memory timings | 14-14-14-34 1T | |||||
System drive | Samsung 970 EVO 500 GB NVMe SSD |
Processor | Core i7-7820X | Core i9-7900X | Core i9-7960X | Core i9-7980XE | Core i9-9980XE | |||
CPU cooler | Corsair H100i Pro 240-mm closed-loop liquid cooler | Corsair H110i GT 280-mm CLC | ||||||
Motherboard | Gigabyte X299 Designare EX | |||||||
Chipset | Intel X299 | |||||||
Memory size | 32 GB | |||||||
Memory type | G.Skill Flare X 32 GB (4x 8 GB) DDR4 SDRAM | |||||||
Memory speed | 3200 MT/s (actual) | |||||||
Memory timings | 14-14-14-34 1T | |||||||
System drive | Intel 750 Series 400 GB NVMe SSD |
Our test systems shared the following components:
Graphics card | Nvidia GeForce RTX 2080 Ti Founders Edition |
Graphics driver | GeForce 411.63 |
Power supply | Thermaltake Grand Gold 1200 W (AMD) Seasonic Prime Platinum 1000 W (Intel) |
Some other notes on our testing methods:
- We tested the Core i9-9980XE in both stock and overclocked configurations. Our overclocked settings used a 45x multiplier for a 4.5-GHz all-core result.
- All test systems were updated with the latest firmware, graphics drivers, and Windows updates before we began collecting data, including patches for the Spectre and Meltdown vulnerabilities where applicable. As a result, test data from this review should not be compared with results collected in past TR reviews. Similarly, all applications used in the course of data collection were the most current versions available as of press time and cannot be used to cross-compare with older data.
- Our test systems were all configured using the Windows Balanced power plan, including AMD systems that previously would have used the Ryzen Balanced plan. AMD’s suggested configuration for its CPUs no longer includes the Ryzen Balanced power plan as of Windows’ Fall Creators Update, also known as “RS3” or Redstone 3.
- Unless otherwise noted, all productivity tests were conducted with a display resolution of 2560×1440 at 60 Hz. Gaming tests were conducted at 1920×1080 and 144 Hz.
Our testing methods are generally publicly available and reproducible. If you have any questions regarding our testing methods, feel free to leave a comment on this article or join us in the forums to discuss them.
Memory subsystem performance
The AIDA64 utility includes some basic tests of memory bandwidth and latency that will let us peer into the differences in behavior among the memory subsystems of the processors on the bench today, if there are any.
Some quick synthetic math tests
AIDA64 also includes some useful micro-benchmarks that we can use to flush out broad differences among CPUs on our bench. The PhotoWorxx test uses AVX2 instructions on all of these chips. The CPU Hash integer benchmark uses AVX and Ryzen CPUs’ Intel SHA Extensions support, while the single-precision FPU Julia and double-precision Mandel tests use AVX2 with FMA.
Javascript
The usefulness of Javascript microbenchmarks for comparing browser performance may be on the wane, but these tests still allow us to tease out some single-threaded performance differences among CPUs. As part of our transition to using the Mechanical TuRk to benchmark our chips, we’ve had to switch to Google’s Chrome browser so that we can automate these tests. Chrome does perform differently on these benchmarks than Microsoft Edge, our previous browser of choice, so it’s vitally important not to cross-compare these results with older TR reviews.
WebXPRT 3
The WebXPRT 3 benchmark is meant to simulate some realistic workloads one might encounter in web browsing. It’s here primarily as a counterweight to the more synthetic microbenchmarking tools above.
WebXPRT isn’t entirely single-threaded—it uses web workers to perform asynchronous execution of Javascript in some of its tests.
Compiling code with GCC
Our resident code monkey, Bruno Ferreira, helped us put together this code-compiling test. Qtbench records the time needed to compile the Qt SDK using the GCC compiler. The number of jobs dispatched by the Qtbench script is configurable, and we set the number of threads to match the hardware thread count for each CPU.
File compression with 7-Zip
The free and open-source 7-Zip archiving utility has a built-in benchmark that occupies every core and thread of the host system.
Disk encryption with Veracrypt
Cinebench
The evergreen Cinebench benchmark is powered by Maxon’s Cinema 4D rendering engine. It’s multithreaded and comes with a 64-bit executable. The test runs with a single thread and then with as many threads as possible.
Blender
Blender is a widely-used, open-source 3D modeling and rendering application. The app can take advantage of AVX2 instructions on compatible CPUs. We chose the “bmw27” test file from Blender’s selection of benchmark scenes to put our CPUs through their paces.
Corona
Corona, as its developers put it, is a “high-performance (un)biased photorealistic renderer, available for Autodesk 3ds Max and as a standalone CLI application, and in development for Maxon Cinema 4D.”
The company has made a standalone benchmark with its rendering engine inside, so it’s a no-brainer to give it a spin on these CPUs.
Indigo
Indigo Bench is a standalone application based on the Indigo rendering engine, which creates photo-realistic images using what its developers call “unbiased rendering technologies.”
V-Ray
Handbrake
Handbrake is a popular video-transcoding app that recently hit version 1.1.1. To see how it performs on these chips, we converted a roughly two-minute 4K source file from an iPhone 6S into a 1920×1080, 30 FPS MKV using the HEVC algorithm implemented in the x265 open-source encoder. We otherwise left the preset at its default settings.
SPECwpc WPCcfd
Computational fluid dynamics is an interesting and CPU-intensive benchmark. For years and years, we’ve used the Euler3D benchmark from Oklahoma State University’s CASElab, but that benchmark has become more and more difficult to continue justifying in today’s newly-competitive CPU landscape thanks to its compilation with Intel tools (and the resulting baked-in vendor advantage).
We set out to find a more vendor-neutral and up-to-date computational fluid dynamics benchmark than the wizened Euler3D. As it happens, the SPECwpc benchmark includes a CFD test constructed with Microsoft’s HPC Pack, the OpenFOAM toolkit, and the XiFoam solver. More information on XiFoam is available here. SPECwpc allows us to yoke every core and thread of our test systems for this benchmark.
SPECwpc NAMD
The SPECwpc benchmark also includes a Windows-ready implementation of NAMD. As its developers describe it, NAMD “is a parallel molecular dynamics code designed for high-performance simulation of large biomolecular systems. Based on Charm++ parallel objects, NAMD scales to hundreds of cores for typical simulations and beyond 500,000 cores for the largest simulations.” Our ambitions are considerably more modest, but NAMD seems an ideal benchmark for our many-core single-socket CPUs.
Digital audio workstation performance
After an extended hiatus, the duo of DAWBench project files—DSP 2017 and VI 2017—return to make our CPUs sweat. The DSP benchmark tests the raw number of VST plugins a system can handle, while the complex VI project simulates a virtual instrument and sampling workload.
A very special thanks is in order here for Native Instruments, who kindly provided us with the Kontakt licenses necessary to run the DAWBench VI project file. We greatly appreciate NI’s support—this benchmark would not have been possible without the help of the folks there. Be sure to check out their many fine digital audio products.
A very special thanks also to RME Audio, who cut us a deal on one of its Babyface Pro audio interfaces to assist us with our testing. RME’s hardware and software is legendary for its low latency and high quality, and the Babyface Pro has exemplified those virtues over the course of our time with it.
We used the latest version of the Reaper DAW for Windows as the platform for our tests. To simulate a demanding workload, we tested each CPU with a 24-bit depth and 96-KHz sampling rate, and at two ASIO buffer depths: 96, the lowest our interface will allow at a 96 KHz sampling rate, and 128. In response to popular demand, we’re also testing two buffer depths at a sampling rate of 48 KHz: 64 and 128. We added VSTs or notes of polyphony to each session until we started hearing popping or other audio artifacts.
Apologies for the lack of results at 96 KHz and a buffer depth of 96 here. Thanks to something in the chain of Reaper, Windows 10, and our ASIO driver, our many-core CPUs couldn’t run the 96-96 test at all—we got popping and crackling from the get-go.
Conclusions
Let’s summarize the reams of data on the preceding pages using one of our infamous scatter plots. To more accurately represent each chip’s price-to-performance ratio, we used real-world pricing data from Newegg where it was available and manufacturers’ suggested prices where it wasn’t.
For the straight-and-narrow stock-clocked system, the story of the Core i9-9980XE is a simple one. Where Intel CPUs were already superior to the competition, the i9-9980XE offers some nice performance improvements. In applications that can light off Threadripper WX CPUs’ rocket boosters, the i9-9980XE’s under-the-hood refinement can’t overcome the Threadripper 2990WX’s sheer core-count advantage.
Overclock the i9-9980XE, though, and the 18-core chip both extends its leads considerably and closes some of the gaps the 2990WX opens. Part of that performance comes courtesy of Intel’s decision to resume soldering the heat spreaders to the die on its refreshed Core X chips. We couldn’t take our i9-7980XE past about 4.4 GHz without fighting thermal limits, but thanks in part to the reintroduction of solder TIM, we were able to push the i9-9980XE to an impressive 4.5 GHz on all 18 cores without calling the fire department—all with attainable, off-the-shelf cooling hardware. Hallelujah for that.
Presuming the potential errors that overclocking might introduce are tolerable in an i9-9980XE system, those willing to deploy custom liquid cooling loops may find that they can take all 18 cores of this chip to the limits of the underlying silicon, not just what paste thermal interface material might allow. Our casual overclocking suggests those could be some exciting limits to probe. Folks who don’t want to push the limits of their i9-9980XEs may find it easier to cool their chips quietly, too.
The only complaint some may harbor about Intel’s refreshed Core X chips is that the company didn’t see any cause to cut prices on its high-end chips this time around. The reason, I imagine, is that AMD and Intel are trying to sell high-end desktop buyers two different stories of performance as competition heats up at the top of the CPU heap.
In the $1200-and-up range we’re concerned with today, AMD seems plenty willing to make chips that really rip in workloads where sheer core count dominates, like rendering and some varieties of scientific computing. The tradeoff is that Threadripper WX chips can fall far behind in other tasks. That inconsistency leads to Threadripper WX chips’ lower-than-might-be-expected standings in our overall value chart, even as they excel in a few particular workloads.
Intel, for its part, seems unwilling to create chips with any corner cases, even if that means it can’t offer as many raw cores and threads as AMD can for the dollar. Core X CPUs may not take home the gold in every workload, but they’ll never put owners in a position where they can’t run certain tasks (or subsets of tasks) acceptably well, either. The extra money Intel CPUs command per core, then, is essentially insurance that you won’t be left hanging if the idea of inconsistent performance in any potential workload bothers you.
If you know that a Threadripper WX CPU benefits your work and don’t care about the cases where it might not, then those chips can still be screaming performance bargains. Every person needs different things from a PC, so check around and see how your work maps onto AMD and Intel’s high-end desktop platforms before making the leap.
If you have less than $1000 to spend on a CPU, Intel’s latest Core X chips might not move the needle much (presuming you even need four channels of memory and gobs of CPU-powered PCIe lanes to begin with). We already know that the Threadripper 2920X and Threadripper 2950X offer better performance than the Core i9-7900X in many heavy-duty workloads for less money, and 10 higher-clocked Skylake-X cores may not close that gap much in the case of the $889 i9-9820X or $989 i9-9900X. That’s before we consider the $650-ish price tag on the remaining stock of Threadripper 1950X chips, as that processor remains quite formidable in its own right.
If the Core i9-9980XE’s time on our test bench is any indication, though, the reintroduction of solder TIM, the advantages of the 14-nm++ process, and the end of PCIe lane segmentation from the CPU should make all Core X CPUs more attractive to those that want or need what they offer. Intel’s revised 10-core chips will almost certainly overclock better than Threadripper parts, and they should still hold a slight edge in whatever heights of high-refresh-rate gaming a builder might foolishly try to scale on high-end desktop platforms. It’s a shame we didn’t get some of those cheaper Core X chips to test, as we imagine many high-end desktop-building enthusiasts will be more interested in their performance than that of the halo i9-9980XE.
Ultimately, AMD’s X399 platform still offers higher core counts for the dollar and fewer restrictions on RAM capacity or types (including ECC support), but thanks to that renewed competition, fans of the blue team are undoubtedly getting more for their money than they were a year ago, whether in higher clocks, a better process technology, the potential for cooler operation, or more CPU-connected PCIe lanes in lower-end parts.
Whether those improvements are enough to draw buyers back into the Skylake-X fold as excitement builds around AMD’s next-generation server CPUs—and the potential Threadrippers derived from them—is a chapter yet to be written. For the moment, though, Intel is doing its best to put a happy ending on this chapter of its highest-end desktop chips, and a Core i9-9980XE ticking away at 4.5 GHz or better will make even the most demanding enthusiasts quite happy indeed.