Intel’s 10th gen Core i9-10900Ok is—indubitably—precisely as Intel has described it: “the world’s fastest gaming CPU.”
Intel’s drawback has been weaknesses outdoors of gaming, and its total efficiency worth in comparison with AMD’s Ryzen 3000 chips. With the Core i9-10900Ok, Intel doesn’t seem like eliminating that hole, nevertheless it may get shut sufficient that you simply won’t care.
What is Core i9-10900Ok?
Despite its 10th-gen naming, Intel’s latest desktop chips proceed to be constructed on the corporate’s growing older 14nm course of. How previous is it? It was first used with the 5th-gen Core i7-5775C desktop chip from 2015. Many tips, optimizations, and far binning later, now we have the flagship client Core i9-10900K, announced April 30. The CPU options 10 cores and Hyper-Threading for a complete of 20 threads, with an inventory value of $488.
The Core i9-10900Ok does convey a number of adjustments. Intel officers stated the chip makes use of a thinner die and thinner solder thermal interface materials (STIM) to enhance thermal dissipation. Intel additionally needed to make a thicker warmth spreader (that little steel lid to maintain you from crushing the fragile die).
Gordon Mah UngA 10th gen Core i9-10900Ok (proper) subsequent to an Eighth-gen Core i7-8700Ok (left). There are delicate variations within the “wings” that the load plate clamps onto, and the notches within the substrate are on reverse corners.
Why make the die and STIM thinner, however the warmth spreader thicker? The purpose is value. Intel stated it needed to hold the peak of the warmth spreader on all of its CPUs the identical so that they’d be suitable with current cooling {hardware}. Intel officers did say the supplies used for the warmth spreader assist compensate for that compromise, so total the brand new chip is best at energy dissipation.
IntelThe 10th-gen Comet Lake S CPUs characteristic a thinner die and thermal interface materials, plus a thicker warmth spreader to assist enhance warmth dissipation.
A brand new socket?!
It’s true: Intel’s new 10th-gen CPUs convey with them a brand new LGA1200 socket that’s—after all—incompatible with the earlier Ninth-gen CPUs. Intel took flack for introducing a brand new chipset with its Eighth-gen desktop chips that was incompatible with the earlier era, so you may perceive the anger for individuals who simply wish to improve the CPU, not additionally the motherboard.
Gordon Mah UngThe bottom of an LGA1151 Eighth-gen Core i7-8700Ok (left), and an LGA1200 10th-gen Core i7-10900Ok (proper).
The LGA1200 socket and new X490 don’t appear to vary a lot. You nonetheless set up the CPU nearly the identical manner, and in case you have an current LGA1151 cooler, it’ll nonetheless match. Sadly, rumors of PCIe 4.0 on the X490 proved unfaithful, leaving Intel at an obstacle in comparison with Ryzen 3000 chips which have the quicker interface for SSDs and GPUs.
Gordon Mah UngA brand new LGA1200 socket is required to run a 10th-gen Comet Lake S CPU.
How We Tested
For this evaluation, we persist with Intel’s flagship, the $488 Core i9-10900Ok. Its pure competitor is AMD’s Ryzen 9 3900X with 12 cores and 24 threads. Its checklist value is $499, however its road value as of this writing is truly extra like $410 on Amazon. The Ryzen 9 3900X comes with an honest air cooler, too.
The solely different CPU we’d evaluate can be the Ryzen 9 3950X, but with a street price of $720 on Amazon (as of this writing) the mathematics doesn’t work. So we’ll persist with the 12-core Ryzen 9 3900X. It was examined on an MSI X570 MEG Godlike with 16GB of DDR4/3600 in dual-channel mode. We usually use the identical SSD on all platforms, however we really feel that’s unfair to AMD, which might run PCIe 4, so we used a Corsair MP600 PCIe 4.0 drive.
For the Core i9-10900Ok, we used an Asus ROG Maximus VII Extreme board with 16GB of DDR4/3200 in dual-channel mode and a Samsung 960P SSD.
Both methods used Windows 10 1909, equivalent GeForce RTX 2080 Ti Founders Edition playing cards, and NZXT Kraken X62 coolers with followers set to 100 p.c. Both boards have been run open-case, with matching desk followers blowing cool air immediately onto the graphics playing cards and the boards’ socket space. All methods used the identical drivers, the most recent UEFI’s, and the most recent Windows safety updates.
Due to time and different constraints, we ran the boards with MCE and PBO set to Auto, and 2nd-level XMP and AMP profiles chosen. While these manufacturing unit settings are past what’s inventory, we expect it’s near what a client see will see out of the field.
Gordon Mah UngThe 10th-gen Core i9-10900Ok sits within the LGA1200 socket of an Asus ROG Maximus Extreme VII.
Core i9-10900Ok Rendering Performance
We’ll kick this off the place we usually do: Maxon’s Cinebench R20. It’s a 3D modelling check constructed on the corporate’s Cinema4D engine, and it’s built-in into such merchandise as Adobe After Effects. Like most 3D modelling apps, extra cores and extra threads usually yield extra efficiency.
Our outcomes for the Core i9-10900Ok and Ryzen 9 3900X are recent, however we determined to sprinkle in outcomes from earlier evaluations for extra context. Although these older outcomes aren’t utilizing the most recent model of Windows, Cinebench may be very dependable. The R20 model makes use of AVX2 and AVX512 and takes about 3 times as lengthy to run because the older R15 model. That means any enhance efficiency ought to matter much less.
Remember these outcomes, as a result of for essentially the most half it received’t change an excessive amount of as we transfer by means of multi-threaded efficiency: Cores matter. The Ryzen 9 3900X’s 12 cores prevail over the the Core i9-10900Ok’s 10. If it’s any comfort, the most recent Core chip fares noticeably higher than the Core i9-9900Ok and Core i9-9900KS, which have been each handcuffed by their “mere” 8 cores.
IDGThe 10-core wasn’t anticipated to out gun AMD’s 12-core however at the least it’s nearer.
Switching Cinebench R20 to single-threaded efficiency, constrain the load to a single CPU core. The outcomes are so shut that no would or ought to care. We anticipated the Core i9-10900Ok to have a bit extra of an edge, however possibly it’s the luck of the draw.
IDGSingle-core efficiency among the many client chips is just about dead-even.
Corona is an unbiased photo-realistic 3D renderer, which suggests it takes no shortcuts in the way it renders a scene. It loves cores and threads, so the outcomes right here comply with the development, however the Core i9 finishes simply 7 p.c shy of the Ryzen 9. In Cinebench R20, the Ryzen 9 had a bigger 15-percent benefit.
IDGThe Corona renderer closes the hole shut between the Ryzen 9 and Core i9 to about 7 p.c.
The Chaos Group’s V-Ray Next is like Cinema4D’s engine, and it’s a biased rendererer—it takes shortcuts to complete initiatives so you may, , win an Academy Award like the V-Ray has. It loves thread rely, so guess what: The Ryzen 9 comes out about 14 p.c quicker than the Core i9.
IDGV-Ray is a biased renderer like Cinema4D and efficiency tracks carefully with the Ryzen 9 about 14 p.c quicker than the most recent Core i9.
Our final rendering check measures ray tracing efficiency utilizing the most recent model of POV Ray. The Ryzen is available in nearly 17 p.c quicker than the Core i9. That’s fairly near the thread-count benefit the Ryzen 9 has, which equals 20 p.c extra.
IDGPOV Ray places the Ryzen 9 forward of the Core i9 by nearly 17 p.c.
Switching POV Ray to single-threaded efficiency, the Ryzen 9 squeezes by the Core i9, which shocked us—we thought the Core i9 would take the lead.
IDGSingle-threaded efficiency of the Ryzen 9 is just below 3 p.c quicker than the Core i9.
Core i9-10900Ok Encoding Performance
Video encoding wants quick CPUs, too. That’s why we use the most recent model of HandBrake to transform a 4K video brief to 1080p utilizing H.265. Using the CPU to complete the duty, the 12-core Ryzen 9 completed with a 16-percent benefit over the 10-core Core i9. So far, that’s just about every thing we’ve anticipated.
IDGWe use Handbrake to transcode a 4K video utilizing the HEVC/H.265 codec.
Our subsequent check makes use of Cinegy’s Cinescore to evaluate CPU and GPU efficiency throughout a number of dozen broadcast business profiles from SD to 8K, utilizing codecs from H.264, to MPEG2, XDCAM and AVC Ultra in addition to Nvidia H.265 and Daniel 2. It runs completely in RAM to take away storage as a bottleneck. (Note: The model of Cinescore we use is older and now not works with out setting an older date on the PC—the model has timed out the codec license.)
While the Core i9 doesn’t win, it will get awfully near the Ryzen 9. This may imply Cinescore and its CODECs don’t care that a lot about thread rely, or the upper clock speeds of the Core i9 could also be of extra worth. Sorry, Ryzen 9 followers.
IDGIt’s too shut for consolation for the Ryzen 9: The Core i9 comes inside 3 p.c of its efficiency.
Our subsequent check is a video check, however not in a conventional sense. While normal encoding or transcoding isn’t all that good in the way it downsamples or upsamples video, Topaz Lab’s Video Enhance AI claims to have a look at each body and use machine studying inferencing to determine what is going to make every body look higher on the upscale, based mostly on finding out different movies. For the check, we take a two-minute household video shot on a Kodak video digicam and upscale it from 720p to 1080p, utilizing the app’s Gaia HQ preset.
This upscale would usually be executed on a GPU, the place it could be considerably quicker, however we ordered it to make use of the CPU cores for the upscale. Topaz Labs makes use of Intel’s personal OpenVINO for the Deep Learning. Doing a frame-by-frame upscale isn’t straightforward and nonetheless takes actually hours to finish. The Ryzen 9 finishes with a couple of five-percent benefit over the Core i9. Too shut for consolation, however a win is a win.
IDGTopaz Lab’s Video Enhance AI makes use of machine studying to determine the way to upscale video.
Core i9-10900Ok Compression Performance
Moving on to compression and decompression efficiency, we first use RARLab’s WinRAR. As with prior Ryzen CPUs, the result’s dangerous—a loss for the Ryzen 9 and an enormous win for the Core i9. The Ryzen structure has lengthy carried out poorly right here. In the built-in benchmark, the Core i9 is 82 p.c quicker.
IDGNo shock, the Core i9 walks away with an enormous win right here as a result of the Ryzen half simply runs poorly in WinRAR.
Switching WinRAR to single-core efficiency, nothing adjustments besides the Core i9’s win grows to a 194-percent benefit. We use WinRAR as a result of it’s price stating that some software program will closely penalize Ryzen’s microarchitecture. Intel has fielded software program assist to corporations for for much longer than AMD, and it exhibits.
IDGThe hole closes with the way more common (and free) 7Zip. We set the built-in benchmark to make use of the variety of threads obtainable to the CPU—on this case, 24 for Ryzen 9 and 20 for Core i9. The first result’s multi-core.
IDGThe multi-core efficiency of the Core i9 can’t sustain with the Ryzen 9 in 7Zip.
Decompressing efficiency, in response to the developer, leans closely on integer efficiency, department prediction, and instruction latency. Compressing efficiency leans extra on reminiscence latency, cache efficiency, and out-of-order efficiency. It doesn’t matter both manner, because the Core i9 falls to the Ryzen 9 in each areas. The Ryzen 9 is about 21 p.c quicker in decompression, and a whopping 44 p.c quicker in compression.
Moving to single-threaded efficiency, we see the Core i9’s fortunes reverse, with about 7 p.c quicker decompression and 17 p.c quicker compression.
IDGSingle-threaded 7Zip efficiency is a strong win for the Core i9.
Core i9-10900Ok Gaming Performance
Intel didn’t name the Core i9-10900Ok ‘the best CPU for multi-threaded performance’ as a result of it possible knew it wasn’t going to squeeze out the Ryzen 9 3900X. On the opposite hans, Intel’s chips have lengthy led in gaming efficiency, ever for the reason that first Ryzen was launched.
Rather than have you ever scroll by means of 16 charts, we mixed 16 outcomes into one megachart, from an inventory of video games run at various resolutions and settings. We’ll run by means of it from prime to backside.
In Far Cry New Dawn we see the Core i9 range from about 9 p.c to about 14 p.c over the Ryzen 9, relying on the decision and recreation setting. As you jack up the decision or the sport setting, the check is more and more GPU-bound.
Ashes of the Singularity: Escalation, lengthy the poster little one for DX12 efficiency, is hailed for truly utilizing the additional CPU cores obtainable to players right this moment. For this run, we use the Crazy high quality preset and choose the CPU-focused benchmark, which is meant to throw extra models into the sport. The outcome: a couple of 7.5-percent benefit for the Core i9-10900Ok.
Chernobylite, an early-access recreation, incorporates a benchmark to showcase its stunning graphics. Set to excessive, the place the sport will not be restricted by GPU efficiency, we see that acquainted 7.9-percent benefit for the Core i9 over the Ryzen 9. As we crank up the graphics from excessive to extremely, it turns into an more and more GPU-bound check.
The solely purple (AMD) bar longer than a blue (Intel) bar is in Civilization VI Gathering Storms—however sadly for the Ryzen 9, this explicit check measures how lengthy it takes for the pc to make a transfer, and shorter time is best. The Core i9 is about 6.5 p.c quicker.