The Intel Kaby Lake-X i7-7740X Review
This launch-day coverage isn’t going to be easy for us due to the fact that it entails a number of firsts, some of which are positive, others certainly aren’t. Parts of this introduction are also going to be a bit more personal than is typical for me but there’s good reason for that too: Intel’s launch of the Core-X series is unique in terms of scope, market conditions and their approach to press engagement.
These also happen to be some of the most controversial processors launched in recent memory and many of you are likely here to see whether or not we’re going to give Intel’s latest endeavor a failing grade since anything else would be unacceptable. Don’t deny it. If you’re here for fireworks, and I promise you there will certainly be some.
Before I get too far into this let’s get let the cat out of the bag: this is the first time in more than a decade we almost missed a launch-day review of a major new processor evolution. To say I’m disappointed would be an understatement of epic proportions but I won’t dwell upon it other than to simply explain the situation. However, we do have content here and now due to a benefactor who will remain anonymous. It just won’t be of the processor that Intel was hoping they would get coverage for. Instead of focusing primarily on the top-of-the-line i9-7900X, my focus here is solely upon the oddball 7740X, the Kaby Lake-X black sheep runt of this litter.
Due to a number of factors –some of which will be made apparent by the time you finish this introduction- Intel’s sampling of Core-X CPUs to the press was spotty at best. I’m writing this a mere two business days away from launch and I still don’t have Intel’s samples in-hand. Maybe by the time you read this we’ll be busy beavering away running even more benchmarks and figuring out where the flagship Core-X CPU lands in the grand scheme of things. However, rushing out benchmarks of that processor would be a disservice to our readers and could very well give false first impression (be it positive or negative) of a platform that will be a pretty big deal for enthusiasts.
Over the course of the last few weeks after Computex, I’ve been toying with a thought: have we reached “peak Intel”? Is this a time when the buying public is so fatigued by forced market segmentation, overly high pricing structures and overlapping platform capabilities that they look elsewhere? For the last decade or so, folks simply didn’t have a choice; either choose Intel or get saddled with an AMD architecture that just wasn’t all that competitive. Yet now AMD is back in the game with Ryzen and the upcoming high end desktop platform code-named Threadripper and the alternatives to the status quo look very promising. Our Computex coverage certainly pointed to that possibility since X399 articles and videos received triple the traffic of their X299 competitors.
Obviously these are the concerns that should have kept Intel up at night. However I get the distinct feeling they couldn’t have been bothered until the last possible moment, when AMD’s Threadripper salvo was already on its way downrange. Not only has Intel’s press engagement fallen by the wayside, but their Core-X parts and X299 platform are obviously being launched in an overly hurried manner. This is what happens when complacency sets in and one day you suddenly realize the mice have overrun your kitchen.
The ramifications of Intel’s knee-jerk reaction to Threadripper are evident everywhere today, with the launch initially slated for later this year being pulled in to mid-June. Core-X series processors are only available for preorder today while actual inventory will only arrive on June 26th or thereabouts. A key feature –that being VROC- won’t be rolled out until later this year or early next year. A large number of the announced i9 processors won’t even be available until next year, pointing towards them being a simple paper-bound hedge against whatever AMD has in store.
Adding insult to injury, some motherboards don’t even feature support for Kaby Lake X processors while others will have hallmark capabilities of Intel’s new X299 platform missing in action. There weren’t even enough processors to go around so only select sites and YouTube channels were targeted for sampling in time for today’s performance evaluations. The end result of all of this could be a lack of broad coverage, short term availability problems and the potential for significant confusion among end users. It’s a mess but one of Intel’s own making.
However none of this means you should automatically hop onto the X299 hate train like so many online personalities seem wont to do. There’s actually quite a lot to like here, such as expanded chipset capabilities (finally!), platform longevity and the possibility for some truly titanic performance metrics. What remains to be seen is how the associated processors line up in relation to their predecessors and AMD’s new stable of competitors.
Intel’s Skylake-X Under The Microscope
A few days ago I mentioned on Twitter this is likely one of the most confusing launches in recent history. I wasn’t kidding. Not only is Intel launching a brand new lineup of high end desktop (HEDT) processors in Skylake-X but they are also rolling the mid-level Kaby Lake architecture into the mix in the form of Kaby Lake-X. This is the first time their HEDT platform will feature two different processor architectures and that causes some issues due to wildly differing capabilities offered by each.
To give you a quick idea of this situation before I go into a few more of its nuances, everything from the PCI-E lane allotments to the PCH’s I/O layouts to the platform’s feature set will change depending on which processor you have installed. You will also need to be careful about memory installation since the underlying architecture of each processor handles DIMM assignments differently. I’m going to do my best to go through this one blow at a time but keep in mind there are still a lot of parallels you can draw between these new processors and their Broadwell-E predecessors.
One of the hallmarks of Skylake-X series processors is a carry-over from the aforementioned Broadwell-E generation: Intel Turbo Boost Max 3.0. Basically ITBM 3.0 determines which specific cores on your particular CPU are capable of running at higher speeds during lightly threaded workloads and accelerates their frequency level. It does so by actively monitoring temperatures and how many cores are being engaged in order to estimate input current and power consumption. If there is available headroom, the active core(s) will have more leeway to operate faster.
In essence this grants a third gear above and beyond the typical Base and Turbo rates we are all used to seeing but where Skylake-X processors differ is how many cores will boost to the stated ITBM 3.0 frequency. Whereas Broadwell-E featured a single “best” core, there’s now the possibility that up to two cores will reach significantly higher speeds. That in itself could prove to be a key differentiating factor in gaming where many of today’s most-popular titles tend to benefit from higher speed, lower core count designs. It also happens to highlight why CPUs in the Skylake-X family may prove to be poor value companions for gamers.
Skylake-X actually has several core architecture updates including a new Xeon-derived “Mesh” topology which seems to be very much like AMD’s Infinity Fabric. This so-called mesh effectively replaces the less scalable ring bus architecture with a highly adaptable interconnect backbone that uses a series of more direct paths between the I/O controllers, memory, cache and cores. This is supposed to not only reduce on-chip latency but it will also help Intel expand and contract Skylake-X’s core layout as needed with a minimum of intervention.
The move to this layout also improves accessibility to the chip’s on-die last level cache (LLC). Efficiently communicating with this large block of cache is a key aspect for application developers and in this case all o the cache banks can be accessed with a bare minimum of latency fluctuation. As a result, Intel has also been able to revise their caching hierarchy in a pretty meaningful way.
Previous generations of HEDT processors featured a shared last level cache size of up to 2.5MB per core and 256KB per core of integrated mid level cache. That structure is now changing with an LLC layout of up to 1.375 MB/core which may cause some issues for highly threaded workstation applications that tend to mercilessly pound away at that shared LLC block. However, the new low latency mesh interconnect could theoretically offset some of that bottleneck through more efficient communication.
In order to combat any perceived caching discrepancy, Intel has quadrupled the MLC within each core to a whopping 1MB. That means a 10 core processor will have a full 10MB of MLC to draw upon in demanding situations.
Let’s start things off with the thousand pound gorilla in the room: the four processors which sit atop the chart above but don’t have any real specifications other than core counts and stratospheric prices. In many ways these were announced as a hedge against AMD’s impending Threadripper and nothing more. At this point it looks like the i9-7920X, i9-7940X, i9-7960X and i9-7980XE will have their launches split between the second half of 2017 and Q1 2018. In the meantime, their true performance is anyone’s guess since final frequencies are still being determined by Intel’s labs. I won’t spend any more time on those until we’re closer to launch.
Moving towards the processors that will actually be available come launch and we have the i9-7900X which is meant to take over from the i7-6900K while offering more cores, additional threads, 4 more PCI-E lanes and higher clock speeds. As a matter of fact, above all else this processor highlights the downwards pressure a resurgent AMD is having upon Intel’s pricing structure. A cool thousand bucks is a lot to ask for any piece of computer hardware these days but the 10 core, 20 thread i9-7900X does seem to have a lot more performance on tap than its predecessor while costing about $100 less. With that being said I think this looks like a “fair” price simply because the 6900K was so over priced.
The i9-7900X’s feeds, speeds and power consumption point towards a broad refinement in Intel’s 14nm manufacturing process. Base, Turbo and Turbo 3.0 frequencies are respectively 300MHz, 800MHz and 500MHz faster than the similarly laid out i7-6950X. Meanwhile, supported memory speeds have increased to 2666MHz and TDP remains at 140W which is quite impressive for such a capable CPU.
Whereas Intel’s goal with the i9-7900X seems to be pretty clear cut, venture into the i9-78xx series lineup and things get muddy very, very fast. As a matter of fact, the first time I saw the specifications for the i7-7820X and i7-7800X, I thought there was a misprint but my fears were quickly confirmed after a quick email to Intel.
Let’s start with the i7-7820X, a processor that at first seems to follow directly in the footsteps of the i7-6850K. As is the trend for Skylake-X parts, the $600 7820X gets an upgrade to 8 cores and 16 threads but and operates at much higher Turbo and Turbo 3.0 frequencies. It also gets the benefit of that double-fisted best core ITBM 3.0 technology. Unfortunately, while the i7-6850K had access to 40 PCIe lanes, this new CPU gets castrated down to a mere 28 lanes which makes dual x16 graphics operation impossible without an expensive lane multiplier on the motherboard. Its officially supported memory speed gets cut to 2400MHz as well. So the question is simple: are you willing to sacrifice potential dual GPU throughput for a higher core count? If only Intel didn’t ask you to make that decision.
The i7-7800X had some massive shoes to fill since the i7-6800K was arguably one of the most popular SKU’s in the Broadwell-E lineup. Well it looks like when creating the replacement i7-7800X, Intel chould have trimmed a bit off of the i7-7820X but instead used a gas powered weed whacker. This CPU is cut down to the same 6 cores and 12 threads as its forefather, has 28 PCIe lanes and –no, that isn’t an error in the charts- even gets Turbo Boost Max 3.0 cut out of its equation. Luckily it does operate at up to 4.0GHz but if you are currently rocking an i7-6800K it looks like you’ll be safe from the upgrade bug for now.
The 7800-series are extremely important to the success of this platform for Intel since their specifications lead to direct competition against AMD’s higher-end Ryzen processors. The $599 i7-7820X will find itself going head to head against the less expensive yet frequency deficient 1800X and 1700X. Meanwhile the $389 6 core, 12 thread i7-7800X could find itself in tough company against the 16 thread 1700X and 1700.
While it does look like Intel is opening the door to a few more SKUs (looks like there will be an i5 lineup as well) by taking a hacksaw to their once-sensible HEDT lineup, the reason behind this seemingly nonsensical move may be straightforward. It can be summed up in two and a half words: Kaby Lake-X. Let’s jump ahead and see why that is.
Explaining the Unexplainable; Say Hello to Kaby Lake-X
If that past was any indication, our introduction to Intel’s new HEDT lineup would have started and stopped on the previous page. But this is 2017 and sowing confusion among your potential clients seems like the “in” thing to do among companies. Enter Kaby Lake-X, a series of processors that are being shoehorned into this launch in truly odd fashion and with capabilities that are nothing if not bewildering on a high end platform. You may have noticed that the features I mentioned on the last page were listed as “Skylake-X” and that’s because none of those improvements are making their way to Kaby Lake-X chips.
The reason for my slightly salty introduction to Kaby Lake-X is because adding them into an X299 system will cause all manner of changes to the platform’s core feature sets. Even though these CPUs still utilize the LGA 2066 socket, they only have dual channel memory so four of a typical X299 motherboard’s eight DIMM slots will become non-functional. They also have just on-chip 16 PCIe lanes so connectivity options are further limited beyond even the diminished i7-78xx series. Turbo Boost Max 3.0 isn’t available either and the new baseline architectural improvements of Skylake-X like its mesh interconnect and updated caching hierarchy have been consigned to the dustbin as well.
If all of these cuts make Kaby Lake-X sound strangely familiar, that’s because they should. You see, these new processors are carried over lock, stock and barrel from their LGA1151 stablemates. Think of them as standard Kaby Lake processors simply soldered onto an LGA2066 die and then given a slight bump in official memory support to 2666MHz.
Sitting atop of this particular lineup is the i7-7740X which, like the 7700K has four logical cores which account for eight processing threads. The base clock has received an insignificant bump of 100MHz to 4.3GHz in an effort to slightly differentiate this processor from one that has been on the market for the better part of six months now. At $339 even its cost aligns perfectly with the LGA1151-bound alternative.
The i7-7640X’s story is pretty much the same as its sibling. There’s a very minor 200MHz improvement in the Base Clock but nothing has been touched within the Turbo rates. Pricing aligns with the i7-7600K as well so there really isn’t anything here to differentiate the two other than the platform upon which they are used.
TDP will be an interesting factor with the Kaby Lake-X processors since they’re rated for 112W versus the frugal 91W of Kaby Lake-S. Naturally there will be some efficiency losses when moving to such a large die package but don’t take this to mean they’re going to consume significantly more power. Rather, the additional TDP headroom could conceivably given these two CPUs the ability to maintain higher speeds more frequently and also offer more overclocking. Maybe.
At this point you are probably wondering why these processors are even a thing. I actually don’t have any direct answer to that but Intel’s official line is that they give additional choices to buyers who may want to buy into the 2066 platform and do so at a reduced cost. That explanation provides few insights into the situation given the fact that installing a Kaby Lake-X processor onto an X299 motherboard won’t net you any additional features over Z270.
The addition of Kaby Lake-X has also necessitated the aforementioned nerfing of i7-7800-series chips since Intel sort of backed themselves into a corner. You see, if they carried over the 6800-series PCIe lane allocation but maintained a pricing structure that competed against AMD’s Ryzen CPUs (make no mistake about it, the 7800X’s $389 price point is still quite tempting), sales of Kaby Lake-S processors like the 7700K could have been compromised.
Now with Kaby Lake-X processors the lines of distinction between product lineups are further blurred and there’s something of a logjam between the $299 and $399 brackets. Their addition has also forced Intel’s hand to make a CPU like the i7-7820X which features 28 PCIe lanes lest it becomes overly popular from a price / performance standpoint.
It is hard to predict where we go from here too. Any faster Kaby Lake CPUs or upcoming 14nm refreshes will invariably compete with some pretty major elements in Intel’s HEDT lineup, particularly now that Kaby Lake-X is a thing. Meanwhile, Intel has obviously left some space open for future expansion. It could very well be that any higher end additions to the Kaby Lake lineup will be made on the 2066 socket rather than 1151. That may go for the upcoming Cannon Lake and Coffee Lake as well. Only time will tell.
With all of this being said, I do think there’s a silver lining on Kaby Lake-X. Whereas the 7700K and 7600K represent the very pinnacle of Intel’s 1151 platform, the i7-7740X and i7-7640X could conceivably act as a great gateway into a platform that’s rife with upgrade possibilities down the road. X299 will stick around for a while folks. That means you can buy a sub-$400 processor and X299 board now and then simply drop in a higher end CPU sometime down the road rather than requiring a massive system upgrade. It’s a tantalizing possibility for people who want to hedge their bets against future requirements in their workflow.
The X299 Chipset Under the Microscope
Before we get too far into this section, it is important to go back in time and discuss X99. Back when it was released I asked a simple question: did that chipset have more lives than disco? Since it was relatively untouched from the X79 generation despite advances like U.2 and other PCIe-based storage being brought to the forefront, the answer to that question was a pretty straightforward “yes”. However, X99 was obviously a chipset behind the times since it was bound to PCI-E 2.0 and thus actually struggled to incorporate several next-generation storage interfaces.
X299 goes about things a bit differently…by borrowing a page or three from Intel’s own Z270 chipset and then adding in a good dose of CPU-bound PCIe lanes. As a matter of fact, if you look closely enough there’s very little to distinguish this chipset from the one being used by less expensive Kaby Lake processors.
There’s plenty to talk about on the CPU front so I’ll keep that until a bit later and for now focus exclusively upon the X299’s capabilities. In short they are a copy / paste from Z270 which is both a good thing and a bad thing. On one hand there are some significant upgrades when compared to the lackluster X99 but marketing this as a high end desktop platform is a bit of a misnomer. Actually, let’s put that a different way: rebranding Z270 as X299 (minus a few very basic differences) and then charging a premium for it won’t make Intel any friends.
Regardless of how close this platform mirrors Z270, it still represents a significant upgrade over X99 since there’s (finally!) a move towards a PCIe 3.0 backbone rather than the 2.0 interface of yesteryear. Unfortunately, there’s still no native support for USB 3.1 or NVMe-based storage devices but the PCH’s 24 PCIe lanes will provide more than enough bandwidth for the controllers required for high level IO functionality.
There are also eight native SATA 6Gbps ports and ten USB 3.0 ports which should provide adequate throughput for connected devices and more basic storage options. Unfortunately neither of those interfaces is particularly forward looking which makes this whole PCH feel like a transitional evolution between X99 and whatever next generation offering Intel is currently working on. Maybe that’s why it feels like a rebadge of Z270.
Like with Z270 Intel is utilizing their DMI 3.0 connection between the PCH and attached processor. This four lane interface offers almost double the throughput of the DMI 2.0 link which graced X99 which is good news considering this new platform is expected to keep pace with high bandwidth storage. However, we do have to wonder whether it will be sufficient if users decide to install multiple NVMe storage drives.
Any of the Skylake-X processors which get installed onto the X299 will bring with them additional capabilities but before we get into those, let’s talk about the layout of general graphics slots and memory support. At the high end, there is currently a single CPU with 44 PCIe 3.0 lanes which will have two different slot layouts: either x16 / x16 with two GPUs installed or the potential to run in x16 / x16 / x8 mode for triple graphics support. That leaves four additional lanes for future I/O functions like direct-to-CPU NVMe RAID setups through Intel’s upcoming VROC technology. These processors support quad channel memory.
28 lane processors like the i7-7820X and i7-7800X get a slightly different implementation due to their unique PCIe allocation. When two graphics cards are installed the lane speeds switch to a x16 / x8 layout while triple GPU setups aren’t supported without a secondary lane multiplier which some higher end motherboards may have. Like the i9 series processors, memory for these runs in quad channel mode.
Finally there’s Kaby Lake-X’s 16 native lanes which act exactly like the LGA1151 platform’s Z270 motherboards. With a single GPU installed the primary slot runs at x16 speeds while the secondary slot is disabled completely. Pop in a second graphics card and an internal switch splits the available lanes into a simple x8 / x8 layout. Memory support for these processors is somewhat unique as well with only dual channel kits being compatible. That means you will only have access to the rightmost four DIMM slots on many motherboards as the left slots will be disabled.
Test Setups & Methodology
For this review, we have prepared a number of different test setups, representing many of the popular platforms at the moment. As much as possible, the test setups feature identical components, memory timings, drivers, etc. Aside from manually selecting memory frequencies and timings, every option in the BIOS was at its default setting.
For all of the benchmarks, appropriate lengths are taken to ensure an equal comparison through methodical setup, installation, and testing. The following outlines our testing methodology:
A) Windows is installed using a full format.
B) Chipset drivers and accessory hardware drivers (audio, network, GPU) are installed.
C)To ensure consistent results, a few tweaks are applied to Windows 10 and the NVIDIA control panel:
- UAC – Disabled
- Indexing – Disabled
- Superfetch – Disabled
- System Protection/Restore – Disabled
- Problem & Error Reporting – Disabled
- Remote Desktop/Assistance – Disabled
- Windows Security Center Alerts – Disabled
- Windows Defender – Disabled
- Screensaver – Disabled
- Power Plan – High Performance
- V-Sync – Off
AIDA64 Extreme Edition
AIDA64 uses a suite of benchmarks to determine general performance and has quickly become one of the de facto standards among end users for component comparisons. While it may include a great many tests, we used it for general CPU testing (CPU ZLib / CPU Hash) and floating point benchmarks (FPU VP8 / FPU SinJulia).
CPU PhotoWorxx Benchmark
This benchmark performs different common tasks used during digital photo processing. It performs a number of modification tasks on a very large RGB image:
This benchmark stresses the SIMD integer arithmetic execution units of the CPU and also the memory subsystem. CPU PhotoWorxx test uses the appropriate x87, MMX, MMX+, 3DNow!, 3DNow!+, SSE, SSE2, SSSE3, SSE4.1, SSE4A, AVX, AVX2, and XOP instruction set extension and it is NUMA, HyperThreading, multi-processor (SMP) and multi-core (CMP) aware.
CPU ZLib Benchmark
This integer benchmark measures combined CPU and memory subsystem performance through the public ZLib compression library. CPU ZLib test uses only the basic x86 instructions but is nonetheless a good indicator of general system performance.
CPU AES Benchmark
This benchmark measures CPU performance using AES (Advanced Encryption Standard) data encryption. In cryptography AES is a symmetric-key encryption standard. AES is used in several compression tools today, like 7z, RAR, WinZip, and also in disk encryption solutions like BitLocker, FileVault (Mac OS X), TrueCrypt. CPU AES test uses the appropriate x86, MMX and SSE4.1 instructions, and it’s hardware accelerated on Intel AES-NI instruction set extension capable processors. The test is HyperThreading, multi-processor (SMP) and multi-core (CMP) aware.
CPU Hash Benchmark
This benchmark measures CPU performance using the SHA1 hashing algorithm defined in the Federal Information Processing Standards Publication 180-3. The code behind this benchmark method is written in Assembly. More importantly, it uses MMX, MMX+/SSE, SSE2, SSSE3, AVX instruction sets, allowing for increased performance on supporting processors.
FPU VP8 / SinJulia Benchmarks
AIDA’s FPU VP8 benchmark measures video compression performance using the Google VP8 (WebM) video codec Version 0.9.5 and stresses the floating point unit. The test encodes 1280×720 resolution video frames in 1-pass mode at a bitrate of 8192 kbps with best quality settings. The content of the frames are then generated by the FPU Julia fractal module. The code behind this benchmark method utilizes MMX, SSE2 or SSSE3 instruction set extensions.
Meanwhile, SinJulia measures the extended precision (also known as 80-bit) floating-point performance through the computation of a single frame of a modified “Julia” fractal. The code behind this benchmark method is written in Assembly, and utilizes trigonometric and exponential x87 instructions.
CineBench R15 64-bit
The latest benchmark from MAXON, Cinebench R15 makes use of all your system’s processing power to render a photorealistic 3D scene using various different algorithms to stress all available processor cores. The test scene contains approximately 2,000 objects containing more than 300,000 total polygons and uses sharp and blurred reflections, area lights and shadows, procedural shaders, antialiasing, and much more. This particular benchmarking can measure systems with up to 64 processor threads. The result is given in points (pts). The higher the number, the faster your processor.
PCMark 8 is the latest iteration of Futuremark’s system benchmark franchise. It generates an overall score based upon system performance with all components being stressed in one way or another. The result is posted as a generalized score. In this case, we didn’t use the Accelerated benchmark but rather just used the standard Computational results which cut out OpenCL from the equation.
wPrime is a leading multithreaded benchmark for x86 processors that tests your processor performance by calculating square roots with a recursive call of Newton’s method for estimating functions, with f(x)=x2-k, where k is the number we’re squaring, until Sgn(f(x)/f'(x)) does not equal that of the previous iteration, starting with an estimation of k/2. It then uses an iterative calling of the estimation method a set amount of times to increase the accuracy of the results. It then confirms that n(k)2=k to ensure the calculation was correct. It repeats this for all numbers from 1 to the requested maximum. This is a highly multi-threaded workload. Below are the scores for the 1024M benchmark.
Single Thread Performance
Even though most modern applications have the capability to utilize more than one CPU thread, single threaded performance is still a cornerstone of modern CPU IPC improvements. In this section, we take a number of synthetic applications and run them in single thread mode.
At face value, 7-Zip is a simple compression/decompresion tool like popular applications like WinZip and WinRAR but it also has numerous additional functions that can allow encryption, decryption and other options. For this test, we use the standard built-in benchmark which focuses on raw multi-threaded throughput.
Adobe Premier Pro CC
Adobe Premier Pro CC is one of the most recognizable video editing programs on the market today as it is used by videography professionals and YouTubers alike. In this test we take elements of a 60-second 4K video file and render them out into a cohesive MP4 video via Adobe’s Media Encoder. Note that GPU acceleration is turned on.
Blender is a free-to-use 3D content creation program that also features an extremely robust rendering back-end. It boasts extremely good multi core scaling and even incorporates a good amount of GPU acceleration for various higher level tasks. In this benchmark we take a custom 1440P 3D image and render it out using the built-in tool. The results you see below list how long it took each processor to complete the test.
3ds MAX Corona Renderer
Autodesk’s 3ds MAX is currently one of the most-used 3D modeling, animation and rendering programs on the market, providing a creative platform for architects to industrial designers alike. Unfortunately its rendering algorithms leave much to be desired and third party rendering add-ons are quite popular. One of the newest ones is called Corona.
In this test we take a custom 3D scene of a room with global illumination enabled and render it out in 720P using Corona’s built-in renderer.
While it may be open source, GIMP is actually one of the most popular free photo editors available right now. It uses both CPU and GPU acceleration for certain tasks. In this test we use an 8K image and use a script to run eight different filters in succession. This is considered a lightly threaded workload since the memory, CPU and storage drive can all play a role in performance.
Video conversion from one format to another is a stressful task for any processor and speed is paramount. Handbrake is one of the more popular transcoders on the market since it is free, has a long feature list, supports GPU acceleration and has an easy-to-understand interface. In this test we take a 6GB 4K MP4 and convert it to a 1080P MKV file with a H.264 container format. GPU acceleration has been disabled. The results posted indicate how long it took for the conversion to complete.
POV Ray 3.7
POV Ray is a complex yet simple to use freeware ray tracing program which has the ability to efficiently use multiple CPU cores in order to speed up rendering output. For this test, we use its built-in benchmark feature which renders a high definition scene. The rendering time to completion is logged and then listed below.
WinRAR is one of those free tools that everyone seems to use. Its compression and decompression algorithms are fully multi-core aware which allows for a significant speedup when processing files. In this test we compress a 3GB folder of various files and add a 256-bit encryption key. Once again the number listed is the time to completion.
3DMark Fire Strike (DX11)
3DMark Time Spy (DX12)
Battlefield 1 will likely become known as one of the most popular multiplayer games around but it also happens to be one of the best looking titles around. It also happens to be extremely well optimized with even the lowest end cards having the ability to run at high detail levels.
In this benchmark we use a runthough of The Runner level after the dreadnought barrage is complete and you need to storm the beach. This area includes all of the game’s hallmarks in one condensed area with fire, explosions, debris and numerous other elements layered over one another for some spectacular visual effects.
Call of Duty: Infinite Warfare
The latest iteration in the COD series may not drag out niceties like DX12 or particularly unique playing styles but it nonetheless is a great looking game that is quite popular.
This benchmark takes place during the campaign’s Operation Port Armor wherein we run through a sequence combining various indoor and outdoor elements along with some combat.
Deus Ex – Mankind Divided
Deus Ex titles have historically combined excellent storytelling elements with action-forward gameplay and Mankind Divided is no difference. This run-through uses the streets and a few sewers of the main hub city Prague along with a short action sequence involving gunplay and grenades.
Not many people saw a new Doom as a possible Game of the Year contender but that’s exactly what it has become. Not only is it one of the most intense games currently around but it looks great and is highly optimized. In this run-through we use Mission 6: Into the Fire since it features relatively predictable enemy spawn points and a combination of open air and interior gameplay.
Grand Theft Auto V
In GTA V we take a simple approach to benchmarking: the in-game benchmark tool is used. However, due to the randomness within the game itself, only the last sequence is actually used since it best represents gameplay mechanics.
Overwatch happens to be one of the most popular games around right now and while it isn’t particularly stressful upon a system’s resources, its Epic setting can provide a decent workout for all but the highest end GPUs. In order to eliminate as much variability as possible, for this benchmark we use a simple “offline” Bot Match so performance isn’t affected by outside factors like ping times and network latency.
I don’t typically dedicate a whole page to power consumption but there’s a pretty substantial story lurking behind the numbers you see below and how they directly relate to TDP claims from both Intel and AMD. Without getting too technical, the way these two companies go about measuring TDP is fundamentally different from one another. Intel themselves published a very comprehensive and quite neutral White Paper (Word doc download) about the differences a few years ago and its worth a quick read if you have a chance.
What you need to know is that TDP values are a universally poor way to determine actual power consumption for end users since they are simply thermal design guidelines that are given to system integrators. As I say in every review, TDP is not actual power consumption so don’t take it as such.
As both Intel and AMD recommend, the best way to measure true power deltas between processors is via a simple (yet calibrated) power meter plugged into the wall outlet. That’s exactly what we do but add in a controlled 120V power input to eliminate voltage irregularities from impacting the results.
You may remember that earlier in this review I mentioned the Kaby Lake-X processors were given higher TDP ratings due their upscaled packaging and some other factors. Those TDP ratings my not directly translate to power consumption but there’s no denying the i7-7740X and X299 platform consume more power than the i7-7700K and Z270. It isn’t a massive amount but its still measurable.
Some of this could also be due to platform maturity since I still don’t feel like the processor’s C-States and Turbo modes are properly implemented, even after numerous revisions in the days leading up to launch. Hopefully as time goes on additional refinements will be built in and we’ll see the gap narrow somewhat.
I’m going to keep this section a lot shorter than I normally would since with the rushed nature of this launch I haven’t had much time to play around with the lone Kaby Lake-X processor I have in hand. With that being said, if there is one thing Intel did well on this Core-X lineup it’s overclocking. The i7-7740X has a good amount of additional TDP headroom and according to our contacts at various motherboard vendors, these chips should easily hit between 5.0 and 5.5GHz with proper cooling.
Now the term “proper cooling” is a loose one since it doesn’t necessarily mean water cooling. Rather, I was able to easily overclock my engineering sample with a relatively simple Noctua NH-U12S. Naturally the larger 10 core CPUs will require substantially more heatsink mass but don’t start thinking that you’ll need to spend $100 or more on a cooler to achieve optimal clock speeds.
After retesting a number of Ryzen processors over the last few weeks, I have to say that overclocking on this X299 platform is like breath of fresh air. Instead of a struggle ending with smashing face first into a wall at 4GHz, the i7-7740X easily blazed through the 5GHz mark on its way to an even 5.1GHz. And this was after a mere 15 minutes of trial and error and an hour of stress testing to insure the overclock was stable. I’m sure that given a few more hours I would have been able to go quite far beyond that point.
Conclusion: What a Tangled Web We Weave
The lead up to Intel’s Core-X launch has been filled with critique, some of it well placed and the rest based off of a simmering anger against Intel which has in many cases led to very real bias. A good amount of that angst stems from the fact that many enthusiasts feel Intel’s HEDT platform’s cost structure has reached the point where it is no longer a viable solution. The i7-7740X featured in this review goes a long way towards lowering the cost of entry into X299 territory but that doesn’t necessarily make it a good value.
At first glance the i7-7740K seems to be exactly what the doctor ordered: it has a solid architectural pedigree, high clock speeds and a price that looks pretty appealing given the tangible benefits we’ve been led to believe X299 brings to the table. After crunching the numbers it looks pretty competitive as well, that is until there’s further analysis and you realize that not much has changed with Intel’s approach to segmentation.
What could have been a compelling product ended up being a complete re-launch of the i7-7700K at the exact same price. But in this case it comes with lower performance and not one additional feature to distinguish the whole X299 + Kaby Lake-X combo from Z270. Drop an i7-7740X processor into an X299 motherboard and everything that makes the platform unique is thrown out the airlock. If you want additional PCI-E lanes, quad channel memory, compatibility with VROC or Turbo 3.0, look elsewhere. As a matter of fact, the “new” chipset’s capabilities are no different from Z270 either so nothing is gained there either. This whole situation is simply mind boggling.
The lower performance part of that equation is likely due to the platform’s immaturity and the obviously rushed pace at which Intel’s motherboard partners had to cobble together BIOSes. As a matter of fact, just three days before launch I received key BIOS updates from ASRock, ASUS and Gigabyte which optimized Turbo responsiveness and netted performance gains of between 2% to 5%. Unfortunately, that still didn’t help the i7-7740X win against the i7-7700K but it does show there may be a bit more in the tank.
Whether Intel wants to admit it or not, outside of the obvious parallels you can draw between their i7-7700K and the i7-7740X’s, this processor’s main competitor will be the Ryzen 7 1700X. Against AMD’s $399 alternative I feel the i7-7740X falls flat from an overall value standpoint. While it competes well in gaming (though not as well as I would have expected), it is comparatively terrible for multitasking due to the lower core and thread counts. The only saving grace is that the Kaby Lake architecture can run at much higher core frequencies than anything in the Zen stable right now.
The main challenge facing the i7-7740X isn’t the cost of the processor itself but rather the overall platform’s pricing structure. X299 motherboards are being marketed as premium products even though dropping in a Kaby Lake-X CPU knocks their capabilities down to near-Z270 levels. Let’s take Gigabyte’s Aorus Gaming 7 series as a good example of this: the Z270 version retails for about $240 whereas the X299-based board will hit $399. That’s a $160 premium for no mentionable benefits. I’m picking on Gigabyte here but the situation will repeat itself with every vendor.
So is there any silver lining here? I happen to think the i7-7740X does provide a balance of positive and negative points but a very specific set of requirements are needed for it to be a viable solution. Unlike the 7700K which locks you into a pretty narrow performance bracket, you can buy an i7-7740X now and then upgrade to a more capable LGA2066 CPU sometime down the road. Kaby Lake-X does indeed bring down the price of entry into the HEDT space and a savvy buyer can certainly use that to their advantage.
I also have to commend Intel on the ease at which their processors overclock. With AMD’s Ryzen, overclocking feels like a fight to the death against an unwilling platform whereas the 7740X I have on hand blazed through the 5GHz mark and didn’t show any sign of letting up. The only thing that limited my overclock to 5.1GHz was time rather than any limitation in the motherboard or chip. Overclocking is what the HEDT platforms are all about and this is a key win for Intel in my books.
However, this doesn’t make the i7-7740X a great processor, nor is this a smooth launch worthy of Intel’s name. With the i7-7740X Intel is forcing you to make a decision: either buy a Kaby Lake-X and lose all of the features which make X299 an enthusiast-level platform or buy a processor that costs a thousand bucks. There is no middle ground anymore since every one of the i7-7800 processors is seriously nerfed in some way.
Meanwhile, the advent of the i7-7740X and i7-7640X does bring down the cost of entry but due to their truncated feature sets the situation is literally no different from a year ago when the Z170 versus X99 debate was raging. I’m actually quite sure many will end up sticking to Z270 or jumping on the Ryzen bandwagon due to the significantly lower cost of their motherboards. What this leads to isn’t necessarily a failed launch but rather one which has introduced a pair of misplaced Kaby Lake-X processors that feel a bit like orphans in their own house.