Intel Core i9-7900X review: The fastest chip in the world
arstechnica.com
arstechnica.com
I'm still running a i7-4790K at home, and though I'd like something with more cores... nothing is compelling enough to bring me to switch given the costs involved. If I were building new, would most likely go with an AMD solution.
I know many people still sitting on anywhere from a 2xxx gen to a 5xxx gen and just don't feel compelled to upgrade from a CPU perspective. Those that eventually do, do so for the motherboard features more than the CPU - that's just a necessary cost for a small benefit.
This is marginally different for laptops where perf/watt becomes more important, of course. However for desktops, I certainly wouldn't be troubled by a 4xxx gen. I upgraded last year and it was from a 920 to a 6600K. Even the 920 did much of what I wanted, honestly, it was more a luxury upgrade.
a 386 machine would be ~$10k (in 80s $s!) and more like a high end multichip Xeon system or something today - they actually had a compaq 386 of that vintage in my first job and it was still in active service in 2001.
Just wanted to add a little historical context.
There's no perceivable difference.
If you have enough RAM to avoid swapping and cheap SATA SSDs, all the requirements for fast/large CPUs or graphics cards are special applications of some sort. Browsers and editors and terminal windows and video playback don't count.
http://www.gamersnexus.net/guides/2773-intel-i5-2500k-revisi...
http://www.eurogamer.net/articles/digitalfoundry-2016-is-it-...
For "work" workloads, Sandy Bridge is still quite potent though. Honestly in that segment the gains have mostly been coming from moving to more cores. A hexacore or octacore with hyperthreading/SMT knocks the stuffing out of an i5.
While NVMe benchmarks are stunning I really am curious whether in an ordinary setting the difference to SATA can be felt at all. Not measured -- felt. When the era of swap ended because CPUs / chipsets finally could handle enough RAM and when we went from HDD to SDD, those could be felt for sure.
> This is marginally different for laptops
In laptop world, the problem is that with the spread of thin craze most laptops are now running 15W CPUs instead of 35W like in the old age and so there is little performance increase, no core count increase.
I sometimes tweak stuff on both laptops at the same time, and doing a side-by-side restart or even waking up from sleep makes the older model's lesser performance painfully obvious
Today I can't even tell if my 13" dell xps is in it or not.
Not much difference in daily usage under Windows 7.
I have an NVMe and a SATA SSD in my workstation. I do most of my work off the NVME, but occasionally work on projects that live on the SATA disk. I definitely notice a difference in things with a lot of disk access.
The 920 is a solid performer, but oh my does it need efficient cooling. At launch it almost felt like a return to the Pentium 4 era in terms of power consumption.
If I were to upgrade my 920 anytime soon, it would be to reduce the need for cooling fans and to get a quieter system. As a programmer, I don't need more ultra-cores or jigga-hertz.
I also have multiple SSDs and an NVMe I recently put in... that seems to have made the most difference. I'm running 32GB ram now, and don't tend to bump into that as a limitation.
I'm not a fan of either megacorp (though I prefer Intel because their stuff works much better historically on GNU/Linux), but it should be painfully obvious that "i9" is just a reaction to the threat of ThreadRipper from AMD (which is still going to be more powerful and far more affordable than i9 when it launches).
EDIT: To be fair, they do mention this in TFA:
> That these chips are currently little more than a product name and a price [...] is a strong indication that Intel was taken aback by AMD's Threadripper, a 16-core chip due for release this summer.
This is incorrect. As Ryan Shrout from PcPer notes,[1]
> In some circles of the Internet, the Core i9 release and the parts that were announced last month from Intel seem as obvious a reaction to AMD’s Ryzen processor and Threadripper as could be shown. In truth, it’s hard to see the likes of the Core i9-7900X as reactionary in its current state; Intel has clearly been planning the Skylake-X release for many months. What Ryzen did for the consumer market was bring higher 4-count core processors to prevalence, and the HEDT line from Intel has very little overlap in that regard. Threadripper having just been announced in the last 60 days (even when you take into account the rumors that have circulated), seems unable to have been the progenitor of the Core i9 line, isn't its entirety. That being said, it is absolutely true that Intel has reacted to the Ryzen and Threadripper lines with pricing and timing adjustments.
[1] : https://www.pcper.com/reviews/Processors/Intel-Core-i9-7900X...
I probably should've been clearer, but I assumed it was obvious you can't do R&D on new silicon in 60 days and already have an announcement for it.
I thought the development cycle for a new CPU was between 2-3 years (hence the 2 teams, and tick tock thing).
It's like adding new features late in the game for a webapp. Some things would be very easy (we need a new page for X). Some things would be very hard (I want you to rewrite the entire app from angular to react). My guess is something like changing core count is very very hard to do late in the game. Changing price, maybe changing overclock settings, those would be fairly easy.
So I suspect if Intel didn't see threadripper coming (which I doubt) the thread counts were set years ago. However threadripper is making Intel drop the prices a bit.
Intel is afraid of these HCC chips cannibalizing their sales of more expensive Xeons, so they're holding back as long as possible and crippling key features. If you want ECC, for example, you have to buy a Xeon - or more realistically for many people, a Threadripper.
But it's the new, very-high-core-count CPUs Intel hastily announced with no real details and without even warning their partners first that look like a reaction to Ryzen, specifically Threadripper.
Those CPUs might be a problem for Intel. They can only make so many high-core-count Xeons and now some of the best chips will be headed to the HEDT market rather than servers. Also, thrusting chips with almost twice as many cores as originally planned onto motherboards which aren't designed for them and which already have seen overheating problems might not end well.
Maybe the branding but the HEDT lineup has been around since 2010. Shockingly, this lineup escaped the notice of AMD loyalists, or was assumed to be equivalent to Bulldozer's CMT (which is much more like a SMT/hyperthread than most people would admit at the time) - however they have always been excellent at gaming, particularly compared to Bulldozer's miserable IPC.
The high-core-count lineup are probably a reaction to Threadripper but these chips have been around in the Xeon lineup since forever and really should have stayed there. These chips (including Threadripper) are really multi-socket-in-a-package and don't perform particularly well in gaming workloads (despite AMD advertising them for such). Games don't scale well given the latency. They will be nice for things like CAD rendering workstations but that's a much narrower niche.
The i9 branding also slices off ECC, which is a significant feature for many things these will actually be good at. Threadripper will be at a significant advantage for actual server usage (even home-server) as a result. Intel is trying to avoid cannibalizing sales in their more expensive Xeon lineup but it does kill a bunch of the utility of the processor as a result.
> Intel Core i9-7900X review: The fastest chip in the world, but too darn expensive
> When eight-core Ryzen costs £300, do any of these new Intel chips make sense?
But for that extra cost you get four more threads at a higher clock rate, twenty extra PCIE lanes, a 500 MHz higher turbo clock speed, and double the memory bandwidth.
Even with the lesser Intel Core i7-7820X you will get the same thread count but with a higher clock rate, four extra PCIE lanes, still a 500 MHz higher turbo clock speed, and double the memory bandwidth for only $140 more.
Now, of course, the AMD Ryzen Threadripper 1950X comes much closer to the i9-7900X price point.
However, you will sacrifice single core performance to gain twelve more threads at a lower clock rate. But, you will receive twenty more PCIE lanes, over twice as much L3 cache, and the same memory bandwidth as the i9-7900X.
So if your plan is to build a 3D render farm, the Threadripper seems quite appropriate.
Although, if you plan to build a workstation on which to model 3D assets and to perform preview renders--the Intel i9 series seems more apt.
My guess would be PC users by count are
gamers > programmers > 3d renderers
For most gamers, seems like i7 or maybe i9 wins in current benchmarks. For programmers maybe ryzen is a better fit, but I bet it depends on your language.
Google right now, tells me there are:
155 million "gamers" in the US
3.6 million programmers in the US
I cannot guess that the CGI industry is higher than either of these numbers.
Now very much they may pay for bleeding edge, and spend more dollars on hardware than programmers. But I am at least 95% confident there are less people in the US running a 3d program on their desktop compared to running eclipse/visual studio/atom/vim.
I invoked the example of 3D rendering as it's the origin of most of these discussions among my co-workers.
Not as much as you would think, I have a ryzen 1700 I use for dev, mostly PHP (not a language that threads well) in an Enterprise environment.
Those extra cores/threads come in incredibly handy for virtualization, both dev environments and things like running windows for testing
Even VS2017 inside virtual box absolutely flies when given 16GB of RAM and 4 cores (8T).
For my day to day the cores are useful even if my primary language isn't using all of them.
I'm currently in the strange situation where my desktop running the main system in virtual box is faster than our production/spare servers.
1) How many are there?
2) it seems like you are usually not CPU bound. The people I know doing this spend 20 minutes waiting for a video to render than 6 hours uploading it to youtube. In most cases their bandwidth is a limiter 10x over their CPU. (Plus most of those people are on Mac anyway, so they don't even get this choice)
The GIGABYTE GA-AB350M-Gaming 3 is the cheapest Ryzen board I can find at NewEgg and costs only $94.99 by comparsion.
Therefore, there's an obvious savings of $124. Although, you can easily spend $189.99 or more on a higher-end Ryzen board given NewEgg's offerings.
So while there is a discrepancy, it's not massive.
But I agree that a $200 - $400 difference may not be substantial for a workstation in heavy use. Personally, I do still have concerns about availability over the life of the CPU. I currently have an X79 and was looking to replace the motherboard in its 2nd year of ownership. eBay is really the only option available and with a scarce 2nd-hand market, the boards don't depreciate much. It's gotten better over time, but I was looking at paying $400+ for a used motherboard. I decided to just deal with the quirks of my current one. The X299 is early in its lifecycle, of course, so I'd hope availability for a few years.
Agreed.
> The X299 is early in its lifecycle, of course, so I'd hope availability for a few years.
Same, here. I've seen ASRock Intel boards vanish from the market only a year after they debut.
It's disturbing. And moreover, it's detrimental to the lifetime of the board as driver and BIOS updates cease.
Hopefully this chipset and socket will last a bit longer.
Chipset cost will come down in 6-12 months after launch like it always does. This is part for the course, at launch X370 boards for Ryzen were going for well over $250 as well.
Power consumption is a consequence of AVX512 and the mesh interconnect along with raw core count. Everyone wants higher clocks, more cores, and more functional units. There are no easy efficiency gains anymore, and this is the price - power consumption. This is the "everything and the kitchen sink" processor and it runs hot as a result - but it absolutely crushes everything else on the market. This is no Bulldozer.
Board partners with insulators on top of their VRMs was going to come to a head sooner or later. This is the natural outgrowth of form over function, RGB LEDs on everything and stylized heatsink designs that insulate the board instead of actual cooling. The terrible reviews on those boards will sort this problem right out, they will be unusable in their current form.
Intel has been cruising for issues with their TIM for years (since Ivy Bridge), this time they finally have a chip that puts out enough heat they can't ignore it. Intel can get away with making you delid a $200 i5 or a $300 i7, it's not acceptable on a $1000 processor.
There is still a market for a 6-12C HEDT chip that can hit 5 GHz overclocked. This thing absolutely smokes Ryzen in gaming at stock clocks let alone OC'd - single-thread performance is still a dominant factor in good gaming performance and this chip delivers in spades. Combining its leads in IPC and clocks, it's fully 33% faster than Ryzen in single-thread performance. This is just a brutal amount of performance for gaming. Unfortunately without delidding you're not going to hit good OC clocks given the current TIM. And delidding is a dealbreaker on a $1000 CPU.
TIM is the actual core problem with Skylake-X - everything else will sort itself out. Skylake-X with solder would be a winner and Intel would be wise to turn the ship as fast as possible. The 6C and 8C version are priced much more reasonably and will sell great as long as they fix the TIM problem.
Intel claims they have problems with dies cracking, but AMD manages to solder much smaller dies, so IMO Intel just doesn't have a leg to stand on here. This is not something that should be pushed onto the customer with a $1000 processor - you're Chipzilla, figure something out.
TIM = Thermal Interface Material
http://i.imgur.com/7BIJmxS.png
http://www.tomshardware.com/reviews/intel-core-i9-7900x-skyl...
Heat just is not getting to the IHS properly on these chips, and heavy overclocking just makes the whole thing worse.
It's really been a slow-burning problem since Ivy Bridge, where Intel switched from soldering the lid to TIM + an adhesive. Solder has long been preferred for its superior heat transmission, but Intel says say smaller dies have problems with cracking over time due to thermal cycling. However, AMD has been happily soldering the lid on much smaller Ryzen dies, so apparently it's not all much of an issue in practice.
Well, you can get away with that on a processor that puts out 50 watts during normal operation. It's been an issue for a while on the unlocked/overclockable SKUs, particularly on the latest 7700Ks, but even an OC'd 7700K only puts out ~100W, so it was relatively manageable. Extreme overclockers could delid and replace the thermal paste with something better (often liquid metal like Conductonaut), which does help performance quite a bit.
But, with the higher TDP of Skylake-X, this has become a pressing issue just for normal operation. Things change when we're talking about a $1000 processor that needs to be delidded to sustain boost at stock settings. That's just not acceptable.
I almost would rather have a bare die at this point. Mounting pressures are no longer insane so it wouldn't be as terribe an ordeal to mount as as Athlon XPs were back in the day (god forbid your screwdriver slip on that bracket, with 50+ pounds of force you are guaranteed to gouge something).
For a sense of perspective here, going from a circa-2012 2600K to a current 7700K is a 40% jump in performance, so it's roughly equivalent to 4-5 years of gains at Intel's usual tempo - only you also have 10 cores on this platform. This thing is an absolute monster for gaming or other tasks that lean heavily on single-threaded performance.
But the power consumption is really the triggering issue for the problems with shitty partner-boards overheating and the TIM. The TIM is really the showstopper right now.
https://www.hardocp.com/article/2017/01/13/kaby_lake_7700k_v...
Actually, air-cooling is marginal with these CPUs, even at stock frequencies. There is no air-cooler which allows them to sustain their frequencies under load.
Water-cooling is pretty much required, and an AIO does not cut it. Still, even with water-cooling you can't really overclock these. They are pretty much at their limit out of the factory.
This could conceivably solved by Intel switching away from silicon TIM to e.g. solder, since the Rth(jc) of the CPUs is much worse at ~0.3 K/W than the thermal resistance of a big CPU air cooler (~0.1 K/W).
The overclocking problems wouldn't be fixed though; there is no easy fix for a CPU that jumps to 400+ watts.
-
An entirely separate issue is that you need to get the heat out of your office. A human dissipates around 50-100 W; you can imagine that a small office crowded with half a dozen people is not pleasant in the summer.
If you have a hall full of developers, having lots of noisy PCs can be annoying.
Where I work, we optimized for more silent PCs, because all these things do add up, and given the perf/watt-ratio you can get out of modern CPUs, there's no reason for people to need to have noisy PCs.
Even at home, optimizing watt-usage, even for a desktop build, is not completely without merits. All my future projects are planned as fanless as possible. And I know others who do the same. And if Intel can't deliver that, they'll just go buy something Arm-based, like an Rpi 3, which these days are getting good enough to actually do production loads.
I'm not even getting close to wanting a system where the CPU alone can draw 400+ watts.
They really aren't, CPU performance is really irrelevant given the RPi's architectural weaknesses. USB was never meant as a system bus and everything has to loop through the kernel stack. Having every single peripheral hanging off a single USB 2.0 bus is crippling for performance. A Pi can't even serve a share at full 100 mbit speed due to bus contention let alone do anything more intensive.
It's very similar to one of Apple's more famous goofs, the Performa 5200/6200 with its left-hand/right-hand bus split that forces the CPU to handle everything.
http://lowendmac.com/2014/power-mac-and-performa-x200-road-a...
Some of the clone boards have USB 3.0, SATA, gigabit ethernet, etc and are much better performers in practice despite having slower CPUs "on paper". Or there are little mini-PCs using 5-15W laptop processors that are really nice and run x86 distros/binaries.
All of these are at roughly comparable TCOs to a Pi (they include things like AC adapters that must be purchased separately for the Pi). The RPi is a bad choice for server usage.
Obviously "production loads" is an undefined term and as such we can discuss infinitely back and forth exactly how much these cheap ARM machines can actually handle.
I also didn't mean to single out the Rpi3 as a universal performer, optimal for everything out there.
My point was that I'm seeing an increased amount of people who are happy with what these cheap boards can do, who 10 years ago would have been forced to buy a server of sorts to cover the same needs.
So now they don't buy servers. Instead they buy cheap, tiny and fan-less ARM-based machines and they're perfectly happy. They even think running ARM is cooler than running Intel, so it's something they brag about.
I'm absolutely not saying I'm going to replace my company's server-farm or my dev-computer with these anytime soon, but Intel cannot completely ignore the power-efficiency aspect either if they want to keep their dominance in the market.
There is no such alternative for Skylake-X - Intel charge you an arm and leg for its half decent products.
Yeah, absolutely agree. What's sad is there are fanboys defending this and saying this makes direct die cooling easier through delidding, which saves 1-2 C° over solder, and is the beast idea Intel had in recent years.
That if you are limiting yourself to 1080p... At the resolution that I game (3440 x 1440) those performance differences disappear very fast. And even at 1080p 150 vs 180 frames don't matter that much for the majority of people.
The cost of this chip alone is the same as some Ryzen builds. There is a point where it financial doesn't make sense (price/performance) even if it's the fastest chip around.
I remember I used to advice people to spend half on monitor, half on the pc system (and out of that, maybe close to half on the gpu). That used to mean a ~1000 USD monitor(s), and a ~500 USD GPU - for a total system price of ~2000 USD.
Today, most people would probably aim for a lower total cost, but it's still silly to sink a lot of cash into getting a great system, only to have a crappy monitor ruin the experience.
(Another caveat, I'd guess a high-end monitor should be able to survive/remain usable for closer to a decade than to 3-5 years -- which would be more typical for a pc. Of course, part of the reason for getting a pc system would be the possibility of ~incremental upgrades)
Unless you want ultrawide that is - but there are some caveats there with game compatibility due to the aspect ratio.
With this much of a lead in single-thread performance (~33%) a 6C Intel is actually outperforming an 8C Ryzen even in multi-thread performance and it's stomping it in games because single-thread performance is still so critical.
And the 8C Intels are just 33% faster than Ryzen across the board.
High-refresh gaming requires excellent single-threaded performance regardless of resolution, and 144 Hz is basically the new standard for midrange/high-end gaming builds at this point. A 144 Hz monitor starts at literally $150 and a very nice IPS 144 Hz FreeSync/GSync monitor can be had for $400-600.
It's not just 1080p - CPU single-thread-performance requirements scale with the framerate, it's just easier to hit higher framerates at lower resolutions. So 1080p benchmarks are a "leading indicator" of future high-refresh gaming performance as GPU tech improves and you upgrade in a year or two.
On the flip side, 4K benchmarks really mean almost nothing for CPUs. A Pentium G4560 is within a stone's throw of a 7700K at 4K because everything is GPU-bottlenecked at a very low framerate that virtually any processor can deliver. But that G4560 will fall behind in no time at all as GPU performance continues to improve and its actual performance (or lack thereof) is laid bare.
Intel clearly rushed this out to prevent AMD from having the perception of leading the space, at least in terms of the largest core count. The article even references this.
I just hope that it stays competitive, as this is clearly a win for consumers.
The rest of the i9 lineup, thus far unreleased with only the core count actually known, is Intel's rushed response to Ryzen. Not this chip.
People will attribute anything one of AMD's competitors does to fear of AMD - 780 Ti, 980 Ti, 1080 Ti, Skylake-X, you name it. It's frankly a little comical given the actual amount of competition AMD put up with Bulldozer and Fiji/Polaris/Vega against Haswell/Skylake and Maxwell/Pascal - which is to say, hardly any. NVIDIA and Intel both have their own yield strategies and release schedules that are largely independent of what AMD does. Intel and NVIDIA are tweaking prices and specific launch dates - that's about the extent of AMD's impact on their competitors so far.
(although Intel is definitely paying attention to Threadripper/Epyc now, and will be pushing core count up on consumer chipsets starting with Coffee Lake and presumably Cannonlake as well)
The real problems with Skylake-X are chipset cost, power consumption, shitty partner boards with insulators on the VRMs, and TIM. None of those have anything to do with a "rushed launch", all of those are issues that have been slow-boiling for years now.
They have already been selling vastly superior Skylake-SP (and -EP) Xeons to Google/Amazon/Facebook since 2015.
Intel: "why so serious!?"
I suspect that benchmarks do a really poor job of measuring worst case performance... which is what users notice. Things like UI lag and audio skipping. I suspect double the memory bandwidth (assuming a nice fast M.2 SSD) is the limiting factor for heavy workloads made up of independent tasks.
Impressive for a x86 as this is, some POWER and SPARC users may disagree with this assessment. In fact, some Xeon users will doubtlessly scratch their heads too.
Power9 is competitive in performance/Watt, and in some weird benchmarks which no one cares about. I haven't seen anything competitive in any way from SPARC for a long time.
If I think Microsoft isn't free enough for me, then I can remove Windows and install Linux (or BSD). If I think Chrome is sending my data to Google, then I can remove it and install Firefox.
But if don't like that Intel can take over my PC at any time, watch my screen, log my keystrokes, prevent me from installing another operating system, manipulate what I see on the screen, and much much more, then there is nothing I can do. I can not remove the second chip or remove the code. I must have a proprietary blob [4] (of who's source code no one can see to audit) running on my Intel PC.
But the worst thing has to be Intel and AMT's complete refusal to provide a clean chip to companies that are trying to provide backdoor free computers. Look at http://puri.sm [5] for example. They are trying to provide a PC that does not restrict what operating system or BIOS you run, and have repeatedly contacted Intel to ask them to provide a batch of chips with no ME or AMT installed. Even Google, which sells millions of chromebooks (coreboot preinstalled) have been unable to persuade them. [6]
As Intel and AMT are the biggest players and arguably a monopoly of the microprocessor market, they have a responsibility to provide safe and clean processors that customers can truly own. Please try your best not to buy these products until they resolve these issues.
[1] https://libreboot.org/faq.html#intelme
[2] https://www.theregister.co.uk/2017/05/01/intel_amt_me_vulner...
[3] https://www.intel.com/content/www/us/en/architecture-and-tec...
[4] http://boingboing.net/2016/06/15/intel-x86-processors-ship-w...
So you can view the code, verify it's what is installed but for security purposes can't change it.
The way this reads it seems deliberately suspicious. Having a HTTP server so low in the hardware stack feels wrong.
Also, I'm under impression that [6] refers not only to firmware blobs running on various auxiliary coprocessors but also the machine's firmware, i.e. BIOS, and in particular the CPU initialization component which had been provided only in binary form by both Intel and AMD for a few years now.
AMT is the poorly written and proprietary firmware Active Management Technology. This allows you to remotely control a computer.
AMD is a company and competitor to Intel.
I myself am thinking of buying one of these new AMD processors.
It's not that notebookcheck is benchmarking something outlandish: it shows 668 while this Ars article claims 637 for the 6700K, Notebookcheck benchmarked the 6950X to 1859 while Ars has 1786, both are very close.
I see a 7600K with the 637 score, but that lacks hyperthreading and has 25% less L3 cache compared to the 6700T, so it makes sense that the 17% frequency advantage is mostly balanced out (there's little IPC difference between Kaby Lake and Skylake).
You don't have notebookcheck's numbers matched up with the right CPUs either: https://www.notebookcheck.net/Mobile-Processors-Benchmark-Li...
To be fair, Ars has the i5-7600K listed as an i7, and Notebookcheck has the cache sizes wrong: http://ark.intel.com/compare/88200,97129,97144,88195
So there is plenty of confusion to go around.
Edit: Actually the frequency difference may be a bit off from 17%, that was based off the max single core turbo frequencies. I don't know what the all core turbos are.
I suspect moving from a ring bus to a mesh helps as well.
The CPUs will generally use far less power than the TDP might suggest.
Plus, this is Intel's 3rd iteration on the same process; even if the feature size doesn't change you can extract a bit more power efficiency with 2 years of feedback.
what a huge load of biased non-sense! I have a machine pretty similar to the one mentioned in the article below, it is using Intel processors released ages ago, in fact they were from decommissioned servers from some random data centres. I'd willing to bet that machine with "the fastest chip in the world" is significantly slower/more expensive than mine when it comes to my long list of day to day development tasks.
Oh, don't forget to mention the fact that the "fastest chip in the world" can reach >100 degrees when fully loaded. Maybe Intel should pay some review sites to claim it to be the processor most suitable for cooking a meal.
https://www.techspot.com/review/1218-affordable-40-thread-xe...
In case you want to argue that my machine has two Xeon - you can actually order one single Xeon from newegg.com which is more recent, put it into a consumer motherboard and beat the xxx out of i9-7900x. There is no way a 10-core Intel processor could possibly be the "fastest chip in the world".
Also, have both of these setups run a mixed workload that isn't absolutely parallelizable and watch the Xeon struggle.
2. you can buy a pair of such 10-core Xeon for almost the same price of a single i9-7900x.
3. i9-7900x faces the exact same problem when the workload can not be paralleled - you can buy a much cheaper quad core intel processor that overclock well, you can push it to say 4.5 or 5G and beat the "fastest chip in the world".
I'll humour you however.
Your Xeon turbos to 3.1 if only 1 core is stressed, but it's all core boost is 2.4 as per https://www.pugetsystems.com/blog/2015/07/09/Actual-CPU-Spee...
I don't see how your 10core part boosting to 2.4 Ghz is "much faster in the day to day development tasks" given that it's 2 CPU architectures behind and clocks at almost half the 4.0 Ghz achieved by the 7900x (all-cores boost).
But even a pair of those 10 cores is probably slower (caches are not shared, all-core boost almost half while roughly 5% slower in IPC performance due to the jump from Broadwell to Skylake).
So I don't think you're actually trying to argue that this isn't the fastest chip yet, but that it's a bad deal compared to looking around and buying some used server parts.
And yes, that's a better deal, but also a used i9-7900x is a better deal than a new i9-7900x...
i9-7900x is _NOT_ the fastest chip in the world, not even the fastest intel chip. Consumer/server difference is purely a marketing thing, my Xeon based workstation running CS:GO on daily basis is not a server.
maybe if it ended with "for very specific cases" it would be more true.