Specifications and Architecture
It has been an interesting 2017 for Intel. Though still the dominant market share leader in consumer processors of all shapes and sizes, from DIY PCs to notebooks to servers, it has come under attack with pressure from AMD unlike any it has felt in nearly a decade. It started with the release of AMD Ryzen 7 and a family of processors aimed at the mainstream user and enthusiast markets. That followed by the EPYC processor release moving in on Intel’s turf of the enterprise markets. And most recently, Ryzen Threadripper took a swing (and hit) at the HEDT (high-end desktop) market that Intel had created and held its own since the days of the Nehalem-based Core i7-920 CPU.
Between the time Threadripper was announced and when it shipped, Intel made an interesting move. It decided to launch and announce its updated family of HEDT processors dubbed Skylake-X. Only available in a 10-core model at first, the Core i9-7900X was the fastest tested processor in our labs, at the time. But it was rather quickly overtaken by the likes of the Threadripper 1950X that ran with 16-cores and 32-threads of processing. Intel had already revealed that its HEDT lineup would go to 18-core options, though availability and exact clock speeds remained in hiding until recently.
|i9-7980XE||i9-7960X||i9-7940X||i9-7920X||i9-7900X||i7-7820X||i7-7800X||TR 1950X||TR 1920X||TR 1900X|
|Base Clock||2.6 GHz||2.8 GHz||3.1 GHz||2.9 GHz||3.3 GHz||3.6 GHz||3.5 GHz||3.4 GHz||3.5 GHz||3.8 GHz|
|Turbo Boost 2.0||4.2 GHz||4.2 GHz||4.3 GHz||4.3 GHz||4.3 GHz||4.3 GHz||4.0 GHz||4.0 GHz||4.0 GHz||4.0 GHz|
|Turbo Boost Max 3.0||4.4 GHz||4.4 GHz||4.4 GHz||4.4 GHz||4.5 GHz||4.5 GHz||N/A||N/A||N/A||N/A|
|Memory Support||DDR4-2666 Quad Channel||DDR4-2666 Quad Channel||DDR4-2666 Quad Channel||DDR4-2666 Quad Channel||DDR4-2666
|DDR4-2666 Quad Channel||DDR4-2666 Quad Channel|
|TDP||165 watts||165 watts||165 watts||140 watts||140 watts||140 watts||140 watts||180 watts||180 watts||180 watts?|
Today we are now looking at both the Intel Core i9-7980XE and the Core i9-7960X, 18-core and 16-core processors, respectively. The goal from Intel is clear with the release: retake the crown as the highest performing consumer processor on the market. It will do that, but it does so at $700-1000 over the price of the Threadripper 1950X.
Subject: Processors, Chipsets | September 24, 2017 - 11:03 PM | Scott Michaud
Tagged: Z370, Intel, coffee lake
The official press deck for Coffee Lake-S was leaked to the public, so Intel gave us the go-ahead to discuss the product line-up in detail (minus benchmarks). While the chips are still manufactured on the 14nm process that Kaby Lake, Skylake, and Broadwell were produced on, there’s more on them. The line-up is as follows: Core i3 gets quad-core without HyperThreading and no turbo boosting, Core i5 gets six-core without HyperThreading but with Turbo boosting, and Core i7 gets six-core with HyperThreading and Turbo boosting.
While the slide deck claims that the CPU still has 16 PCIe 3.0 lanes, the whole platform supports up to 40. They specifically state “up to” over and over again, so I’m not sure whether that means “for Z370 boards” or if there will be some variation between individual boards. Keep in mind that only 16 lane of this are from the processor itself, the rest are simply a part of the chipset. This unchanged from Z270.
Moving on, Intel has been branding this as “Intel’s Best Gaming Desktop Processor” all throughout their presentation. The reasoning is probably two-fold. First, this is the category of processors that high-end, mainstream, but still enthusiast PC gamers target. Second, gaming, especially at super-high frame rates, is an area that AMD has been struggling with on their Ryzen platform.
Speaking of performance, the clock rate choice is quite interesting compared to Kaby Lake. In all cases, the base clock had a little dip from the previous generation, but the Turbo clock, if one exists, has a little bump. For instance, going from the Core i7-7700k to the Core i7-8700k, your base clock drops from 4.2 GHz to just 3.7 GHz, but the turbo jumps up from 4.5 GHz to 4.7 GHz. You also have a little more TDP to work with (95W vs 91W) with the 8700k. I’m not sure what this increase variance between low and high clock rates will mean, but it’s interesting to see Intel making some sort of trade-off on the back end.
(Editor's note: the base clock is only going to be a concern when running all cores for a long period of time. I fully expect performance to be higher for CFL-S parts than KBL-S parts in all workloads.)
The last thing that I’ll mention is that, of the two i3s, the two i5s, and the two i7s, one is locked (and lower TDP) and one is unlocked. In other words, Intel has an unlocked solution in all three classifications, even the i3. Even though it doesn’t have a turbo clock setting, you can still overclock it by hand if you desire.
Prices range from $117 to $359 USD, as seen in the slide, above. They launch on October 5th.
Subject: Processors, Chipsets | September 23, 2017 - 06:52 PM | Scott Michaud
Tagged: Z370, z270, kaby lake, Intel, coffee lake
According to the Netherlands arm of Hardware.info, while Kaby Lake-based processors will physically fit into the LGA-1151 socket of Z370 motherboards, they will fail to boot. Since their post, Guru3D asked around to various motherboard manufacturers, and they claim that Intel is only going to support 8th Generation processors with that chipset via, again, allegedly, a firmware lock-out.
Thankfully, it's not Chocolate Lake.
Image credit: The Red List
If this is true, then it might be possible for Intel to allow board vendors to release a new BIOS that supports these older processors. Guru3D even goes one step further and suggests that, just maybe, motherboard vendors might have been able to support Coffee Lake on Z270 as well, if Intel would let them. I’m... skeptical about that last part in particular, but, regardless, it looks like you won’t have an upgrade path, even though the socket is identical.
It’s also interesting to think about the issue that Hardware.info experienced: the boot failed on the GPU step. The prevailing interpretation is that everything up to that point is close enough that the BIOS didn’t even think to fail.
My interpretation of the step that booting failed, however, is wondering whether there’s something odd about the new graphics setup that made Intel pull support for Z270. Also, Intel usually supports two CPU generations with each chipset, so we had no real reason to believe that Skylake and Kaby Lake would carry over except for the stalling of process tech keeping us on 14nm so long.
Still, if older CPUs are incompatible with Z370, and for purely artificial reasons, then that’s kind-of pathetic. Maybe I’m odd, but I tend to buy a new motherboard with new CPUs anyway, but I can’t envision the number of people who flash BIOSes with their old CPU before upgrading to a new one is all that high, so it seems a little petty to nickel and dime the few that do, especially at a time that AMD can legitimately call them out for it.
There has to be a reason, right?
Subject: Processors | September 18, 2017 - 05:13 PM | Jeremy Hellstrom
Tagged: linux, EPYC 7601, EPYC
Phoronix have been hard at work testing out AMD's new server chip, specifically the 2.2/2.7/3.2GHz EPYC 7601 with 32 physical cores. The frequency numbers now have a third member which is the top frequency all 32 cores can hit simultaneously, for this processor that would be 2.7GHz. Benchmarking server processors is somewhat different from testing consumer CPUs, gaming performance is not as important as dealing with specific productivity applications. Phoronix started their testing of EPYC, in both NUMA and non-NUMA configurations, comparing against several Xeon models and the performance delta is quite impressive, sometimes leaving even a system with dual Xeon Gold 6138's in the dust. They also followed up with a look at how EPYC compares to Opteron, AMD's last server offerings. The evolution is something to behold.
"By now you have likely seen our initial AMD EPYC 7601 Linux benchmarks. If you haven't, check them out, EPYC does really deliver on being competitive with current Intel hardware in the highly threaded space. If you have been curious to see some power numbers on EPYC, here they are from the Tyan Transport SX TN70A-B8026 2U server. Making things more interesting are some comparison benchmarks showing how the AMD EPYC performance compares to AMD Opteron processors from about ten years ago."
Here are some more Processor articles from around the web:
- Core i7 vs. Ryzen 5 with Vega 64 & GTX 1080 @ TechSpot
- AMD Threadripper 1950X Linux Benchmarks @ Phoronix
- The Top 5 Best CPUs of All Time @ [H]ard|OCP
Subject: Processors | September 6, 2017 - 12:06 PM | Ryan Shrout
Tagged: Intel, 6700k, 6600k
Initially launched in August of 2015, the Skylake consumer desktop processors are finally ramping down at Intel production facilities. Based on this Intel Product Change Notification, the widely coveted Core i7-6700K and Core i5-6600K, along with several other parts, are being EOL'd (end of life).
If you are an OEM or just really love building around these parts, you still have some time to get your orders in. You have until March 30, 2018 to place your final requests and the final shipment date will be a year from today. If you ever were curious how complex the ramp down on parts that ship to thousands of different customers in consumer, enterprise, and embedded markets, the table above should give you a glimpse.
(Probably that 3017 date is a typo...)
With the focus for Intel squarely on the Kaby Lake Core i7-7700K and Core i5-7600K, in addition to the upcoming Coffee Lake refresh 8000-series, enthusiasts might be wondering why it took Intel so long to shut things down on this set of Skylake parts.
I think it is safe to say that this marks the end of an interesting window of time for Intel, where it had clear and uncontested dominance of the consumer processor market. Re-reading my conclusion to the 6700K review reveals almost no mention of a relevant AMD competing part, making today's situation with Ryzen and Threadripper all the more impressive.
Subject: Processors | September 5, 2017 - 11:16 AM | Ryan Shrout
Tagged: skylake-x, Intel
We are just starting to ramp back up here after the long holiday weekend, so let's start with something that is both interesting and easy to absorb. High-profile overclocker Der8auer has gotten his hands on an 18-core Skylake-X processor and did exactly what you would expect - delidded it.
The takeaway from this is two-fold. First, the die appears very clean, indicating that Intel has still not decided to solder these high-end processors and is going with a standard thermal interface between the die and the heat spreader.
Also...it's friggin huge. Look at the 10-core die from the Core i9-7900X that was observed earlier this year and compare it to the image above.
Though the camera angles aren't ideal, comparing the layout of the die to the physical substrate, which IS the same size between all the Skylake-X processors, you can see how much larger this 18-core die truly is. Expect to see the 18, 16, 14, and even the 12-core processors to use the same physical die.
Subject: Processors | August 29, 2017 - 12:00 PM | Sebastian Peak
Tagged: Xeon W, xeon scalable, xeon, workstation, processor, Intel, cpu
Intel has officially announced their new workstation processor lineup, with Xeon Scalable and Xeon W versions aimed at both professional and mainstream workstation systems.
"Workstations powered by Intel Xeon processors meet the most stringent demands for professionals seeking to increase productivity and rapidly bring data to life. Intel today disclosed that the world-record performance of the Intel Xeon Scalable processors is now available for next-generation expert workstations to enable photorealistic design, modeling, artificial intelligence (AI) analytics, and virtual-reality (VR) content creation."
The first part of Intel’s product launch announcement are the new Xeon Scalable processors, first announced in July, and these are dual-socket solutions targeting professional workstations. Versions with up to 56 cores/112 threads are available, and frequencies of up to 4.20 GHz are possible via Turbo Boost. Intel is emphasising the large performance impact of upgrading to these new Xeon processors with a comparison to older equipment (a trend in the industry of late), which is relevant when considering the professional market where upgrades are far slower than the enthusiast desktop segment:
“Expert workstations will experience up to a 2.71x boost in performance compared to a 4-year-old system and up to 1.65x higher performance compared to the previous generation.”
The second part of announcement are new Xeon W processors, which will be part of Intel’s mainstream workstation offering. These are single-socket processors, with up to 18 cores/36 threads and Turbo Boost frequencies up to 4.50 GHz. The performance impact with these new Xeon W CPUs compared to previous generations is not as great as the Xeon Scalable processors above, as Intel offers the same comparison to older hardware with the Xeon W:
“Mainstream workstations will experience up to a 1.87x boost in performance compared to a 4-year-old system4 and up to 1.38x higher performance compared to the previous generation.”
Full PR is available from Intel's newsroom.
Subject: Processors | August 24, 2017 - 12:43 PM | Jeremy Hellstrom
Tagged: XSPC, amd, Threadripper, overclocking, Raystorm
For those convinced that the Threadripper is being held back by poorly endowed partners, [H]ard|OCP received the new XSPC RayStorm which has a cold plate as large as Threadrippers heatspreader. As you can see from the picture, new habits will need to e learned when spreading the TIM on such a large area so keep an eye out for tips or carefully experiment on your own. The heatsink let [H] reach a solid 4GHz on all 16 cores with a 3200MHz memory clock, at significantly lower voltages than Ryzen required to reach the same frequency. Even better news is that this is not the limit, [H] intends to test again using a more powerful radiator and expects to see an even better overclock.
"XSPC got us over one of its first waterblocks specifically designed to help handle Ryzen Threadripper CPU's heat while overclocking. We give you a quick unboxing, break down the block itself, and then we look at Threadripper long-term performance. We finally get it dialed in at 4GHz."
Here are some more Processor articles from around the web:
- AMD's Ryzen Threadripper 1920X and Ryzen Threadripper 1950X @ Tech Report
- Ryzen 3 vs. Core i5-2500K vs. FX-8370 @ Techspot
- AMD Ryzen 3 1300X and Ryzen 3 1200 @ Modders-Inc
- Athlon II X3 vs. Ryzen 3: How AMD's Performance Has Evolved & Performance-Per-Watt @ Phoronix
Subject: Processors | August 10, 2017 - 03:55 PM | Jeremy Hellstrom
Tagged: Zen, X399, Threadripper, ryzen, amd, 1950x, 1920x
When you look at the results Ryan posted, it was clear that when it comes to video rendering and other content creation it is AMD's chip which comes out ahead in performance, and at a better price point that Intel's Core i9. Don't just take our word for it, many others reviewed the new chips, including [H]ard|OCP. Their results agree, showing that the only advantage Intel has is in single threaded applications, in which case the frequency of the 4.6GHz Intel part can outpace the 4GHz Threadripper. Those picking up Threadripper have no interest in single threaded applications, they prefer their programs to be spread across multiple cores and not only does Threadripper have the most cores, it allows you to flip between NUMA and UMA depending on your preference. Check out [H]'s review here before continuing below the fold.
"The day is finally upon us that many CPU enthusiasts have been waiting for. We get to see what AMD's new Threadripper CPU is all about in terms of performance, and in attempts to cool the beast. There has been no lack of hype for months now, so let's see if it is all justified."
Here are some more Processor articles from around the web:
- AMD Threadripper 1950X review: Better than Intel in almost every way @ Ars Technica
- AMD Ryzen Threadripper 1950X @ Kitguru
- AMD Ryzen Threadripper 1950X & 1920X @ Techspot
- AMD Ryzen Threadripper 1920X & 1950X Review @ Hardware Canucks
- Intel Core i7 7820X Linux Benchmarks @ Phoronix
Who is this for, anyway?
Today is a critically important day for AMD. With the launch of reviews and the on-sale date for its new Ryzen Threadripper processor family, AMD is reentering the world of high-end consumer processors that it has been absent from for a decade, if not longer. Intel has dominated this high priced, but high margin, area of the market since the release of the Core i7-900 series of Nehalem CPUs in 2008, bringing workstation and server class hardware down to the content creator and enthusiast markets. Even at that point AMD had no competitive answer, with only the Phenom X4 in our comparison charts. It didn’t end well.
AMD has made no attempt of stealth with the release of Ryzen Threadripper, instead adopting the “tease and repeat” campaign style that Radeon has utilized in recent years for this release. The result of which is an already-knowledgeable group of pre-order ready consumers; not a coincidence. Today I will summarize the data we already know for those of you just joining us and dive into the importance of the new information we can provide today. That includes interesting technical details on the multi-die implementation and latency, overclocking, thermals, why AMD has a NUMA/UMA issue, gaming performance and of course, general system and workload benchmarks.
A Summary of Threadripper
AMD has been pumping up interest and excitement for Ryzen Threadripper since May, with an announcement of the parts at the company’s financial analyst day. It teased 16 cores and 32 threads of performance for a single consumer socket, something that we had never seen before. At Computex, Jim Anderson got on stage and told us that each Threadripper processor would have access to 64 lanes of PCI Express, exceeding the 40 lanes of Intel’s top HEDT platforms and going well above the 28 lanes that the lower end of its family offers.
In mid-July the official announcement of the Ryzen Threadripper 1950X and 1920X occurred, with CEO Lisa Su and CVP John Taylor having the honors. This announcement broke with most of the important information including core count, clock speeds, pricing, and a single performance benchmark (Cinebench). On July 24th we started to see pictures of the Threadripper packaging show up on AMD social media accounts, getting way more attention than anyone expected a box for a CPU could round up. At the end of July AMD announced a third Threadripper processor (due in late August). Finally, on August 3rd, I was allowed to share an unboxing of the review kit and the CPU itself as well as demonstrate the new installation method for this sled-based processor.
It’s been a busy summer.