Intel Releases Final Core i9 Specs and Release Dates -- And Threadripper Is Faster (Sometimes) (pcworld.com) 91
On Monday, Intel took the wraps of final details of its Core i9 microprocessors. From a report: Remember that Intel's Core X-series family (also called the Core i9) was announced with several key omissions: namely the clock speeds of the 12-core Core i9-7920X and above, as well as the thermal design power, or TDP. On Monday, Intel filled those in. The 12-core Core i9-7920X launches Aug. 28 while the 14-, 16-, and 18-core Core i9 chips ship on Sept. 25. Perhaps most important, though, is that we now know how fast Intel's Core i9s will run. When Intel inadvertently revealed that its 12-core Core i9-7920X was 2.9-GHz -- slower than the comparable AMD Threadripper -- a subset of the internet had a small freakout. We now know that that will be true for the remaining Core i9s as well, but with a big caveat. Here are the remaining speeds and feeds for the high-end Core i9 chips:
Core i9-7980XE (18 cores, 36 threads): 2.6GHz; Boost, 4.2GHz to 4.4 GHz.
Core i9-7960X (16 cores, 32 threads): 2.8GHz; Boost, 4.2GHz to 4.4 GHz.
Core i9-7940X 14 cores, 28 threads: 3.1GHz; Boost: 4.3GHz to 4.4GHz.
Core i9-7920X (12 cores, 24 threads): 2.9-GHz; Boost: 4.3-GHz to 4.4GHz.
Note that the boost speeds refer to both Intel's Turbo Boost Technology 2.0 and 3.0. [...] Essentially, both Intel and AMD can claim the title of fastest processor. Threadripper's base clock speeds are faster, but Intel's boost speeds climb higher than Threadripper can. It's also important to note that while Threadripper consumes 180 watts, even the fastest Core i9 chips Intel has announced have a lower TDP of 165 watts.
Core i9-7980XE (18 cores, 36 threads): 2.6GHz; Boost, 4.2GHz to 4.4 GHz.
Core i9-7960X (16 cores, 32 threads): 2.8GHz; Boost, 4.2GHz to 4.4 GHz.
Core i9-7940X 14 cores, 28 threads: 3.1GHz; Boost: 4.3GHz to 4.4GHz.
Core i9-7920X (12 cores, 24 threads): 2.9-GHz; Boost: 4.3-GHz to 4.4GHz.
Note that the boost speeds refer to both Intel's Turbo Boost Technology 2.0 and 3.0. [...] Essentially, both Intel and AMD can claim the title of fastest processor. Threadripper's base clock speeds are faster, but Intel's boost speeds climb higher than Threadripper can. It's also important to note that while Threadripper consumes 180 watts, even the fastest Core i9 chips Intel has announced have a lower TDP of 165 watts.
Re: Accumulation of bad karma... (Score:2, Interesting)
Intel compilers are used for a lot of backend libraries (eg. parts of Matlab), and tens to turn off some SSE extensions even on AMD processors that support them. Lack of vectorization leads to worse performance, 'nuff said.
Re:Accumulation of bad karma... (Score:4, Interesting)
Depends on longevity in the market.
Back in the Netburst days, you marvelled at how Intel did so poorly despite 'looking' like it should be faster.
From Conroe to about Bulldozer, things were about the same.
The bulldozer screwed up AMD in the same way that Netburst screwed up Intel for a while. Meanwhile Intel progressed well.
Now with Zen, at least on desktop it's back to mostly neck and neck. In high end server, it's a mixed bag, Epyc having more memory channels means better capacity, but individual memory performance is equal to their desktop product. This is fine for a lot of applications (e.g. VDI, similar virtualization) and gets more aggregate performance and capacity, though single thread/process memory throughput takes a hit.
The Megahertz Myth is alive and well (Score:4)
... on slashdot no less. How about some real workload comparisons?
Re: (Score:1)
Software is the real bottleneck (Score:2, Interesting)
We're talking about CPUs with 24 or more threads of execution here. The bottleneck isn't hardware, it's the software. Most software just doesn't support that high degree of threading. That's why we need to start using programming languages built around concurrency, like Rust and Erlang. Rust is a great example. Look at the innovation we're seeing from Mozilla with their new Servo web browser engine. It is being designed from the ground up using Rust so that it can make full use of CPUs with many simultaneou
Re: (Score:1)
Stop trying to make Rust happen.
TRUE @ an application level, but... apk (Score:1)
See subject - The OS process scheduler itself can use the extra cores for BIG overall gains ala e.g. in Windows:
Increase cpu core count @ hardware level
Windows Registry Editor Version 5.00
[HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Control\Session Manager\Executive]
"AdditionalCriticalWorkerThreads"=dword:00000008
"AdditionalDelayedWorkerThreads"=dword:00000008
* I.E. - How much extra cores will help BEYOND today's CPUs for the OPERATING SYSTEM itself (in Critical Worker Threads) in juggling threads in itself
Re: (Score:1)
The internets are stupid. Intel have been way below their Pentium 4 clock frequencies ever since they proved it was a terrible optimization goal.
Re: (Score:2)
that's probably because, to first order, MHz does often roughly correspond to real processing power.
(or at least it does in my comparisons for scientific floating point number crunching for a variety of systems)
Re: (Score:2)
So I shouldn't upgrade and just stick with my 6 core Sandy Bridge E ?
I paid $600 about 5 years ago.
Re: (Score:2)
So to a first order, floating point units and SIMD extensions are irrelevant?
You might want to check your calculations again.
Re:The Megahertz Myth is alive and well (Score:5, Insightful)
In real workloads, the 16-core Threadripper (16 cores, 3.4GHz) @ $1000 destroys the 10-core i9 (also $1000)
The problem for Intel is it also destroys the the 12-core i9 ($1200) and 14-core i9 ($1400). In all likelihood it's going to match the 16-core i9 ($1700) and lose overall (by a small margin) to the 18-core i9, which is $2000.
The fact that PCWorld haven't explictly mentioned price, or even implicitly mentioned TR's 64 PCIe lanes and the other benefits (a stable socket, cheaper boards, lower power draw than the i9s by a huge distance, etc.) is just testamanet to how much the mainstream tech press shill for Intel.
There is essentially no reason to buy the i9s. If you want the best workstation performance, you buy Threadripper. If you want the best gaming performance, you buy the i7-7700K. The i9s are an absurdity, and a panicked reaction to AMD's massive performance gains.
People forget that Intel had only planned to release a 10-core i7 based on Skylake-EP. Once they got wind of Threadripper they bolted on a 12-core, then a 14-core, then a 16-core, and finally (when TR was shown to have incredible workstation performance) the $2000 waste of space that is the 18-core 2.6GHz i9-7980XE with only 44 PCIe lanes.
Meanwhile, for half that cost AMD give you 16 cores @ 3.4GHz, 64 PCIe lanes, cheaper motherboards, and most importantly amazing workstation performance. It's not even a contest.
Re: (Score:2)
On what basis do you say that Intel hasn't designed for greater core scaling?
Re:The Megahertz Myth is alive and well (Score:5, Interesting)
AMD dumps FSB in favor of HT in 2001
Intel dumps FSB in favor of QPI in 2009.
AMD switches from HT to Infinity Fabric in 2017.
Intel switches from QPI to ????? in ????
Intel had a few good wins in the past several decades, especially getting 14nm out the door so much sooner than any other fab company could get 14nm workable, but Intel has not had any wins afaict in actual processor design in that same period. When AMD went dual core, Intel had to go frankenchip with double-die solutions as well as hyper-threading in order to come close to keeping up in processor design.
The processor design space of Intel is littered with failures starting with Pentium 4. The start of the Core series was just Intel scrapping the Pentium 4 design completely and going back to the Pentium 3 design.
Intel needs a lithography breakthrough, and they need it yesterday. They have proven that they dont do actual design well.
Re: (Score:2)
They have proven that they dont do actual design well.
Seeing as how Intel's performance has been crushing AMD for almost a decade, that's the stupidest thing I've read this month.
Re: (Score:2)
But AMD's newest have an issue. 0-60 in so many seconds sounds good...however 0-60... segfault doesn't. Until AMD fixes some of the shortcoming on these new chips, there's some risk. Fast is nice, but fast and reliable is better.
I like that AMD is making a attempt at getting back in the game, but I can't afford a system to be down when under heavy load. And arguably it's the point of going with a higher end chip to begin with.
Re:The Megahertz Myth is alive and well (Score:5, Informative)
The bottom line is that they are both blazingly fast, especially for multi-threaded workloads. So when building a workstation there are more important things to worry about than a few percent in this benchmark or that benchmark.
Threadripper has more PCIe lanes, a definite bonus. Not only useful now for things like NVMe SSDs and RAID cards, but in future for adding a USB 4.0 card or whatever. If you want your system to last longer then expandability is key.
Threadripper motherboards also seem to be a bit cheaper, and of course the CPUs themselves are. For the longer term, AMD tend to stick by their sockets for longer so chances are if there is a major CPU upgrade option available (unlikely) you will be able to drop it in.
Re: (Score:2)
The problem for AMD is that most desktop workloads aren't multi-threaded, and many that are (like HandBrake) show diminishing returns after 4 threads.
Heck, Firefox still only uses a few threads.
Six cores (from AMD, or four from Intel) really does seem to offer the best performance for "average" users. (I won't speak to CAD, Mathematica, etc. users.)
Re:The Megahertz Myth is alive and well (Score:4, Interesting)
Handbrake's outdated documentation says it has diminishing returns past 6 cores. Current users post that the diminishing returns start after 10 cores. The highest core count for Threadripper is 16, so there's definitely some overkill there on the high end... but, that just means that you've got 6 cores to do something else with. Maybe even transcode more than one video simultaneously.
You're absolutely right that 4-6 cores is perfectly fine for most people, but in order to make the leap to the next generation of computing with the next generation of applications, we need to move beyond that.
A/V, 3D-rendering, VR, and AI all benefit from multi-core... and Firefox is being re-written in Rust to better take advantage of multi-core.
Personally, I'd love to have my own Jarvis AI straight out of Iron Man on my 64-core supercomputer I got in the $2,000 range and runs google's open-source AI. I'd like to have it power my VR setup right out of Lawnmower Man, but with modern graphics. Love to have my own Industrial Light & Magic / Pixar 3D animation studio that my nephews can play with to make movies for our entertainment... and render things that would normally take months or years on today's average computers.
CPUs aren't going to get much faster GHz-wise with current techniques, but we can cram in more cores to do things that no one ever dreamed they could do with a home PC. These threadrippers are game-changers. It's not fair to judge their usefulness by what people use their current CPUs for. That's sort of like judging modern 4 and 6 core CPUs by what single-core Pentiums were used for 20 years ago back when multimedia on PCs was still in its infancy.
Re: (Score:2)
in order to make the leap to the next generation of computing with the next generation of applications, we need to move beyond that.
I'm pretty sure I read that exact quote 10 years ago when dual-core CPUs came out.
Re: (Score:2)
Current users post that the diminishing returns start after 10 cores. The highest core count for Threadripper is 16, so there's definitely some overkill there on the high end...
As far as gaming goes, one word: Vulkan.
Re: (Score:2)
Assumes that all new games will be written in Vulkan instead of DX10.
(DX12 has the same feature set as Vulkan regarding multi-threading, but I'm not holding my breath on it's wild adoption either.
Re: (Score:2)
Assumes that all new games will be written in Vulkan instead of DX10.
(DX12 has the same feature set as Vulkan regarding multi-threading, but I'm not holding my breath on it's wild adoption either.
Then you haven't been paying attention. Every major game engine has a Vulkan prototype up and running, or (like idsoft) already shipping. DX10 is Microsoft's Vulkan clone, so yes, the whole graphics world is going to Vulkan, the process is well advanced. Anything that wants performance or (for mobile) power saving is going Vulkan.
Re: (Score:2)
DX10 was release 10 years ago with Windows Vista, whereas DICE started working on the predecesor to Vulkan just four years ago.
Thus... I'm a bit skeptical about your claims.
Re: (Score:2)
Simple typo, meant DX 12, which you should have known, so I am skeptical about your moral backbone.
Re: (Score:2)
I can't read your mind.
Re: (Score:2)
Where did I even hint that I'm going to buy an i9 or TR? My old FX-6100 is perfectly adequate to the tasks I set it to.
Re: (Score:2)
Firefox still only uses a few threads.
That's due to change in the very near future.
Re: (Score:2)
The bottom line is that they are both blazingly fast, especially for multi-threaded workloads.
And that's enough for me to happily abandon aging monopolist Intel at the first viable opportunity.
Re:How could this possibly happen? (Score:4, Insightful)
Intel can't help themselves can they? (Score:5, Insightful)
That market segmentation is like a parasite to their thinking. They can't do anything without chopping the product line into as small as bites as they're able.
Re: (Score:2)
I will freely admit that sticking it to the man is on the list of considerations for me. Only as long as price/performance makes sense, though.
Intel's TDP lies (Score:3, Interesting)
"It's also important to note that while Threadripper consumes 180 watts, even the fastest Core i9 chips Intel has announced have a lower TDP of 165 watts."
The actual power draw of even the 10-core i9 is >200W. Intel are deceiving us yet again.
Re: (Score:3)
Re: (Score:2)
This is not new. The definition that Intel uses for TDP is different from that used by AMD so you can't compare raw TDP numbers across manufacturers. AMD's numbers are more conservative so I'd expect an AMD 165W product to use less power than an Intel 180W product in most situations.
Threadripper is also cheaper. (Score:1)
The 16-core i9-7960X is ~$1700, while the 1950X Threadripper is $999. And the Threadripper X models are the higher-clocked/premium version. If AMD has a non-X model in the works, the pricing difference could be even greater (AMD didn't reveal the X/non-X version of the R3 until way later, so it's possible).
Threadripper Is half the cost with more pci-e lane (Score:5, Informative)
Threadripper Is half the cost with more pci-e lanes.
On Intel it's about $1000 min to get 48 pci-e lanes. AMD $550 for there 64 pci-e lane one.
Re: (Score:2)
Threadripper Is half the cost with more pci-e lanes.
On Intel it's about $1000 min to get 48 pci-e lanes. AMD $550 for there 64 pci-e lane one.
In addition to the base price, what matters here (for the datacenter market) is how much computing can be done per Watt with your average server application. We need a third-party review for that because AMD exaggerates while Intel lies it's ass off about everything.
Re: (Score:2)
Re: (Score:2)
It's pretty close depending on application and usage not really.
AMD also has usb 3.1 on die.
Intels 24 pci-e on chipset is also stacked with sata / usb / MB IO / sound pci-e / network link and do lower end cpus only having as low as 16 pci-e lanes. Some boards also wire storage M.2 slots to the chipset as well.
Core i9 is just too dam hot (Score:2, Interesting)
Re: (Score:2)
Re: (Score:2)
Data centers don't care about noise or heating like an oven, they have massive power supplies and massive cooling. At worst you'll replace one 4P server with one 2P server with 2x the cores. Or two 2P VM platforms with one 2P VM platform with the same number of cores. I very much doubt they'll ever feel it's too much processing power in one place.
Re:Core i9 is just too dam hot (Score:4, Informative)
Data centers don't care about noise or heating like an oven, they have massive power supplies and massive cooling. At worst you'll replace one 4P server with one 2P server with 2x the cores. Or two 2P VM platforms with one 2P VM platform with the same number of cores. I very much doubt they'll ever feel it's too much processing power in one place.
This is just plain wrong -- data centers care very deeply about power consumption. Since data centers pack thousands of CPUs into warehouse-size rooms, even a single extra watt per CPU is a big deal. Higher power consumption means a greater power delivery requirement for the room, more cooling, a bigger UPS requirement, and a bigger power bill. Operating costs are driven by transactions per second per watt.
High performance computing and high-end desktops are pretty much the only remaining markets where power consumption is a secondary concern. This has not been true for data centers in a long time.
Is this really an upgrade? (Score:2)
Great timing for spreading that FUD, Intel... (Score:5, Interesting)
Re: (Score:2)
I thought the review embargo was lifted tomorrow or something. They've had the unboxings now and there's been a fractal design system IIRC reviewed, but just the one system
Faster? (Score:1)
And Threadripper Is Faster
And the performance of a processor is not simply clock speed X cores.
Re: (Score:2)
Point is, we will know when someone runs real world tests across a full suite of applications and benchmarks.
Re: (Score:2)
The Zen architecture IPC is pretty much on par with the Core one as long you're not using the intel AVX extensions.
Good lord that's fast! (Score:2)
Cool! CNN's terrible advertising scripting can lock up Chrome even faster now!