I'm not really part of the "prosumer" market (what a word!), since I don't have a burning need for that much power, but it helps move prices lower in the middle and bottom of the performance range when the top end improves.
Is there any chance that they are not having all OEMs by the balls this time too?
They were bribing and blackmailing OEMs.
Which to me says that Intel is not competitive. If you can only release a slightly better product at twice the price, then you failed. Of course Intel's products have some branding inertia behind them, too, but that will be short-lived if the products don't become more competitive in the near future.
[1] Kaby Lake-X, 4/8 cores/threads, 4.3/4.5 freq/turbo, $329
https://www.anandtech.com/show/11549/the-intel-kaby-lake-x-i...
https://www.anandtech.com/show/11697/the-amd-ryzen-threadrip...
> And, in theory, I suspect that a dual-7740X build might match or beat TR in multithreaded, too.
Is that possible? And I really doubt you'll get 8 cores to beat 16.
My point is it's the category most enthusiasts would be happiest with, even if they think otherwise.
>>>Is that possible?
Looks like I was mistaken, I thought the Sky/Kaby Lake-X chips used the same socket as the Xeons. So no.
>>>And I really doubt you'll get 8 cores to beat 16.
If it were possible, it would be 8 cores operating at a 50% higher frequency (with substantially more room for overclocking, too) and that support AVX-512 instructions. And you get quickly-diminishing returns from throwing more threads at a program (just mathematically, even if it's embarrassingly parallel and CPU-bound, each time you double the threadcount you get half the improvement of the previous doubling).
That's fine, but use that category for both manufacturers.
> 8 cores operating at a 50% higher frequency
26% higher base, 10% higher boost.
With 3 glaring factual errors across two posts, I'm going to go out on a limb and suggest to you that the average enthusiast purchasing Threadripper is in fact more aware of the trade offs between the products available on the market than you are.
In generally it actually doesn't. The AMD Ryzen 7 1800x is selling for less than half the price of an Intel i7-6900k ($420 Vs $1010), and in benchmarks the Ryzen system beats the i7 or practically matches the performance.
Getting the same performance for 40% of the cost is something that does not depend on your use case.
Sure, but in many cases the practical difference between an Intel i7-6900k and for example an Intel i3-7XXX (or a 4 year old i5/7) is basically negligible, and the i3 is less than half the price of the Ryzen
What makes AMD more interesting?
The fact that if you want a chip that powerful, there is only Intel. An example in GPUs are the NVIDIA Quaddro cards. If your frame needs more than 16GB memory to render, there is no other card available. So they sell the same chips packaged as a Quaddro at 5x the price of a Titan.
[1] http://www.nvidia.com/object/quadro-design-and-manufacturing...
[2] http://www.nvidia.com/object/media-and-entertainment.html
How does, e.g., POWER8 compare? (And POWER9, soon?)
That said, POWER tends to be even more expensive than this.
POWER8 is not well suited for HPC workloads as their vector instructions (VMX/AltiVec) are only 128-bit wide vs Intel's latest Skylake Xeon CPUs have vector instructions (AVX512) that are 512-bit wide.
For non HPC workloads (e.g. integer arithmetic) the POWER8 CPUs were somehow competitive with Intel's CPUs when they were first released in 2015 but today AMD and Intel's CPUs are clearly faster.
From what I can tell CPU enthusiasts were not impressed by the POWER8 CPU architecture.
In the case of nVidia the quadro cards are the only ones sold as manufactured by the company and are incredibly high binned.
Lots of people on this thread are conflating a few things such as ML, but I work in an industry where we deal a lot with GPUs and its certainly a fact that flipped bits are significantly more common than many think. If you do any work that requires the output to be nonvolatile then you can run into issues.
http://ieeexplore.ieee.org/document/5493404/?arnumber=549340...
And this specific example here is clearl in the high-end enthusiast space, where whoever has the best can charge the highest price and the target market will pay it.
There may be some that are down to the wire per unit cost (typically the cheaper ones) but I have a feeling Intel operates with plenty of margin on most of their higher end products. And plenty of them are the same chip repackaged.
The only thing that Intel surpassed them is single-thread performance. It's important even on workstation, IMO, that's the main reason I'm thinking to stick with Intel for now, everything else AMD made better.
If anything AMD should follow NVIDIA in supporting OpenACC which is the way to go as far as true probability goes.
It doesn’t matter that the rocm libs are optimised for AMD.
In all effectiveness ROCm is AMD's version of CUDA it includes it's own run time and driver stack which effectively replaced the OpenCL frontend for AMD GPUs, basically it gives AMD the ability to do with their GPUs what CUDA allowed NVIDIA to do and it's not to be dependent on a camel (OpenCL).
This is what ROCm is: https://rocm.github.io/images/ROCm_Stack.png or at least this is what AMD is pushing it to be, currently a lot of things are still unfinished or non-existent.
The representation of it being open is misleading, while more of the source code is available than for CUDA, ROCm isn't vendor agnostic on any tangible level. In 2-3 years once it's somewhat in a production ready state it will lock you into a single vendor just like CUDA does.