Wow. AMD has really pulled through on this one. They are really giving Intel a run for their money. This is going to be great for consumers. What is interesting is the HUGE jump in multi-threaded performance on AMD. Very interesting.
Indeed. Single threaded performance, however, seems bad. Virtual cores most likely share less of the actual core than on Intel parts, allowing one to compete less with the other and get more done. On my i7 laptop, on anything more demanding, I see half the "cores" idling, probably because one is using the actual underlying core while the other is waiting.
Still, while AMD obviously did a good job, it will not make the Intel guys lose any sleep for now.
The server parts are another story. If they manage to reliably outperform Xeons (even if it's only on a per watt basis), it'll be fun to watch.
Right, thats exactly what happens with Hyper-Threading. If you can fully utilize the physical core, Hyper-Threading can't do much.
From the benchmarks, it seems a single thread can fully utilize all resources from an Intel processor while it can't on the new AMD ones and that's why their single-thread performance is lower and their multi-thread performance is higher.
We should keep in mind the benchmarks know nothing about the new AMD parts and how to keep all parts of the processor busy.
Yeah, as would be expected from most people on this site. Sad thing is, there's a bunch of pissed-off gamers right now who were on the hype train expecting Ryzen to be the next big thing in gaming, but it's far from it. Ryzen is great, just not amazing for gaming.
EDIT, in response to the other edit: sorry about that, I wasn't sure which comparison you were making. Still, the difference is slim enough that if I were in the market, I'd go with this one just to put the heat on Intel.
[1] http://tvtropes.org/pmwiki/pmwiki.php/Main/SeinfeldIsUnfunny (apologies for the TVTropes link at night!)
http://timmaurer.com/2012/01/16/horse-sense/
https://www.quora.com/Is-it-true-that-two-horses-pulling-tog...
I'm about to buy a beefy machine for FPGA development, I'm definitely considering AMD. Seems like I'd get better performance for a much lower price. I'm still waiting for more in-depth reviews though.
[1] http://core0.staticworld.net/images/article/2017/03/ryzen_ci...
Web and office? Single-threaded?
Why are these benchmarks, specially in 2018?
I do understand where you're coming from, but real-world performance is important, especially when that world is imperfect.
That's a load of bullshit. No one ever complained that the web is unbrowsable or an excel spreadsheet is unworkable without a high-end boutique processor. Those benchmarks are simply a load of bullshit that evaluate something that is utterly irrelevant.
With AMD's previous offerings, they benched well but ran like crap because so many applications are single-threaded. These single thread tests are important to know how these chips will work in real world scenarios.
This is the data enterprise buyers want to hear. If AMD can match Intel at a much lower cost, expect enterprise buyers to switch.
That said, those numbers aren't particularly bad for a new CPU running un-optimized software. All the important software will be updated soon enough once the compilers are in order.
https://www.pcper.com/reviews/Processors/AMD-Ryzen-7-1800X-R...
Perhaps you'd like to restrict yourself to 32 bit software just to be standards conformant?
If you do more than games, the Ryzen seems like a a good choice. It has serious power in applications, and is still fast in Games - which is a big step up compared to the FX. But no, it is not the goto choice for a gaming PC.
I'd take Intel CPU that runs 120 FPS average over AMD that runs 130 FPS average, when lowest FPS for Intel drops to 100 FPS and for AMD it drops to 80 FPS...
The Ryzen quads will come. I'm not at all sure what your point is here.
By ~15% compared to >160% for Bulldozer.
And this is day zero before any applications have been optimized for the new microarchitecture, compared with Core which everything has been targeting for more than a decade.
They did good.
Would you do it if you had the choice between that or adding new money making features/fixing bugs?
Of course I would, because people are going to benchmark my software against my competitor's software and choose the one that runs faster on their hardware.
Especially if my software is a compiler and people are benchmarking the programs it produces against other compilers, which is where most of the optimizing actually happens these days.
As for the priorities of the compiler devs, I think it will get some love.
Besides, if your application is one where performance is a serious issue, then you are going to spend some time optimizing rather than adding features. For some applications, performance is a feature.
Do you really care if your webpage renders for half a second longer?
If I was after encoding performance, I wouldn't buy a Ryzen. If I was after rendering performance, I wouldn't buy a Ryzen. Throw an mid-range nVidia on there for OpenCL, and the only CPU that makes sense for that price range is the 7700K.
If you're after pure performance, then you simply go with the best for your use case, whether it is Ryzen or not.
If you're on a tight budget, however, things get interesting.
hence, "things get interesting"
I expect if AMD went back through again, they could improve them even more.
(In the handbrake case, most likely it's half-speed AVX2 that is hurting them)
But why? http://www.anandtech.com/show/11170/the-amd-zen-and-ryzen-7-...
The benchmarks show Ryzen being the obvious choice other than cinebench single threaded.
Only if you are using a shitty software renderer or are completely stupid would you do rendering on the CPU.
The Ryzen performs nicely on the encoding tests, though that will be performed by the GPU soon. It already can be (NVENC and AMD VCE), but the quality isn't quite as good as everyone's favorite h.264 encoder, x264.
You're also missing that even on chrome the js and layout engines run different threads. Same with audio and video rendering. Same with downloads. The average page is probably running things on 10-20 random threads. And the average user has multiple tabs open.
Multicore CPU is so common these days that any performance critical code that can be parallelized, is. This includes GZIP and HTTP which are used heavily by web browsers. The network stacks underneath those in the OS are also multi core capable.
Single thread performance is not all that important anymore. It's better to have 2x the number of fast cores than half the number of slightly faster cores in almost every situation these days.
Not even. Video acceleration, browser acceleration, flash acceleration, and much more, are becoming increasingly GPU-oriented.
For the first time in a decade, AMD are seriously competing at the high end. That's great news for everyone except Intel.
Technically this is true, but unlike the desktop market where you can buy stuff in pieces, it really depends on how the laptop manufacturers adopt them. As in how fast they show up in product lines, and in which laptop segments they show up in.
If, for example, Intel gives laptop manufacturers pricing incentives (legally or not) to stick with their chips, it may not cause any problems at all with respect to actual customer purchases.
Cheap, power efficient, eight cores. It may not win everything, but it works for me.
It also has 1/2 cores, which doesn't show up on the single threaded benchmarks but explains the price and TDP differences.
The 7700k is losing in all aspects below w.r.t. the Ryzen 7 1800X, except absolute single-threaded perf. Intel is losing in:
- absolute multi-threaded performance
- relative multi-threaded performance per $
- memory bandwidth
- L3-cache bound workloads (AMD has twice more L3)
- price of motherboards
> - relative multi-threaded performance per $
Highly depends on the particular workload. Given the few benchmarks we have right now, it seems to go 50-50 once you factor in the lower price of the 7700k. At the very least, this one is close.
> - memory bandwidth
Seems about equal from the measurements I've seen (the theoretical numbers are irrelevant).
> - L3-cache bound workloads (AMD has twice more L3)
Cache-size bound workloads to be precise. Cache latency still seems higher, even on L3.
Overall, the 1800X seems to make the entire Broadwell-E series redundant on price/performance (at least at current price), no argument there. Against the 7700k, however, it's a very interesting decision, entirely based on the distribution of your workloads. And if Intel drops the prices on Kaby Lake, the value equation can change very quickly indeed.
http://www.guru3d.com/articles_pages/amd_ryzen_7_1800x_proce...
[1] https://www.pcper.com/reviews/Processors/AMD-Ryzen-7-1800X-R...
What's happening here is that AMD has priced Ryzen so inexpensively that it actually broke the usual market segmentation by driving high-thread-count workstation CPUs (usually $1k+) into the same price range as high-end gaming CPUs (~$300-400). Think of it as being able to buy an industrial dump truck for the same price as a pickup truck. The dump truck is clearly a better value in terms of hauling capacity, but most people would be far better off with a pickup truck for their daily traveling.
AMD's pickup truck (the 1400X) doesn't launch until later this year. The 7700K will probably still beat it on single-threaded workloads due to the difference in clock speed, but it should get you most of the way there for substantially less money, around $200 if the leaks are accurate.