Not that I think this is the case but remember how Fury was awesome right up to a certain wattage? I wonder...
RV770 was never a massively performant product, man.Remember the last time AMD had a massively performant product
You a word, and frankly, the longer the wait goes, the less people in the market for it remain without any information.I dont know why people AMD is going to show there hand 1 and a bit months before launch!
We know there are 3.6//4.0ghz 95 tdp ES's.
One mentioning of a sample is not enough to get a clear picture here. This is a matter of what comes out of the fab. There will be a lot of dies, and likely also many which reach these speeds at the targeted voltage (for example: 10%). But if only a fraction of them would stay below the targeted power consumption levels at the same time due to leakage (let's say half of the 10%), then AMD might still offer this model. 5% of all dies produced in volume (for 4C, 8C, Naples, and Snowy Owl) might still end up being a lot.While I hope that the 3.6/4.0 is at 95TDP, technically we don't know for sure. The New Horizon demo would point to at least 3.4 base being achievable.
RV770 was never a massively performant product, man.
Just to add for clarification of my short statement: For SKL-X I thought in the lines of IPC improvements over BDW-E, especially as the BDW core is more power optimized which becomes more important with each additional core. Zen+ has been speculated to have 15% higher IPC. This is not unlikely, but remains to be seen. Somewhat different tradeoffs could be made (e.g. targeting HPC, while Zen is there to cover the lower power end). So my statement hinted at a possibly useless wait for an intermediate solution. Of course, platform features do matter and might be more important to many, than the CPU performance alone.It's more dependent on whether or not the product is possibly disruptive. Zen and Coffee Lake are possibly disruptive products in terms of what they offer and pricing. In both cases you could be looking at 50% or more cores at the current mainstream pricing levels with roughly comparable or even improved ST performance, so potentially 50% or more MT performance for the same price which some same is becoming more relevant for certain workloads. It's unlikely Zen+ and SKL-X will be disruptive, they'll be the same iterative performance improvements (10% faster is really unnoticeable for most people in practice) at the same price levels (or whatever new price levels the the market has already adjusted to from previous disruptive releases).
So in Zen's case it will come down to pricing given what we know. If they price a 6 core build (entire platform) at minimum against the 7700k (entire platform) that would be exciting. 8 core as an alternative the 7700k and 6 core against the 7600k would be a large game changer. If it's just going to priced and positioned as a slightly better option than HEDT it'll be a rather meh release, as you could've been using a Intel HEDT build for quite some time now, and the argument for sure it isn't a strong jump in point and might as well wait.
So is the situation for Zen. I think the sweet spot prices for it are maybe 10-20% below Intel's for comparable products, as demand for Zen might significantly increase at these levels. Going lower would eventually lower this gradient in demand, but it certainly would reduce AMD's gross margins (and future demand too).
You mean that RV770 had perf crown? Not quite. It sure exceeded expectations by a country mile, though, making the surprise justice... And Zen will not, simply because of where expectations of most presently are.I recommend you read up on your GPU history.
You mean that RV770 had perf crown? Not quite. It sure exceeded expectations by a country mile, though, making the surprise justice... And Zen will not, simply because of where expectations of most presently are.
Just to add for clarification of my short statement: For SKL-X I thought in the lines of IPC improvements over BDW-E, especially as the BDW core is more power optimized which becomes more important with each additional core. Zen+ has been speculated to have 15% higher IPC. This is not unlikely, but remains to be seen. Somewhat different tradeoffs could be made (e.g. targeting HPC, while Zen is there to cover the lower power end). So my statement hinted at a possibly useless wait for an intermediate solution. Of course, platform features do matter and might be more important to many, than the CPU performance alone.
Regarding pricing I think it wouldn't be difficult to create disruptive products. Even Intel could technically and process wise do so by reducing prices of their i-series processors and even chip sets by 30-50%. But that might not be the optimal pricing structure regarding earnings, and it would instantly cause reduced future earnings and/or increased R&D to stipulate enough fresh demand with interesting products.
So is the situation for Zen. I think the sweet spot prices for it are maybe 10-20% below Intel's for comparable products, as demand for Zen might significantly increase at these levels. Going lower would eventually lower this gradient in demand, but it certainly would reduce AMD's gross margins (and future demand too).
1. ZEN IPC is lower than Kabylake
Does anyone here will consider buying a 8C 16T Ryzen at 20% lower price ($880) vs Core i7 6900K ($1100) ???
Something like Bluestacks, I guess you must've tried that a long time back, if at all, on an AMD system?$880 for 8c/16t? Well, I would go buy 6800k instead, and this is coming from a current i5 3570k owner.
Perf/price aside, I guess there could be some compatibility issues when it comes to development stuffs. One of them is the fact that you cannot use x86 android emulator on Windows. I might be wrong about this but these chip's value could be lowered quite a bit if that is the case, in developer's point of view.
Something like Bluestacks, I guess you must've tried that a long time back, if at all, on an AMD system?
Most Android emulators I've used work fine on AMD & only a handful of them don't run under Windows.
Virtualization extension requirements
Before attempting to use acceleration, you should first determine if your CPU supports one of the following virtualization extensions technologies:
Most modern computers do. If you use an older computer and you're not sure, consult the specifications from the manufacturer of your CPU to determine if it supports virtualization extensions. If your CPU doesn't support one of these virtualization technologies, then you can't use VM acceleration.
- Intel Virtualization Technology (VT, VT-x, vmx) extensions
- AMD Virtualization (AMD-V, SVM) extensions (Linux only)
Virtualization extensions are typically enabled through your computer BIOS and are frequently turned off by default. Check the documentation for your motherboard to find out how to enable virtualization extensions.
I think you misunderstood my point.1. ZEN IPC is lower than Kabylake
2. RYZEN clocks could be lower than Kabylake
Unlikely. I would compare it to lower i7 KBL variants, except a single CCX has some clocking headroom (how did BDW-E clocks look compared to the 4C parts...)Does anyone here will consider buying a 4C 8T Ryzen at 20% lower price ($280) vs Core i7 7700K ($350) ???
This 6C part might even not exist.Does anyone here will consider buying a 6C 12T Ryzen at 20% lower price ($350) vs Core i7 6800K ($425) ???
$220 off -> maybe. But then it also depends on if this chip is fully comparable to the 6900K, i.e. has about the same avg. performance, and desired platform I/O capabilities present or even more than that.Does anyone here will consider buying a 8C 16T Ryzen at 20% lower price ($880) vs Core i7 6900K ($1100) ???
I wouldn't, too. But there are also new platform capabilities, and likely also lower priced 8C models with high turbo boost for maybe $350. So if we do comparisons and analyses based on only cherry picked examples (like SB-HSW users would only consider 4C KBL), we might end at the wrong place.It will not even make people considering upgrading. I know i wouldnt, why anyone would even consider upgrading from an older Intel Core i Sandy/Ivy/Haswell for 10-20% performance at $280 ?? Its the same reason people dont upgrade to Intel CPUs.
How about earning money? Think big -> Zen will stay for a while and doesn't need to be sold completely during Q1 2017. Also there are such unimportant things like lengthy fab production ramps and capacity constraints (GloFo already making Polaris, PS4+XBOne SoCs IIRC, and other 14LPP products).As i have said before, today AMD have less than 10% market share in Desktop and almost ZERO is Servers, 10-20% lower price for 10-20% lower performance will not start a revolution
4-5 months? OK, what do you think how long we'll have to wait for Zen+ after SKL-X?Not only that but Intel will launch a new socket with 4-6-8-10 Core CPUs in August 2017, just 4-5 months after RYZEN. If RYZEN prices are that high, I would wait to see the 4C 8T Kabylake-X that will be faster than any RYZEN 4C 8T and have a more feature rich platform.
That is only how i see it, AMD may choose to go for high margins low volume but that will not make them earn nothing.
Indeed, it makes sense for AMD to address only the $600-$900 and the $150-$300 segments. /sThis 6C part might even not exist.
What Zen part do you see offering similar performance to i7 7700(K), even with accent on throughput rather than ST performance?I might be off by a few percent, then let's assume 10-30%. But I don't expect them to throw away similar performance for 50% the price.
http://www.zolkorn.com/news/amd-ryzen-sr5-will-be-8c-8t/One of the reasons the issue became controversial because of that. Once we have come to acknowledge that AMD will have to produce a DIE CPU family ZEN or RYZEN only two forms namely DIE is 8-core and 4-core, so initially we would. I guess in a small RYZEN SR3 will be a four-core CPU in the course will be open SMT or simply said that Hyper-Threading. To make a trade work eight trade or not. No confirmed But something suspicious was the SR5 we are looking at how it will come out in two formats is 6 core 12 thread or 8 cores, 8 threads, and by including the speculations of the media. he came out of it looking like that will come in SR5 6 core 12 thread, but today I had a definite answer increasingly come together.
For a summary of the number of CPU cores, the number of traders in each class then. To come out here, I can not confirm 100% full, but it was the information I received is something that would be in accordance with the table top that's it, that's OK then RYZEN SR5 come. in eight cores, 8 threads, not 6 cores, 12 threads, according to the assessment, but somehow the SR3 will be the four cores and eight threads from where we see it. It is a remake cents to me. Because of that, you can choose to SR5 with 8 cores, it is ideal as a whole comes out, it will become one of SR7 fit by closing the Hyper-Threading to only have to sit off the court hassles. What
How about earning money? Think big -> Zen will stay for a while and doesn't need to be sold completely during Q1 2017. Also there are such unimportant things like lengthy fab production ramps and capacity constraints (GloFo already making Polaris, PS4+XBOne SoCs IIRC, and other 14LPP products).
I might be off by a few percent, then let's assume 10-30%. But I don't expect them to throw away similar performance for 50% the price.
ZoLKoRn wrote. SR5 = 8c/8t !?
http://www.zolkorn.com/news/amd-ryzen-sr5-will-be-8c-8t/
BIOSTAR AM4 Motherboards Specifications
http://www.zolkorn.com/news/biostar-second-gen-racing-motherboards-launch-event/
lists Mini-ITX boards (X370GTN & B350GTN)
ZoLKoRn wrote. SR5 = 8c/8t !?
http://www.zolkorn.com/news/amd-ryzen-sr5-will-be-8c-8t/
BIOSTAR AM4 Motherboards Specifications
http://www.zolkorn.com/news/biostar-second-gen-racing-motherboards-launch-event/
lists Mini-ITX boards (X370GTN & B350GTN)
hmmm no 6C 12T SKUs ??
I am starting to think there may not be a hex-core option as well, at least not initially. It would be easier, and a better use of resources, to make a 4C/8T, 8C/8T, and 8C/16T. the next question is how close can a 4C/8T come to an 8C/8T?