AMD what will happen in 2015?

[cbn]

Pointless, absolutely pointless. AMD FX would be so irrelevant in terms of efficiency and performance in average joe's mind... That's exactly my point. How many people think that much when they make a purchase: only a few. Not mentioning business sales. Business sales are zero. Hell, it is getting hard even for myself to keep up with this cherry pickings.

I'm not sure I know what your post means exactly......but I think a lot of people do research the cpu before making an enthusiast level desktop purchase even if it just a low end one.

In fact, I wouldn't surprised if this is even more so in other countries where money is tighter.

P.S. According to Tom's hardware monthly "Best Gaming CPU for the money" guide AMD has lost every price point from the lowest to highest, only making honorable mention with the FX-6300 @ $100:

http://www.tomshardware.com/reviews/gaming-cpu-review-overclock,3106.html

But if they can figure out a way to realistically make price cuts then I think they can at least solidly regain a position or maybe even two.

 

[monstercameron]

Intel's iGPU may not be all that flash but a basic locked i5 or i7 with that iGPU up against an APU - who wouldn't take the extra CPU grunt every time? APUs are a tiny niche. Plus Intel constantly improves that iGPU year on year, and its not a 5% CPU IPC bump. Come Skylake low end 720p gameplay with some settings turned up shouldn't be a problem.

the FUD is real, apus take up a majority share of amd revenues in nearly every market, Enterprise, desktop, mobile and embedded.

 

[cbn]

the FUD is real, apus take up a majority share of amd revenues in nearly every market, Enterprise, desktop, mobile and embedded.

Hopefully that changes for desktop (other than Jaguar/Puma APU), so AMD's revenue can actually grow to something that can sustain the company.

 

[monstercameron]

Hopefully that changes for desktop (other than Jaguar/Puma APU), so AMD's revenue can grow to something that can actually sustain the company.

they are at break even right now, the products are good, they just need to get a few more design wins.

 

[cytg111]

Intel's iGPU may not be all that flash but a basic locked i5 or i7 with that iGPU up against an APU - who wouldn't take the extra CPU grunt every time? APUs are a tiny niche. Plus Intel constantly improves that iGPU year on year, and its not a 5% CPU IPC bump. Come Skylake low end 720p gameplay with some settings turned up shouldn't be a problem.

True. Well, for the "who wouldn't"-part. As we are all agreeing that cpus today is plenty fast for office apps, what is left?
But yea, i get what you are saying and I agree to some point.
You say skylake and 720p. - in what formfactor? A tablet-hybrid? Is the haswell trend gonna continue (where even desktop sku's throttlle at stock)? Intel may have a massive process lead but what happens when you halve the die but not halve the watts? My logic dictates that it is gonna get harder to dissipate the heat.. and maybe that is some of the trouble we're seing with core-m in early designs?

 

[mikk]

Is the haswell trend gonna continue (where even desktop sku's throttlle at stock)?

This is nonsense.

 

[frozentundra123456]

There is a lot of that going around in several threads lately.

 

[Enigmoid]

HP Pavilion 13,3" with A8-6410 28nm SoC

HP Pavilion 13,3" with Core i3 4030U 22nm FF Haswell

One of the lowest binned 2+2U intel chips vs. the highest binned beema chip. ST performance is 40% higher, MT performance is around 0-10% higher.

Notebookcheck doesn't contain load data for the intel system but the beema chip is running at 1.4 ghz during their max load.

What Notebookcheck had to say about the issue.

While idle, the power consumption of the Pavilion is below 10 Watt - just like the sister model's. But, they perform differently during the stress test. The Pavilion's consumption of 23.8 Watt is lower than the sister model's (29.7 Watt). The reason is found quickly: The processor of the AMD model is throttled, while the Intel model's isn't.

Broadwell dual core that was just released, has 50% or more of its die dedicated to the iGPU.

A lot of that is video display and media capabilities. The actual execution units are relatively small. You can also see the non-slice area (rectangle between EUs and CPU cores).

 

[2is]

2015 will pretty much showcase whatever mark of leadership brilliance Rory imparted on AMD.

For better or worse, Lisa will have to manage the company with a product portfolio and employee morale that was essentially 100% created by her predecessor.

Given his ungraceful exit from the building, it is difficult to imagine 2015 will be a great year for AMD or Lisa.

That said, with the unquestionably delayed 14nm situation at Intel, plus the equally delayed rollout of anything post-28nm that isn't a mobile phone part, it isn't clear to me that 2015 will be all that exciting for customers in general, be they AMD's or some other business.

We might see some low-clocking or otherwise unremarkable CPUs from Intel, and possibly some expensive 20nm GPUs, and that is about all I'm expecting from 2015 to be completely frank and honest. Not going to defend that unqualified statement though.

I agree, 2015 looks like it's going to be a rather slow year, at least the first 3 quarters anyway. That said, it's probably exactly what AMD needs to attempt to make up some ground while the competition is somewhat stagnant.

 

[escrow4]

the FUD is real, apus take up a majority share of amd revenues in nearly every market, Enterprise, desktop, mobile and embedded.

Only APUs I see here in Australia are dirt cheap and turtle slow bargain basement AMD laptops. Mobile? Where? Enterprise? Where?

 

[DrMrLordX]

Companies are still migrating from 6 to 7. Few are on 8, and even fewer will be moving to 9 immediately. It will take a few years for this stuff to settle in even if Oracle is on time.

I am not certain that the enterprise users of whom you speak will be the ones to first readily adopt JDK 9. Actually, given how poorly AMD has done in expanding their installed base of Kaveri, it should be interesting (at best) to see who finally takes advantage of widely-available HSA support in software.

There should also be some improvements to C++ Amp sometime in 2015 to make it easier to use, which would be important since some users just don't want to mess with the JRE at all.

 

[witeken]

A lot of that is video display and media capabilities. The actual execution units are relatively small. You can also see the non-slice area (rectangle between EUs and CPU cores).

That's not true. If you take the processor graphics from left to right, at about 4/5 you have a vertical line. Everything left from that is GPU, and probably also some portion from the 1/5 part.

 

[Sequences]

I am not certain that the enterprise users of whom you speak will be the ones to first readily adopt JDK 9. Actually, given how poorly AMD has done in expanding their installed base of Kaveri, it should be interesting (at best) to see who finally takes advantage of widely-available HSA support in software.

There should also be some improvements to C++ Amp sometime in 2015 to make it easier to use, which would be important since some users just don't want to mess with the JRE at all.

Those who are already programming for GPU aren't using JAVA and those who don't will struggle to either find a good use case or come up against a high learning curve. Not to mention lack of availability for consumer and enterprise chips. Developing for HSA will take many years to mature and even longer for widespread JRE version adoption.

I really hope they spend 2015 improving their software. It is one of the few things they should have complete control over.

 

[schmuckley]

2015 is going to be a long, boring year for AMD watchers. Any excitement is saved for 2016 when new CPU architectures arrive, and 14nm becomes viable. 2015 is 28nm and placeholder products, with the one interesting part being the arrival of stacked memory.

What..another year and again nothing new from AMD?
seriously?
They've had since 2011

 

[mrmt]

the FUD is real, apus take up a majority share of amd revenues in nearly every market, Enterprise, desktop, mobile and embedded.

Except for embedded, what is happening with AMD market share in every market where they fielded APUs?

 

[DrMrLordX]

Those who are already programming for GPU aren't using JAVA and those who don't will struggle to either find a good use case or come up against a high learning curve.

Well, here's the thing: the whole idea behind Sumatra (which some of the GPGPU guys really hate) is that it'll be just another optimization tool in the JVM's belt. The JVM may already be using any number of SIMD instructions if the host hardware supports them and if the code being executed can be realistically converted to SIMD operations. Who knows what Oracle JDK8 already supports? SSE4? AVX2?

So, in theory, JDK9 will be able to take mathematical operations that both can and should be offloaded to the GPU/iGPU and let er rip without the coder really having to know what's going on. Naturally it will help if the coder is at least vaguely aware that there is GPU compute capability available so they can structure operations in such a way that using the GPU makes sense. The JVM will have to be sensitive to the differences in return times between iGPUs and GPUs, since the PCIe bus is going to impose some latency that you won't see when using an iGPU.

I really hope they spend 2015 improving their software. It is one of the few things they should have complete control over.

No arguments here. AMD has struggled with that for years.

 

[MisterLilBig]

The HSA play seeks to increase the number of things that fall into category 3 but it has so far produced very few results.

I think that we have seen the results and know what to expect in most areas, I haven't seen many examples tho, but, I am not looking for them and that could explain part of it.

Spending generations getting APUs to where they provide tangible benefits to a tangible portion of the market is R&D budget, software budget, development time and actual silicon chip transistor budget that could have been spent on other goals (like entering the Phi/CUDA space more aggressively, or lower power CPUs, or more aggressive ARM adoption, or any other number of things).

AMD has been doing that, better OpenCL support, unmatched dGPU compute performance, the "first" ARM 64 bit server, HSA development and more, with less.

Come Skylake low end 720p gameplay with some settings turned up shouldn't be a problem.

720p...? Trinity could do that. Haswell can do that. Aim higher! Hopefully 1080p gaming with some High settings will finally be possible by then?

Those who are already programming for GPU aren't using JAVA and those who don't will struggle to either find a good use case or come up against a high learning curve. Not to mention lack of availability for consumer and enterprise chips. Developing for HSA will take many years to mature and even longer for widespread JRE version adoption.

You don't target "those who are already programming for the GPU", you target people interested in doing so, without having them to learn new things, some who did so natively might convert to higher level code development because it would be faster to develop a product, a service, a system, a prototype. This is the Intel mindset of Phi, but even higher level, and for the same reasons that Phi gets support, this will get support.

HSA hardware support, like everything else, it takes time, if AMD decides to kill the cat cores and only sell Carrizo this year, with its full HSA 1.0 support, its a start for full AMD support in the next 3-5 years, which I think they should do. (Along with Mantle and TrueAudio.)

JRE support, if you need it, you will get it, if they decide to auto update it even sooner, better. Not much different than any other software.