[updated] AMD lowers revenue expectations because of poor APU sales - Q2 earnings call

[patrickjp93]

http://www.hsafoundation.com/

AMD does have a footprint on the compute market.. Intel is focusing on Unified Memory for their iGPU's now and are willing to put IGPU's in their Xeons with skylake-E, NVidia with NVlink for better compute with CPU like IBM's PowerPC. Intel doesn't Have geometry engines like AMD does which is where 3D performance can suffer at times, where this falls in line with the rest of performance because Kaveri is starving for bandwidth.

Right now it's just starting to get a footprint in computing.

Forgive me but I think he was referring to market presence. AMD has practically no presence in servers and supercomputers. What it had on the CPU side is rotting away. It has struggled and pretty much failed to gain a toehold on the accelerator side. The only machine of note is that top 500 green supercomputer, and that award primarily lends itself to the highly innovative cooling and electrical system for the building, not the raw performance or efficiency of the hardware itself in compute/Watt.

[Deletive]

Forgive me but I think he was referring to market presence. AMD has practically no presence in servers and supercomputers. What it had on the CPU side is rotting away. It has struggled and pretty much failed to gain a toehold on the accelerator side. The only machine of note is that top 500 green supercomputer, and that award primarily lends itself to the highly innovative cooling and electrical system for the building, not the raw performance or efficiency of the hardware itself in compute/Watt.

 

Oh well my bad then but I did mention that in my first point. Their opterons completely suck right now which is true but their Arm servers and their server APU's are on their websites almost no where to be found which I think no one has really even put into a Higher end servers and Supercomputer yet. Their server business is as unorganized as their mobile seems. 

hopefully something is fixed with K12 which is pretty much what we say over anything related to AMD cpu being brought up, =) once Skylake-e with it's iGPU's AMD probably won't have much to work with.

[Fred Castellum]

Meh not worried at all, if they fall too far a company will just buy them up and continue to compete. 

[evilarceus]

They are great for budget builds IMO, but the G3258 became a better option. Pair that up with something like a 750 Ti and you're good to go.

[zMeul]

My Dad and I are living proof you're full of crap. His Q6600 and 2 SLI GT 9800s lasted almost 8 years before he actually needed something new, and yes, the machine was used to game when he wasn't doing office work.

and again, quoting me out of context; because the issue at hand was building a small form factor PC with just the APU alone, no discrete graphics at all where the focus was graphics performance of the said APU

 

let's go back to my OP:

what's the reason for APUs in desktops? people that are looking for a small factor PC will most likely not care about the graphics core processing power; those who will, will most likely buy a discrete graphics card

[Popingheads]

Well there were a few other causes as well aside from just desktop APUs, mainly the issue with 20 nm node not working and lower OEM APU sales (so still APUs).

 

http://seekingalpha.com/news/2614505-amd-issues-q2-warning-shares-halted-updated

 

 

But anyway back to the other point, there are some reasons you might want an APU, a powerful GPU isn't needed just for gaming, it is also important to have for media (ie videos) especially high resolution ones. I'm pretty sure the Carrizo chip has 4k hardware decoding now as well doesn't it? Plus remember the way DX 12 works means you will be able to use the APU and a dedicated GPU to play games, with the work being divided up between them based on how powerful they are. In that case it would likely make more sense in the future for budget builds to use both an APU and a dedicated GPU, which would likely end up offering more performance than the currently very common budget builds that use the G3258 Pentium as the CPU.

 

AMD also intendeds for them to be much more useful in the future for general purpose tasks than they are right now. In other words the whole HSA thing. I imagine that eventually APUs will be the only thing they make and, assuming the support is there, it would work out very well for them and everyone else.

 

Ultimately it seems to me like AMD is playing the long game with APUs, as with a number of their technologies. In fact it does remind me of someones comment I heard on the whole AMD and Intel competition, how it is impossible for AMD to beat Intel outright (considering the huge difference in revenue) but instead they just keep cutting them off to the next new technology and/or trying to push the industry in their favor. With Intel usually being slower to adapt to large shifts.

[Prysin]

wtf do you expect when Intel does this

and completely wrecks AMD at their own game. I know there is a large price difference, but prebuilt PC makers know consumers like Intel Inside.

unfortunately, you could get a 7870k + mobo + ram + cheap cooler for the cost of that. Except, the cheapest retail price i can find for the 5675c is around 300 USD. Which makes an AMD system even cheaper to set up, and all you really get from intel is an absurdly expensive part with little real value for money.

APUs are great, but they need DDR4 (since GDDR5 is based in DDR4). Thus they could probably CF with GDDR5 cards with more ease.

[Dabombinable]

unfortunately, you could get a 7870k + mobo + ram + cheap cooler for the cost of that. Except, the cheapest retail price i can find for the 5675c is around 300 USD. Which makes an AMD system even cheaper to set up, and all you really get from intel is an absurdly expensive part with little real value for money.

APUs are great, but they need DDR4 (since GDDR5 is based in DDR4). Thus they could probably CF with GDDR5 cards with more ease.

2 words: Price/Performance

Intel as they stand make APU's with an iGPU that is superior to the ones in AMD APU, and have a far stronger CPU (lets face it, Intel makes APU, but hasn't bothered with a name to separate them from CPU with no iGPU).

[deviant88]

AMD has chose very wrong going only APU.

People want strong CPU's, and it finally shows in their sales for the non believers.Intel caught up iGPU wise now, who the hell would buy an A10 7850k when you can buy an infinitely better i5.

If they dont comeback with strong Zen AMD is finished, GPU division will have to break off else they will be dragged down too it already shows with FuryX poor quality/performance.

[Robin88]

The problem with AMD for me is that they focused on the low end almost exclusively, and that's great and all assuming that your competitor doesn't try to get into the same market, but AMD forgot that Intel wishes to get into the mobile space, and the only way in is to go through AMD by getting power consumption down and performance up at the low end and destroying AMD's market share.

Another thing that hurt AMD is public perception, if you have the most badass flagship part that is miles ahead of your competition, then that will trickle down into increased sales of your low end parts, if you're seen to have the better equipment at the high end, people will assume that means that everything you sell is better, regardless of facts. But AMD for far too long has focused on the low end and on price per dollar instead of focusing on the areas where it matters, the high end enthusiast and server/workstation markets, if you're king there, you're king everywhere.

Sure the 390 beats out the 970, and the Fury X trades blows with the 980Ti but who cares, nVidia has the Titan X, and sure AMD has the 9590 for less than the price of a mainstream i7, but who cares when Intel has the 5960X? AMD has it's Opteron lineup, but who would buy one when Intel has it's Xeon E5 and E7 range.

AMD needs to keep selling APU's no doubt, they're good value and perform extremely well compared to the competition, and the only way Intel can compete is with a part that costs more than 2x the cost of AMD's flagship APU's, but they desperately need Zen and if it fails to blow away Intel and only trades blows with an i7 instead of outright beating it then AMD are in trouble and nothing will save them except a buyout.

I'm seriously considering an R9 390 though as my next upgrade, I don't need 4K right now and it performs better than a 970 for a not insignificant amount cheaper, and if Zen is better than at least a 5820K then I'm gonna upgrade to that when it's released, but if it only competes with a i7 5775C then Intel will get my money, I need more performance than I currently have now, and I want to support AMD, but only if they offer a tangible increase in performance compared to my current setup.

[Sauron]

APUs are actually a good option for budget builds. They were selling pretty well last year, not sure what happened here.

[UnknownEngineer]

I would buy AMD's stock now, so I could sell it when someone buys AMD, but I don't know how.

[Trik'Stari]

AMD....please just wake up and realize you can't get ahead by being the cheaper option that offers less performance.

 

Stop focusing on three different markets at once. Pick a direction, put everything into it, make something amazing, make money, and then keep doing that for a while. THEN you can have the fun little side projects like APU's and HSA

[HippY]

Wow, I was not aware that LTT is full of financial analysts, market researchers, microprocessor engineers, CEOs and have a globe that shows the future!

 

AMD made a shitty choice with Bulldozer and they are aware

 

They have the 300 series that seems to be a competetive refresh, they have the Fury X that just appeared on the market with HBM, and keeps up with the 980Ti (voltage OC will be sorted, they had a problem with pumps, retail version is sorted, and remember, the card is fresh, the price will drop over time) so they are keeping up well in GPU segment

 

at CPU segment they are doing the best they can: call in a big gun, Jim Keller (he creates mobile CPUs for Apple, and those 2 cores are competing with 4 core alternatives, commonly topping performance benchmarks and low power consumption), creating a new architecture reusing some good parts from Bulldozer, but without any of the features that were made there, and they will sell it in a year. 

 

"All" they have to do is to last until Zen, make sure, Zen is good, then sell good APUs and processors. It has the possibility that if they manage to survive for a year, make Zen based processors competitive that the competition comes back

 

And this "say no to APUs" attitude should stop, as it is shown, you can have a high end CPU with a iGPU like 5775C that has good performance for cheaper than a high end CPU and same cat GPU, or a cheap APU that has similar performance as the high end 5775C yet half the price. If people (programmers) can exploit APUs well, they have the ability to shine

[Albatross]

If only they would just get their shtako together they wouldn't be slowly and painfully dying.

[vm'N]

at CPU segment they are doing the best they can: call in a big gun, Jim Keller (he creates mobile CPUs for Apple, and those 2 cores are competing with 4 core alternatives, commonly topping performance benchmarks and low power consumption), creating a new architecture reusing some good parts from Bulldozer, but without any of the features that were made there, and they will sell it in a year. 

 

And this "say no to APUs" attitude should stop, as it is shown, you can have a high end CPU with a iGPU like 5775C that has good performance for cheaper than a high end CPU and same cat GPU, or a cheap APU that has similar performance as the high end 5775C yet half the price. If people (programmers) can exploit APUs well, they have the ability to shine

People have been well over-praising Jim Keller. After all, he is a single guy, and it is his team in entirety that is going to matter.

He cant bring out a miracles on his own.

APUs shine (currently) where the throughput required is low enough not to use the dGPU and the acceptable latency is high enough to not use integrated SIMD units. When APUs gone bring on more GPU performance, there is a bigger roof for when iGP will have it advantages. Now, it makes little sense for a developer. In the future, it will make more sense.

[Maslofski]

no shit, noone wants APUs, we need them CPUs

[That Norwegian Guy]

Well what do you expect when Carrizo has been hyped forever and isn't even out yet?

 

Once Carrizo hits, I'll buy my first laptop since 2010.

 

"Coming in first half of 2015" they said. FAIL. Bet they're "waiting for Godavari stock to deplete". FAIL.

[patrickjp93]

People have been well over-praising Jim Keller. After all, he is a single guy, and it is his team in entirety that is going to matter.

He cant bring out a miracles on his own.

APUs shine (currently) where the throughput required is low enough not to use the dGPU and the acceptable latency is high enough to not use integrated SIMD units. When APUs gone bring on more GPU performance, there is a bigger roof for when iGP will have it advantages. Now, it makes little sense for a developer. In the future, it will make more sense.

No, Keller makes all the difference. His entire original team has been with AMD since he left, and where have they gotten without him?

[vm'N]

No, Keller makes all the difference. His entire original team has been with AMD since he left, and where have they gotten without him?

Okay, if that is what you think  :lol:

There are simply to many factors to take into play, to even consider than 1 man is the key to success.

Also, they did fire some engineers after the bulldozer failure, and have also brought in some new IIRC.

 

They have done quite well, considering the transformation of the bulldozer-family. From bulldozer to carrizo, that is some fine work, considering what they had to work with.

[othertomperson]

This is like Intel expecting most of its CPU revenue to come from sales of Celerons and Pentiums... it's not going to happen. On the consumer end of things they probably make a decent amount from i3s and i5s, and then I'd imagine most of their sales are i3s and Xeons in the corporate sector. This then also allows Xeons to be re-purposed for the enthusiast market at practically zero additional R&D cost.

 

It's absolute madness to bet your company on the extreme budget PC side of things.

[Dabombinable]

It's absolute madness to bet your company on the extreme budget PC side of things.

So was bringing their CMT architecture out of servers.

[Mulsanne]

This is like Intel expecting most of its CPU revenue to come from sales of Celerons and Pentiums... it's not going to happen. On the consumer end of things they probably make a decent amount from i3s and i5s, and then I'd imagine most of their sales are i3s and Xeons in the corporate sector. This then also allows Xeons to be re-purposed for the enthusiast market at practically zero additional R&D cost.

 

It's absolute madness to bet your company on the extreme budget PC side of things.

 

Intel, along with Nvidia and IBM, have the benefit of being so heavily in use by HPC - something AMD doesn't have on any level. They make a hefty amount from supercomputing and the contracts they land from that...AMD fell behind on that front. 

[Berkut99]

Poor APU sales shouldn't come off as a surprise when you promise full HSA support for your flagship but never moved to deliver on it until a refresh over a year later. But that's just a drop in the bucket compared to what AMD has been doing.

 

I'm a huge supporter of what AMD APU's represent, and I completely see what they were after. The idea of depreciating dedicated FPU performance to focus on ALU performance, relegate that floating-point power to the iGPU, and take advantage of the high bus-speed that HyperTransport offered all during a time when Intel was keeping themselves occupied with clock speed and an off-chip FSB should have completely locked them into the mainstream market by the early 2010's. AMD has been moving towards a full desktop SoC for nearly a decade now--Everything from their aquisition of ATi to their push to be the first to market with HBM shows this.

 

But compared to Intel they Simply. Can't. Deliver. Without the same R&D towards scaling what can they ever hope to expect other than entry-level CPU performance while trying to keep chip yields high on a small die that's stuck at the 28nm node? They're at the mercy of a mostly third-party fabrication company that refuses to agressively ramp up their 14nm process. Their inistance on public hype generation also allows Intel to brute force their way to leap frog AMD. AMD moves towards an SoC with an MCM that features a large, fast shared cache, Intel uses their low production node and insanely quick ability to shift fab production to simply stuff a bunch of tiny stream processors and move fast SRAM off chip to beat their deployment of HBM to APU's. All during this time Intel's able to stuff more, faster cores that stay cool into their chips which appeals to big data and cloud services. This forces AMD to react, abandonding their whole inital roadmap to that point to move back to R&D, meaning in this case the abandonment of CMT and HSA and movment towards SMT, which they have never done and Intel has them beat on by over a decade.

 

And it's frustrating as heck from a solutions standpoint. AMD's APU's could make excellent mainstream and workstation solutions for large companies as well as completely eliminate the low-end video card market. But between their transparent and wild roadmap strategy, reliance on third-party fabs, and the juggernaut that is Intel I just don't know what they can do at this point. The hate towards APU's here is unfortunate, but irrelevant--they have always been the future and are what the vast majority of users need. AMD recognized that immediately and acted, but just can't get their resources aligned well enough to punch through the market.

[patrickjp93]

Okay, if that is what you think :lol:

There are simply to many factors to take into play, to even consider than 1 man is the key to success.

Also, they did fire some engineers after the bulldozer failure, and have also brought in some new IIRC.

They have done quite well, considering the transformation of the bulldozer-family. From bulldozer to carrizo, that is some fine work, considering what they had to work with.

Carrizo (minus the HDL) is what Bulldozer V1 should have been. Keller is also the only electrical engineer who ever outfoxed Paul Otellini who was legendary in his tenure at Intel. Keller is one of three grand authorities of CPU design. Intel and IBM have the other two.