AMD looks at sinking sales, gulps: It's worse than we thought Struggling chipmaker AMD warned investors on Monday that its second-quarter results, which are due to be announced on July 16, will be lower than previously expected. During a conference call with financial analysts at the end of the company's previous quarter in April, AMD chief financial officer Devinder Kumar said he …
Unfortunately 'good enough' may not be good enough here. If I'm looking at a new pc, core i5/i7 jumps to mind. Rightly or wrongly. I suspect it's the same with a lot of pc buyers. I want amd to succeed. I think they need to spend some $s establishing a 'brand' line of chips in the minds of people.
'Good enough' doesn't cut it anymore. In custom builders market they focussed on, only the best is important. 'Good enough' is fine for lowcost OEM, but this doesnt earn money without large numbers. And if they want to play in this business, they have to provide unique business relevant features on all final products, such as those Intel has with AMT/vPro (while there are not many companies actively using it, those functions make Intel 'better for business'). OPMA is just a fantasy no one offers in volume and they don't have anything else in this segment Intel can't trump.
They focussed on declining markets such as traditional home DIY PC, so now they have even less money for development. And it shows. Before, at least in GPUs they had fiercely competitive cards in top end, which is a must from marketing perspective. Today, they are lagging. They are nowhere to be seen in ultramobile, server, compute acceleration. In their major PC related markets they compete almost exclusively on price. This won't end well for them. They need to find their long lost competitiveness and find new markets or they'll end like many others - either part of bigger company, or pure SoC player in OEM business. Or even worse.
This is hardly surprising as Intel just gave the bullet to its execs. as both Intel and AMD rely on Microsoft software to generate sales and as Microsoft missed the boat in 2008 with the advent of tablets, Relying on Microsoft to generate sale seem a rather misguided business plan.
Not relying on Microsoft == certain death for a PC component manufacturer. Anyone who tried this switched back or did not live long.
Microsoft system reqs have been dropping since Vista. So it is not particularly surprising that the PC industry is suffering. The "upgrade every 2-3 years" demand cycle created by the ever growing resource hog that was Windows 95-98-Me-2000-Xp-Vista train is just not there any more.
Even games find it hard to continue pushing the hardware to new horizons. There is very little new hardware can add in terms of visual reality beyond what has been achieved already. We are clearly in diminishing returns territory here. While AI can be (it is not today) an even bigger CPU hog making the user shell out extra dosh for a more bastardly computer opponent is a questionable proposition. The user will prefer to shoot at other users with a similar level online instead.
Even the developing world is no longer a market driver. Whoever wanted to have a PC there got one. The rest may never purchase one and go straight to "consumption" mode on a tablet or phone.
So any PC component manufacturer which pretends that there will be growth in the future is not sharing whatever their marketing team has been smoking. PC is a legacy business now and should aim for what it can deliver - steady slowly declining or "business as usual" revenue stream with a steady profit margin. From there on the infestors can take it or leave it - you cannot get growth where there are no conditions for growth to be found.
I think you make a good point, how many people are left who a) want a full Wintel machine, and b) don't already have one which meets their performance requirements? It's difficult to find a truly crap PC these days, even a £50 Intel-powered Windows tablet is arguably "good enough" for your average YouTube-and-Facebooker and even for some beyond that, but so is an ARM Android tablet or phone if you don't need legacy compatibility. Even in the case of gaming there are plenty of people content with consoles, machines which by design are out of date before they even hit the shelves and lag further and further behind gaming PCs by the second. There's clearly more to buying decisions than "I want all all the MHz!" and some companies (Apple, for example) don't even bother to list half the specs because they don't expect their users to care how many googleflops or petathingies are inside the box.
There will always be those of us who want expandability, raw power and "because I can" functionality, but I'd have to grudgingly admit that there's an awful lot more people who just want "I don't care, just give me something that works". I'd rather not admit that, because the idea of "post-PC"/"mobile-first"/"nobody gives a damn if it's all glued together" makes me want to do bad things to the people who vomit forth those words without consideration for the power users and the people who's jobs it is to create the media and apps consumed by owners of content-slurping devices. Trouble is, there's some truth to it, and most people don't need a full ATX tower with upgradeable everything and top whack parts.
Also forgot to add that regarding Microsoft, for general purpose everyday computing who needs x86 if not for Windows? Sure, that badass i7 rig with the tri-SLI GPUs and 2TB worth of SSDs is great and all, whatever OS you choose to put on it, but for everyone who doesn't need that class of hardware there's really nothing wrong with a half-decent ARM SoC as long as you have a suitably usable OS and app selection. So with that in mind I think if the trend is towards embedded/low end "just let me do stuff" type devices then x86 has a very, very grim outlook without Microsoft. AMD's recent work with ARM stuff seems like a very smart addition to its stable of tech and I hope it serves them well if x86 does eventually start circling the plughole, I'd like to think it could provide them something of a life raft to jump on whilst Intel continues to try and shrink x86 to compete.
I was hoping for more from WinRT, because the idea of a low power ARM laptop which retained source compatibility with Windows was an attractive one, but alas, t'was not to be, not in terms of desktop apps anyway. There's universal apps and whatnot, but Android already fills my need for big fullscreen poke-friendly apps and there's more choice, so if Microsoft want to continue to offer something non-x86y then they'll have to step it up.
For years, Intel has left just enough room in the market for AMD to survive, so they wouldn't be seen as a monopoly by the FTC. If they really wanted to kill them, they could have done so anytime in the past 20 years.
Now they don't have to leave enough room on the low end for AMD to stick around, because they can argue that the "relevant market" for monopoly purposes includes not only PC/server CPUs but also mobile SoCs. Since they have almost no share of the mobile market, they can easily argue they are far from a monopoly in the relevant market when AMD is dead and buried (with the carcass perhaps bought by Apple for the GPU expertise)
When I built my own desktop I picked an AMD chip as it gave the best CPU/GPU performance for my budget. When I went to buy laptop last year the only options with AMD chips were either under powered or more expensive than identical laptops which had Intel inside. Manufacturers or retailers are pushing Intel or they just don't know what to do with AMD.
AMD are perceived (rightly or wrongly) as having a power management issue compared to Intel. Quite simply, the battery life on an AMD powered notebook is less than that of an Intel powered notebook unless the Intel-powered notebook has a discrete GPU.
That's quite a big deal.
If they crack that, there's still the perception gap to bridge. I hope they manage it.
I also sort of secretly hope those rumours about Microsoft buying AMD are true because it would piss off Intel beyond belief and a) they're fucking asking for it and b) my ex-wife works for Intel so I totally wish them ill, irrational though that may be.
When I went to buy laptop last year the only options with AMD chips were either under powered or more expensive than identical laptops which had Intel inside
That is one of the main problems of AMD. They never really could gain a foothold in the Laptop market. And unfortunately this is where the margins are. This and of course the server chips.
For some time they had a competitive server product with the Opteron, partly thanks to Intel's Itanium stumble, but since Intel got serious on X86-64 server iron, AMD is having a hard time, indeed.
Their APUs are interesting, but I am still not sure, what market they are aiming for. People who don't care about graphics acceleration at all are probably happy with chipset integrated graphics and would rather have more computing power, and who does care about graphics power will likely buy a discrete graphics card, anyway.
I have always bought AMD, although Intel are slightly better (i'm a gamer) they have never justified the extra cost.
Really want to go out and buy a new AMD chip to support them, but I have to admit, My finance forecast is slightly lower than theirs !!!
also waiting for this a little bit :)
http://techreport.com/review/28228/amd-zen-chips-headed-to-desktops-servers-in-2016
*Thumbs up to wish them Luck !!!
AMD still offer some pretty good chips if you're on a tight budget, you can pick up a half-decent AMD chip with 8 cores for less than a low end quad core i5. Sure, it runs a little hotter and single-threaded performance is a tad weaker, but you get twice as many cores and it's cheaper, so it's not all bad. That says to me that unless you're bothered about power consumption or having the best rig money can buy, AMD's still got something for you.
I may be a bit biased, I currently have a Phenom II X4 965 in an AM3+ motherboard and I'm looking to upgrade to the 8320E, one of the aforementioned octocores (because yay more cores for less power, the 125W 965 is a bit on the warm side). I considered switching back to Intel, but I can't justify the cost, which is what caused me to choose AMD when I originally built this machine.
Having said all that, when it comes to mobile I haven't got a clue what AMD has to offer in the low end space, my mind automatically jumps to Atom/Celeron/Pentium. I keep hearing about AMD APUs and whatnot, but where's all the tablets? I've yet to stumble across a rock bottom AMD tablet anywhere, nor have I seen an AMD "compute stick" or smartphone*. Intel seems to have those all sewn up pretty tightly.
In the end it comes down to the right tool for the job, but I do hope that AMD can keep pace and shed some of that "eh, it's nice that you suck less vs Intel than you maybe did before, but you're still second rate" image. I like choice. Choice is good.
* Ok, so Intel don't exactly have the smartphone market, but they're at least in a few phones, I've never heard of an AMD-powered phone.
Probably what you would term lapsed PC enthusiast here and I have absolutely no idea what AMD's line-up or product hierarchy is nowadays.
Intel? Sure, I get the Celeron/Pentium/i3/i5/i7 progression and even the nuances of the HT/VT/turbo/cores/cache differentiation, but AMD?
Nope.
Where's the consistent branding? Where's the killer price points that they can tackle Intel on, no quarter given?
Chucking a few extra cores in and/or some headline clock speeds doesn't compete when you're lumbered with a confusing nomenclature and up against the might of Intel's marketing. Lingering doubts over power consumption and per-core performance don't help.
@Lunatik re "Where's the consistent branding?" - Intel's branding is not consistent, Re-using old brands like Celeron / Pentium for chips using recent cores is confusing. You have to resort to checking the exact model number to find out what you're really buying. If they scraped the old brands and renamed the Atom to the i1, then they'd have consistent branding.
If only AMD had called time on FX and started on Zen sooner then it could have been timed to release with Windows 10, as the Athlon XP coincided with Windows XP. To keep things ticking over in the short term, I would have thought there was space in the market for an AMD equivalent to the Intel NUC / Mac mini. A competitively priced device of that size with "good enough" gaming performance, space for a couple of SODIMMS and 2.5" internal bays could help revive their fortunes. The era of full size desktop PCs seems to be drawing to a close, and the remaining power users / pro gamers are more likely to pick Intel (for now).
Well AMD forced Intel to retain full 8/16/32 bit compatibility in their 64 bit chips by preempting them. Intel had wanted to remove the legacy junk. I don't think it is a big deal to change the instruction set. The only real concern is that Oracle would do a version of Java for it. AMD should copy the basic ARM 64 instruction set but provide a simplified version of the ARM SIMD instruction set.
It should include fast sine,cosine,exp like the NVidia CUDA instruction set and maybe think about 128 bit floating point to capture some of the scientific computing market place.
I think you could put together a very nice chip using less transistors per core than even the current ARM 64 CPUs.
The Register 
Biting the hand that feeds IT
Copyright. All rights reserved © 1998–2024
