|
And so the PC ends with a whimper, not a bang. Broadwell will be available in a ‘desktop’ variant as well as a laptop version, but neither will be socketed. There are a lot of good technical reasons to release it only as an embedded and mobile CPU, but not for anyone other than Intel. They want more of the PC ecosystem, and are taking it. Enthusiasts have been written off, and the rest of the ecosystem is being preemptively kneecapped in case they try to step out of line. The desktop is dead, and with it, PCs become irrelevant, mobile or not. Broadwell CPUs will end the long run of socketed, DIY computer building.
|
|
|
|
|
In other news, AMD got relevant again?
|
|
|
|
|
|
Yes.. but they could probably change that, if they wanted
|
|
|
|
|
Laying off thousands of engineers at AMD is not going to help the situation. Neither is their stock price decline or them considering using ARM.
John
|
|
|
|
|
Even if they wanted to, AMD's much slower cadence of major architecture shifts means it would be a long time out. BullDudzer was AMD's first major (high performance) architecture change since the first A64 operons in 2003. They spent most of the 8 years between doing minor tweaks that, after Intel started smoking them with Core2, were always massively overhyped in advance as being able to close the gap before release hardware showed it actually widened. BullDudzer's architecture (and the new ARM license) was throwing in the towel and admitting that they're not going to be able to beat Intel at it's own game. Quite frankly I'm not convinced that AMD will still be in the x86 business (or in business at all for that matter) in 2019.
Did you ever see history portrayed as an old man with a wise brow and pulseless heart, waging all things in the balance of reason?
Is not rather the genius of history like an eternal, imploring maiden, full of fire, with a burning heart and flaming soul, humanly warm and humanly beautiful?
--Zachris Topelius
Training a telescope on one’s own belly button will only reveal lint. You like that? You go right on staring at it. I prefer looking at galaxies.
-- Sarah Hoyt
|
|
|
|
|
That's what I thought before I got this news - but this might change things. After all, with Intel gone from that niche, AMD only needs to come up with something "not completely laughable". They haven't succeeded at that lately, but it doesn't look to me as though they were even trying.
|
|
|
|
|
They'd need to do a lot better than just "not completely laughable"; Intel's lead is large enough that I strongly suspect even their mobile chips at stock speeds would beat an OCed AMD desktop chip for gaming.
I also don't see anything in the SA article that indicates that the gaming mobo makers won't be able to produce OC friendly boards with soldered 45xx/46xx/47xx chips. Having to swap the two parts as a whole will hurt repairability; but with the incremental gains on Intel's Tock's having been so minor and each Tick needing a new socket anyway, swapping CPUs as an upgrade path is much less important than it was in the LGA775 era anyway.
SA is IMO massively overstating the impact from the switch from LGA1366 leading with the enthusiast product and the alleged enthusiast product LGA2011 lagging badly behind the mainstream part. Except for the 3 GPU user and LN2 benchmarking crowds LGA1155 has met most enthusiast needs quite well. The 3rd memory channel almost never mattered; and the extra PCIe lanes almost never matters for 2 GPU gaming use while the southbridge has enough lanes for almost all in addition to a high performance GPU needs.
Did you ever see history portrayed as an old man with a wise brow and pulseless heart, waging all things in the balance of reason?
Is not rather the genius of history like an eternal, imploring maiden, full of fire, with a burning heart and flaming soul, humanly warm and humanly beautiful?
--Zachris Topelius
Training a telescope on one’s own belly button will only reveal lint. You like that? You go right on staring at it. I prefer looking at galaxies.
-- Sarah Hoyt
|
|
|
|
|
Thinking about this a bit more, I'm coming to the conclusion that SA has it backwards and it might actually end up being good for mobo-vendors, if not necessarily for consumers. A common thread I've seen in Anandtech motherboard reviews over the last year or two is that there's increasingly little reason to buy a top of the line motherboard instead of a midrange one (or a brandname budget board if you historically have shopped in the middle tier) because the increased integration of subsystems and tighter QA have meant that even budget boards are highly reliable and that the additional features on high end boards are increasingly less likely to be useful to enthusiasts and instead only serve to pad profit margins.
It's unlikely we'll see OEMs bloat their SKUs from ~10 to ~100 to include each current possible mobo+CPU pair. Instead what will probably happen is that the number of base mobo designs will drop since they no longer need to come up with minimally different boards to populate $20 increment between $100 and $300 to meet market segmentation/pricepoint matching goals. Instead there will probably be only a few base boards at low/medium/high prices matched with Intel's low/medium/high price CPUs (and only limited overlap between the CPU buckets). The lower number of base board designs from each vendor will probably boost differentiation between them; which currently is little more than plastic color and heatsink shape. Assuming they don't all make the same set of tradeoff's we'll actually have real competition in board vendors again. The only groups of consumers almost certain to lose are those who currently pair i7 CPUs with very low end mobos or who put Celerons in kitchen sink mobos.
Did you ever see history portrayed as an old man with a wise brow and pulseless heart, waging all things in the balance of reason?
Is not rather the genius of history like an eternal, imploring maiden, full of fire, with a burning heart and flaming soul, humanly warm and humanly beautiful?
--Zachris Topelius
Training a telescope on one’s own belly button will only reveal lint. You like that? You go right on staring at it. I prefer looking at galaxies.
-- Sarah Hoyt
|
|
|
|
|