Read more.Major Scandinavian retailers list and price these unannounced 'F' suffix processors.
Read more.Major Scandinavian retailers list and price these unannounced 'F' suffix processors.
Intel iGPUs are power gated if disabled anyway, so this shouldn't make any difference to overclocking etc. If true it's just a way of selling off silicon with broken iGPU, but the prices I've seen are hardly worth it.
People have been making the argument that integrated GPU either increases cost or holds back performance for ages but they're based on wrong assumptions. As the article says, dark silicon helps with thermals if anything, and economies of scale mean there's likely very little potential for saving by removing it, in fact a specialised, low-volume iGPU free die would likely be more expensive overall, hence it makes no sense to produce one.
Question: the total 8 cores have less transistor count than the iGPU combined? if yes why not simply use those resources elsewhere? Question 2: is it true the iGPU being squeezed with the cores is the main reason for the many socket changes that haunts intel every generation compared to AMD?
I guess we'll need to see how the final retail (Rather than RRP) prices shake out on these - if its 10-20% cheaper that can be a significant saving on a £550 CPU.
What else would you use the resources for? Adding cores means higher power dissipation, more complex binning, extended ring bus, etc. And as I said earlier if it's a lower volume part, you won't necessarily get the cost savings people seem to assume by stripping out the IGP.
The socket changes that haunt Intel every generation are pretty much just because they want to, there's little in the way of technical reasons they need to happen as often as they do. Don't forget Intel sell those chipsets too! Higher power dissipation of the latest 6 and 8 core parts have more merit than the usual reasons, but AMD are retaining backwards compatibility with AM4 despite higher core counts and a PCIe revision (though it will likely fall back to PCIe 3.0 on existing boards).
Maybe it's me but at those sort of minimal price differences I'd just buy the one with a gpu purely as a 'fall back' if a gpu fails.
I also think it's handy for that reason, but only if the price difference is small. It also means the CPU could be moved into a second system without a discrete GPU when you upgrade to something else.
yeah, sockets.
I would have upgraded 2 or 3 times were it not for the mobo tax.
Still rocking the ivy bridge
I do wonder how the balance looks for Intel i.e. potential lost sales (e.g. due to inconvenience, competition or cost) vs profit from selling more chipsets when people do choose to upgrade? Anecdotal I know, but I would have considered a 6/8 core Skylake were it not for the needless socket change. As it stands now I'd rather wait and see what AMD bring first, if I'm going to be changing sockets anyway! Likewise on my second system, I would have considered a quad i3 to upgrade the 6100 currently in there, but forcing a socket change means a Ryzen APU looks like a more obvious choice for what it's used for. Chances are I'll probably hold off on that one too, but the 8100 would have been very temping were it a supported upgrade path. I don't imagine I'm the only one with that option.
I would have thought the risk of losing profits due to upgrades would only apply to the enthusiast market anyway - business and retail systems ship with a new motherboard regardless so they're getting a sale there either way?
There are currently 1 users browsing this thread. (0 members and 1 guests)