Link here:
http://www.techpowerup.com/192627/intel-delays-14-nm-broadwell-chips-to-q1-2014.html
You thoughts in the comments below:
Link here:
http://www.techpowerup.com/192627/intel-delays-14-nm-broadwell-chips-to-q1-2014.html
You thoughts in the comments below:
My thoughts is that Q1 2014 could mean late march... It also means we all miss out on new gen christmas sales which is very disappointing. I don't know about intel now, it would have to be pretty good performance and be reasonably priced to tempt me. There are alot of reasons not to buy intel now such as cost, heat, lack of virtualization or overclocking support depending on if you get the K part or not.... Oh and there is that thing where they do a socket every year...
Correct me if I am wrong in this line, but I think that given lack of competition (in terms of performance) that Intel will just do a die shrink on the consumer parts, produce more chips per wafer but maintain transistor counts due to higher densities and charge the same amount? Maybe a slight increase in transistor count? There is no news on even 20nm AMD CPUs even though TSMC is going online soon with 20nm and 16nm after a while?
Overall I feel very negative, the dark days of processors is upon us! Hopefully 20/16nm comes online soon...
Not sure about that. I've been skimming through a lot of PC news these days, and avoiding the semi-conductor fab news altogether pretty much.
I think Intel wants to make CPUs more energy efficient, so they can go mobile. The increases in CPU performance per generation are almost meaningless. Only motherboard and chipset features seem to improve, and by chipset I just mean PCI-Express, USB and SATA internal connectors. Things like virtualization and other things are going away, so that way Intel can put them in a more expensive segment to extort customers out of their money for features they could just leave on if they gave a d*** about giving customers any bang for their buck. But Intel cares more for it's shareholders than their customers.
I hope AMD comes back with something amazing in terms of CPUs, and makes a good chipset with great connectivity and PCI-Express standards. If AMD can die shrink, we might see some really awesome stuff!
Remember, Intel is now on their second generation of 22nm CPUs with Haswell. But AMD is on their second generation of 32nm CPUs. Considering Intel's second generation 32nm CPU was the i7 2700K as their top-tier part, and AMD's top-tier CPU (that was mass-marketed) was the FX 8350, I'd say AMD can compete if it's fab gets better and if it's chipset actually improves a lot. But AMD needs to pick up the pace, and semi-conductor fabs need to offer compelling, competitive solutions to fight Intel's own fabs.
If GPUs were on 22nm or 20nm by now, we'd be doing really well in gaming! And think of it this way, considering that Ivy Bridge was the first release of 22nm parts, imaine if 22nm/20nm GPUs where available when Ivy Bridge was released. Something to think about. ;)
Probably will be delayed even further. These will be BGA, so will pretty much mean a death sentence for many OEMs because all they can do is deliver a reference board. Intel is not in a hurry, the market is not in a hurry, the OEMs are not in a hurry, and there are plenty of stocks to be cleared before orders of Broadwell will pick up, and since Intel is not going to be making their own mobos anymore, they will not be controlling the market entry.
Sounds like Intel is just milking the market until they don't profit and then they'll launch broadwell to bring back the profit. Sounds like broadwell won't be much of an improvement on performance. In all fairness performance is not what matters, it's about making software more efficient. The hardware is already there, it's the software that needs to do a couple of generation upgrades. Multi-threading is finally becoming something usefull in gaming and with Battlefield the 8350 is finally starting to get harnessed. Mantle is getting implemented which will make the AMD's 7970 last another generation. The evidence is clear. Take a look at the i7 4930k vs i5-4670k...the difference is not that drastic to justify the price difference but it tells that the performance really lies more on software than hardware. I know a lot of people will disagree but this is proven everywhere. Power isn't everything, efficiency in power is more important. So i'd say next gen CPU's won't be much of an improvement unlike APU's which are implementing different types of systems like HuMA and HSA which are making bigger difference over raw power.
I thought Broadwell was always going to be released in 2014? I'm not so sure this is a delay. And I would rather see it pushed back further, since I built a Haswell system a matter of months ago.
Upon further reading, it is a delay in production? Which means we will not see Broadwell until Q3 possibly Q4. Much more acceptable release schedule.
They should focus on developing current hardware, I agree. It would stop wasting resources on new platforms with minor improvements.
You know, HSA and HuMA are going to be huge.
And I agree. In gaming and in most applications, the i7 4930K versus the i5 4670K seem very close indeed. That's because multi-threading doesn't solve the underlying issue that one core can only do so much work at a time. And multi-threaded software needs to exist to solve this.
This is why CPUs don't run 1080p gaming without a GPU or even integrated graphics; because it'd freeze the CPU in a split second. Likewise, I think we do need better software, which means we need better programmers and we need management that doesn't say "get it done fast" instead of what they should be saying, which is "get it done right".
I hope the changes AMD is implementing might get passed on to a new generation of CPUs, APUs and GPUs. And I hope that 22nm and 20nm from TSMC might be able to help push AMD to make a comeback. I know AMD can compete with Intel, in silicon versus silicon... but could AMD use better software, and better design to overcome the difference that low-power CPUs have and also to overcome the semiconductor manufacturing node that Intel has, versus what AMD has through GlobalFoundries and TSMC? Time will tell...
There needs to be more competition. AMD is not challenging Intel enough to push Intel deeper into R&D. Intel still has the upper hand with significantly higher IPC. AMD can add all the cores they want, it still won't be enough. IMO, AMD needs to work on their architecture and improve their IPC capabilities. If they can do that with 8 true cores and keep it affordable, then things might get interesting.
In the mean time, I think each manufacturer is comfortably enjoying success in their respective areas of the market. But until there is some serious competition, Intel will only continue with minor incremental improvements.
I sincerely hope AMD can really give Intel a stronger run for their money with the next gen non-APU CPU's. Would be cool to see the CPU battle heat up like the GPU battle that's raging on as we speak. :)
oh no, delaying a die shrink, whatever will we do
Given that die shrinks don't really improve performance (at least with Intel), I anticipate this launch with whole lot of "meh".