Haswell may be the last upgradeable mainstream CPU from Intel

[Sparky]

I'm thinking about the OEMs like HP, Dell, etc who have a few different CPU options for each model computer. This would greatly increase everyone's cost with little benefit to anyone (including Intel).

[Tenknics]

Charlie's reporting on it too and his story matches up with what hornett said...just in different words..


Now that I think about it Hornett's interpretation sounds more realistic. Intel necessarily isn't killing off the enthusiast per say, they're just changing up their release cadence and probably changing tick-tock. Broadwell will just be a mobile only cpu, just like the original banias/centrino was. ..And we will continue to use Haswell-DT for a very long time..Instead of the usual 1 year..more like 2+..Do we really need a new desktop cpu every 12 months in this day and age?..We've already seen the gaps in release dates between the high-end sockets, 1366 and 2011, get longer. Now they're doing it with the desktop.

Also with that being said, I'm sure Ivy-E being so late fits in with this strat. We will probably have Ivy-E at the high-end, haswell-dt for mainstream desktop. And Broadwell for mobile+.

We've seen intel do this before, many times (merom, conroe, woodcrest) (penryn, wolfdale, yorfield) (bloomfield, lynnfield, clarksfield)... It was only recently we had SB/IB be the same up and down just with different blocks added/removed..The only difference now is the node and uarch's will be mixed at matched at each segment, rather than all being based on the same core.

If you look at that picture, (Ivy-E, Haswell, Broadwell) it doesn't seem that bad, its just we're getting less frequent updates to the non-mobile sector..

half full

[Warboy]

Charlie's reporting on it too and his story matches up with what hornett said...just in different words..


Now that I think about it Hornett's interpretation sounds more realistic. Intel necessarily isn't killing off the enthusiast per say, they're just changing up their release cadence and probably changing tick-tock. Broadwell will just be a mobile only cpu, just like the original banias/centrino was. ..And we will continue to use Haswell-DT for a very long time..Instead of the usual 1 year..more like 2+..Do we really need a new desktop cpu every 12 months in this day and age?..We've already seen the gaps in release dates between the high-end sockets, 1366 and 2011, get longer. Now they're doing it with the desktop.

Also with that being said, I'm sure Ivy-E being so late fits in with this strat. We will probably have Ivy-E at the high-end, haswell-dt for mainstream desktop. And Broadwell for mobile+.


If you look at that picture, it doesn't seem that bad, its just we're getting less frequent updates..

half full

Yea, I saw charlie's post on it also...I was gonna make a thread with [FUD] in the title but then I saw this one

[BababooeyHTJ]

They never really marketed a westmere cpu for the desktop either. I'm not too surprised.

[Movieman]

They never really marketed a westmere cpu for the desktop either. I'm not too surprised.

Yea, I think Intel always figured that quads were more than enough for desktops and in 99% of the users they are correct.
I mean who really needs a 16 core/32 thread machine anyway...

[zalbard]

I mean who really needs a 16 core/32 thread machine anyway...

Is that an offer? ^_^

[Movieman]

Is that an offer? ^_^

As my Father used to say :" Your too willing!" LOLOLOLOL

[Warboy]

As my Father used to say :" Your too willing!" LOLOLOLOL

And he said

"Why System On a Chip? when you can have Supercomputer On a Chip?"

That's my opinion on this news.

[Iconyu]

Charlie's reporting on it too and his story matches up with what hornett said...just in different words..


Now that I think about it Hornett's interpretation sounds more realistic. Intel necessarily isn't killing off the enthusiast per say, they're just changing up their release cadence and probably changing tick-tock. Broadwell will just be a mobile only cpu, just like the original banias/centrino was. ..And we will continue to use Haswell-DT for a very long time..Instead of the usual 1 year..more like 2+..Do we really need a new desktop cpu every 12 months in this day and age?..We've already seen the gaps in release dates between the high-end sockets, 1366 and 2011, get longer. Now they're doing it with the desktop.

Also with that being said, I'm sure Ivy-E being so late fits in with this strat. We will probably have Ivy-E at the high-end, haswell-dt for mainstream desktop. And Broadwell for mobile+.

We've seen intel do this before, many times (merom, conroe, woodcrest) (penryn, wolfdale, yorfield) (bloomfield, lynnfield, clarksfield)... It was only recently we had SB/IB be the same up and down just with different blocks added/removed..The only difference now is the node and uarch's will be mixed at matched at each segment, rather than all being based on the same core.

If you look at that picture, (Ivy-E, Haswell, Broadwell) it doesn't seem that bad, its just we're getting less frequent updates to the non-mobile sector..

half full

Yeah I think Intel have noticed a backlash against their release strategy. How often on these very board have people questioned how close together these micro architectures coming out. One comes out, and you're waiting for the one afterwards, it should always be a question of 'do I need this upgrade over what I have', not 'I'm six months late for x chip, y chip will be out in another five months'.

[To(V)bo Co(V)bo]

This shouldnt be a big surprise. Especially with the widely announced lack of future competition from AMD. AMD already has shifted gears, giving up to this modern day cpu space race. In reality though, as processes nodes shrink the return is getting a lot harder to quantify. From a business perspective, you only have a limited number of process shrinks left to iterate. The cost for every new process is a immensely huge amount of money. It's a lot of wasted business opportunity to not maximize your profits. AMD has done a much better job of doing just that. AMD has done some remarkable things considering their disadvantage to Intel's process advantage. Sadly, the mobile devices have consumed entire markets. I am a developer, and will always be in need of a workstation environment. For average people this just isn't the case anymore. Phones are ever consuming the laptop, tablet, PC markets. What ever happened to netbooks, that market died before it even happened.

[Aberration]

This shouldnt be a big surprise. Especially with the widely announced lack of future competition from AMD. AMD already has shifted gears, giving up to this modern day cpu space race. In reality though, as processes nodes shrink the return is getting a lot harder to quantify. From a business perspective, you only have a limited number of process shrinks left to iterate. The cost for every new process is a immensely huge amount of money. It's a lot of wasted business opportunity to not maximize your profits. AMD has done a much better job of doing just that. AMD has done some remarkable things considering their disadvantage to Intel's process advantage. Sadly, the mobile devices have consumed entire markets. I am a developer, and will always be in need of a workstation environment. For average people this just isn't the case anymore. Phones are ever consuming the laptop, tablet, PC markets. What ever happened to netbooks, that market died before it even happened.

Tablets are what netbooks should have been.

[Skrips]

Yea, I think Intel always figured that quads were more than enough for desktops and in 99% of the users they are correct.
I mean who really needs a 16 core/32 thread machine anyway...

Ummm, i do? Rendering a double hd res image on anything less than 8 cores is misable. And i only have 4 cores.