New GloFlo Plans

>>61054146
AMD is BANKRUPT!!!!!!!

>>61054146
For all intents and purposes 40% perf=fmax uplift since no one autistically calculates gate delay.
If nominal target frequency for 14mm LPP was 3.5ghz, then GloFo is talking about a target of 4.9ghz.
Another bullet point they've shown is that they're targeting frequencies up to 5ghz.

Nazi science at work. You can't beat IBM.

>>61054193
>nazi science destroys jewtel
This is antisemitism

>>61054207
Oy vey, shut it down.

Its the Common Platform Alliance showing fruit the likes of which will upset the industry. That in itself is wholly impressive. Bringing EUV light source to mass production tears before anyone thought it was possible is also incredible. The fact that beyond this coming 7nm node, they have a developed GAA process, is mind blowing. That is such a radical change in routing and design methodology someone deserves to get a medal for it.

If you are Intel you have to be shitting bricks and reevaluating your goals at 10 and 7nm. I mean, sure their 10nm is pretty good, requires less circuits, and sips power. But they have hit a wall with where they want their process to be.

Meanwhile IBMs process is killing it for every other silicon chip.

>>61054193
IBM gets billions of dollars from the pentagon to develop high-tech stuff.

So really, this is jew science.

>>61054321
All of DARPA, just like 90% of the Air Force, is super gentile. IBM does have tons and tons of Indian employees that it shuffles in and out constantly, though thats pretty much the state of tech at the moment.
Regardless IBM is an utter behemoth of hard R&D. They've developed two totally new CPU architectures for DARPA projects that have been declassified, they have enterprise level Qbit systems for sale in production levels which is something Dwave can't even do yet. They developed self assembling nano structures to replace photolitho techniques, they have workable DNA storage, they have working ReRAM, magnetic spin storage, all kinds of things that will undoubtedly trickle out into the market over the coming decade. Whatever they're working on right now in those classified labs is probably world changing.

>>61054280
>heir 10nm is pretty good
But it's actually gonna be worse than 14nm at the start. Intel themselves have confirmed it.

>>61054370
>Source: Inside my ass. Take a look if you find it.

>>61054650
Read the news once in a while, kid.
http://www.eetimes.com/
http://www.siliconsemiconductor.net/

Based Nazi science

This is coming half a year earlier than anyone expected it to.
Guess I'm skipping 14nm, 7nm is the real gamechanger for AMD, 14nm is the lube preparation phase

is it possible to have one thread on /g/ that isn't packed full of shitposts?

File: ArmstrongPP.png (487KB, 972x807px) Image search: [Google]

487KB, 972x807px

>>61054370
>They developed self assembling nano structures to replace photolitho techniques

Nanomachines?

>>61056230
Not quite, but they're getting there.

File: b7d.png (66KB, 1000x1000px) Image search: [Google]

66KB, 1000x1000px

>>61054370
>yfw IBM spent over 100 billion dollars on stock buy backs
>100 billion could have been spent on R&D

>>61056264
Maybe they didn't want any uppity members of the board making demands as they continually exit various market segments and strengthen their marriage to the government.

>>61054193
>Nazi science at work. You can't beat IBM.
Oy vey.

What happens when we reach 1nm?

>>61056716
Ultimately linear area scaling does stop at some point, matter isn't getting smaller after all. Though that point may not be on the horizon just yet. There is a field of study that already has the industry getting balls deep and going for the reach around. Quantum junctions.
Don't let the popsci name make you think its some black science man tier grandiose horseshit. Its finding ways to manipulate subatomic particles and measuring their positions as states instead of controlling current in a channel to perform logic.
This could potentially give a venue of conventional scaling far beyond the nanometer scale, but after that we're shooting blacks. Transistor density would have to come from new routing techniques and 3D stacking of logic.

>>61056716
I wouldn't count on reaching 1nm. You get quantum magic shit at those sizes.

File: amadaexciteULTRA.png (210KB, 422x537px) Image search: [Google]

210KB, 422x537px

>>61054146
>>61054155
>>61054164
>>61054171