Economy Google about to change the game in process of designing any chip

[PEB]

Google seems to have done something completely in secret unlike anything they have done before and making available to the world. In the past chips where designed by hand, then by EDM software tools now by AI and software tools. But even that not nearly as efficient and cost effective to bring these insanely complex chips to market. The time it takes now to bring out 4nm and under chips to market is still measured in years and are much more prone to have problems during the development. Things like chip layout, wire lengths on the chip, thermal issues, stray electrons crossing over interconnects and tons more. They developed an extremely large and expensive chip that is used to accelerate the processing of the EDM tool designing and testing the chip. They claim like a 2 to 5X improvement in designing and testing before sending it out to the process of making the actual chip. It reduces the chances of first time failures and shortening the lead time. It will accelerate the chip war by a lot even if its still not completely finished technology.

More reason why the US needs to keep up investment in chip manufacturing no matter who is involved as long as its with US and their partners.

"
This week, Google unveiled its AlphaChip reinforcement learning method for designing chip layouts. The AlphaChip AI promises to substantially speed up the design of chip floorplans and make them more optimal in terms of performance, power, and area. The reinforcement learning method, now shared with the public, has been instrumental in designing Google's Tensor Processing Units (TPUs) and has been adopted by other companies, including MediaTek.


Chip design layout, or floorplan, has traditionally been the longest and most labor-intensive phase of chip development. In recent years, Synopsys has developed AI-assisted chip design tools that can accelerate development and optimize a chip's floorplan. However, these tools are pretty costly. Google wants to democratize this AI-assisted chip design approach somewhat.

Nowadays, designing a floorplan for a complex chip — such as a GPU — takes about 24 months if done by humans. Floorplanning of something less complex can take several months, meaning millions of dollars in costs, as design teams are usually quite significant. Google says that AlphaChip accelerates this timeline and can create a chip layout in just a few hours. Moreover, its designs are said to be superior as they optimize power efficiency and performance. Google also demonstrated a graph showing wire length reduction across various versions of TPUs and Trillium compared to human developers."

You know the Chinese will do whatever in their power to get their hands on this technology its that big a deal.


By now, AlphaChip has been used to develop a variety of processors, including Google's TPUs and MediaTek's Dimensity 5G system-on-chips, which are widely used in various smartphones. As a result, AlphaChip is able to generalize across different types of processors. Google says it has been pre-trained on a wide range of chip blocks, which enables AlphaChip to generate increasingly efficient layouts as it practices more designs. While human experts learn, and many learn fast, the pace of learning of a machine is orders of magnitude higher.

Extending usage of AI for chip development​

Google says AlphaChip's success has inspired a wave of new research into using AI for different stages of chip design. This includes extending AI techniques into areas like logic synthesis, macro selection, and timing optimization, which Synopsys and Cadence offer already, albeit for a lot of money. According to Google, researchers are also exploring how AlphaChip's approach could be applied to even further stages of chip development.

Google unveils AlphaChip AI-assisted chip design technology — chip layout as a game for a computer

Google and MediaTek already use it.

www.tomshardware.com

This is what they dedicated to chip designing.

"

 

[avenue94]

Google seems to have done something completely in secret unlike anything they have done before and making available to the world.

Nvidia already announced it was testing ways to optimize floor planning over a year ago.

Having said that, the Mediatek angle is interesting since Google seems to be backing them as a ticket to give chromebooks a much needed shot in the arm.

 

[IDGETKTFO]

Starts off saying "in secret" then finishes with its still not completely finished......

 

[SandisLL]

Starts off saying "in secret" then finishes with its still not completely finished......

There actually can't be secrets.
Ability to reduce semiconductor element size is limited by real physic and there is limit no one will be able to beat because you always will need insulation layer between elements of chip...cos physical reasons.
So ofc there also are abilities to change logic architeture and here x86 type stuff really isn't energy saving wise.
Ofc there are experiments with quantum computing and still it is more like lab type stuff today..

 

[avenue94]

So ofc there also are abilities to change logic architeture and here x86 type stuff really isn't energy saving wise.

Most of the difference between ARM and x86 is down to use case optimization, not the actual ISA at this point. But yes, we're getting close to the limit of what we can do with node shrinks, next advancements will mostly be around packaging and power delivery. Also keep in mind that most consumers don't care much about processors at this point since for the vast majority of the market we're past "good enough."

 

[PEB]

Starts off saying "in secret" then finishes with its still not completely finished......

By not completely finished they still are constantly refining the design. People bring up something they have not thought of an thanks to the design they can update the design without the usual heavy lift required.