DARPA Launches $1.5 Billion Program to Reinvent Chip Technology

cageymaru

cageymaru

Fully [H]
Joined
Apr 10, 2003
Messages
22,141
The Defense Advanced Research Projects Agency or DARPA has launched a new $1.5 billion fund to reinvent the way that we think of computers and electronics at the base level. The military is cognizant of the death of Moore's Law and are now seeking alternative designs and techniques to keep America ahead of the rest of the world. They want to focus on chip design, architecture, materials and integration. Some goals are chip and circuit designs that can be completed in 24 hours by a person with little knowledge of the inner workings of each by using machine learning, open source hardware design, and more.

In its architecture efforts, DARPA is aiming for a future where systems-on-chips contain a lot of specialized hardware--such as accelerators hardwired to do particular functions fast--instead of focusing on generalized computing. Software for these chips must still be easy to write and yet be able to take the best advantage of the accelerators involved. According to DARPA, the SDH program aims to develop hardware and software that can be reconfigured in real time based on the sort of data being processed, adapting the computing architecture for the workload in milliseconds.
 
Nice, I don't mind throwing some cash at that problem as long as the research is shared with US companies to be commercialized.

Kinda sounds like they're trying to work out gpgpu though.
 
Some goals are chip and circuit designs that can be completed in 24 hours by a person with little knowledge of the inner workings of each
Click to expand...

Who is going to be stupid enough to trust a chip designed by someone who doesn't understand how to design chips?
Well, probably a lot of people. But it's still a bad idea, which is why the government (not private industry) is funding it.
Because it's always fun to waste Other People's Money on your favorite constituents.
 
Keep America ahead of the rest of the world. News to me they were ahead in the first place.
 
Here is their first working prototype consisting of 5 fifty core modules each at 10 Ghz capable of 10 billion qubits each ….:D
 
The pic did not upload the first time...
W8t2abQH.jpeg
 
all I can say is, "Good luck with that"

While modern chips are designed with a pseudo like circuit coding language, you have to know why things are designed that way to get the most benefit. Designing things like pipelines and prefetches, branch prediction and types of cache is very difficult.

I have hard enough time getting programmers who are good down at the hardware level and good coders. It takes a very unique engineer who can master both.
 
In three years, Cyberdyne will become the largest supplier of military computer systems. All stealth bombers are upgraded with Cyberdyne computers, becoming fully unmanned. Afterwards, they fly with a perfect operational record. The Skynet Funding Bill is passed. The system goes online August 4th, 1997. Human decisions are removed from strategic defense. Skynet begins to learn at a geometric rate. It becomes self-aware at 2:14 a.m. Eastern time, August 29th. In a panic, they try to pull the plug.
 
cageymaru said:
The Defense Advanced Research Projects Agency or DARPA has launched a new $1.5 billion fund to reinvent the way that we think of computers and electronics at the base level. The military is cognizant of the death of Moore's Law and are now seeking alternative designs and techniques to keep America ahead of the rest of the world. They want to focus on chip design, architecture, materials and integration. Some goals are chip and circuit designs that can be completed in 24 hours by a person with little knowledge of the inner workings of each by using machine learning, open source hardware design, and more.

In its architecture efforts, DARPA is aiming for a future where systems-on-chips contain a lot of specialized hardware--such as accelerators hardwired to do particular functions fast--instead of focusing on generalized computing. Software for these chips must still be easy to write and yet be able to take the best advantage of the accelerators involved. According to DARPA, the SDH program aims to develop hardware and software that can be reconfigured in real time based on the sort of data being processed, adapting the computing architecture for the workload in milliseconds.
Click to expand...
It's called a "memory processing unit." Now can I have $1.5B?
https://www.sciencedaily.com/releas.../top_science+(ScienceDaily:+Top+Science+News)
 
Cool! More tech development with public funds so they can sell it to the private sector who can in turn sell it back to the taxpayers or use it against them somehow.
 
Prisoner849 said:
Cool! More tech development with public funds so they can sell it to the private sector who can in turn sell it back to the taxpayers or use it against them somehow.
Click to expand...

You say it sarcastically, but that's pretty much how most underlying technology is initially developed. The overwhelming majority of technology in use today was developed on taxpayers dime. Industry just turned that technology into some sort of product.
 
Prisoner849 said:
Cool! More tech development with public funds so they can sell it to the private sector who can in turn sell it back to the taxpayers or use it against them somehow.
Click to expand...
Well, you know that GPS, mobile communications and the internet come from military research right?
 
  • Like
Reactions: risc
like this
BloodyIron said:
Moore's Law has been "dead" for literally over 40 years now.
Click to expand...

does this mean that we can finally accept moores law for what it is? a cutesie offhand observation made by moore many years ago in a time that no longer applies to ours?
 
Dekoth-E- said:
Only 1.5b? I'm pretty sure Intel's Ceo wipes his ass with more than that.
Click to expand...

Agree, $1.5B doesn't sound like an effort for the US government.

To put some perspective, translating in today's money, the project Manhattan cost about $36B and the more expensive development of the B-29 Superfortress bomber cost a whopping sum of $54B.
 
Except it is still relevant and has been since inception.

THRESHIN said:
does this mean that we can finally accept moores law for what it is? a cutesie offhand observation made by moore many years ago in a time that no longer applies to ours?
Click to expand...
 
Atearen said:
Agree, $1.5B doesn't sound like an effort for the US government.

To put some perspective, translating in today's money, the project Manhattan cost about $36B and the more expensive development of the B-29 Superfortress bomber cost a whopping sum of $54B.
Click to expand...
I dunno if the cost of the most destructive force ever used by mankind and the development and production of nearly 4,000 giant planes brings all that much perspective.
 
gamerk2 said:
You say it sarcastically, but that's pretty much how most underlying technology is initially developed. The overwhelming majority of technology in use today was developed on taxpayers dime. Industry just turned that technology into some sort of product.
Click to expand...

Yep...kinda have to know the truth in order to be sarcastic. ;)
 
Let me tell you about a little something we like to call Itanium - Intel
 
"In its architecture efforts, DARPA is aiming for a future where systems-on-chips contain a lot of specialized hardware--such as accelerators hardwired to do particular functions fast--instead of focusing on generalized computing." I do tend to believe as multi-core CPU's threading benefits reach diminishing returns in a major way special brute force task specific function cores within such a design would be the next logical step.
 
for this, I shall quote that Transylvania Vlad guy.... and I quote: "Blah ba blah"

Also... so this is the pre-story in Lost, I think! I never understood that show, but something about DARPA
 
Last edited:
BloodyIron said:
Except it is still relevant and has been since inception.
Click to expand...

It is only relevant because we make it so. The problem is we try to treat it as scientific law.

Here, I just made my own! Threshin's law: every morning before 8am EST threshin will take a dump.

Its currently 6:10am EST and I'm taking a crap while I write this, so its relevant! My law holds true!
 
Hey, a few billion extra gubby bucks so companies don't need to make money to have money and the endless dumping can continue until all the companies that actually have to make money to have money are all bought up. They might have some good stuff to rebrand as our own!

Innovation Through Abusive Trade Practices.
 
FTFA

In the materials and integration arm of the $1.5 billion effort, DARPA introduced the wining teams in two new programs, Foundations Required for Novel Compute (FRANC) and Three Dimensional Monolithic System-on-a-Chip (3DSoC). 3DSoC aims to gain a 50-fold power computation-time improvement and power reductions by growing multiple layers of interconnected circuitry atop a CMOS base. A group led by MIT’s Max Shulaker gained the lion’s share of the research funds ($61 million) for an ambitious project that aims to show that a monolithic 3D system using 90-nanometer manufacturing technology can match the performance of a CMOS chip made with today’s most advanced process technology.
Click to expand...

Interesting
 
I thought that is why artificial intelligence was important to solve certain aspects that makes mankind run in circles when facing certain problems...
 
So MIT is letting everyone know their junk is too small and his big old 90nm CMOS will set you straight again 50 times over
 
You must log in or register to reply here.
Back
Top