Intel has moved a step closer to commercial neuromorphic computing

So I guess this is where Intel has been focusing it's resources instead of the desktop CPU market. They have been pioneering the next level of computing. Not surprised, it seemed to me they didn't really care about the challenge from AMD since they had bigger fish to fry.
 
Dr. Righteous said:
So I guess this is where Intel has been focusing it's resources instead of the desktop CPU market. They have been pioneering the next level of computing. Not surprised, it seemed to me they didn't really care about the challenge from AMD since they had bigger fish to fry.
Click to expand...

That leap triggered my acrophobia.
 
just imagine what they have been doing the past 40 years over at Los Alamos, Lawrence Liverrmore, Sandia Labs ... you know, all the Military research divisions
 
Dr. Righteous said:
So I guess this is where Intel has been focusing it's resources instead of the desktop CPU market. They have been pioneering the next level of computing. Not surprised, it seemed to me they didn't really care about the challenge from AMD since they had bigger fish to fry.
Click to expand...
No excuse, they can tackle both, with next to no impact.
But hey, not enough billions this quarter!!! The humanity!!! Cut x86 development, its not like we have competition anyway.
Sadly in a way it makes sense... Just that 'the way' is a shit way.
 
You can kiss quantum computing bullshit good bye, this is what make sense. This and eventually this in organic "cpu"s
 
Eighit million neurons is what, a zebrafish brain, according to Wikipedia?

I think given the limits of fabrication process it might be a while before we get anything with a useful amount of brainpower simulated.

Something tells me AMD is not losing their shit over this.
 
deton8 said:
Eighit million neurons is what, a zebrafish brain, according to Wikipedia?

I think given the limits of fabrication process it might be a while before we get anything with a useful amount of brainpower simulated.

Something tells me AMD is not losing their shit over this.
Click to expand...

AMD won't be a player in this arena.
 
5150Joker said:
AMD won't be a player in this arena.
Click to expand...

Or they'll wait until the market is viable and then fly their own design. Being a research pioneer in a new market, especially a market that won't be viable for another 5, 10, 15 years, means that you put up the research costs and then fight to keep your design principles out of the hands of your competitors. Reverse engineering is cheap, and if you spend too much money pursuing early market possibilities you can find yourself damaged instead of being the market leader (i.e. 3dfx).
 
Mighty Man said:
Wonder how long until a beach on T.A.H.I.T.I.?
Click to expand...

Its where Elon Musk is right now. His body is under the control of Intel/Darpas Mauna Loa neuromorphic super cluster. (there a volcanic shield joke riffing on a code name inside a geek joke inside a geek culture joke about an secret organization called shield) MetaMetaMeta joke.

Really though Elon talking about neuralink... guy does look like he needs some more sleep. The combo of these 2 techs though does scare the shit out of me. We will soon be living in a Gibson novel no one wants to be living in. Worried about loosing your job to AI today... how about the AI generation of kids growing up with implants in their heads. AHI artificial human intelligence is coming for all us human 1.0s.
 
seanreisk said:
Or they'll wait until the market is viable and then fly their own design. Being a research pioneer in a new market, especially a market that won't be viable for another 5, 10, 15 years, means that you put up the research costs and then fight to keep your design principles out of the hands of your competitors. Reverse engineering is cheap, and if you spend too much money pursuing early market possibilities you can find yourself damaged instead of being the market leader (i.e. 3dfx).
Click to expand...

This is why patents are a thing and to this day things like x86 are only in the hands of 3 companies. The same will be true of whatever Intel discovers and patents. It takes enormous R&D but you can protect it and lock in a market.
 
seanreisk said:
Or they'll wait until the market is viable and then fly their own design. Being a research pioneer in a new market, especially a market that won't be viable for another 5, 10, 15 years, means that you put up the research costs and then fight to keep your design principles out of the hands of your competitors. Reverse engineering is cheap, and if you spend too much money pursuing early market possibilities you can find yourself damaged instead of being the market leader (i.e. 3dfx).
Click to expand...

Intel being a pioneer for so many technologies is the reason why they can push AMD around, even with an inferior product. This is NOT a good way to go about business. AMD historically does innovate and push itself forward but it has hardly ever pioneers entirely new concepts.

Also, 3dfx failed as a company because they stopped innovating and tried to play the catch-up game to Nvidia.
 
Dr. Righteous said:
So I guess this is where Intel has been focusing it's resources instead of the desktop CPU market. They have been pioneering the next level of computing. Not surprised, it seemed to me they didn't really care about the challenge from AMD since they had bigger fish to fry.
Click to expand...
Like anything else I'll wait for more than a companies press release.
 
Trimlock said:
Also, 3dfx failed as a company because they stopped innovating and tried to play the catch-up game to Nvidia.
Click to expand...

I think it had more to do with trying to dominate the supply chain instead of just being an OEM chip supplier. Buying STB ended up being a disaster. They threw away well over $100 million for nothing.

Intel is having to spend a whole lot of money getting their fabs back on track, money that could have been spent designing new chips. Fortunately for Intel they've got a lot more money to burn through than 3dfx ever did.
 
DeathFromBelow said:
I think it had more to do with trying to dominate the supply chain instead of just being an OEM chip supplier. Buying STB ended up being a disaster. They threw away well over $100 million for nothing.

Intel is having to spend a whole lot of money getting their fabs back on track, money that could have been spent designing new chips. Fortunately for Intel they've got a lot more money to burn through than 3dfx ever did.
Click to expand...

It helps that even with all the 10nm fuckups they have had good product to sell. As nice as it is for AMD to have 7nm parts shipping now... it might honestly be a bit late. AMD is going to grab a ton of enthusiast market for sure with Ryzen2. However the majority of consumer parts are mobile and Intel will keep winning those OEM contracts with what they have. If their 10nm really does get back on track for Q1 they could all but erase the gains AMD has made. Especially if their XE parts are not all smoke. Even if AMD gets a good ryzen2 mobile part shipping before xmas. OEMS probably won't have their heads turned as quickly as the enthusist market. Intel starts shipping 10nm mobile parts with much improved XE graphics mid next year and AMD is shut out again.

AMDs window is pretty small... they need to execute at the server and mobile level if they really want to hurt Intel for real. 2019 will be a good year for AMD... but 2020 is going to have some real heavy hitting competition from Intel and Nv. If NV went the right way around and spun Ampere as a little->big design NV could come for their RX/5600/5500 bread and butter. Hopefully AMD has some more cards up its sleeve.
 
DeathFromBelow said:
I think it had more to do with trying to dominate the supply chain instead of just being an OEM chip supplier. Buying STB ended up being a disaster. They threw away well over $100 million for nothing.

Intel is having to spend a whole lot of money getting their fabs back on track, money that could have been spent designing new chips. Fortunately for Intel they've got a lot more money to burn through than 3dfx ever did.
Click to expand...

Can't dominate the supply chain if you don't have a product the OEM's want. Intel has a mixed history of listening to OEM's but they do mostly deliver on what the OEM's demand.
 
DeathFromBelow said:
I think it had more to do with trying to dominate the supply chain instead of just being an OEM chip supplier. Buying STB ended up being a disaster. They threw away well over $100 million for nothing.
Click to expand...

Closer to $150 million, and they also painted a target on themselves by becoming the main competitor to all the graphics card makers, which caused those same card makers to turn to Nvidia for 3D chips. But 3dfx was also trying to push Glide onto everyone. They had some wins in the arcade cabinet market and used that to court all the console makers (who weren't able to afford 3dfx chips) in the hopes they could lock game developers into Glide. They were simultaneously trying to limit the scope of OpenGL (and PowerVR) in the courts, but this didn't really work for them. Glide was a boutique subset of same API calls found in OpenGL, trimmed down to make it work on a PC's limited memory and bandwidth. For a few years that worked great, but computers grew, and throughout the Voodoo series 3dfx never added new features back in. And since Glide was limited to 3dfx chips, and 3dfx chips were limited to Glide, by the time Voodoo (3?) came out they were pooched.

And that's the point I was making with Intel. Intel tried to control the field when they took the math co-processor off the motherboard, when they made exclusive deals with Rambus, when they made their sockets and board designs 'Intel only', and when they stopped licensing development of their chipsets.

I don't know if neuromorphic computing will be a thing, or if Intel will be the industry leader, but I do think that if the market becomes real and Intel is the only useful dog in the stockyard, the courts might break up Intel. AMD is still trying to get back on their feet, but I don't think this is what threatens them.
 
You must log in or register to reply here.
Back
Top