Slashdot Mirror
Intel's "Terascale" Vision
Vigile writes, "Intel is pushing the envelope with its latest vision — 80 cores on a single processor. Dubbed 'Terascale' computing, Intel aims to bring low-powered, massively interconnected cores and unleash a new era in data-mining, media creation, and entertainment." For balance, read Tom Yager over at InfoWorld imploring AMD to stop at 8 cores while everybody gets the architecture right.
Now I can run 80 instances of Doom at the same time. Nothing quite like heavy multitasking.
Palm trees and 8
we are on our way to L-Cars computers i can feel it.
CH
This processor must already be submitting stories... If it is there should be 78 more dupes just like it.
I like the idea of an 80 core processor. Multithreaded applications will work better. Why are people afraid of multiprocessors? Systems with dozens of processors are not uncommon. I dont see why it would be bad for the desktop.
http://github.com/gbook/nidb
Anyone else first read that as "Intel's Testicle Vision"?
Man, it's been a long day.
What if the Hokey Pokey really is what it's all about?
When you can have 80 underfed chickens?
Reality is nothing but a collective hunch.
Gilette is releasing a new shaver called the "Plutonium Mach80", a razor with 80 blades. Each blade has a separate distinct function, and you can get even closer shaves with the synergistic cuisinart action. Also comes in a "For Women" model for "sensitive areas". "Basically, 5 blades isn't enough. I mean, really, more is better, right?", says Gilette CEO James Kilts. Schick is reportedly working on a competitor blade that may exceed the legendary "100 blade barrier".
And slashdotters will still be overclocking the sumbitch.
Insert witty sig here.
If they succeed, does this meen the tera-rists have won?
A lab prototype like this can help them with something important: Given multi-core processors look to be the way future computers will be built, how do you feed them data? The current paradigm won't scale past 4 cores on a single chip's worth of FSB, and there are folks who don't think that even 4's going to be a useful increase over 2.
Even if Intel never sells a chip bigger than 16 or 32 ways, an 80 core lab mule will teach them many things about how to get information to a processor and keep those caches full of appropriate data.
-F
Your comments are a lot more true than many people realize. Specialized hardware always wins.
As an example, people talk about using using multi-GHz machines for TIVO-type appliances, and "getting away" with 600 Mhz or so if your card has hardware MPG encoding. Some of the original TIVOs, because of their reliance on specialized chips and ASICs, used measly 33 MHz CPUs - and worked just fine.
steve
Oh, you're not stuck, you're just unable to let go of the onion rings.
Practicality and usefulness problems aside, you can fit over 6,000 6502 processors in the space of a P4, each running at several ghz.
The first 80-core chip will actually look live a conventional kitchen hotplate. You add a pot of cold water on top of the chip, then with a dial on the unit you determie how much heat you want to produce. The CPU will automatically run the correct number of instances of Seti@Home to generate the desired level of heat.
The 4 X 80 "stove top" model will come out later that year. It will include an "oven" that has its own chip and convectional cooling.
AMD: We now have two cores, so there!
Intel: Oh yeah, well we now have four cores- losers!
AMD: Oh yeah, well we're coming out with eight cores next. Ha beat that!
Intel: We can and will! We're going to come out with, with EIGHTY cores! Yeah that's right, eighty cores!
Disclaimer: I've not kept up on the Core War, so any inaccuracies are for dramatic effect...
If brevity is the soul of wit, then how does one explain Twitter?
I'm a video guy. I can't render video fast enough. I can't do transcoding fast enough. My video is getting larger and deeper in color, and i need more power.
all of that is threadable.
so is photographic processing. You can divide a picture 80 ways and have each processor do whatever it is you want to do on it.
Gamers? Fscking a.... i'm so SICK of hearing hiow everything is for them. Just because something isn't going to help Halo Life 3 run faster is not any of my concern.
There are lots of people working on their computers that want to see more cores because it will make our lives better.
guns kill people like spoons make Rosie O'Donnell fat.
That didn't work because AMD worked out that architecture can trump speed. They innovated, and then did it again with decent dual-core (as in NOT the two-dies-on-one-chip cack that you churned out at first).
So, you improved your architecture and implemented dual-core properly, to produce the fantastic Duo. You got back in the race.
And then there was talk of more cores. And you went "Fuck that, bitches, stay DOWN - we is gon' fuck you up good with 80 cores, bitch, an' dat hard!". Yes, you decided to try and dominate the pissing contest of multi-core instead of megahurtz.
Jesus guys, didn't you learn a fucking thing? STOP trying to turn out something that little bit "more" than the competition, just get on with innovating and coming up with damn good chips. That's how AMD threatened you and, if you go on with this "anything you can do" shit again, you'll be back to square one.
Meta will eat itself
What's a memory bus? Oh right, that thing you use to access the DDR4 swap device when the page you want to access is no longer in the on-CPU RAM. ;-)
Seriously, look at the growth of L2 caches, and tell me the day isn't coming when they just call it "RAM" instead of "cache." If Intel and AMD want to keep piling transistors onto their chips, this'll give 'em something to do.
As copyright owner of this comment, I authorize everyone to defeat any technological measure which limits access to it.
Maybe we'll get 80 copies of this article.
Video Production Support
640 cores should be enough for anyone.
Why is it that with intel talking about a radical change in consumer hardware the level of comments on /. is barely higher then that on AOL.
:-) Of course a nice message passing symbolic language might score big.
We have had multi processor machines for ages. This is not a sudden unknown. Look up transputer, connection machine, beowulf, cray. There is still ground to be covered but it's not unkown territory. The difference is this is intel, intel needs a big market to sell to.
This is not going to make significant difference to the end user, most of them will still write letters, calculate spreadsheets and browse the web. It might be enough to finally expose MS et al for what they have always been, the parasites.
Where this is going to hit home is in the realm of programming and OS.
Want to run an OS primarily designed for uniprocessing on a multi way architecture? Look at the issues Win&Lin have with SMP, limited to 16 processors I believe. Numa and beowulf are a different kettle of fish. So what will we have on these massive SMP architectures?
Programming, at last we might be getting out from under VonNuman. Progress might be possible after 30+ years of stagnation. The symbolic/functional languages are going to start to move forward. Hell we might even get to run on stack based cpus with energy reclamation automated
But given then history of software we'll have a bunch of ignorant, loud mouth idiots running around telling everybody the one true way is Java with mutex and semaphores. PHBs will grab at the first thing that has enterpise written on it and is 'guaranteed'. Most programmers will code how they have always coded head down, ass up. The number of processors will double every two years and the speed of software will continue to halve in the same period.
Of course nobody will suggest that a staged conversion should take place. There will be all these reasons to throw everything away and start over. Because this time we'll get it right!
if you read his article, you would know that he wants them to create cpus that actually perform in the real world, not just add marketing numbers that will have very little effect.
from http://www.informationweek.com/shared/printableArt icle.jhtml?articleID=191901844
"I've always been amazed at the Apollo spacecraft guidance system, built by the MIT Instrumentation Lab. In 1969, this software got Apollo 11 to the moon, detached the lunar module, landed it on the moon's surface, and brought three astronauts home. It had to function on the tiny amount of memory available in the onboard Raytheon computer--it carried 8 Kbytes, not enough for a printer driver these days. And there wouldn't be time to reboot in case of system failure when the craft made re-entry. It's just as well Windows wasn't available for the job. The Apollo guidance system probably seems like routine software to technology sophisticates. Far more complex navigational systems are in operation today. The system's essentials were a few well-known algorithms based on proven logic. But to me, it's still rocket science. Great software dazzles us by virtue of what it does correctly in the face of everything that could go wrong."
Whow, can you do that without 80 cores?
2+2 = 5 (for very large values of 2)
I like the idea of an 80 core processor. Multithreaded applications will work better
Multithreading models from the Windows/Unix/Linux community all assume equal access to system resources such as memory across all threads. They like Uniform Memory Architecture models.
An 80 core system can't really provide a uniform memory access model, as it runs into severe switching and coherency problems. (You want to snoop HOW MANY L1 caches?!??). Fancy interconnects like hyperchannel and Monte Carlo stochastic schemes start getting pinched for bandwidth around 8 cores. With this many cores, you'll wind up with computing meshes of local processors and memory interconnected using some interesting switching scheme. The article even mentions this, with a bit of hand-waving over the issues of bandwidth in shared system resources. "Intel's answer is to attach 256 Mbit of SRAM directly to EACH core. " Interconnect topology is left at a simple tiling scheme, but they are exploring ring topologies.
The result looks remarkably like a transputer mesh. I've programmed these in the past, and the model is rather different than simple multithreading. Being able to decompose the programming problem into a number of independent steps with relatively low communications demands is essential. The ability to reconfigure the interconnect topology to match the problem's data flow is essential to being able to get as much out of the processor set as possible. Without this, one can wind up with lots of idle processors, blocked on data starvation.