Multi-Core Research and Development

[piasabird]

So we have Built this marvel of technology and have no Idea what to do with it??? 2 Cores, 4 Cores, 8 Cores, 16 Cores, and there is no software that can even utilize this new technology.

I was just thinking if we can have multiple cores on one motherboard why not multiple motherboards on one motherboard? Instead of Intel and Envidia calling each other names, why not find out how to make games run better???



http://www.newsfactor.com/news...?story_id=01200172S2YO

Last month, Intel and Microsoft announced they were jointly financing new labs at the University of California, Berkeley; and the University of Illinois at Urbana-Champaign to tackle the same problem.

All three efforts are in response to a growing awareness that the software industry is not ready for the coming availability of microprocessors with 8 or 16 or more cores, or processing units, on a single chip . Computer and chipmakers are concerned that if software cannot use the new hardware efficiently, customers will have little reason to upgrade.

So we have Built this marvel of technology and have no Idea what to do with it???

 

[Bateluer]

Software R&D always lags hardware. There is a great deal of software that can utilize 2 and 4 cores right now, just a lot of it is in the scientific, distributed computing, workstation, and server applications.

 

[JackMDS]

Software Businesses do not march to the Drum Beat of enthusiasts (especially that most of us usually post, "I want X and please find one for me that is Free").

The applications for the advanced technology are No tout there because in general there is No functional demand for them.

WWE is doing well with DueCore, and tunes for Smart Phones are fine with one Core.

 

[bendixG15]

When the first telephone was built, nobody wanted it either. They figured the best it would be was a toy for kids.

 

[JackMDS]

Originally posted by: bendixG15
When the first telephone was built, nobody wanted it either. They figured the best it would be was a toy for kids.

On the other hand IBM invested hundred of Millions in the IBM PCjr.

 

[Bateluer]

Originally posted by: bendixG15
When the first telephone was built, nobody wanted it either. They figured the best it would be was a toy for kids.

You have a source for that? Seems like a device that enabled long distance communication at the touch of a few buttons would have staggering practical applications for nearly every one, private, civilian, government, etc.

 

[bendixG15]

Originally posted by: JackMDS

Originally posted by: bendixG15
When the first telephone was built, nobody wanted it either. They figured the best it would be was a toy for kids.

On the other hand IBM invested hundred of Millions in the IBM PCjr.

The IBM PcJr was a marketing ploy to sell a crappy computer with a chiclet keyboard to the public. They did not invest hundreds of million of dollars on that thing. Whoever told you that they did was spinning a yarn ...........

 

[bendixG15]

Originally posted by: Bateluer

Originally posted by: bendixG15
When the first telephone was built, nobody wanted it either. They figured the best it would be was a toy for kids.

You have a source for that? Seems like a device that enabled long distance communication at the touch of a few buttons would have staggering practical applications for nearly every one, private, civilian, government, etc.

Just go read up on your history .........

 

[Schadenfroh]

Originally posted by: Bateluer

Originally posted by: bendixG15
When the first telephone was built, nobody wanted it either. They figured the best it would be was a toy for kids.

You have a source for that? Seems like a device that enabled long distance communication at the touch of a few buttons would have staggering practical applications for nearly every one, private, civilian, government, etc.

We have the telegram and an established infrastructure for it. The telegram serves this purpose already, no need to spend the extra money on such a gimmick as voice transmission. The quality of sound is rather poor and the telegram is much less prone to mistakes in interpretation (when operated by a skilled technician).

 

[RebateMonger]

Originally posted by: Bateluer
You have a source for that? Seems like a device that enabled long distance communication at the touch of a few buttons would have staggering practical applications for nearly every one, private, civilian, government, etc.

In the real world, NOTHING has staggering implications. Even totally free energy wouldn't affect life as much as you might think. And every new development has infrastructural changes necessary to make it useful.

My only dream is a tiny 8080 chip on each motherboard, dedicated to making the mouse and menu system run smoothly even when the PC is busy. THAT would be staggering to me.

 

[JackMDS]

Originally posted by: bendixG15
The IBM PcJr was a marketing ploy to sell a crappy computer with a chiclet keyboard to the public. They did not invest hundreds of million of dollars on that thing. Whoever told you that they did was spinning a yarn ...........

I guess that you never signed a check paying for advertizing, TV commercial, and other marketing ploys.
I did not specify that the money was only for the technology part.

Down the line I meant that Nothing is totally sure in developing technology, but some kind of projection is needed for a decision to develop.

If it currently Not in high priority development it means that investigation with the business community shows that there is Not a big demand for such application to run on work stations.

 

[ichi]

Originally posted by: piasabird
So we have Built this marvel of technology and have no Idea what to do with it??? 2 Cores, 4 Cores, 8 Cores, 16 Cores, and there is no software that can even utilize this new technology.

I was just thinking if we can have multiple cores on one motherboard why not multiple motherboards on one motherboard? Instead of Intel and Envidia calling each other names, why not find out how to make games run better???

Well, there is a lot of software that can use multi-core machines, and there has been for many years.
Intel is pushing developers hard now to get them to build more multi-core friendly software.
I think it is really a problem with developers, not with Microsoft or Intel. Now that multi-core machines are in the mainstream, we will see more and more games taking advantage of it.

 

[Blain]

Originally posted by: piasabird
So we have Built this marvel of technology and have no Idea what to do with it???

So we have Built this marvel of technology and have no Idea what to do with it???

You are of course, wrong.

 

[hans007]

intel honestly needs to stop whining.

cpu development has always been a , if you build it they will come up with killer apps thing.
it will always be that way.

software engineers need to be given more time to come up with "killer apps". and coming up with an entire new field of software to take advantage of cpus no one really needs righ tnow, is much harder than, cranking out chips with more and more and more cores. one requires creativity the other is just proces shrinks.

i mean intel sounds like they want us to basically make ray traced word processors or something just so that they can continue making money selling 5000 core cpus.

 

[Modelworks]

I brought this up to people that were saying buy a quad core all the games will be using it.
I kept telling them that programming games that would use a quad core was not easy and that even full dual core usage isn't quite there yet.
They all replied back, yeah but supreme commander did it, so it just programmers don't know what they are doing if they can't write games for quad cores.

I wish it was as easy as saying, hey , lets support quad core.
I use smp every day because I do 3d graphics, but its a lot harder to use smp in realtime applications.

 

[JackMDS]

Originally posted by: ichiWell, there is a lot of software that can use multi-core machines, and there has been for many years.

Do down the line what the explanation is?

1. Software developers are misanthropes that do not want us to enjoy new applications.

2. Software developers hate money and do not want to sell new software.

3. Software developers are fools that cannot develop any new applications.

4. Research and surveys indicate that the general market is not yet of need of hyper multi-core software applications

P.S. Right now, Niche market users develop their own propriety application if needed.

 

[gsellis]

If you were a code monkey, you would understand. Define a task that you do that would be rapidly finished by a massively parallel operation.

Doing your checkbook balance? Serial as it needs the balance value to subtract or add the new ledger item. Fail

Building a house? Some parallel processes, but each has a dependency (wait/semaphore) on some other process (foundation must be dug, then poured to put framing on. Framing can occur with some independence, but must wait for the foundation before assembly. Roof joists can be built, but must wait for the framing to be in place, etc.) Some, but 4-8 cores is efficient.

Calculate all the permutations of a chemical composition and see if it will plug into a cancer cell to block its ability to absorb nutrients. Pass! A master thread can tell each to do an independent task and pass the results to a test process and back to the master.

Most of the stuff we do is serial. There is a lot of research into predictive branching, but it is not always efficient. Like playing chess. While your opponent is considering his move, you can calculate all of the possibilities and then the possibilities of your counters and the counters to that. When your opponent moves, you can quickly move as you have already gotten ahead of him. But all the rest has to be thrown out.

 

[JackMDS]

Originally posted by: gsellis
If you were a code monkey, you would understand. Define a task that you do that would be rapidly finished by a massively parallel operation..

In my case it is more than a code monkey. I am a brain Specialist; I know how the Dominant Left Verbal/Serial hemisphere works, how the Right Visual/holistic hemisphere works, and the interaction between them.

I do not think that we are short of developers that can take advantage of combo serial - parallel processes, it seems that we are just short of over all commercial need for it.