even better than memory controler on dice, the memory on dice: TeraScale
(Using multipackaging technics already used in Core 2 ... terribly efficent!
This is working, up and running, Tera is the next step, we should all focus on this big leap.


who?

Who?, do you think you could get me one of those Tera's? :-D

I am sure those will be main stream in few years.
so, yes, I think you will get one.

who?

If one were allowed to ask out of curiousity, what were the dimensions of the prototype?

---

even better than memory controler on dice, the memory on dice: TeraScale
(Using multipackaging technics already used in Core 2 ... terribly efficent!
This is working, up and running, Tera is the next step, we should all focus on this big leap.

Very intersting, especially for getting stacked die packaging to work. Cray was trying that decades ago.

All old ideas, but still really neat to now get it to work all in one package.

Also similar to Tilera announces 64-core processor, the newish nVidia 8xxx chipset series, and the old Transputer of very old!

Good for the low power, small space, and big numbers.

When and how much? Why +5 years away?...

(OK, so more to do with 'good business' rather than improving the world... :-( )

Happy crunchin',
Martin

---

See new freedom: Mageia Linux
Take a look for yourself: Linux Format
The Future is what We all make IT (GPLv3)

even better than memory controler on dice, the memory on dice: TeraScale
(Using multipackaging technics already used in Core 2 ... terribly efficent!
This is working, up and running, Tera is the next step, we should all focus on this big leap.

Very intersting, especially for getting stacked die packaging to work. Cray was trying that decades ago.

All old ideas, but still really neat to now get it to work all in one package.

Also similar to Tilera announces 64-core processor, the newish nVidia 8xxx chipset series, and the old Transputer of very old!

Good for the low power, small space, and big numbers.

When and how much? Why +5 years away?...

(OK, so more to do with 'good business' rather than improving the world... :-( )

Happy crunchin',
Martin

well Tirela can not even beat a Core 2 Quad at SPEC_fp , we are speaking TERA scale. Big difference, very , very big.
the nVidia G80 is far from tera scale too, come on, give us some credit, it is working today, in one little packaging , a little bigger than a core 2.

don't you think Intel's process technology is helping the world? To stay competitive with us, all the other companies will have to invent, be creative, as we did! it is all about progress! So, come on Martin, remove your hat few seconds and reconized that.
What ever i can come up with, you always come up with something to make it looks bad, please readjust your judgement to something a little big more fair. All the tech in the tera scale project are novel, way a head of what ever you came up with. It is nice to have old idea, but the credits come to the people who made it work and can make millions of units. You are not going to change the world if you can only make 20.

when I work on processors, we keep in mind that we have to be able to make enough that the entire planet can get it. Yes, it is a business, but it is as well a duty to the world, and you never give me the credit for it. I am serious,

(OK, so more to do with 'good business' rather than improving the world... :-( )

: here you offended me very seriously, I am an enginneer, and as it, I create, invent, while having fun, because my work is my hobby. I got a ZX80 when I was 13, I could never disconnect from computers, and I still can't. my company is full of people like this, born to live and breath computers, it is our passion. I think you read too many web page and forget that behind the company, there are many men working hard, I do it for my pay check, but I could work only 40 hours per day, instead, I work endless hours, because I love it! Fornutatly for me, I have a fire in my head, with many idea about computer architecture, and I squize this every day to the best computer possible. At least, give us the credit of our passion and dedication, forget the brand 5 minutes and see the humans working behind it, and give them some respect. if you can't do this, well, you are an an objective person, and I can't help you.

Martin, what do you do yourself to improve the world? seriously?

who?

Thanks for the informative posts who,

I think that Intel push the others to keep up and then the compitition puts out a better CPU then it pushes intel. Its known as compitition. Its great for what it does to performance and reliability in the PC world.

Also with projects like rosetta@home and WCG any 1 processor can change the planet (even a 300Mhz!). With a 1.81 Teraflops chip for a project like those could REALLY change the planet.

Thank you for your work in making the planet better.

BoB

---

Do you Good Search for Seti@Home? http://www.goodsearch.com/?charityid=888957
Or Good Shop? http://www.goodshop.com/?charityid=888957

I think that Intel push the others to keep up and then the compitition puts out a better CPU then it pushes intel. Its known as compitition. Its great for what it does to performance and reliability in the PC world.

Nicely put BoB. If Intel were the only chip maker, I think that there would be many people throughout the world who would not be able to afford a computer because of the high price of equipment.

With competition, not only are prices reasonable, but it forces the competing companies to better their current kit.

This in turn, provides people with ability to purchase/build the computer of their choice, as well as within their budget.

When Intel/AMD release their new chips, it enables me to purchase several of the previous chips for the cost of 1 when 1st released, which provides me with the ability to build a nice little network at a reasonable cost.

---

I think that Intel push the others to keep up and then the compitition puts out a better CPU then it pushes intel. Its known as compitition. Its great for what it does to performance and reliability in the PC world.

Nicely put BoB. If Intel were the only chip maker, I think that there would be many people throughout the world who would not be able to afford a computer because of the high price of equipment.

With competition, not only are prices reasonable, but it forces the competing companies to better their current kit.

This in turn, provides people with ability to purchase/build the computer of their choice, as well as within their budget.

When Intel/AMD release their new chips, it enables me to purchase several of the previous chips for the cost of 1 when 1st released, which provides me with the ability to build a nice little network at a reasonable cost.

I did not say that Intel will be the same without competition. I love competition, and we all do, it allows me to go and put more candy into the CPU, and request more.

well, I remember the time when a FX was 1000$ ... I think it is a natural market behavior, if you have a nice and leading technology, it is natural to get some cash for it. after all, a Ferrari cost more than a honda. they are both good, but one is performing better. see the point? So, now that Core 2 is a head by much more than the FX ever was, don't complain about the premium of it, it is the market behavior. And a Core 2 Q6600 at his price is really a good deal!!!!

Competition is good, I love it.

who?

<snip>
And a Core 2 Q6600 at his price is really a good deal!!!!

Competition is good, I love it.

who?

That is an understatement. I don't know how Intel is selling the Q6600's for under $300, but I'm glad they are. I just picked one up and it is amazing. Over twice the horsepower of the P4-D it replaced, and uses about the same power.

---

Boinc Button Abuser In Training >My Shrubbers<

I did not say that Intel will be the same without competition. I love competition, and we all do, it allows me to go and put more candy into the CPU, and request more.

well, I remember the time when a FX was 1000$ ... I think it is a natural market behavior, if you have a nice and leading technology, it is natural to get some cash for it. after all, a Ferrari cost more than a honda. they are both good, but one is performing better. see the point? So, now that Core 2 is a head by much more than the FX ever was, don't complain about the premium of it, it is the market behavior. And a Core 2 Q6600 at his price is really a good deal!!!!

Competition is good, I love it.

who?

Agreed. The point I was making that it is all down to a person's budget.

Most of my customers are parents buying for their kids. Main use is Internet, music, school work, e-mail etc. My builds are dictated by their budget, which is why I mainly use AMD because they are cheaper than Intel.

As my own systems are good enough for the use I put them to, I have no need of a Ferrari.

I do realise that there are people out there who want the best & the latest & there are companies out there that know this.

At the end of the day, it is all down to an individual's choice.

---

Phew! Quite a lot to answer.

Who?: In brief, I also work for the love of what I do. It's also good that I get paid for what I would do as a 'professional hobby' in any case. Life is far too short not to enjoy what you spend your time doing!

That doesn't stop me from making sure that my high ideals are not taken for granted, or that I'm otherwise taken too much advantage of...

It's also very good to see your enthusiasm!

... Tera is the next step, we should all focus on this big leap.

Very interesting, especially for getting stacked die packaging to work. Cray was trying that decades ago.

All old ideas, but still really neat to now get it to work all in one package.

Also similar to Tilera announces 64-core processor, the newish nVidia 8xxx chipset series, and the old Transputer of very old! ...

(OK, so more to do with 'good business' rather than improving the world... :-( )

well Tirela can not even beat a Core 2 Quad at SPEC_fp , we are speaking TERA scale. Big difference, very , very big.

That cpu array is designed more for multiple server threads rather than FPU number crunching...

the nVidia G80 is far from tera scale too,

And that is now already old technology dedicated to GPU functionality. The multiple "FPU" functionality is an 'aside', but also a good example of how 'general' that new GPU arrangement is.

come on, give us some credit, it is working today, in one little packaging , a little bigger than a core 2.

Yes, that is very good, but tuned to hit just that one benchmark to get the headline hitting numbers. Very neatly done and very good 'proof of principle'. Also, rather good for the packaging.

Next is to overcome the hype to come to something that is realistic in servers and home PCs. Which is where the business compromises come in to play...

don't you think Intel's process technology is helping the world? To stay competitive with us, all the other companies will have to invent, be creative, as we did! it is all about progress!

Very much agreed, yes. That part is indeed good for progress, as seen with the recent spectacular improvements in price/performance for CPUs.

Very good kudos for your part in helping that.

What ever i can come up with, you always come up with something to make it looks bad, please readjust your judgement to something a little big more fair. All the tech in the tera scale project are novel, way a head of what ever you came up with. It is nice to have old idea, but the credits come to the people who made it work and can make millions of units. You are not going to change the world if you can only make 20.

Agreed.

If you look up the Transputer, despite its more 'forward looking' architecture, a grid of Transputers couldn't keep up the performance race with the PC CPU developments. However, that was due to fab limitations rather than architecture limitations. The question there is whether that design failed the race due to slow design, slow adoption, poor business, or whatever? I guess that the 'old-style' single thread PC CPU at the time 'looked better'.

Interestingly, many years later, we're revisiting the Transputer ideas... Except that now, Intel and nVidia, and perhaps others, have more of the capability and a more receptive market to leap into the future...

when I work on processors, we keep in mind that we have to be able to make enough that the entire planet can get it. Yes, it is a business, but it is as well a duty to the world, and you never give me the credit for it. I am serious,

With a large dependent market completely dependent on continued supply, that is indeed a very important consideration and responsibility. I guess the USA would be brought down to collapse if a few fab plants were wiped out by severe weather, or even if who knows who wired up a few transistors out of a few hundred million wrongly so that they failed too quickly... That would be bad, but the loss of the company and future developments would be even worse...

And that is just one of the dangers with dependence on just one or two 'big players'...

(OK, so more to do with 'good business' rather than improving the world... :-( )

: here you offended me very seriously, I am an enginneer, and as it, I create, invent, while having fun, because my work is my hobby. ... At least, give us the credit of our passion and dedication, forget the brand 5 minutes and see the humans working behind it, and give them some respect. if you can't do this, well, you are an an objective person, and I can't help you.

Very definitely no offense intended. To get the latest CPUs to work as well and as fast as they do, and to continue to push the physical limits in the fab plants to make them, does indeed take some passion and dedication to make it all work. Everyone is certainly pushing entropy to the extreme limits! Full kudos to you and all the others involved.


The other side of all that is that the big business then takes those very clever toys to then market them like some glorified child's game of 'oneupmanship' and marketing hype to confuse the customers out of their money...

The technology is fantastic.

The application of that technology is usually far from fantastic.

It's good in some ways, but not all.


The present ferociously competitive market seems to produce boom and bust cycles and extra expense due to deliberately designed-in incompatibilities. And all entangled by traps in an elaborate and carefully laid (abused) patents minefield.

Would not development be speeded yet further, and interoperability improved for all, if the main players could agree at least on using common interconnect standards?

The PC as released by IBM has been a phenomenal success because of this. Using a common interconnect for multiple parallel CPUs that includes direct connect to peripherals would be a great improvement. For example, why the mix of PCIe, SATA, HyperTransport, Ethernet, all with very different protocols, interfaced by a Northbridge and Southbridge? Just one standard hierarchical protocol and one interface could be used...

Or is the "Microsoft" style "Embrace and Extend" to deliberately break interoperability, and so to stifle competition, the more profitable path towards world domination?

There is a wider context.

Martin, what do you do yourself to improve the world? seriously?

By trying to be myself...


Keep up the good work,

Regards,
Martin

---

See new freedom: Mageia Linux
Take a look for yourself: Linux Format
The Future is what We all make IT (GPLv3)

Phew! Quite a lot to answer.

Excellent post Martin.

I for one don't always rely on "brand" names as sometimes the product is just not up to scratch.

Most system builders on the forum have the same problems - people comparing our systems against HP,Dell,Compaq etc & trying to get our systems as cheaply as possible.

I have always been amazed by the engineers who developed products for consumers but unfortunately, it has been the Financial/Marketing departments of these companies that promote the products with no reference to the developers.

If all engineers took umbrage at the current state of affairs, we, the consumers would be the ones who lose out.

I for one, appreciate people such as yourself actually posting on forums such as this & providing comments/information to help those same comsumers out.

Many thanks & stay with it. There really are people out there who appreciate people like you. We have the desire to do somethimg like you but do not have the ability.

Regards

PJ

---

Phew! Quite a lot to answer.

Who?: In brief, I also work for the love of what I do. It's also good that I get paid for what I would do as a 'professional hobby' in any case. Life is far too short not to enjoy what you spend your time doing!

That doesn't stop me from making sure that my high ideals are not taken for granted, or that I'm otherwise taken too much advantage of...

It's also very good to see your enthusiasm!

... Tera is the next step, we should all focus on this big leap.

Very interesting, especially for getting stacked die packaging to work. Cray was trying that decades ago.

All old ideas, but still really neat to now get it to work all in one package.

Also similar to Tilera announces 64-core processor, the newish nVidia 8xxx chipset series, and the old Transputer of very old! ...

(OK, so more to do with 'good business' rather than improving the world... :-( )

well Tirela can not even beat a Core 2 Quad at SPEC_fp , we are speaking TERA scale. Big difference, very , very big.

That cpu array is designed more for multiple server threads rather than FPU number crunching...

the nVidia G80 is far from tera scale too,

And that is now already old technology dedicated to GPU functionality. The multiple "FPU" functionality is an 'aside', but also a good example of how 'general' that new GPU arrangement is.

come on, give us some credit, it is working today, in one little packaging , a little bigger than a core 2.

Yes, that is very good, but tuned to hit just that one benchmark to get the headline hitting numbers. Very neatly done and very good 'proof of principle'. Also, rather good for the packaging.

Next is to overcome the hype to come to something that is realistic in servers and home PCs. Which is where the business compromises come in to play...

don't you think Intel's process technology is helping the world? To stay competitive with us, all the other companies will have to invent, be creative, as we did! it is all about progress!

Very much agreed, yes. That part is indeed good for progress, as seen with the recent spectacular improvements in price/performance for CPUs.

Very good kudos for your part in helping that.

What ever i can come up with, you always come up with something to make it looks bad, please readjust your judgement to something a little big more fair. All the tech in the tera scale project are novel, way a head of what ever you came up with. It is nice to have old idea, but the credits come to the people who made it work and can make millions of units. You are not going to change the world if you can only make 20.

Agreed.

If you look up the Transputer, despite its more 'forward looking' architecture, a grid of Transputers couldn't keep up the performance race with the PC CPU developments. However, that was due to fab limitations rather than architecture limitations. The question there is whether that design failed the race due to slow design, slow adoption, poor business, or whatever? I guess that the 'old-style' single thread PC CPU at the time 'looked better'.

Interestingly, many years later, we're revisiting the Transputer ideas... Except that now, Intel and nVidia, and perhaps others, have more of the capability and a more receptive market to leap into the future...

when I work on processors, we keep in mind that we have to be able to make enough that the entire planet can get it. Yes, it is a business, but it is as well a duty to the world, and you never give me the credit for it. I am serious,

With a large dependent market completely dependent on continued supply, that is indeed a very important consideration and responsibility. I guess the USA would be brought down to collapse if a few fab plants were wiped out by severe weather, or even if who knows who wired up a few transistors out of a few hundred million wrongly so that they failed too quickly... That would be bad, but the loss of the company and future developments would be even worse...

And that is just one of the dangers with dependence on just one or two 'big players'...

(OK, so more to do with 'good business' rather than improving the world... :-( )

: here you offended me very seriously, I am an enginneer, and as it, I create, invent, while having fun, because my work is my hobby. ... At least, give us the credit of our passion and dedication, forget the brand 5 minutes and see the humans working behind it, and give them some respect. if you can't do this, well, you are an an objective person, and I can't help you.

Very definitely no offense intended. To get the latest CPUs to work as well and as fast as they do, and to continue to push the physical limits in the fab plants to make them, does indeed take some passion and dedication to make it all work. Everyone is certainly pushing entropy to the extreme limits! Full kudos to you and all the others involved.


The other side of all that is that the big business then takes those very clever toys to then market them like some glorified child's game of 'oneupmanship' and marketing hype to confuse the customers out of their money...

The technology is fantastic.

The application of that technology is usually far from fantastic.

It's good in some ways, but not all.


The present ferociously competitive market seems to produce boom and bust cycles and extra expense due to deliberately designed-in incompatibilities. And all entangled by traps in an elaborate and carefully laid (abused) patents minefield.

Would not development be speeded yet further, and interoperability improved for all, if the main players could agree at least on using common interconnect standards?

The PC as released by IBM has been a phenomenal success because of this. Using a common interconnect for multiple parallel CPUs that includes direct connect to peripherals would be a great improvement. For example, why the mix of PCIe, SATA, HyperTransport, Ethernet, all with very different protocols, interfaced by a Northbridge and Southbridge? Just one standard hierarchical protocol and one interface could be used...

Or is the "Microsoft" style "Embrace and Extend" to deliberately break interoperability, and so to stifle competition, the more profitable path towards world domination?

There is a wider context.

Martin, what do you do yourself to improve the world? seriously?

By trying to be myself...


Keep up the good work,

Regards,
Martin

I think you hate big companies.

who?

For AMD enthusiasts (and ignoring some recent static in the thread):

x86Watch.com

Drilling down through the relevent links (I'm sure you can figure out which ones they are) results in some interesting tidbits.

Alinator

I think you hate big companies.
who?

I don't think it's a case of hating big companies, I think it's a case of big companies letting down their customers.

Take Norton as an example - pretty good software, however, since it was bought out by Symantec it has been nothing but problemtic. So much so that -I no longer install it in any of my customer's machines. I also know numerous system builders who does exactly the same, & that is just in my area. Multply that worldwide & that is a major loss for any company.

When innovative companies grow, they somehow tend to forget the very people that helped them grow.

Another example - quite a few of my "upgrading" customers purchased their systems from PC World. I have no problem with that as it is their choice.

Since their warranties have expired, they came to me for upgrades. However, I am finding it difficult to upgrade their bios as the mobo's are Soltek's.

You try going to Soltek's website to download latest bios!!!

These companies are letting down their consumers big time!!!

This is the main reason why I will never buy a ready built "Brand". This does not make me a big company hater!!!

---

For AMD enthusiasts (and ignoring some recent static in the thread):

x86Watch.com

Drilling down through the relevent links (I'm sure you can figure out which ones they are) results in some interesting tidbits.

Alinator

Great link Alinator. Can I have your view - Is it worth waiting until the end of the year for the 3.0ghz chip?

Regards

PJ

---

Great link Alinator. Can I have your view - Is it worth waiting until the end of the year for the 3.0ghz chip?

Regards

PJ

Well, the answer is, it depends. If your need is in the traditional Desktop space, then the truth of the matter is Intel has AMD covered and will continue to in the short term. So the best move is to build now while the great deals on the current Intel product lines are around. This is assuming you're looking to get the best bang for the buck in a 'do everything including crunching' wonder box.

If your need is in Server space, then the choice really boils down to how urgent the need is. There may be an advantage in performance/price if you can hold out until the end of the year.

Alinator

Great link Alinator. Can I have your view - Is it worth waiting until the end of the year for the 3.0ghz chip?

Regards

PJ

Well, the answer is, it depends. If your need is in the traditional Desktop space, then the truth of the matter is Intel has AMD covered and will continue to in the short term. So the best move is to build now while the great deals on the current Intel product lines are around. This is assuming you're looking to get the best bang for the buck in a 'do everything including crunching' wonder box.

If your need is in Server space, then the choice really boils down to how urgent the need is. There may be an advantage in performance/price if you can hold out until the end of the year.

Alinator

Thanks. I do want a server but am in no rush to finalise yet. Quite happy to wait for the new year - there should be plenty of bargains by then.

---

You try going to Soltek's website to download latest bios!!!

These companies are letting down their consumers big time!!!

I was going to ask what is wrong with Soltek but then I went to their website - or actually tried. .de is still alive but quite useless. Looks like the others are gone.

Now, tell me, how is Soltek letting down their customers if the company went bankrupt? (Well, that's it seems like.)

-Juha

---

You try going to Soltek's website to download latest bios!!!

These companies are letting down their consumers big time!!!

I was going to ask what is wrong with Soltek but then I went to their website - or actually tried. .de is still alive but quite useless. Looks like the others are gone.

Now, tell me, how is Soltek letting down their customers if the company went bankrupt? (Well, that's it seems like.)

-Juha

That I was/am unaware of. However, a customer recently bought a system from PC World - Mobo=Soltek. If they are bankrupt, then PC World are not treating their customers fairly!! & PC World is a BIG Company!!!!

This I believe, proves my point with regards to big companies!!!

---

I think you hate big companies.

Not at all.

Big (cooperative) companies can bring great positive advantages. Just as cooperative societies can advance civilization for the good of all.

What I do 'hate' are the (anticompetitive) 'dirty tricks' that are often perpetrated by those in power in big companies. Also the blatent Marketing lies that are perpetrated.

We all pay dearly for that.


In other areas (for example food, medicine, automotive, others), some people are sent to a premature death and even still believe to the very end the Marketing message that it is all good for them.

Unfortunately, there is enough ignorance and lack of interest that many of the dirty tricks can remain unidentified until long after the damage is done.


Regards,
Martin


(Special note: These comments are completely off-topic and completely general.)

---

See new freedom: Mageia Linux
Take a look for yourself: Linux Format
The Future is what We all make IT (GPLv3)