Tolliman X3 gets a name

[MU_Engineer]

Wow, I didn't see the post you quoted. I guess whoever said that wasn't here last year when I posted some sample code that everyone and their mother Googled. I wonder if they felt stupid when it turned out to be mine? I was an automation programmer in test at Micorsoft. Now I mostly do websites.

Nah, nothing personal. I just saw the opportunity for a good joke (linux_0 still come around here? He'd laugh!) and took it.

 

[yomamafor1]

I posted my guess based on the data available, but you didn't post anything to support your claim. Where is "I heard one place"?

humm...I guess you were too busy figuring out 283mm^2 = 28.3cm^2, that you simply ignored my argument?

 

[zenmaster]

I for one like the ideas of 2.2 and 2.4Ghz Quad Core Barcelonas.

I highly suspect they will beat the existing Q6600 on a clock for clock basis before any OCs. (Check the Server Chip Reviews that show the much slower chips doing just fine in many tests.)

This may give Intel a little shove to release cheaper faster processors sooner.

It's nice to seem some dates and speeds listed.

Sure I would have loved to see them 6months ago, but get over it people.

 

[BaronMatrix]

...and I remembered AMD charging 300 USD for a X2 3800...

AMD's a year late on 45nm, a year late on quad core, a year late on HK/MG, a year late on "tick-tock" strategy....

As I said, AMD64 doesn't mean much to most people, as they're still using x86. Native quad cost them 0.6 billion per quarter, lowered stock, possibility of going bankrupt, and nothing else.

Time to look at the future, Baron, i mean, Christian.

So do I. Intel was setting prices even with HeatBurst. Since AMD was faster they could charge more. With the way Intel dropped the bottom of the CPU industry, NO NEW MANUFs can get by the barrier to entry. No new architecture cost less than the previous, except Core 2.

I wonder why?

Barcelona's design principals had nothing to do with the losses. The price war did. If AMD had made an MCM quad, Intel would have dropped prices further. Just face it, Intel doesn't care about anyone but Intel.

 

[onestar]

Hey guys....in the end it will not matter. AMD and INtel will charge according to demand and yield is just a tiny part of the equation.

So...demand more and pay more. The equation stands.

 

[BaronMatrix]

I posted my guess based on the data available, but you didn't post anything to support your claim. Where is "I heard one place"?

humm...I guess you were too busy figuring out 283mm^2 = 28.3cm^2, that you simply ignored my argument?

Google Barcelona wafers.

 

[yomamafor1]

So do I. Intel was setting prices even with HeatBurst. Since AMD was faster they could charge more. With the way Intel dropped the bottom of the CPU industry, NO NEW MANUFs can get by the barrier to entry. No new architecture cost less than the previous, except Core 2.

I wonder why?

So Intel price Core 2 according to their cost, so what? Even when priced competitively, Intel's margin maintained. Its standard business practice to compete.

Barcelona's design principals had nothing to do with the losses. The price war did. If AMD had made an MCM quad, Intel would have dropped prices further. Just face it, Intel doesn't care about anyone but Intel.

I did not say Barcelona's design is directly related to AMD's loss, but it will be for quarters to come. Currently AMD's loss is because of its poor execution.

Just face it, Intel's a company, and AMD is a company. No one is better than the other. If AMD can charge you an arm and a leg for a processor, they will do so.

 

[BaronMatrix]

You're right, I don't. But I can guess, based on the data available, right?

Based on AMD's own slide, current 65nm Barcelona are yielding at 30%, maybe a little more. Then AMD came out and said, "look, we really can't make 2.5Ghz Barcelona now, but 2.0Ghz will do". After a while, AMD said, "hey, let's launch a new processor that was never on the roadmap."

Don't you think 30% yield has a little merit now?

My point was that I wouldn't be happy with 30% yields and AMD says they're happy with the yields. If they really wanted to harp on yields they would have rather than defect density. I still don't think the yields are anywhere near that low. The higher bins could be lower for steppings prior to B2F or B3, but you are saying that ONLY 3 out of 10 chips EVEN WORK.

That's nearly impossible. They would never have enough chips for Newegg, much less Dell HP, and Sun. Get yields out of your head. When AMD or a representative that has been to Fab 36 says that, I'll believe it. Until then their yields in my estimation, with the near maturity of 65nm, are in the high 60-70%.

The B2F stepping will be the release of Phenom. B3 will probably be 3GHz+ Kuma and Tolliman.

 

[BaronMatrix]

Nah, nothing personal. I just saw the opportunity for a good joke (linux_0 still come around here? He'd laugh!) and took it.

Hey I can take a joke. I almost bought Vista.

 

[BaronMatrix]

So Intel price Core 2 according to their cost, so what? Even when priced competitively, Intel's margin maintained. Its standard business practice to compete.


I did not say Barcelona's design is directly related to AMD's loss, but it will be for quarters to come. Currently AMD's loss is because of its poor execution.

Just face it, Intel's a company, and AMD is a company. No one is better than the other. If AMD can charge you an arm and a leg for a processor, they will do so.

I give up. You win. K10 sucks. AMD 65nm sucks. Dell, HP, Sun and the others have a sucky piece of crap fused sand. It melts cases and can't add. AMD should be sued for releasing it. I mean it's only 20%+ faster than Opteron clocked 1GHz higher. Those rip-off artists.

The X2 prices were great. Everyone had a fast proc for the price they wanted to pay. And yes I'm pissed that my 4400+ is worth nothing now when it should be worth at least $250.

 

[yomamafor1]

K10 will definitely perform in scaled environment, but I'm not so sure about IPC though...

http://realworldtech.com/forums/index.cfm?action=detail&id=83485&threadid=83478&roomid=2

Ooops... Barcelona 1.9Ghz perform slightly better in INT than Opteron 2.0, but perform similarly in FP.

So where is your "only 20% faster than opteron clocked 1Ghz higher"?

 

[yomamafor1]

My point was that I wouldn't be happy with 30% yields and AMD says they're happy with the yields. If they really wanted to harp on yields they would have rather than defect density. I still don't think the yields are anywhere near that low. The higher bins could be lower for steppings prior to B2F or B3, but you are saying that ONLY 3 out of 10 chips EVEN WORK.

By stating 30%, it means AMD can only yield 30% of the functional Barcelona, with four cores intact. Adding Tolliman, I'm not sure about the yield, but if the defects are not clustered, then I'm sure the yield can go up to 50% or more. Then, adding Kuma, that adds another 20%.

So for AMD to yield all processors lineup with K10 core, I would suspect about 70~80%, which is reasonable. But for Barcelona and Phenom, I'm suspecting no more than 40% as of now.

That's nearly impossible. They would never have enough chips for Newegg, much less Dell HP, and Sun. Get yields out of your head. When AMD or a representative that has been to Fab 36 says that, I'll believe it. Until then their yields in my estimation, with the near maturity of 65nm, are in the high 60-70%.

Let me see, after Barcelona went on sale in channel, the chips were gone within several days, and most vendors do not get another shipment until November.

On the other hand, several OEMs complained about their availability.
Dell still has no Barcelona available for their servers. As of now, you cannot buy a Dell server with Barcelona. Same with HP and Sun.

The B2F stepping will be the release of Phenom. B3 will probably be 3GHz+ Kuma and Tolliman.

Wow!....but so far AMD only plan to release them at 2.5Ghz maximum for Tolliman....

Guess AMD isn't too confident about B3 either.

 

[chunkymonster]

...and I remembered AMD charging 300 USD for a X2 3800...

AMD's a year late on 45nm, a year late on quad core, a year late on HK/MG, a year late on "tick-tock" strategy....

As I said, AMD64 doesn't mean much to most people, as they're still using x86. Native quad cost them 0.6 billion per quarter, lowered stock, possibility of going bankrupt, and nothing else.

Time to look at the future, Baron, i mean, Christian.

That's a very two dimensional view on semi-conductors and the business world as a whole.

My favorite overpriced processor launch was the 840 Extreme Edition for a mere $1030! :lol:

AMD is a year late?! Late with what? Did the the SIA establish an schedule to reach specific fabrication processes? I never read that in the newsletter!

As a head's up, Intel is "still using" x86 too. It's a pretty interesting story on how Intel came into their x86-64 instruction set, you should research that.

I realize you get some perverse enjoyment sticking it to BaronMatrix but it doesn't really make for an interesting thread and certainly has taken this one way off topic.

 

[BaronMatrix]

By stating 30%, it means AMD can only yield 30% of the functional Barcelona, with four cores intact. Adding Tolliman, I'm not sure about the yield, but if the defects are not clustered, then I'm sure the yield can go up to 50% or more. Then, adding Kuma, that adds another 20%.

So for AMD to yield all processors lineup with K10 core, I would suspect about 70~80%, which is reasonable. But for Barcelona and Phenom, I'm suspecting no more than 40% as of now.


Let me see, after Barcelona went on sale in channel, the chips were gone within several days, and most vendors do not get another shipment until November.

On the other hand, several OEMs complained about their availability.
Dell still has no Barcelona available for their servers. As of now, you cannot buy a Dell server with Barcelona. Same with HP and Sun.


Wow!....but so far AMD only plan to release them at 2.5Ghz maximum for Tolliman....

Guess AMD isn't too confident about B3 either.

Newegg has both 2347 and 8347. 2347 is selling out every few days and they get more every week. They were out last week, got some more and ran out again.

Why would they release the highest clocked chips first? Kuma was said to go from 2.2-2.9GHz. Phenom from 2.2-2.7GHz this year. I have heard no one complaining about availability. Dell is not gong to releae until they have at least enough for 10,000 servers maybe even 30,000. People who buy from Dell, HP, etc buy in quantity.


Again, what does "intact" mean? All cores work or work at a given speed?


And checking Newegg, they now have 2344HE, 2346HE, 2347 and 8347. that is the sweet spot right now as they will fit all the way up to 2.4GHz K8 Opteron at 95W.

And PriceWatch has all of the SKUs listed at several different OEMs. So that means basically full retail availability in less than a month. They could not provide that many if they were at 30%. They can only get a maximum of 215 chips per die. That means 1000 wafers for an average of 150,000 at 69% yield. 30% yield means twice as many. They can start about 20K wafers a month. That means Brisbane, Turion and K10 all have to use Fab 36 (Fabtech reported not long ago that Chartered was not receiving orders). The link is dead now unless you register.


www.fabtech.org

 

[yomamafor1]

That's a very two dimensional view on semi-conductors and the business world as a whole.

My favorite overpriced processor launch was the 840 Extreme Edition for a mere $1030! :lol:

Your point?

I do not know why Intel priced their Pentium EE 840 at 1030USD, but simply claiming that Intel's using aggressive pricing by basing on the fact that Core 2 Duo cost less than Pentium D is moot. Core 2 Duo features less transistors than Pentium D, so why should it costs more? (Conroe: 291mil, Presler: 376mil)

AMD is a year late?! Late with what? Did the the SIA establish an schedule to reach specific fabrication processes? I never read that in the newsletter!

AMD is a year late to releasing quad core. Guess what, when AMD's busy trying to get Barcelona to work, Intel has already shipped 1 million quad cores by June. If we divide that with the market share, AMD would've at least sold 250,000 quad cores, and save their bleeding company.

SIA didn't establish the fabrication process schedule, but as AMD's main competitor move to 45nm, AMD would be in a financially disadvantage to not implement 45nm. With die size of 283mm^2, how much more dies can AMD produce
if they move to 45nm? Concerning the current yield of Barcelona, wouldn't it make sense to move to 45nm?

Doug Freedman has already expressed his opinion on this, calling Barcelona a "mismatch with AMD's current 65nm".

An analyst said that AMD's quad-core "Barcelona" Opteron processor is an "architectural mismatch" with the 65-nanometer process it is being built on. "We believe the company's late Barcelona introduction and disappointing early performance are an early indication of a bad marriage of process technology and design that will be hard to fix before a move to 45-nm is required," said analyst Doug Freedman of American Technology Research, in a report cited by several news outlets Friday.

http://www.x86watch.com/articles/amd_barcelona_process_mismatch_120.html

I'm sure he won't be the last one expressing that opinion.

As a head's up, Intel is "still using" x86 too. It's a pretty interesting story on how Intel came into their x86-64 instruction set, you should research that.

Let's see... AMD and Intel both developed their x86-64, but AMD got it working first. Intel then sued AMD for not sharing the technology under their license agreement.

Your point? I never discredited AMD64 for being the first instruction set to support both x86 and x64, but it's too early for its time. Please tell me how many computer users, besides servers, are using x64 programs? How many users actually see improvement in 64bit environment?

I realize you get some perverse enjoyment sticking it to BaronMatrix but it doesn't really make for an interesting thread and certainly has taken this one way off topic.

I'm sorry that you felt that way, but I do this because I hate it when someone just outright spreading FUD, with little to no backups. I was not the one who started to go off topic, but I was the one who responded when he's giving out mis-information.

 

[BaronMatrix]

K10 will definitely perform in scaled environment, but I'm not so sure about IPC though...

http://realworldtech.com/forums/index.cfm?action=detail&id=83485&threadid=83478&roomid=2

Ooops... Barcelona 1.9Ghz perform slightly better in INT than Opteron 2.0, but perform similarly in FP.

So where is your "only 20% faster than opteron clocked 1Ghz higher"?

From The Tech report

Now look closely at how much faster Barcelona is core for core. I tried to pick tests with the same amount of cores.


Let's look at Anand's findings:

Well, it looks to me like K10 is much faster than K8 for server. Maybe I'm reading them wrong. face it. As the clock speeds increase so do the IMC speeds, so the near 100% clock scaling will do wonders at 2.5GHz+ as everyone said. but the fact that you can pop two of those in at 2/3 the clock, nearly double your computational power and use the same power envelope means that Barcelona is already a success.

Phenom with 1066 and 3.4GHz HT3 should cancel all L3 latency or at least significantly lower it.



oh yeah,


ALL HAIL THE DUOPOLY!!!

 

[BaronMatrix]

Your point?

I do not know why Intel priced their Pentium EE 840 at 1030USD, but simply claiming that Intel's using aggressive pricing by basing on the fact that Core 2 Duo cost less than Pentium D is moot. Core 2 Duo features less transistors than Pentium D, so why should it costs more? (Conroe: 291mil, Presler: 376mil)


AMD is a year late to releasing quad core. Guess what, when AMD's busy trying to get Barcelona to work, Intel has already shipped 1 million quad cores by June. If we divide that with the market share, AMD would've at least sold 250,000 quad cores, and save their bleeding company.

SIA didn't establish the fabrication process schedule, but as AMD's main competitor move to 45nm, AMD would be in a financially disadvantage to not implement 45nm. With die size of 283mm^2, how much more dies can AMD produce
if they move to 45nm? Concerning the current yield of Barcelona, wouldn't it make sense to move to 45nm?

Doug Freedman has already expressed his opinion on this, calling Barcelona a "mismatch with AMD's current 65nm".

An analyst said that AMD's quad-core "Barcelona" Opteron processor is an "architectural mismatch" with the 65-nanometer process it is being built on. "We believe the company's late Barcelona introduction and disappointing early performance are an early indication of a bad marriage of process technology and design that will be hard to fix before a move to 45-nm is required," said analyst Doug Freedman of American Technology Research, in a report cited by several news outlets Friday.

http://www.x86watch.com/articles/amd_barcelona_process_mismatch_120.html

I'm sure he won't be the last one expressing that opinion.

As a head's up, Intel is "still using" x86 too. It's a pretty interesting story on how Intel came into their x86-64 instruction set, you should research that.

Let's see... AMD and Intel both developed their x86-64, but AMD got it working first. Intel then sued AMD for not sharing the technology under their license agreement.

Your point? I never discredited AMD64 for being the first instruction set to support both x86 and x64, but it's too early for its time. Please tell me how many computer users, besides servers, are using x64 programs? How many users actually see improvement in 64bit environment?


I'm sorry that you felt that way, but I do this because I hate it when someone just outright spreading FUD, with little to no backups. I was not the one who started to go off topic, but I was the one who responded when he's giving out mis-information.

You're something else. FUD is not being supportive of a company it's negative towards another with false statements. Kind of like your yield statements. Intel DID NOT DEVELOP anything 64bit other than Itanium. EMT64 is just a rebadged AMD64. Microsoft uses X64 for ALL server apps. As of now you can't get 32bit Exchange, SQL or any other MS server apps. Obviously X64 has a myriad of uses.

The version of Windows you're using was built using Opteron. As games become larger and textures more unmanageable you will see a real desktop move to X64 to allow more than 4GB RAM and more than 2GB per process.

I would rather have the chip and OS ready than have the apps ready first. Cost has more to do with number of transistors. C2D should have slotted in at PD prices. PERIOD. It didn't because Intel wanted to raise the barrier to entry.

 

[yomamafor1]

From The Tech report

This benchmark is used to test the total throughput of the server. In other words, this benchmark favors scaling. Not surprised here, Barcelona does indeed scale very good.

Again, this benchmarks favors scaling.

Techreport: This isn't a real-time process, and it doesn't reflect the performance one would experience while playing a game. Instead, it shows how multiple CPU cores can speed up game development.

Again, POV-RAY is known for its near perfect scaling.

Let's see single threaded performance, shall we?

Oops... looks like 2360HE can't even outperform 2218 in single threaded performance.

Again, 2360HE is outperformed by 2220 in single threaded performance. This pretty much concurs with the SPEC_int and SPEC_fp scores IBM posted, that IPC wise, Barcelona falls short of expectation.

Now look closely at how much faster Barcelona is core for core. I tried to pick tests with the same amount of cores.

Please don't tell me you divided the number... please tell me you didn't divide the number...

...and you did...

humm... that shows that Barcelona can definitely perform when scaled.

Exactly! But what about IPC? What about single threaded performance? Single threaded performance is relevant to Phenom's performance number, and with reduced IPC, I won't be surprised if Phenom comes up short.

Let's look at Anand's findings:

This benchmark generates 4 threads of codes to be ran. As you can see here, Barcelona still scales pretty well.

Again, it generates 8 threads, so what we see here is a scaled performance.

As you can see here, Opteron 2350 gains an advantage of scaling by going from single processor to dual processor. It maintained a near perfect scaling. Given that Barcelona's FP unit is twice the bandwidth than K8, Barcelona can now execute 128 bit SSE instruction, which improves FP performance. Aside from that, the newer OoO engine might also played a part in this.

Well, it looks to me like K10 is much faster than K8 for server. Maybe I'm reading them wrong. face it. As the clock speeds increase so do the IMC speeds, so the near 100% clock scaling will do wonders at 2.5GHz+ as everyone said. but the fact that you can pop two of those in at 2/3 the clock, nearly double your computational power and use the same power envelope means that Barcelona is already a success.

A success as in, compared to AMD's older Opteron line, Barcelona is a success in scaling. With Barcelona, AMD can still hold its fortress in the MP area. As SPEC results pointed out, however, Barcelona isn't as successful as AMD has hoped. It will come up short when competing against Xeon in the single to dual socket arena, and definitely so when its competing against Core 2.

Phenom with 1066 and 3.4GHz HT3 should cancel all L3 latency or at least significantly lower it.

No offense, but you can dream on. The L3 latency issue is actually a by-product of split power plane. Since IMC and L3 can be clocked differently from the cores, and its completely shared, there must be something in between L3 and cores to sync up the data.

As David Kanter puts it,

...since the L3 is shared between four different cores, access to the L3 must be arbitrated. A round-robin algorithm is used to give access to one of the four cores each cycle. The latency to the L3 cache has not been disclosed, but it depends on the relative northbridge and core frequencies %u2013 for reasons which we will see later.

http://www.realworldtech.com/page.cfm?ArticleID=RWT051607033728&p=7

Again, you demonstrated no knowledge in bandwidth and latency. Latency and bandwidth are close to each other, but with completely different meaning. By having a wider bus won't always guarantee you a faster bus.

The last line has been used too many time with too little meaning, I just omitted it.

 

[yomamafor1]

You're something else. FUD is not being supportive of a company it's negative towards another with false statements. Kind of like your yield statements. Intel DID NOT DEVELOP anything 64bit other than Itanium. EMT64 is just a rebadged AMD64. Microsoft uses X64 for ALL server apps. As of now you can't get 32bit Exchange, SQL or any other MS server apps. Obviously X64 has a myriad of uses.

Actually, Intel did start working on x86-64 when AMD was working on AMD64, with the codename Yamhill.
http://www.theregister.co.uk/2003/08/04/intel_yamhill_coming/
I'll do a little more research on this. If I'm wrong, I stand corrected.

The version of Windows you're using was built using Opteron. As games become larger and textures more unmanageable you will see a real desktop move to X64 to allow more than 4GB RAM and more than 2GB per process.

You mean Windows XP? The OS I am using now released back in 2001. Opteron released in 2003. Now, are you telling me that AMD has time machine, so they could send those AMD64 capable Opterons to M$ to develop their Windows XP? :lol: :lol: :lol:

As for games, we won't see the need to switch to x64 in a short while. Most current games are built without 64bit support (BF2142 for instance). Assuming that games do need to take advantage of x64 in 2009, AMD64 is early by.... 6 years, for desktop users. As for sever applications, AMD64 indeed helped them in boosting performance without resorting to RISC structure. In a sense, AMD64 did help transform the industry.

Therefore, I was wrong to say that AMD64 was completely ahead of its time. I apologize for it.

I still do not see the need to implement AMD64 in desktop applications though.

I would rather have the chip and OS ready than have the apps ready first. Cost has more to do with number of transistors. C2D should have slotted in at PD prices. PERIOD. It didn't because Intel wanted to raise the barrier to entry.

So by your definition, since FX-62 was at 1000USD a pop, are you telling me that E6700 should cost 1200USD? :lol: :lol: :lol:

Ridiculous.

 

[BaronMatrix]

So by your definition, since FX-62 was at 1000USD a pop, are you telling me that E6700 should cost 1200USD?


That's exactly what I'm saying. There was still a 6300 that could have slotted right above 3800+ in price. Besides, there was a 6800 that would have been above the FX62.

Never in the history of X86 has the new arch come out costing less than the old.

 

[yomamafor1]

Newegg has both 2347 and 8347. 2347 is selling out every few days and they get more every week. They were out last week, got some more and ran out again.

Why would they release the highest clocked chips first? Kuma was said to go from 2.2-2.9GHz. Phenom from 2.2-2.7GHz this year. I have heard no one complaining about availability. Dell is not gong to releae until they have at least enough for 10,000 servers maybe even 30,000. People who buy from Dell, HP, etc buy in quantity.

And the reason why OEM still don't have enough processors several weeks after launch because?

Again, what does "intact" mean? All cores work or work at a given speed?

Intact means, a die with four working cores. The clockspeed is hampered by AMD's 65nm process.

And checking Newegg, they now have 2344HE, 2346HE, 2347 and 8347. that is the sweet spot right now as they will fit all the way up to 2.4GHz K8 Opteron at 95W.

And PriceWatch has all of the SKUs listed at several different OEMs. So that means basically full retail availability in less than a month.

They should! This is the first time I've heard that several weeks after launch, there were still no availability from OEMs.

Oh wait.. quad fx?

They could not provide that many if they were at 30%. They can only get a maximum of 215 chips per die. That means 1000 wafers for an average of 150,000 at 69% yield. 30% yield means twice as many. They can start about 20K wafers a month. That means Brisbane, Turion and K10 all have to use Fab 36 (Fabtech reported not long ago that Chartered was not receiving orders). The link is dead now unless you register.

I know that Chartered is not receiving orders from AMD as of now. I'm not too familiar with the fabrication process. I'll dig a little more to answer this.

 

[yomamafor1]

...and why do us, as consumers, get screwed over because you want Intel to fit nicely into AMD's pricing structure?

 

[NMDante]

So by your definition, since FX-62 was at 1000USD a pop, are you telling me that E6700 should cost 1200USD?


That's exactly what I'm saying. There was still a 6300 that could have slotted right above 3800+ in price. Besides, there was a 6800 that would have been above the FX62.

Never in the history of X86 has the new arch come out costing less than the old.

DELETED Who freaking cares if Intel introduced the C2Ds lower than the PentiumD? Only YOU. Everyone else liked the lower price, but you. AMD didn't have to lower it's prices, but they did - BEFORE C2D was released, and with no new products being released. WHY?!? Why didn't they keep their prices up? I didn't see you complaining when the 6400+ was released at such a low cost. Why not? Why shouldn't it have been priced at $700, and not $350?

Here's a suggestion - Intel makes their C2D CPUs for $100. They price it according to demand, and to help regain market share. Well, that's freaking amazing. They don't need to price what they consider their mainstream CPU at over $1k, just cause you think they should. Is Intel losing money? Are they hurting themselves by pricing their products at the prices they believe is good for them?

Quit crying about the same thing over and over. Who's fault is it your 4400+ isn't worth $250? Intel's? Please. Next time, get a new thing to cry about.

 

[weskurtz81]

NM

 

[chunkymonster]

chunkymonster wrote :That's a very two dimensional view on semi-conductors and the business world as a whole.


My favorite overpriced processor launch was the 840 Extreme Edition for a mere $1030! :lol:


Your point?


I do not know why Intel
priced their Pentium EE 840 at 1030USD, but simply claiming that
Intel's using aggressive pricing by basing on the fact that Core 2 Duo
cost less than Pentium D is moot. Core 2 Duo features less transistors
than Pentium D, so why should it costs more? (Conroe: 291mil, Presler:
376mil)

Actually, I was wondering what your point was with bringing up the launch price of the 3800? But is doesn't change the fact that both AMD and Intel are "guilty" of "overpricing" processors at first launch. I just think the 840 pricing was ridiculous especially considering how quickly it was dropped from the line-up. might as well have

AMD is a year late to releasing quad core.
Guess what, when AMD's busy trying to get Barcelona to work, Intel has
already shipped 1 million quad cores by June. If we divide that with
the market share, AMD would've at least sold 250,000 quad cores, and
save their bleeding company.

SIA didn't establish the
fabrication process schedule, but as AMD's main competitor move to
45nm, AMD would be in a financially disadvantage to not implement 45nm.
With die size of 283mm^2, how much more dies can AMD produce if they move to 45nm? Concerning the current yield of Barcelona, wouldn't it make sense to move to 45nm?

Guess what, Barcelona does work and is available on the retail market. Now it may "not work" like you think it should with the current stepping, you have to give AMD the benefit of the doubt and allow for new steppings. I mean c'mon, C2D went thru at least 3 steppings to reach it's current level of performance. As far as shipping and market share goes, you could use your same logic to say that if Intel used their R&D
dollars to develop an IMC two years ago, they could have shipped more
Xeons by now and never lost all that server market share.

Of course the SIA didn't establish a fab process schedule, that's not the purpose of the SIA, but it does illustrate that there is no set schedule to release a specific fab process. If 45nm was not on the AMD roadmap at all, that comment might hold water;s o going on how "AMD is behind schedule" is really a non-issue.

Let's see... AMD and Intel both developed their x86-64, but AMD got it
working first. Intel then sued AMD for not sharing the technology under
their license agreement.


Your point? I never discredited AMD64 for being the first
instruction set to support both x86 and x64, but it's too early for its
time. Please tell me how many computer users, besides servers, are
using x64 programs? How many users actually see improvement in 64bit
environment?

I'm not really sure where this is going or even the point your trying to make by going on about x86-64 instruction sets. It's almost as if you "blame" AMD for the lack of 64bit applications. The x86-64 implementation allows for complete backwards compatibility with both Intel and AMD processors. Having a x86-64 implementation and 64bit OS'es is what let's workstations and power users break the 4GB RAM barrier. Given that, x86-64 is enough of a reason regardless of how widely accepted it is or how many 64bit apps there are.

I'm sorry that you felt that way, but I do this because I hate it when
someone just outright spreading FUD, with little to no backups. I was
not the one who started to go off topic, but I was the one who
responded when he's giving out mis-information.

Regardless of the reasons, the Baron seems to be everyone's favorite whipping boy. Not for nothing, everyone does their fair share of talking out of their a$s in these forums and mine is the only one so far that honestly says so under their avatar.