Is AMD planning a BIG surprise before years end?

[HDTVMan]

So you have the Opterons dissapearing.
You have vendors saying demand is high but supply is low.
You have the rumor of AMD stockpiling chips for Dell.
Add onto that a rumor of early M2 Release.
Also New substrate material announcement. (40% claimed speed improvement)
You have the talks of 65 and 45nm productions but not from AMD facilities.

What is AMD up to? Is this nothing more than not keeping up with demand by diverting attention to something else or is AMD planning some big announcement before years end?

Anyone. Speculation.
Im thinking supply is low because they are ramping up for something else. AMD would like to end 2005 with a Bang announcement going into 2006. Could this be the Dell launch? Could this be an early 65nm launch? Could AMD deliver a quad core and give Paul at Intel a heart attack?

What is everyones thoughts?

 

[davegraham]

M2 is on track for release in Q1. Socket F (or 1270; whatever it's called) is on track for June '06. There haven't been significant speed bumps in the product line, much less a 1.2ghz bump. soooo, i'm going to assume, they're trying to close the year on a strong note and then open the gates wide next year.

 

[HDTVMan]

If they achieved a 40% speed increase like they say with 65nm and the new process that puts them just shy of 4Ghz. Wow would the world take notice is AMD even paper launched they reached 4Ghz before Intel. Especially since the AMD chips already outperform the intel chips at lower speeds. This is interesting to say the least.

Intel stated the 45nm current leakage problem is resolved and that Ghz speeds will increase dramatically when they get to 45nm. Could IBM/AMD have solved it at the 65nm level?

 

[RaynorWolfcastle]

Originally posted by: HDTVMan
If they achieved a 40% speed increase like they say with 65nm and the new process that puts them just shy of 4Ghz. Wow would the world take notice is AMD even paper launched they reached 4Ghz before Intel. Especially since the AMD chips already outperform the intel chips at lower speeds. This is interesting to say the least.

Intel stated the 45nm current leakage problem is resolved and that Ghz speeds will increase dramatically when they get to 45nm. Could IBM/AMD have solved it at the 65nm level?

AMD/IBM have just announced their 65nm process, it'll be a few months before they shrink and validate any CPUs at 65nm.

 

[HDTVMan]

Originally posted by: RaynorWolfcastle

Originally posted by: HDTVMan
If they achieved a 40% speed increase like they say with 65nm and the new process that puts them just shy of 4Ghz. Wow would the world take notice is AMD even paper launched they reached 4Ghz before Intel. Especially since the AMD chips already outperform the intel chips at lower speeds. This is interesting to say the least.

Intel stated the 45nm current leakage problem is resolved and that Ghz speeds will increase dramatically when they get to 45nm. Could IBM/AMD have solved it at the 65nm level?

AMD/IBM have just announced their 65nm process, it'll be a few months before they shrink and validate any CPUs at 65nm.

This is really devestating to Intel as their 65nm process isnt producing higher clock speeds just lower power requirements and reduced heat.

AMD on the other hand is talking a 40% transistor speed increase which widens the gap even further if Intel cant increase the CPU speed and AMD can.

2006 could turn into a bloodbath for Intel on the performance charts. Not that the second half of 2005 hasnt been already but this will really set AMD as a leader.

Lets just hope they live up to the hype and dont take a lets keep 1 step ahead of intel launch schedule. AMD needs to show total dominance overall for once.

 

[coldpower27]

You are aware that the 40% figure comes from comparing the 65nm Silicon Germanium, Dual Stress Liner, Silicon on Insulator against a processor on base 65nm that doesn't contain these 3 technologies. Which is completely hypothetical, as if you compare it to Intels 65nm process which contains Strain Silicon and Silicon Germanium, you will not achieve the 40% gains. Actually when you think about Yonah and it's power requirements you have to see how impressive Intel 65nm process is as they are achieveving these low power processors without the need for SOI to help reduce leakage further.

AMD actually has realrively little on their roadmaps for 2006, basically Socket M2 and 65nm Athlon 64x2 at the latter half of the year, with Socket M2 and 90nm DDR2 Athlon 64x2, Athlon 64 in the first half of 2006, fairly evolutionary nothing revolutionary.

 

[Cooler]

if AMD reaches 4 GHZ next Intel will just die from PR stand point.

 

[Socrilles]

AMD is not just about to release a 4 ghz chip.....................................I mean really, even if they could they would get far more money by releasing the chips in speed steps and just stock piling the best silicone.

 

[Hard Ball]

How about getting to >3GHz first, let not get ahead of ourselves here. I don't think 4GHz is anywhere within the realm of possibility in the near term, since socket M2 chips are said to have higher power consumption than current E stepping chips, even if 65nm arrives early, and has substantial gain in performance, they won't be able to wring that muhc performance out of the hcips while keeping TDP sane at the same time.

That being said, Intel's roadmap

Originally posted by: coldpower27
You are aware that the 40% figure comes from comparing the 65nm Silicon Germanium, Dual Stress Liner, Silicon on Insulator against a processor on base 65nm that doesn't contain these 3 technologies. Which is completely hypothetical, as if you compare it to Intels 65nm process which contains Strain Silicon and Silicon Germanium, you will not achieve the 40% gains. Actually when you think about Yonah and it's power requirements you have to see how impressive Intel 65nm process is as they are achieveving these low power processors without the need for SOI to help reduce leakage further.

AMD actually has realrively little on their roadmaps for 2006, basically Socket M2 and 65nm Athlon 64x2 at the latter half of the year, with Socket M2 and 90nm DDR2 Athlon 64x2, Athlon 64 in the first half of 2006, fairly evolutionary nothing revolutionary.

That being said, I don't think comparing Intel processes and AMD have any validity, since architectural differences also make significant difference in term of power consumption. Since Intel is already moving to 65nm and producing chips, with only marginally better performance; and in the case of Yonah, actually a lower clock speed than what Dothan reached. If there is really 40% increase of transistor performance even on the current 90nm K8 chips, that would vault IBM/AMD chip performance ahead of Merom, and possibly Tigerton. And when AMD's architectural refresh comes (K8L in lat 06 and K9/10 in late 07), there would even be greater problems for Intel, who recently just lost a strong roadmap of Whitefield and derivatives.

 

[BrownTown]

AMDs 65nm transition isn't coming till the second half of 2006 according to AMD themsleves, and the 40% number isn't very likely to translate into anything near 40% faster clockspeeds. Personally I'd say you should take a company talking up their technology that is 6 months away with a grain of salt.

 

[Stoneburner]

So by what date are people expecting the m2 now? I thought it was going to be early q2.

Looks like I may have to delay upgrading again Unfortunate though since I don't think M2 is going to bring much in the way of performance enhancements.

 

[Furen]

Originally posted by: davegraham
M2 is on track for release in Q1. Socket F (or 1270; whatever it's called) is on track for June '06. There haven't been significant speed bumps in the product line, much less a 1.2ghz bump. soooo, i'm going to assume, they're trying to close the year on a strong note and then open the gates wide next year.

You sure about the June release for Socket F? I had heard that it would be released close to the end of 2006...

By the way, a 40% increase in performance would be at transistor level, not necessarily for the whole chip.

 

[coldpower27]

Originally posted by: Hard Ball
How about getting to >3GHz first, let not get ahead of ourselves here. I don't think 4GHz is anywhere within the realm of possibility in the near term, since socket M2 chips are said to have higher power consumption than current E stepping chips, even if 65nm arrives early, and has substantial gain in performance, they won't be able to wring that muhc performance out of the hcips while keeping TDP sane at the same time.

That being said, Intel's roadmap

Originally posted by: coldpower27
You are aware that the 40% figure comes from comparing the 65nm Silicon Germanium, Dual Stress Liner, Silicon on Insulator against a processor on base 65nm that doesn't contain these 3 technologies. Which is completely hypothetical, as if you compare it to Intels 65nm process which contains Strain Silicon and Silicon Germanium, you will not achieve the 40% gains. Actually when you think about Yonah and it's power requirements you have to see how impressive Intel 65nm process is as they are achieveving these low power processors without the need for SOI to help reduce leakage further.

AMD actually has realrively little on their roadmaps for 2006, basically Socket M2 and 65nm Athlon 64x2 at the latter half of the year, with Socket M2 and 90nm DDR2 Athlon 64x2, Athlon 64 in the first half of 2006, fairly evolutionary nothing revolutionary.

That being said, I don't think comparing Intel processes and AMD have any validity, since architectural differences also make significant difference in term of power consumption. Since Intel is already moving to 65nm and producing chips, with only marginally better performance; and in the case of Yonah, actually a lower clock speed than what Dothan reached. If there is really 40% increase of transistor performance even on the current 90nm K8 chips, that would vault IBM/AMD chip performance ahead of Merom, and possibly Tigerton. And when AMD's architectural refresh comes (K8L in lat 06 and K9/10 in late 07), there would even be greater problems for Intel, who recently just lost a strong roadmap of Whitefield and derivatives.

Problem is were talking about the 65nm process here only and not the 90nm process, what is amazing about Yonah is that it debuted at very close to the frequnecies of Dothan of 2.26GHZ vs 2.167GHZ, while making some architectural changes such as shared cache architecture and SSE3, and improved FPU performance. This is on 65nm process that is not quite that mature yet, when AMD debuted their 90nm K8 cores with Winchester clockspeeds were lower then the older 130nm parts at first, it was awhile before the speed ramped up and surpassed the older cores.

It is very doubtful that the 40% increase in transistor performance will traslate directly, to a clockspeed increase of 40% or 40% increased performance. It's too early to make judgements on Merom as we haven't seen performance figures for it yet, as the architectural differences compared to the Pentium M architecture are very great.

It would be interesting to see what AMD has in store for 2006, but I am doubtful if K8L will materialize in 2006.

 

[HDTVMan]

I agree 40% on the overall chip performance is wishfull thinking.

We all have to keep in mind 90nm didnt offer the mhz/ghz increase that was expected like past die shrinks. Intel didnt even see that problem so I have to question their superiority in chip development. This is the main reason AMD is getting the limelight and benefit because if 90nm did scale in performance like previous generations then Intel would probably still lead AMD. Lucky for AMD their design won and Intel mhz/ghz performance came to a screaching halt. That isnt to say that the current leakage problems cant be fixed and were back to mhz/ghz boosts with every die shrink again. Intel has already stated that they wont have the problem at 45nm meaning we will see a significant ghz increase at 45nm from Intel if true. But AMD/IBM seem hell bent on solving the leakage issue and getting the ghz performance solved at 65nm. If they have accomplished this then the 65nm opterons are going to be extreme monsters and 45nm will only really make things interesting as Intel gets their speed boost back where the deeper pipelines will add merit again.

Even if this isnt what is going on I am expecting AMD to announce something big before years end.

 

[BrownTown]

I doubt it, AMD already has new stuff coming out that they have announced, if they do make an announcement it will be something that is at least a year away.

 

[Hard Ball]

Originally posted by: coldpower27

Originally posted by: Hard Ball
How about getting to >3GHz first, let not get ahead of ourselves here. I don't think 4GHz is anywhere within the realm of possibility in the near term, since socket M2 chips are said to have higher power consumption than current E stepping chips, even if 65nm arrives early, and has substantial gain in performance, they won't be able to wring that muhc performance out of the hcips while keeping TDP sane at the same time.

That being said, Intel's roadmap

Originally posted by: coldpower27
You are aware that the 40% figure comes from comparing the 65nm Silicon Germanium, Dual Stress Liner, Silicon on Insulator against a processor on base 65nm that doesn't contain these 3 technologies. Which is completely hypothetical, as if you compare it to Intels 65nm process which contains Strain Silicon and Silicon Germanium, you will not achieve the 40% gains. Actually when you think about Yonah and it's power requirements you have to see how impressive Intel 65nm process is as they are achieveving these low power processors without the need for SOI to help reduce leakage further.

AMD actually has realrively little on their roadmaps for 2006, basically Socket M2 and 65nm Athlon 64x2 at the latter half of the year, with Socket M2 and 90nm DDR2 Athlon 64x2, Athlon 64 in the first half of 2006, fairly evolutionary nothing revolutionary.

That being said, I don't think comparing Intel processes and AMD have any validity, since architectural differences also make significant difference in term of power consumption. Since Intel is already moving to 65nm and producing chips, with only marginally better performance; and in the case of Yonah, actually a lower clock speed than what Dothan reached. If there is really 40% increase of transistor performance even on the current 90nm K8 chips, that would vault IBM/AMD chip performance ahead of Merom, and possibly Tigerton. And when AMD's architectural refresh comes (K8L in lat 06 and K9/10 in late 07), there would even be greater problems for Intel, who recently just lost a strong roadmap of Whitefield and derivatives.

Problem is were talking about the 65nm process here only and not the 90nm process, what is amazing about Yonah is that it debuted at very close to the frequnecies of Dothan of 2.26GHZ vs 2.167GHZ, while making some architectural changes such as shared cache architecture and SSE3, and improved FPU performance. This is on 65nm process that is not quite that mature yet, when AMD debuted their 90nm K8 cores with Winchester clockspeeds were lower then the older 130nm parts at first, it was awhile before the speed ramped up and surpassed the older cores.

It is very doubtful that the 40% increase in transistor performance will traslate directly, to a clockspeed increase of 40% or 40% increased performance. It's too early to make judgements on Merom as we haven't seen performance figures for it yet, as the architectural differences compared to the Pentium M architecture are very great.

It would be interesting to see what AMD has in store for 2006, but I am doubtful if K8L will materialize in 2006.

I don't think too many would disagree with most of your statements. Although the drop in clockspeeds of the Winchester was likely due to capacity constraint, as the conversion in Fab30 took quite a while, while at the same time needed to accomodate the process of tape-out and testing, and OEM sampling of DC processors; whereas intel Fab12 has been changed over, and has had ample time and funds (2Billion) to do so. So Fab12 and D1D should have the same type of capacity constraints that were faced by Fab30 when it had to maintain 130nm process full blast, and making several transitions at the same time. The situation is not comparable here; and whatever deficiencies that 65nm chips that will come from Intel in January will not be due to the lack of capacity, retooling or transition time, but due to the architectural design and the processes themselves.

AMD isn't really going to do much better though in 06 either. They would be lucky to reach >3.5 GHz in 06, since their 65nm process will come at the end of the year, and will not have enough time to ramp up the speeds when 2006 expires, according to a recently updated roadmap.

http://www.mikeshardware.co.uk/RoadmapQ406.htm

Q4 (06)

ATI R600 GPU is expected to be released in Q4, or possibly Q1 2007. The R600 is targeted at the high end market with full support for 'DirectX 10' WGF 2.0, targeting Vista.

AMD Opteron 874 & 872, based on an improved K8 core, are expected to be available in Q4. These processors are expected to released with a clock speed of 2.6Ghz and will be built for the Socket F platform, featuring a 1207 pin interface.

Intel Merom Mobile processor, the successor to Jonah and part of the Santa Rosa platform, is expected to be released in H2 on a 65nm process. Merom is a Dual Core CPU combining the architecture of NetBurst and the Pentium-M to achieve both high performance and lower power consumption. Merom utilises the FSB and EM64T of NetBurst, but is largely based around the Pentium M architecture. The CPU is a 4-issue design (compared to the 3 issue cores of the Athlon 64 and Pentium 4 architectures) with a 14 stage pipeline - significantly shorter than that of NetBurst CPUs (from 20 in Willamette to 31 stages in Prescott). The shorter pipeline will ensure that Merom and it's derivatives will not clock as high as Precott, but it will likely clock as fast or faster than the Athlon 64 - i.e. around 3Ghz. However, the IPC of Merom is likely to be better than the Athlon 64 due to it's 4 issue superscalar design and vastly better than the P4.
Merom will feature 4MB of L2 cache shared between the two cores and will feature a direct L1 to L1 cache transfer system between the L1 caches of each of the cores to improve performance. Merom will also feature a number of enhances prefectching schemes to enhance the use of the caches.

Intel Crestine-GM (Crestline?) chipset for Merom is expected to be released in Late 2006. Crestine-GM is part of the Santa Rosa platform and is expected to support an 800Mhz FSB speed, DDRII 800 SDRAM and PCI Express. Crestine is expected to interface to ICH8-M, featuring support for Serial ATA 300, and the Golan2 / Annadel wireless chipset supporting 80211a/b/g and 11n WiMax.

Intel Conroe desktop CPU is expected to be released in Late 2006. Conroe is a dual core CPU based around the Merom architecture but optimised for the desktop market by removing some of the power constraints from Merom in order to increase performance. Conroe will be available in two forms - with 4MB of shared L2 cache and with 8MB of L2 cache (an 'Extreme Edition' perhaps). Conroe will feature virtualization capabilities, LaGrande technology and 64-bit capability in addition to EDB, EIST and iAMT2.

AMD Athlon 64 X2 (65nm) is expected to be released in Late 2006.

Windows XP SP3 is expected to be released in Late Q4 or Early 2007. Windows XP SP3 is expected to be a fairly significant update containing new features as well as a roundup of security and bug fixes.

But K8L is scheduled to be released with Opteron 872/4, since K9 (some call K10) will be delayed one year to the end of 07. And socket F should arrive in that same timeframe. But by all indications, the real technological improvements, such as HTT3.0, FBDIMM, PCIE on-die controller, wider core will be put off until K9, and K8L will be left with a few incremental updates plus LGA. K9 and FBDIMM should be that in Q4 07, which leave Intel a huge window to make waves, should Merom turns out to be extraordinarily successful. But at this point, Conroe doesn't seem to be doing all that well, given Intel has hardly trumpeted its performance improvements at all, and has focused on its TDP instead.

 

[icarus4586]

Intel is set to be much more competitive in 2006 then they have been recently. Yonah is coming out early in the year. Tt's pretty competitive from a performance standpoint, and when you consider power usage, it really doesn't have a competitor in the x86 market.
Clockspeed is becoming less and less important. If Intel can deliver desktop variants of Yonah with even slightly improved performance, they will be at an advantage even if AMD has better performance. The market is moving more toward mobility; as of May 2005, more PC laptops were sold than desktops. This is an area that AMD really needs to become more competitive in, and an area that Intel is doing very well in. Even if AMD wins from now on in the desktop performance area, they won't be competitive with Intel if they don't improve their mobile line.

 

[dexvx]

Dont you love people with absolutely no background in CE/EE making these assumptions?

 

[Hard Ball]

Originally posted by: dexvx
Dont you love people with absolutely no background in CE/EE making these assumptions?

I think that you really mean ECE/EE, right? CE typically stands for Civil Engineering.

 

[TuxDave]

Originally posted by: Hard Ball

Originally posted by: dexvx
Dont you love people with absolutely no background in CE/EE making these assumptions?

I think that you really mean ECE/EE, right? CE typically stands for Civil Engineering.

Chemical Engineers get no love?

 

[Markfw]

Originally posted by: icarus4586
Intel is set to be much more competitive in 2006 then they have been recently. Yonah is coming out early in the year. Tt's pretty competitive from a performance standpoint, and when you consider power usage, it really doesn't have a competitor in the x86 market.
Clockspeed is becoming less and less important. If Intel can deliver desktop variants of Yonah with even slightly improved performance, they will be at an advantage even if AMD has better performance. The market is moving more toward mobility; as of May 2005, more PC laptops were sold than desktops. This is an area that AMD really needs to become more competitive in, and an area that Intel is doing very well in. Even if AMD wins from now on in the desktop performance area, they won't be competitive with Intel if they don't improve their mobile line.

Have you actually read any good laptop reviews ? Turion runs head to head with the Pentium -M at a less exspensive price point. I say that makes them in the lead. I know this chip has reviews where its equal and some where its not, but its in the ballpark, and it IS less exspensive.

 

[Hard Ball]

Originally posted by: TuxDave

Originally posted by: Hard Ball

Originally posted by: dexvx
Dont you love people with absolutely no background in CE/EE making these assumptions?

I think that you really mean ECE/EE, right? CE typically stands for Civil Engineering.

Chemical Engineers get no love?

Sorry, my sincere apologies,

although I have a roommate who's Chem. E, and he doesn't seem to care what we call him; so we just call him the Flux Capacitor Magician.

 

[glugglug]

Originally posted by: Socrilles
AMD is not just about to release a 4 ghz chip.....................................I mean really, even if they could they would get far more money by releasing the chips in speed steps and just stock piling the best silicone.

Freudian slip?

 

[johnnqq]

i was 99% sure m2 was coming out in q2 or q3...am i wrong??

 

[dexvx]

Originally posted by: Hard Ball

Originally posted by: dexvx
Dont you love people with absolutely no background in CE/EE making these assumptions?

I think that you really mean ECE/EE, right? CE typically stands for Civil Engineering.

Computer Engineering?

I would hestitate to use CS (Computer Science), because of the fact that a lot of low level institutions label their intro-to programming classes as CS. Perhaps a more accurate acronym is CSE (Computer Science and Engineering).