Sign in with
Sign up | Sign in
Your question

Rev G die pic might have given away AMD's Emergency Edition

Last response: in CPUs
Share
May 25, 2006 10:03:44 PM

http://www.dailytech.com/article.aspx?newsid=2489



Look at the die shot. "Higher performance,power efficient 65nm AMD64 processor" and what it is covering. This mixed with anand's quote at the end of his article claiming AMD has 1 more trick up it's sleeve in the high-end.
I would also like to openly state that this is not a fanboy post from either side. Could it be Z-Ram based L3? Please openly discuss.
May 25, 2006 10:11:03 PM

Quote:
http://www.dailytech.com/article.aspx?newsid=2489



Look at the die shot. "Higher performance,power efficient 65nm AMD64 processor" and what it is covering. This mixed with anand's quote at the end of his article claiming AMD has 1 more trick up it's sleeve in the high-end.
I would also like to openly state that this is not a fanboy post from either side. Please openly discuss.


They are implying their .65u process is high performance.
May 25, 2006 10:17:31 PM

And covering up a good section as well. This die shot is interesting though nonetheless because of the asymetrical design.
May 25, 2006 10:18:41 PM

Quote:
And covering up a good section as well. This die shot is interesting though nonetheless because of the asymetrical design.


Copy paste.
May 25, 2006 10:37:52 PM

Quote:
And covering up a good section as well. This die shot is interesting though nonetheless because of the asymetrical design.


Copy paste.
I read the reply's below and saw that mention and thought i'd add it as well. It is not a copy and paste as you claim. Typed it from my computer in front of me just now.
May 25, 2006 11:59:18 PM

AMD doesn't have an Emergency Edition. Their roadmap has not changed. They waited purposely for highe speed DDR2 (Corsair EPP 2-3-2-9). They don't need to be faster when Conroe comes out.

Intel lasted for 3 years with the crappiest platform ever (not a knock). With improvements based on K8L they will win back the desktop ---- as if that matters --- but if anyone saw the last Apache benchmarks they know 40% is not going to help. Everyone (analysts, insiders) knew that AMD could release 65nm this year.

They won't. I remember saying they realize that server share will get them business desktop share since servers are held to a higher standard. I myself don't think they should wait to update the core to K8L but hey one 8xx is up to twice the price of an FX and are sold in at least pairs, so they can make +4x the amount of an FX on each sale.

If you look closely at real world usage in games only FarCry was a clean sweep for Core2. I like X2 because i do a lot of programming on my machine and they are better at compiling.

Anyway, nowadays the video card is more important because Vista is pushing HD gaming so even if AMD releases core updates to FX/X2 it will be sceondary to a DX10 card. Plus AMD is responsible for MS moving ALL ITS APPS to X64 by next year.

EVen Apple used HyperTransport --- emphasis on "USED." AMD will be fine even if they don't overtake Conroe in a few months. Especially since people (Taiwan insiders) are saying they may get "bulldozed" by AMDs new Turion code-named Bulldozer.

Who cares anyway, all of this just keeps Johnny5 with input.
May 26, 2006 12:03:10 AM

Quote:
If you look closely at real world usage in games only FarCry was a clean sweep for Core2.


Divx anyone? Not to mention that the other games were GPU limited.
May 26, 2006 1:16:35 AM

AMD can't release 65nm parts this year because AMD can't make 65nm parts in any sort of reliable quantity. Not because they just decided "we don't need to". that would be kinda stupid meow, wouldn't it?
May 26, 2006 1:37:26 AM

Quote:
If you look closely at real world usage in games only FarCry was a clean sweep for Core2.


Divx anyone? Not to mention that the other games were GPU limited.


I'm trying to think of the last time I needed to encode to DivX. I also don't think I want to turn down my resolution from 1280 to see how fast my proc is at 640 HQ in Quake 4. Let's just agree that everything now is bragging rights except for servers.
May 26, 2006 1:41:15 AM

Quote:
I'm trying to think of the last time I needed to encode to DivX.


Good for you.

Quote:
I also don't think I want to turn down my resolution from 1280 to see how fast my proc is at 640 HQ in Quake 4.


Like I said before, read some CPU benchmarks and see what settings they use!
May 26, 2006 1:46:30 AM

Quote:
AMD can't release 65nm parts this year because AMD can't make 65nm parts in any sort of reliable quantity. Not because they just decided "we don't need to". that would be kinda stupid meow, wouldn't it?


I'm just telling you what has been in the wind. I would think they would wait so that Fab 30 and Chartered coould handle 90nm and Fab 36 would ramp up to do 65nm. Fab 36 has a 30,000 wafer per month capacity, so why rush it. They have until JULY before Core 2 release on the desktop and even then Intel can't stop selling P4s unles they want to throw away Dell's 3rd quarter ( they do have stock). I don't know what their yields are, but I also don't know what Conroe's yields are and Rahul Sood voodoopc.blogspot.com seems to think they may not be great.

The point is that both of these companies will be fine no matter what the consensus on Tom's is.

Again, there is no emergency chip. Dell has already paid for LOTS of 8xx chips at 1400-$2400. If these same chips were rebranded as FX62 they would lose 400-$1400 per chip.

AMDs bread is buttered on the 4Way side. They will keep the push up to get more server share. That will get more business deals than a chip that gets 4 seconds on SuperPi at 1M.
May 26, 2006 2:03:04 AM

Its either the new GDDR3 memory controller or P RAM which is used to speed up pr0n downloads. :wink:

Seriously though I doubt its anything good.
May 26, 2006 2:14:34 AM

Quote:
And covering up a good section as well. This die shot is interesting though nonetheless because of the asymetrical design.


Copy paste.
I read the reply's below and saw that mention and thought i'd add it as well. It is not a copy and paste as you claim. Typed it from my computer in front of me just now.

K.
May 26, 2006 2:20:01 AM

Quote:
Its either the new GDDR3 memory controller or P RAM which is used to speed up pr0n downloads. :wink:

Seriously though I doubt its anything good.
The faster I can get pr0n, the better. I'm an AMD boy now :lol:  .
May 26, 2006 2:59:30 AM

If such a thing existed, I'm sure AMD would be the one to give it to us :lol: 
May 26, 2006 3:57:32 AM

Quote:
And covering up a good section as well. This die shot is interesting though nonetheless because of the asymetrical design.

It's "asymetrical" because that's just a single-core.
May 26, 2006 4:47:48 AM

Quote:
I'm trying to think of the last time I needed to encode to DivX. I also don't think I want to turn down my resolution from 1280 to see how fast my proc is at 640 HQ in Quake 4. Let's just agree that everything now is bragging rights except for servers.
The difference between 140 fps and 170 fps a game may be just bragging rights, but the difference between a 3 hr encode and a 2 hr encode is not only bragging rights, it's productivity.
May 26, 2006 4:59:21 AM

Quote:
And covering up a good section as well. This die shot is interesting though nonetheless because of the asymetrical design.

It's "asymetrical" because that's just a single-core.

Are you sure, I am not so sure.... it could be a real device or it could be a test chip -- if it is a test chip then it would make sense as many companies will put different circuits, packing densities in different quadrants in order to test all aspect of different products as they develop a process.

Hmmmmmmm.....
It's very similar to previous single-core die images. For example:

http://sandpile.org/impl/pics/amd/k8/die_90nm_1024k_sc....
http://sandpile.org/impl/pics/amd/k8/die_90nm_512k_sc.j...
May 26, 2006 5:07:52 AM

I agree with this assessment.
May 26, 2006 5:18:13 AM

Any thought on what AMD's supposed secret weapon is? The inq reported 3.4ghz by the end of the year which would make sense after they ramp 65nm but i just cant see it happening. Its has to be either more cache somewhere or a suprise QC chip that'll fit into AM2.
May 26, 2006 5:21:49 AM

I can't see it being clocked too much higher then it already is, its probably an increase in FSB and DDR2 1066 support, they'd be pretty solid.
May 26, 2006 5:24:19 AM

maybe Brisbane
May 26, 2006 8:34:49 AM

I speculate the "high end secret weapon" might be a 65nm Opteron with a K8 core and an enhanced SSE execution unit, like K8L, but without the other architectural enhancements.
This would at least provide a decent boost for workstation-like applications, as this kind of RISC-like, straight flowing code does not need aggressive HW scheduling/reordering, nor is affected by the issue rate (you'd issue and execute just one 128bit SSE istructions per clock anyway).
May 26, 2006 9:09:06 AM

I'm with you.
Z-ram wouldn't have taped out yet, and AMD has been working on thier SSE performance anyhow.
It would fit, and have a marked result on benchmarks.
May 26, 2006 10:53:21 AM

Quote:
Its either the new GDDR3 memory controller or P RAM which is used to speed up pr0n downloads. :wink:

Seriously though I doubt its anything good.


P RAM? well anything that helps me download pron faster is aces high in my book.....AMD ...keep up the good work
May 26, 2006 3:36:06 PM

it has to be either Sparta or Brisbane... and i think its Sparta...
June 5, 2006 2:53:41 AM

Take a closer look. Rev G is not K-8. It's not K-8l. Looks to be badassed. AMD has 65nm working. Check the analyst day slides. They have the most efficient fab in the world to. At the presentation, they had 65nm machines running outside. AMD is getting the most out of the vanilla k-8. It's called maximizing ROI (Return on Investment).

http://129.15.202.185/athlon_rev_g/wtf_mates.html

marty seyer and all his sexyness:
http://phx.corporate-ir.net/phoenix.zhtml?p=irol-eventD...

slidshow:
http://epscontest.com/presentations/05q2_analyst-day.ht...


investor >> techie
June 5, 2006 3:11:20 AM

The move to dual-core is trivial. Developing a single core makes sense because of this.

It takes patience, but about half-way through the presentation, AMD talks about the fab benchmark and how their fab 30 has been rated the best performing fab six years straight. All fabs submit thier numbers to a third-party who sends them a chart showing where they stand in the industry. Competitor's fabs are not identified, so confidentiallity is preserved. Intel's economies of scale are a myth.

Intel uses the 'copy exact' methodology; essentially, perfect production in one location and then minimize variables at other locations. AMD uses APM (automated precision manufacturing) methodology. Essentially, the fab collects data on yields and all variables involved and then converges on an optimal solution. Think,...Acovea compiler technology. The genetic algorithm is easier than brute force and better than a straight-forward solution for problems with large numbers of interdependencies.
June 5, 2006 3:45:57 AM

Quote:
It could be Z-RAM, but I don't think Z-RAM is the ace up their sleeve -- perhaps, perhaps not -- I would not attempt to guess.


I find that approach refreshing!
June 5, 2006 3:50:28 AM

A certain XIP @ xtremesystem's hinted that K8L might be a helluva lot earlier than expected and AMD is doing everything in their power to get 65nm ramped with the new chip in time for Conroe. It honestly wouldnt suprise me but it's all taken with a fine grain of salt until I see it.
June 5, 2006 4:28:33 AM

I may be pessimistic, but they chances of AMD having 65nm chips for the Conroe launch are extremely low. Even the latest information out of the June 1 analyst meeting indicates that 65nm production shipments won't begin until Q4 2007. This would be inline with the stockup period needed before the planned December launch. Fab36 isn't scheduled to have crossover between 90nm production and 65nm production until Q1 2007. It should also be noted that even 90nm production shipments didn't begin until Q2 2006 which was probably sometime early last month. The fact that 90nm production didn't begin until recently makes it unlikely that 65nm production (not talking about development) would have begun yet for anything in Conroe's launch window. Fab36 is scheduled to reach full 65nm production in July 2007, a full year after Conroe launchs.

http://www.thewalrus.org/amdslides.html

I'm still undecided on whether Rev G really includes the features in the links knapper provided. Adding an extra decoder and new out of order buffers is no small modification. If those were actually part of the Rev G parts that are shipping in December then you'd think AMD would capitalize on the additional performance and not still market the 2.2GHz X2 with 512kB of cache per core as the 4200+. Keeping the same performance rating at the same clock speeds indicates their is no performance improvement which isn't consistant with the extra features pointed out. The only explanation is that either the initial 65nm chips are not Rev G, but Rev F+ or something similar or that die image is either fake or not Rev G. In all likelihood, its that latter with the die image likely being a Rev H K8L prototype, possibly a Deerhound test component.

http://www.hkepc.com/bbs/itnews.php?tid=608673
June 5, 2006 4:43:28 AM

I'm not saying I believe it completely but in order for AMD to continue their growth and secure more Dell shipments they need to come out with something big. The silicon is almost certainly already in the testing labs hidden away being tweaked for efficiency and clockspeed. Intel needs to be open to lure people back with their next gen. AMD is still the market leader and may still have a hell of a hand up their sleeves. Like I've always said, K8 is OLD now and I'm sure they know what they are doing with this new core. Just a matter of getting it out the door.
June 5, 2006 4:50:59 AM

Can some one get me a high res shot of that die ? say something in the 80MB's range.... I would love to see some of the finer details (even if 99% is very repetitive....)
June 5, 2006 5:04:41 AM

I'm unsure about AMD in the short term, but I'm definitely thinking that AMD has something massive up their sleeve with K10. It was originally scheduled for 2007, then they delayed that and put K8L to substitute. K8L is obviously evolutionary rather than revolutionary so I'm not expecting crazy things. It may be better than Core, but it won't be huge. K10 however, has been in the works for a long time, and K8L has bought it an additional 2-3 years. I'm looking at an early 2009 launch, right after Nehalem launches in H2 2008. It's clearly going to be designed from the ground up. We've heard rumours about reverse hyperthreading, but I don't think that's it. Seeing how we're looking at multithreading, and personally I'm hoping for 3D gaming (holodecks please), incorporation of some cell processors or other specialized coprocessors are likely. These can then be further expanded through the HTX slots.
June 5, 2006 5:05:24 AM

Bottom line, AMD is the best semiconductor manufacterer on the planet.

Quote:
Unfortunately, I don't think these are questions we will be able to answer because neither company divulges yield info.


As I said, they all give data to Sematech and Sematech tells them whether or not to fire all their production engineers. Each company gets a letter. They get the graph. They know their letter. They do not know anyone else's. Read more on the benchmark.

AMD is the best:
http://epscontest2.home.comcast.net/presentations/06q2_...

AMD ramp times getting better and better. 65nm ramp should likewise take very little time, and they've had 65nm SRAM's working for over six months:
http://epscontest2.home.comcast.net/presentations/05q2_...

AMD ramped fab 30 way beyond it's original 15k (wafer starts per month) capactiy and was running at 30k by Q1, increasing from 150% they were already running at:
http://epscontest2.home.comcast.net/presentations/06q2_...

AMD moves wafers through the fab generally at the fastest rate in the industry:
http://epscontest2.home.comcast.net/presentations/05q2_...

Over 400 things in APM that Intel does not do:
http://epscontest2.home.comcast.net/presentations/05q2_...

Intel pays the most for AMAT's most advanced parts. This is why they lead process shrinks (geometry only). They are behind the curve in virtually everything else though. They do not do dual-strees liner, SOI, and now they are even balking at the immersion lithography technology and once more proving they're arrogant beyond comprehension. Not invented here. Intel has a history of getting horrible results out of their process shrinks and taking absolutely forever to tweak their processes.

The reason you never saw a 10Ghz Tejas is because Intel never foresaw the end of process shrinks yielding free performance benefits. Process geometries only increase the ammount of chips/wafer lately. It takes other innovations to actually increase the device speed/power characteristics. High outputs of chips/wafer let Intel build up 3.4billion (one entire quarter) in inventory. High performance/power characteristics is part of what allowed AMD to help Intel reach this milestone of mediocrity. There's your best indicator of whether or not AMD is a better manufacturer.

I wasn't annoyed. I am now. Not because you may disagree, but because you expect me to keep spoon feeding you this stuff. Watch the presentation. Read the slides. Then do some of your own D&D and quit expecting me to donate my time to discuss topics you are unwilling to study yourself, even after I have provided you with the resources.
June 5, 2006 5:32:43 AM

Quote:
I'm still undecided on whether Rev G really includes the features in the links knapper provided.


Rightfully so. I've changed my mind about the SRAM (they look like SRAM) banks above the data L1 cache. They look to be rearrangements from the rev-F. Two half-sized peices where there was one long peice. Yes, it's not certain, but that's best I can do for rampant speculation ;) 

Also, it's unclear whether or not AMD has OOO load/store capabilities in mind for K-8l. Maybe just OOO loads, meaning that loads will be re-arranged to reflect OOO instruction processing. As for the performance advantage, OOO load/store and extra decoder were the primary enhancements to Core 2. Check out the synthetic scores though:


from: http://techreport.com/etc/2006q2/woodcrest/index.x?pg=8

Looks like Intel's new core is still having trouble achieving it's maximum throughput due to defficiencies in its memory infrastructure. Woodcrest is way better than Sossoman, but the synthetics are just out of this world. Because of this fact, I expect K8 will scale even better than Core 2 relative to their respective predecessors when the same enhancements are made.
June 5, 2006 5:56:13 AM

I have a question. So, AMD fans seem to enjoy mocking Intel's "Copy Exactly" program. Which is understandable, since they had nothing to copy too. Will AMD's new fab's be producing the same materials? If so, I bet they jump on the CE bandwagon real quick. I will be shocked if they do not use Copy Exactly.
June 5, 2006 6:05:38 AM

Quote:
Bottom line, AMD is the best semiconductor manufacterer on the planet.

Unfortunately, I don't think these are questions we will be able to answer because neither company divulges yield info.


As I said, they all give data to Sematech and Sematech tells them whether or not to fire all their production engineers. Each company gets a letter. They get the graph. They know their letter. They do not know anyone else's. Read more on the benchmark.

AMD is the best:
http://epscontest2.home.comcast.net/presentations/06q2_...

AMD ramp times getting better and better. 65nm ramp should likewise take very little time, and they've had 65nm SRAM's working for over six months:
http://epscontest2.home.comcast.net/presentations/05q2_...

AMD ramped fab 30 way beyond it's original 15k (wafer starts per month) capactiy and was running at 30k by Q1, increasing from 150% they were already running at:
http://epscontest2.home.comcast.net/presentations/06q2_...

AMD moves wafers through the fab generally at the fastest rate in the industry:
http://epscontest2.home.comcast.net/presentations/05q2_...

Over 400 things in APM that Intel does not do:
http://epscontest2.home.comcast.net/presentations/05q2_...

Intel pays the most for AMAT's most advanced parts. This is why they lead process shrinks (geometry only). They are behind the curve in virtually everything else though. They do not do dual-strees liner, SOI, and now they are even balking at the immersion lithography technology and once more proving they're arrogant beyond comprehension. Not invented here. Intel has a history of getting horrible results out of their process shrinks and taking absolutely forever to tweak their processes.

The reason you never saw a 10Ghz Tejas is because Intel never foresaw the end of process shrinks yielding free performance benefits. Process geometries only increase the ammount of chips/wafer lately. It takes other innovations to actually increase the device speed/power characteristics. High outputs of chips/wafer let Intel build up 3.4billion (one entire quarter) in inventory. High performance/power characteristics is part of what allowed AMD to help Intel reach this milestone of mediocrity. There's your best indicator of whether or not AMD is a better manufacturer.

I wasn't annoyed. I am now. Not because you may disagree, but because you expect me to keep spoon feeding you this stuff. Watch the presentation. Read the slides. Then do some of your own D&D and quit expecting me to donate my time to discuss topics you are unwilling to study yourself, even after I have provided you with the resources.

Is it you SexBomb? 8O
Anyhow, nice post.
June 5, 2006 6:12:07 AM

You still pretend sexbomb is someone else. :lol:  :lol: 
June 5, 2006 6:19:38 AM

What makes you think that?
June 5, 2006 6:31:09 AM

Intel is the legacy chip bellweather. In spite of the fact that their stock is the worst performing Dow component over the last year, (lower than even GM last I checked) there is still about 50billion in market capital that has to flow out before institutional investors will have no further reason to continue propping them up. That's just a nasty reality of the market. It affects what analysts will say about the stock. Especially over the last two years, there has been a lot of harping on Intel even while unloading shares. Check out Bank of America's holdings in particular in relation to their interest in Intel. It's sick. That's what's behind the buzz. Another sad reality is that the market is simply not efficient in getting data out. I read an article today from a sideline analyst (not buy or sell-side...basically a nobody) who is still under the impression that AMD produces inferior chips, but sells them at a discount. How 2002 :)  The worst part for me is I actually have to invest in the midst of all this rediculousness. It affects me materially, so I tend to take it very personally. Especially being down 30% off the high this year.

As far as the process shrinks, like I said. Intel essentially drives the AMAT tool production cycles. Other chip manufacturers just wait until AMAT goes into mass production. This gives Intel the smallest processes, but not necissarily the best devices. Whether or not Intel manufactures a better transistor is arguable. What allowed Intel to achieve higher clock speeds with the Net-burst was architectual: deep pipeline. By haveing more steps, the device sizes are broken up into more pieces, and therefore stabalize faster, enabling higher clocks. (I'm assuming you're referring to Intel's higher clocks. Let me know if you're talking about something else.) Intel has recently delayed their 45nm plans a bit and AMD has bumped theirs up, closing the projected gap somewhat.

Intel is ramping their third factory. Only two are currently 65nm. I'm betting on worst case. Intel will have all three of their newest fabs at 65nm, mature, and will hit their 70% woodcrest milestone.

The revenue shipment date discrepency bothers me. I think I may have overlooked it. I'll look into it further, but I suspect it's going to take some resourcefulness to get any answers.

Back to the slides, it was analyst day. Gleaning the important information off of them is somewhat difficult since AMD wanted to highlight specific concepts and demonstrate them very clearly to garner a closer look (guess it does sound like a horrible excuse). Numbers all boil down to earnings, where AMD has been doing spectacularly. If you still want the numbers... ;) 

Wafer output:
-AMD produced 30k wafers Q1 '06. It was originally desined for only 20kwspm. I did not have that resource Q1 '05. In terms of raw capacity, Fab 36 is designed for 20k 300mm wspm, dwarfing fab 30 in terms of silicon area. Important part is AMD squeezed far more silicon through fab 30 than it was designed for. True, it was necessity, but it's still impressive.
Time to mature yield:
-Yield was "mature" for each process in a shorter time interval. Extrapolate the dots to see that they're getting better at being the best.
Production Cycle:
-The number of days/masking layer was simply the best in the industry except for one quarter's data. Whatever that may have been, it couldn't have been much better.
APM slide:
-Intel could have reverse engineered everything and implemented them in secret, but even if they did, they still don't set the bar on the Sematech benchmark.
June 5, 2006 6:32:15 AM

He did pull a lot of stuff from forums and the like before. Besides they're just powerpoint slides mainly. I also noticed that after 9-inch left he signed in and later. The bold statements are also very like him too.
June 5, 2006 6:48:11 AM

I can point out a definite motive for the change. Stating that revenue starts later lowers depreciation on the earnings report, raising profit. This is contrary to what most people try to do with private financials, where showing the greatest loss as early as possible is desirable to lower tax rates. ...anyway, I'm still looking into it.

Nothing can compare to Paul Otellini stating in the Q1 conference call that they "might have lost" a little more market share than they thought in Q4 '05 and "maybe looking back into Q3." The reason that's a golden moment to me is it makes me believe that Intel has potential accounting glitches that will someday send AMD to $100/share within days. When your chief competitor Enron's, lift-off. I hope for it every day I get up.
June 5, 2006 6:53:07 AM

Trust me, it will be a while before any fab's are actually "Light's out". It just isn't practical at this point. Close but still, it will be a while.
June 5, 2006 7:05:37 AM

Yeah, it is a worrysome term for many people. If they were somehow able to achieve this, there would be way fewer jobs. They try to push the idea of a Self Sustaining Technician (a tech that is able to run wip and maintain tools), but the sad reality is that many people simply run wip. The automation is nice, but thankfully it is not at the level that anyone needs to worry about their job.... yet.
June 5, 2006 7:29:15 AM

You caught me. Not that I was trying to hide it. WIP (Work In Progress). I work at D1C.

I find these forums pretty fascinating. Lots of mis-information. Lots of loyal fans on both sides. I don't consider myself a fan-boy, but I am loyal to Intel, since my family depends on Intel's succes and all.
June 5, 2006 7:33:22 AM

Quote:
I work at D1C.

I don't consider myself a fan-boy, but I am loyal to Intel, since my family depends on Intel's succes and all.

He's a "Paycheck" Fanboy.....if you will. :wink:
June 5, 2006 7:38:07 AM

Quote:
Unlike AMD, Intel simply does not dish it out to the masses in eye popping "Look How Great We Are" headlines.


I remember the Athlon XP days and all the forward looking K8 hooplah. I have had to wade through this stuff for several years. I actually dumped AMD at $20 after they failed to gain enough momentum in my mind (even after going black and making me a killing, I was completely disguested with them.) to live up to the hype with Opteron during Q4 '04. I thought I was done with them, but then they filed that huge lawsuit and NetBurst started to hit the wall (definitely a lot more to do with the latter). Now, you can't tell me that Intel hasn't been up to more than their fair share of hooplah for the last year either. Net-burst has been getting pummeled in the server space and is causing Intel's ASP's to crumble. Intel seems to be doing anything and everything in order to hold onto their customers lately.

As for APM, the best bits are not the little robots tracking around the cieling on their own volition, but that the fab tweaks itself based on production data gathered at the different steps. AMD can change product mix, critical dimensions, and bin targets almost instantly. That's power through control.

Believe it or not...well, believe it; fab technology isn't all that important to me. The biggest thing I pay attention to is making sure AMD has room to grow, which they do. Other than that, I try to determine how quickly Intel can ramp good low-power Core 2's. I expect, based on Intel missing their production targets lately for bin splits, they will underwhelm ever so slightly.

Other than that, I plan on seeing 3.33Ghz Woodcrest at year end, side by side with 65nm Opterons, Rev-G or not. I expect AMD to have the edge in mobile and server while Intel will have a lead in the single processor desktop space. Strange people who spend too much money on systems, yet seem to have an overall small impact on the marketplace, may opt for 4x4, but it's not that important to me. I'm very certain I will be all in on AMD during 2H '08. K-8l will trounce everything. Clovertown might as well be marketed as "Return of the Paxville: Disaster Strikes Where Mistakes are Ignored!"

I don't have room for bias. I would be poor. I've already been losing my ass this year. That's a clear indication that being wrong about what the street will think or how the market will move has definite consequences that necessitate me always putting my money where my mouth is. I simply cannot afford to harp on AMD without good reason...or bad reason for that matter.

Now, for something I'm interested in, I'd really like to get back to this Rev-G myths. Does anyone have any likely places to look? I know no one who can tell me anything as to whether it will be in the first mix of 65nm parts in Q4. My nose tells me AMD has clear reason to have been sandbagging PR at this point. It's important because it will determine whether I try to go all in on AMD for Q4. The second Intel starts showing marketshare gains, AMD will trade at a discount. I'm sure it's important to enthusiasts to, since faster chips are always better.

Anyway, I'm tired, and my writing is going downhill. Buenes noches.

(edit) I'm suffering amnesia. It was during Q4 '04 that AMD blew me away with their inability to move a clearly superior product.
June 5, 2006 7:38:22 AM

Quote:
I work at D1C.

I don't consider myself a fan-boy, but I am loyal to Intel, since my family depends on Intel's succes and all.

He's a "Paycheck" Fanboy.....if you will. :wink:


No doubt about it :) 
June 5, 2006 7:50:49 AM

Quote:
Yeah, it is a worrysome term for many people. If they were somehow able to achieve this, there would be way fewer jobs. They try to push the idea of a Self Sustaining Technician (a tech that is able to run wip and maintain tools), but the sad reality is that many people simply run wip. The automation is nice, but thankfully it is not at the level that anyone needs to worry about their job.... yet.





While I love automation, people need jobs so we need to strike a balance between the two.

Semper Fi Carry^H^H^H^H^H Linux on :-D
!