Sign in with
Sign up | Sign in
Your question

Wow, was Matisaro right??

Last response: in CPUs
Share
August 21, 2002 9:52:17 PM

Just finished the new article. Seems to me that with the revision, the Athlon is panning out just like Matisaro said it would. I´m amazed really. It should be an interesting argument now between those who thought the core was dead and those who didn´t.

Judging by the level of AMDmemtdown´s newest thread this must be very frustrating.

Raystonn? You out there?

<font color=red>I have a computer and it does weird stuff. please help.</font color=red>

More about : wow matisaro

a b à CPUs
August 21, 2002 11:24:06 PM

No, because Matisario said it would continue to scale at 66MHz for every 100 XP points, and that those 66MHz would continue to equal 100MHz of the P4.

<font color=blue>You're posting in a forum with class. It may be third class, but it's still class!</font color=blue>
August 22, 2002 12:23:40 AM

lol. persistant arent we. Congrats Crashman, You were clearly right in that sense.
Matisaro, congrats on knowing the core wasnt dead.

:smile:

<font color=blue> If it ain't broke, don't fix it...tweak it.</font color=blue>
Related resources
August 22, 2002 12:59:20 AM

He wasn't the only one telling people of a new revision that would scale better and run cooler? Myself and a couple others have been sayng this since the first one's disappointed everyone, or we're just not worthy of any credit? :o P lol, well ok matisairo was right then... hell with it

"Opinions are like assholes, everyone has one"
August 22, 2002 3:17:33 AM

ok, kudos to u as well then.
i never had an opinion cus i never cared about the tbred.
hehe just waiting for what comes next.

<font color=blue> If it ain't broke, don't fix it...tweak it.</font color=blue>
August 22, 2002 3:51:45 AM

Yeah, where is Raystonn now... hiding as usual...

I and Mat were the only dudes actively here being skeptical and claiming that this is not the Tbred's true performance. I dunno if it has to do with the red and green packaging anymore, but the new Tbreds indeed have proven that both our open minds, were the triumphant one over the thousand closed ones here (no offence anybody). There may have been some who also beleive, though like I stated, we were the only ACTIVE ones in the forum.


Oh and DH, welcome back dude! We miss you!

--
Is the opportunity to earn money by working, free?
August 22, 2002 3:58:33 AM

I'm not exactly sure what you're saying here. Are you saying that your blind devotion to something that didn't seem likely was correct? Regardless of whether or not AMD was indeed able to refine their process (which most people agreed with), and the fact that most people did project 2.4-2.6ish GHz for this future enhancements, "believing" in something that had no basis still doesn't justify any gloating. Although I have to say what I've read of the other guy you speak of does sound like a hothead, most of the people in this forum kept their skepticism and they had good reason to. Unless you actually had prior knowledge that AMD was going to move to a 9-layer redesign of the core, and nobody else believed you, there is really nothing to gloat. Just enjoy the fact that we are seeing better technology and that this will surely mean Hammer will be able to scale at least to 3 GHz.

"We are Microsoft, resistance is futile." - Bill Gates, 2015.
August 22, 2002 4:08:18 AM

Ok sorry if I sounded very gloated. What I was saying, is back at the time the XP2200 Tbred came out,(if you were here back then) everyone thought it was the end of the world for AMD. Nobody kept an open mind about the sudden surprise. Nobody could beleive how the thing OCed. Worse, is nobody could possibly perceive that there is something wrong, and instead just told us to "face it"... (for example: The 1.8GHZ Tbred could not OC past 1.9GHZ, and it has to use water cooling to get 100MHZ more. The Palomino at 1.73GHZ could OC at normal cooling, to 1.9GHZ. Yet nobody could see that and have some skepticism)
No I did not know they would have 9-layers, but what I do know, is for starters, Tbred's small core results in weak cooling, and the fact the core design supposedly WAS flawed, and the need for an extra layer was required. The latter, I did not know, but I am mentioning it. Look, if people are to say: Well it need an extra layer to get those MHZ. I'd suggest you people at least consider the dated 10 stages design, over the NetBurst one, before complaining how they needed to resort to 9 layers.

--
Is the opportunity to earn money by working, free?
August 22, 2002 4:12:56 AM

Well, from most of the posts I've read here and in other forums. Most people did expect the new steppings to improve scalability. Asside from a few rabid fanboys of course. I suppose you feel snubbed that not everyone agreed?

"We are Microsoft, resistance is futile." - Bill Gates, 2015.
August 22, 2002 4:41:47 AM

I myself was suprised to see the T-bred hit these speeds. I think a lot of people are. When you see 66mhz increases, and a t-bredA that didnt hit 2ghz you have every right to be skeptic. The other day there was a thread where one of you said ~"I dont remember AMD ever doing anything wrong, or making mistakes".......Well the t-bred-A had PR100+ speed improvement over the pal. Not so good eh? Future didnt look to optimistic. Yeah, you "rooted" for the next stepping/refinement.....I guess it's like when you hit on a horse? Or a roulette ball falls on your number. Because you really were guessing when you speculate on future t-bred performance. No one KNEW for sure.

You talk about this AMD chip like your favorite team won a championship. "D'backs won! In your FACE Yankee fans!"....You got your fanboy flag pretty high tonight.

This sig runs too hot.
August 22, 2002 5:06:07 AM

yes, mat is right, we should worship him more.

<font color=green> there's more to life than increasing its speed -Ghandi</font color=green>
August 22, 2002 6:31:36 AM

Yes I was, and no one can take that away from me.

:wink: The Cash Left In My Pocket,The BEST Benchmark :wink:
August 22, 2002 6:35:12 AM

Actually, crash Is wrong I never said that.


also, I was not right in mistrusting the origional tbred, I was right about predicting a 2.4-2.6 tbred topspeed on .13 because of the shrink gains.

when the first tbred came out and topped at 2ghz(at best) I was flamed and personally attacked, and this redesign proves my origional prediction true and generally sets things right.


PS: imgod2u, Im not really a hothead, I have just been fed up with the trolling and misinformation on the boards as of late, and all the oldschoolers leaving. I really do post alot and try to help as many people as I can.

:wink: The Cash Left In My Pocket,The BEST Benchmark :wink:
<P ID="edit"><FONT SIZE=-1><EM>Edited by Matisaro on 08/22/02 11:29 PM.</EM></FONT></P>
August 22, 2002 7:29:46 AM

vindication: to clear of doubt with supporting arguments or proof; to provide justification or support for.

indeed matisaro... i can only imagine how much grief you took back then. more power to ya.

<font color=green> there's more to life than increasing its speed -Ghandi</font color=green>
August 22, 2002 5:19:08 PM

Quote:
Just finished the new article. Seems to me that with the revision, the Athlon is panning out just like Matisaro said it would. I´m amazed really. It should be an interesting argument now between those who thought the core was dead and those who didn´t.

And it also proves that I was right when I said that AMD seemed to be hiding some sort of problem when they kept taking their sweet time releasing the T-Bred.

When they <i>finally</i> released the T-Bred, it's crappy quality that was barely even good enough to scale the AXP another two stages in itself proved my point. But then to later release a T-Bred core revision that absolutely proves without a doubt that their first T-Bred design was faulty (I.E. too much signal noise to reach a decent speed or OC worth a darn) by fixing those problems just makes me smile that much more.

I was right. :-p AMD <b>was</b> having some sort of a problem with the T-Bred AXP core redesign. And thank goodness they fixed it or else they'd have been kinda screwed. (Or at least until Hammer.)

<pre><A HREF="http://www.nuklearpower.com/comic/186.htm" target="_new"><font color=red>It's all relative...</font color=red></A></pre><p>
August 22, 2002 6:18:58 PM

seeing that you need to dip the cpu in liqued nitrogen to keep it cool i wouldn't call it amazing as of yet. You need a high performanced cooler otherwise your athlon xp 2600+ will burn up and melt. Hey it's right on tomshardware's article.

I would not recommend an AMD athlon XP over 1800+ for reasons of sanity. Those high performance fans are loud and annoying!

<A HREF="http://www.anandtech.com/mysystemrig.html?id=9933" target="_new"> My Rig </A>
August 22, 2002 6:59:52 PM

I admit this is one of AMD's big mistakes, releasing the Tbred A.

I did not root for the next stepping, I clearly stated that there is something wrong in the core itself, and that it has to be rechecked. AMD didn't just go to redesign for Barton, they cared about their current core and redid the adjustment needed. Intel released the 1.13GHZ, unstable, but they didn't try to fix it or anything, they just stayed at 1GHZ until the 0.13m batches started.

Sadly you are indeed an intel fanboy, because you perceive what I say as fanboyism. I had every right to say "IN YOUR FACE". I stated above an example of how people didn't even beleive it was a Tbred core problem, since the Palominos could OC better, yet everyone thought I and Matisaro were just rabid fanboys. We had an open mind, we won finally.

And BTW, when you think about it, if the NW tops out at around 3.4GHZ and does NOT reach 4GHZ before 0.09m, that proves AMD's process yeilds so far is somewhat better than Intel's. P4 has 20 stages, it SHOULD theoretically top out at twice the speed of an AXP in the same process. So if the AXP tops out at 2.5GHZ with 0.13m, and Intel's stops at less than 4GHZ, AMD has the better yeilds IMO, despite the P4 being a huge scalar processor.
May this sound pro-AMD, I don't care, it's the truth, face it!

--
Is the opportunity to earn money by working, free?
August 22, 2002 8:04:00 PM

Either way, it works as advertised now. The Athlon doubt is over. Im going back to the original claims by Matisaro and how he had to eat his shorts when the results of the T-bred A came out.

Im with you on the noise though. Id rather have a slower CPU than the noise :) 

Sorry about the lack of grammer. I cant find all the keys on this flemish keyboard.

<font color=red>I have a computer and it does weird stuff. please help.</font color=red>
August 22, 2002 8:44:51 PM

ya ... AMD needs to work on the heat issue though. Heat and electronics do not mix. I don't care how much you love a company. It's rediculous that you need to purchase a 3000 dollar drum of nitrogen to keep your cpu cool. (thats a joke).

Another thing is that AMD needs to start producing it's own chipsets.

Until then AMD will always be behind Intel. If AMD can solve those two problems then AMD would be more liked as people hate via and since via is AMD's prime source for chipsets they don't buy an AMD system. that was my case when my motherboard split it's guts.

ANd yes the noise is quite irritating. I can finally think while designing my logo, which is done. It's a shame i can't post pictures on this forum.

<A HREF="http://www.anandtech.com/mysystemrig.html?id=9933" target="_new"> My Rig </A>
August 22, 2002 9:22:15 PM

AMD_Man's P4 is as hot as his Tbird man!
Some Xp1600+ systems, like Metal_Rocker's, using a Volcano 7, OCed to XP2000 or I dunno what, run at below 40ºC. Some P4s go up as high as 57ºC. Each chip will perform differently in temperatures, and it depends on how you set it up.

--
And now, an advice from your friendly Nike shoes slogan: JUST DO HER!
August 22, 2002 10:28:22 PM

Just a thing to mention, a doubling of pipeline length does <b>not</b> guarantee a certain amount of increased scalability. It simply allows for easier scaling. Now AMD had to make some pretty drastic moves to get its K7 core up even to 2.5 GHz. The P4 is scaling very easily towards 3 with 6 metal layers of silicon. The problem would actually be power consumption, not electron flow. If you could keep the P4 fed with enough power it will scale well past 3 GHz and onto 4 GHz (look at the LN2 cooling and increased voltage overclocks, those are usually a sign of the end of a CPU's life, they've been 4 GHz on a 6 layer metal design).
To say AMD's manufacturing process is superior is arguable. As 9 metal layer production is significantly more wasteful than 6. However, I will give them this, they managed to pull off an entire core redesign in a very short amount of time. That's something to be said.

"We are Microsoft, resistance is futile." - Bill Gates, 2015.
August 22, 2002 11:29:39 PM

heh

Im running at approximately XP 2000+ speeds.......
With Volcano 7...
right now im idling at 36 celcious under load....not constant 100% though....she wont go over 40 Celcious ever....

<A HREF="http://www.anandtech.com/mysystemrig.html?id=13597" target="_new">-MeTaL RoCkEr</A>
August 23, 2002 1:10:21 AM

"Sadly you are indeed an intel fanboy, because you perceive what I say as fanboyism. I had every right to say "IN YOUR FACE"."

How am I an Intel fanboy? Because I doubt AMD? Jeez, they spend a whole year losing to Intel in performance and OC'ability. I get called a fanboy for supporting, and arguing for the clearly better processor. Oh wait, AXP is a better choice because it's CHEAP! RIGHT!? (keep tellin yerself that).....next it'll be because it's more square. I still see no reason to buy an AXP other than to take a performance hit, and a little less of a hit in the wallet. You OC the highest P4 now, and that's one less trip to the bank later when you would need to upgrade your AXP......

In everyones face eh nostradamus? You predicted the future. You have the same odds when you flip a coin...50/50. Next stepping good or bad? You also couldnt wait for the T-Bred A, cause it was gonna be the P4 Killer......You're 1 for 2. I hope you went double or nothin on this last one. What's your next prediction? Barton will have a 166 FSB? P4's cant scale passed 3ghz? You'd love that.

This sig runs too hot.
August 23, 2002 1:31:17 AM

oh oh.

<font color=green> there's more to life than increasing its speed -Ghandi</font color=green>
August 23, 2002 2:41:46 AM

I won't waste anymore here with you.
Anybody can see you are pro_Intel at about 90% of the time.
If you look around, with no hard searching, you will see I praised Intel more often the last weeks than ever, so I suggest your fanboy insult be retracted.

I did not predict anything, I was looking at logical stuff. Matisaro's a semiconductor expert, he predicted the Tbred SHOULD be able to clock very high for its dated core. Was he wrong? NO. It was an error by AMD, to release a faulty design, one that needed some improvement. Now they released the clear 0.13m Tbred which was supposed to be. So no one was wrong, we just were surprised by Tbred A because under logical circumstances it is not supposed to happen. Again I stated an example, it was clear my prediction or anybody's should've been that Tbred SHOULD'VE scaled higher than that. We were not proved wrong at all, AMD just wasn't smart enough at first and add layers.

You are pro-Intel, you've proven that many times, I can hardly find a partly good AMD comment from you. And I suggest you don't try retaliating by calling me that, before you find out there are at least 5 messages in the past week from me that are Intel praising.

I will not continue talking here anymore, you are close-minded, and I could clearly see it.

PS: BTW if you look well, 55$ for an AXP1600+ which kicks the life off a P4 Celeron at 79$, is pretty sad. Yeah we're cheap, oh yeaa... The AXP 1600+ (my CPU) was able to compete the upper 2GHZ P4s in FPU benchmarks, Anandtech and THG had tests showing that.

--
And now, an advice from your friendly Nike shoes slogan: JUST DO HER!
August 23, 2002 2:49:12 AM

I do agree, and I recall us talking about this a few days ago.
It is of all truth that you need to be skilled and find ways to scale a processor with the new longer pipeline. It would be hypocritical to think just extending it, without any added process technology, such as Intel's new 0.09m tech in the silicon, and some new packaging, would allow twice more speed.
However the P4's current scaling, its OCing, has proven that it should well succeed to reach BIG speeds. Intel has billions into R&D which could translate and make that speed succesful.
Also, there are two extra factors to why the lower number of metal layers are used on P4: Pipeline as discussed, allows easier scaling. AND the fact Intel has all the money to put into allowing the most off a core. They are deeply convinced they can pull it off, and they certainly do want the public to see that 10GHZ one day, I doth my hat to that, because it's possible with them, and I do beleive they will.
What I was expressing to AMD though, was indeed like you and Slvr, they have been amazing with the fact they kept the AXP line running. They constantly allowed a 3 year old core to scale more and more than it should've lasted.

--
And now, an advice from your friendly Nike shoes slogan: JUST DO HER!
August 23, 2002 4:53:04 AM

I wouldn't say the original T-bred's design was faulty. They simply reached an impass. A point where the design reached a dead stop and they had to go back and rework it (pretty drastically may I add) to squeeze some more life out of it. AMD didn't really do anything wrong. They did what was the natural thing to do in any die shrink, shrink the die. The fact that it didn't turn out as all the fanboys expected was not an error, it was simply an impass in the design.

"We are Microsoft, resistance is futile." - Bill Gates, 2015.
August 23, 2002 5:43:25 AM

i also predicted around 2.5ghz for the 0.13 upper limit based on previous 018 to 0.13 dieshrinks... cept everyone forgot :frown: *sniffles*

i also predicted dualchannel (256bit) memory on graphics cards

<b>MegaHertz Matters! ... But not with SDRAM. :cool: </b>
August 23, 2002 6:31:01 AM

Quote:
Either way, it works as advertised now. The Athlon doubt is over. Im going back to the original claims by Matisaro and how he had to eat his shorts when the results of the T-bred A came out.


::coughs up shorts:: much better.

:wink: The Cash Left In My Pocket,The BEST Benchmark :wink:
August 23, 2002 6:33:26 AM

Quote:
ANd yes the noise is quite irritating. I can finally think while designing my logo, which is done. It's a shame i can't post pictures on this forum.


My heatsink cools better than almost anything and is whisper quiet, ask quezacotl hes heard it.

:wink: The Cash Left In My Pocket,The BEST Benchmark :wink:
August 23, 2002 6:35:55 AM

Quote:
i also predicted around 2.5ghz for the 0.13 upper limit based on previous 018 to 0.13 dieshrinks... cept everyone forgot *sniffles*


Yeah, and when the tbred a came out you didnt have threads titled

"tbred sucks suck it down Matisaro" I get the hazard pay on that prediction lol.

:wink: The Cash Left In My Pocket,The BEST Benchmark :wink:
August 23, 2002 7:02:38 AM

yeah.
hazard pay... lol

i remember the number of times i was saying "relax! let the dieshink evolve and mature"
finally we have vindication.
(allthough no appologies lol)

have a nice weekend. its 5pm friday arvo here in ozland... and im GOOOONNNNEEEEE :smile:

<b>Due to Customer Complaints, this sig has been witdrawn from public use. Thankyou. :lol:  </b>
August 23, 2002 7:29:52 PM

Yeah I remember you were a "fighter".

But I don't think you are right on Dual Channel memory. First of all, that is DDR, dual channel is a different thing man!
Unless you can provide some links that state DUAL CHANNEL.
Also you mean 256-bit as in 512bit SDR?

--
And now, an advice from your friendly Nike shoes slogan: JUST DO HER!
August 23, 2002 7:35:01 PM

I'm not so sure about what you're saying. It seems to me that the heat issue is just part of computers nowadays. P4's are on fire as well.

When I built my T-Bird 1000 system I was told it would melt my house and cows would crap in it but it runs in the 30's. It just took a bit of research and money well spent. Total cost for a heatsink: $10 USD. I don't think I did bad at all. Combined with an OEM processor I got the whole deal for $87 USD. Seriously, I just don't buy the heat issue for us enthusiasts. Even when I overclock it doesn't get beyond 45C.

Can't speak on behalf of the new T-Bred B of course, but I'm going to assume that temps won't be an issue if you know what you're doing.

Noise on the other hand is an issue, but only if you're a die hard overclocker. If you don't overclock there are tons of HSF's out there that are silent, cool, and even cheap.

Is it still so? I don't know...

<font color=red>I have a computer and it does weird stuff. please help.</font color=red>
August 23, 2002 7:38:53 PM

Quote:
But I don't think you are right on Dual Channel memory. First of all, that is DDR, dual channel is a different thing man!

why not mean dual channel ddr?

<font color=green> there's more to life than increasing its speed -Ghandi</font color=green>
August 23, 2002 7:40:54 PM

I don't get what you mean?
I know there is DC DDR coming out for P4, but I was asking how come there is DC DDR on video card memory.

--
And now, an advice from your friendly Nike shoes slogan: JUST DO HER!
August 23, 2002 7:42:46 PM

It seems no one even bothered to mention Tbred Bs now run cooler as well.
If it was up to me, I'd tell AMD to release downbinned Tbred Bs at 1.4GHZ or 1.53GHZ. These would overclock like mad to at least 2.2GHZ, and still not go on fire.
The FSB there would be quite high but sure as hell is a nice OC, almost close to the NW 1.6A. I hope AMD thinks about it.

--
And now, an advice from your friendly Nike shoes slogan: JUST DO HER!
August 23, 2002 7:47:06 PM

ah, sorry 'bout that, I see now (pretty much jumped the gun on that one huh)

if dual channel ddr for cpu ups the bandwidth and performance, i think it's natural for video cards to do that as well just cos they can.

<font color=green> there's more to life than increasing its speed -Ghandi</font color=green>
August 23, 2002 7:50:45 PM

Seems to me that the prices on AMD's camp are low enough NOT to warrant a low ball release. Intel had to do it to appeal to overclockers at the time. AMD isn't in that situation.

<font color=red>I have a computer and it does weird stuff. please help.</font color=red>
August 23, 2002 7:52:10 PM

Sorry I didn't get the statement above!

--
And now, an advice from your friendly Nike shoes slogan: JUST DO HER!
August 23, 2002 7:54:22 PM

Of course, but what I wanted to know is, have they been able to design such an architecture?
Remember, video card memory is nothing like motherboard setup. You don't intentionally position them on a card to be able to produce Dual Channel, get what I mean?
So this is why I find it odd that DC can be implemented on video cards.

--
And now, an advice from your friendly Nike shoes slogan: JUST DO HER!
August 23, 2002 8:36:55 PM

Quote:
Just finished the new article. Seems to me that with the revision, the Athlon is panning out just like Matisaro said it would.

Seems to me that the shrinked Palomino core (aka TBred A) couldn't and didn't. AMD realized this, and had to come up with some fundamental changes (record high 9-layer design), in order keep the Athlon in business until Hammer arrives.

That said, I'm also amazed that they could pull this stunt and get the TBred up in the 2.5 - 2.8 GHz range. I NEVER thought that would be possible, and within such a short time. Kudos to AMD's engineering team in Fab 30, Dresden. BUT don't forget, the TBred-B is only at the sampling stage, not mass production. As some have pointed out, this is nothing but a paper launch. Add at least 1-2 month before the chips are widespread available. At that time Intel will be close to releasing it's 3.06GHz parts with hyperthreading.

One more thing; I'm really disappointed in TBred-B's performance at higher frequencies. Even hefty OC'ing near the 2.6GHz mark does not bring the extra performance one would expect. The P4 performance grows a lot more when the frequency dial is turned. This more than anything else shows the age of the Athlon design. It's near the end of the road, even a 166MHz FSB won't do a lot, but probably a larger L2 cache will help in some applications, just as it helped the memory bandwith limited Tualatin. Finally the lack of SSE2 will make this CPU look bad (relatively) as time goes by.

<b>Conclusion:</b> Considering the changes it required to bring the needed speed into the TBred, I wouldn't exactly say that Matisaro was right. You know, if you really want a Ford Lincoln to fly, I guess with the proper modifications it could be done.

<i>/Copenhagen - Clockspeed will make the difference... in the end</i> :cool:
August 23, 2002 9:15:34 PM

First and foremost of all. Asside from Tom producing a boot screenshot of a 2.8 GHz t-bred, no other site has gotten near that. And the 2.6 T-bred needed ice cooling and 2.2 vcore. So let's not exaggerate ppl. With comparative cooling, you could just look at the P4's on vr-zone.com that are at 4 GHz. Nobody mentions them because, well quite simply, it is not a practical overclock.

"We are Microsoft, resistance is futile." - Bill Gates, 2015.
August 23, 2002 10:10:58 PM

Quote:
Seems to me that the shrinked Palomino core (aka TBred A) couldn't and didn't. AMD realized this, and had to come up with some fundamental changes (record high 9-layer design), in order keep the Athlon in business until Hammer arrives.

hi copenhagen. can you or anyone tell me what the extra layer(s) actually did to get better performance for the athlon. what is the physics behind it. i'm sure adding layer is costly because it adds more manufacturing steps so what's the benefit if there is no architecture enhancement?

(i think i have a clue, but i rather hear other people's comments)..

<font color=green> there's more to life than increasing its speed -Ghandi</font color=green>
August 23, 2002 10:13:20 PM

Quote:
First and foremost of all. Asside from Tom producing a boot screenshot of a 2.8 GHz t-bred, no other site has gotten near that. And the 2.6 T-bred needed ice cooling and 2.2 vcore.

i'm totally with ya on this one.. tomshardware is sorta for news, news is sensational, that's how they get the readers in. but that doesn't make it right.

<font color=green> there's more to life than increasing its speed -Ghandi</font color=green>
August 24, 2002 12:03:26 AM

Basically, adding more layers is "maximizing volume". Realize that CPU's are not perfect squares. in fact, they're relatively flat. This means that it takes a shorter distance for electrons to travel from the top of the CPU to the bottom of the CPU as opposed from one end of the CPU to another end of the CPU. Now, adding metal layers while keeping relatively the same transistor count (although the new stepping did bring about 400 thousand more transistors, out of 37.5 million, that's not really a lot) means that you'll be "balancing out" the distance of traveling between layers. This means that more electrons will be traveling between layers and hence, a shorter distance and you will also reduce the area of each layer, meaning that electrons no longer need to travel as far. Basically, the optimal condition would be if you had a perfectly sphere die, in which each transistor node's distance from any other transistor would be more or less the same and there's less of a chance that you'd get a critical path. Of course, such a design wouldn't have a lot of surface area and would be a bitch to cool off. Unless of course, you put it in some type of ultra-heat conductive fluid like the brain is in.

"We are Microsoft, resistance is futile." - Bill Gates, 2015.
August 24, 2002 12:25:43 AM

"Another thing is that AMD needs to start producing it's own chipsets"

They did for a little bit, but they sucked because someone of there got the "brilliant" idea of using a Via southbridge. This caused some huge problems.

Knowledge is the key to understanding
August 24, 2002 3:01:09 AM

does anyone know if the additional layer is a signal, power, or gnd layer? (changed from 8-layers to 9-layers)

i have no idea what the athlon stackup is, but i'm assuming since it's a 8-layer ic, it would be something like this:

1. signal1
2. gnd
3. signal2
4. gnd
5. power (core)
6. signal3
7. gnd
8. signal4

based on the fact that you should have power/gnd planes next to each other for good decoupling and that all signals should be reference to a gnd plane for consistancy. i'm pretty sure the thickness of the ic is still the same, i bet that the core layer had to suffer some trimming for the new layer, but could be other layers as well.

depending on which layer it is, that pretty much explain what i like to know.

<font color=green> there's more to life than increasing its speed -Ghandi</font color=green>
August 24, 2002 3:57:27 AM

If I recall correctly, 2 layers were added. It was originally 7. I've even heard 6 somewhere.

"We are Microsoft, resistance is futile." - Bill Gates, 2015.
!