Intel users - all ignorant and uninformed!

Page 5 - Seeking answers? Join the Tom's Hardware community: where nearly two million members share solutions and discuss the latest tech.
Toggle sidebar Toggle sidebar
Oops, sorry I just skimmed your reply so i didn't really get all of it. I understand now. You're pretty knowledgable eh? We need some more people like you here, help us noobs out a bit. Thanks
 
I'm not coming from a fab-background; I'm working on semiconductor fundamental research. The stuff we have available is probably much simpler, cheaper and less advanced than what Intel and AMD, etc. have available.

However, we DO use an ECR (electron cyclotron resonance) to etch Gallium Arsenide (GaAs), and this is: "a device used to generate high density plasmas...". A plasma is a hot, ionized gas, but can be used to etch semi-conductors. I think (don't quite remember...) that the plasma is positioned somewhat ABOVE the surface, and accelerates the clusters of atoms that actually crash into the surface and etch it. This would decrease the temperature a great deal. As far as I remember, the wafers don't exceed a couple of 100 degrees C. Of course, if you go much above 3-400C, you'd melt any surface structure on your wafer, and the polymeric resist probably wouldn't be too happy, either.

I know for sure that plasmas can be used in many aspects of semiconductor growth and etching, for example in order to grow dilute nitrides, one uses an RF plasma source that nicely enables concentration of nitrogen in GaNAs of up to a few percent.

:lol: <b><font color=blue>gnintsakgnirkskir ksron</font color=blue></b> :lol:
 
Yes, you are right; I left out doping, and a hundred other concerns in semiconductor growth issues.

What kind of defects were you looking for? How can you see these in an SEM? The resolution of a really good state-of-the-art can't be more than some 50nm?

:lol: <b><font color=blue>gnintsakgnirkskir ksron</font color=blue></b> :lol:
 
<blockquote><font size=1>Svar på:</font><hr><p>The p4 willamette topped out at 2.2ghz...<p><hr></blockquote><p>I believe there's no such thing as a 2.2GHz Willamette.


<blockquote><font size=1>Svar på:</font><hr><p>
now intel is NOT doing anything different or magical over amd, their .13 shrink has no reason to work better. <p><hr></blockquote><p>Yes, actually I think so. The P4 was designed for extreme high speed, while Athlon was not. I think this fact is also what is giving AMD problems with the performance of Throughbred. How well they succeed in the "re-design" or whatever they're doing, time will tell.


<i>/Copenhagen</i>

<b><i>Seagate Barracuda IV.
Bad performance in RAID setups!
</i></b>
 
Yeah i think this is true. The P4s have a lower IPC than the AXPs which would make them less sophisticated, which would in turn let them run up to higher speeds. Although the overall performance isn't as good per clock, they run faster, which kinda negates it a bit.
 
Interesting point that you bring up here: you can't make the chips too small, 'cause then you can't connect it to the outside world.

Makes me think of what might happen when your features (130nm, etc.) approaches the size of the individual atoms; this must surely be the limitation of semiconductor technology, no? What'll be next? Quantum computing (could be semiconductor, I suppose)? DNA computers?

I've heard that human brain neurons are actually quite slow, so the fact that we can process as much information as we can is actually quite a feat! We might not have the raw computing power of a CPU, but our ability to see the connection between different kinds of information, associate, etc. far exceeds that which any computer can do these days.

I'm getting flaky here :smile:

:lol: <b><font color=blue>gnintsakgnirkskir ksron</font color=blue></b> :lol:
 
The P4s have a lower IPC than the AXPs which would make them less sophisticated, which would in turn let them run up to higher speeds. Although the overall performance isn't as good per clock, they run faster, which kinda negates it a bit.
Click to expand...

Think about what you're saying here. P4 less sophisticated because it has a lower IPC?

I guess that makes hydrogen powered cars easier to build because the engine isn't as powerful?

<font color=blue>If you don't buy Windows, then the terrorists have already won!</font color=blue> - Microsoft
 
But by having less IPC, that could make their chips less sophisticated. It's like a V8 engine and a V4 engine. The V4 one could have more horsepower than the V8 but wouldn't be as powerful and would be easier to manufacture.
 
lol, the debate is heating :smile:

actually, the power of the hydrogen powered cars aren't that far off the mark now; they'll probably equal gas-powered cars within a few years, assuming that research is continued (which it is). the problem is that each "power unit" is too heavy, and you need to many of them to get a decent amount of power. however, as the technology is improving, the weight will go down and this shouldn't be an issue anymore.

it's remarkable that we're still using combustion engines, over 100 years after they were invented! we've got to move on...

:lol: <b><font color=blue>gnintsakgnirkskir ksron</font color=blue></b> :lol:
 
But by having less IPC, that could make their chips less sophisticated.
Click to expand...

IPC has just as much to do with the core's sophistication as clockspeed does.

Much more important are things like cache latency, miss prediction, hardware prefetch, etc. These are the things that the P4 is better at.

In terms of "sophistication", I'd actually say the P4 wins that over the Athlon. Does that make the Athlon a bad core? Of course not. Not as refined maybe, but definitely not a bad core.

<font color=blue>If you don't buy Windows, then the terrorists have already won!</font color=blue> - Microsoft
 
I don't know much about the cores and stuff. I just read what people write and I just type it back out.... A lot of people say that AXP cores are in like every superior to the P4. I don't agree, but when you think about it, why is the AXP faster per clock if the P4 is so much superior. You can answer this post and then hopefully we can close out this debate because i hate debates that drag on. I'm not a troll or anything, just a person who is not as knowledgable.
 
This is probably true: the P4 is a more sophisticated/refined CPU than the Athlon. However, AMD seems to be doing more with what they have RIGHT NOW than Intel is. (not including the P4-2.4, perhaps, and not taking into account overclocking!)

Raystonn argued some time back that Intel was basically revving up to get some serious clockspeed going, and with that needed smaller die-size and longer pipelines. This doesn't appear to be significant right now (the long pipeline is probably why AMD is competitive right now!?), but will be as soon as Intel steps on the gas. Question: is it correct that AMD is using a higher number of pipelines, but they're shorter than Intel's?

With Intel's technology edge it remains to be seen for how long AMD can compete with them. Again, right now I'd still go for the Athlon for my use: floating point number crunching.

PS. the ironic thing here is that I don't actually KNOW what the pipeline is (as is evident in this thread), but from what I understand, a longer pipeline causes more discarded data, and hence lower performance potential...(?) DS.

:lol: <b><font color=blue>gnintsakgnirkskir ksron</font color=blue></b> :lol:
 
I don't agree, but when you think about it, why is the AXP faster per clock if the P4 is so much superior.
Click to expand...
Because "per clock" does not matter. If that's the main thing that makes a CPU superior, than the G4 is better than the Athlon, and the Itanium is far better than anything AMD has ever even thought about releasing.

The measure of how good a core is is how much performance can be pulled out of it. If the performance comes from a high IPC, so be it. If the performance comes from high clockspeed, that's fine too. If the performance comes from software optimizations, or really good hardware prefetch, branch misdirection prediction, or whatever, as long as it's the performance leader in the applications it was designed for, then it's the better core.

By "applications it was designed for", I merely mean that you wouldn't compare a Cyrix to a G4 in Photoshop alone, because that's not what the Cyrix was designed for. You wouldn't compare a G4 to an Itanium for scientific apps, because the G4 wasn't designed for 64 CPU clusters.

Of course you also have to take into account when it was released. Simply because the best P3 can out-perform the original Pentium in every way doesn't suddenly mean the Pentium core sucked.


EDIT:
So it's clear where I stand, I think the Athlon's core is better than the P4's overall (at least judging by what we've seen from the P4 so far). AMD has squeezed more performance out of that CPU than most people in the industry thought possible. Granted, the core has changed more than anything else has, but even so, it's quite impressive.

Just be careful when you start throwing around terms like "sophistication".

<font color=blue>If you don't buy Windows, then the terrorists have already won!</font color=blue> - Microsoft<P ID="edit"><FONT SIZE=-1><EM>Edited by FatBurger on 04/13/02 04:50 PM.</EM></FONT></P>
 
What kind of defects were you looking for? How can you see these in an SEM? The resolution of a really good state-of-the-art can't be more than some 50nm?
Click to expand...

The jeol tilt sem we used could resolve defects much smaller than our process's critical dimention(smallest feature sizes).

The KLA scanning equipment we used could not detect things which were very small, our main concern was bridging killer defects anyways.

But the sem I used was very state of the art, I have an image I took somewhere around here, ill post it online again and link tonight.

:wink: The Cash Left In My Pocket,The BEST Benchmark :wink:
 
--------------------------------------------------------------------------------

The p4 willamette topped out at 2.2ghz...



--------------------------------------------------------------------------------

I believe there's no such thing as a 2.2GHz Willamette.
Click to expand...

I believe there is no such thing as a 1.8ghz axp as well, we are (and again I have always made it clear on this) using the top speed of the core, NOT the top SOLD speed. Which is why I have always used 1.8ghz for my equasion on the amd side, and 2.2ghz for the p4(which was its average overclocked speed).


Yes, actually I think so. The P4 was designed for extreme high speed, while Athlon was not. I think this fact is also what is giving AMD problems with the performance of Throughbred. How well they succeed in the "re-design" or whatever they're doing, time will tell.
Click to expand...
Which is one of the reasons I stopped using the northwood shift as evidence, people just cannot understand the fact that core design has little to do with the % gain from a process shift, think of the p3 for all the evidecne you need, now give me one reason the p3>tually shift (asides from copper which I accounted for) should be any different than the axp>tbred shift.

Also, according to texas techie, whom I believe and his assertion makes sense, there was a timing issue in the shrink possibly caused by the gate lengths being so tight on the .18 micron process, and that the problem is fixxed. The reason for the delay is the length of time it takes to get a chip through the process, the fixxed chips should be coming out right around now.(and they are building a stockpile for sales). Intels northwood was also 3 months late, would you have back then attempted to say intels .13 micron process sucks because its 3 months late? It is EXACTLY the same thing as is going on right now, except intel has .13 and is putting pressure on amd, other than that there is no difference.

:wink: The Cash Left In My Pocket,The BEST Benchmark :wink:
 
Looking at AMD's Tbred roadmap indicates that the Tbred is not going to have that much to offer. A 2200+ somewhere in between mid may and end of June, 2400+ can be expected by the end of the year and the top dog is going to be 2600+. If this info is to be believed, AMD is going to have a rough time for the rest of Q2, Q3 and Q4. I also think this backs up the claim that Athlon was not designed for extreme high speed as the P4. Also it indicates that AMD, with the optimizations made in XP and process refinements implemented (0.13 micron Class gates), obtained some of the performance that could be expected from a die-shrink, thus leaving only a very modest gain to be obtained with the introduction of Tbred.

Quote from xbitlabs:
<font color=blue><i>Also they claim that the first Athlon XP processor on Thoroughbred core will be marked as 2200+ and will work at 1800MHz. Then in Q3 the rating of Athlon XP (Thoroughbred) will reach 2400+ and by the end of the year the top Athlon XP will be the 2600+ one.</font color=blue></i>


<i>/Copenhagen</i>

<b><i>Seagate Barracuda IV.
Bad performance in RAID setups!
</i></b>
 
Correction:

2200+ somewhere in between mid may and end June.
2400+ in Q3
2600+ in Q4


<i>/Copenhagen</i>

<b><i>Seagate Barracuda IV.
Bad performance in RAID setups!
</i></b>
 
Are you a PARTY-POOPER??? Why are you coming here? This is a HARDWARE forum, hence the topic of conversation is... HARDWARE!

Have a great day! :wink:

:lol: <b><font color=blue>gnintsakgnirkskir ksron</font color=blue></b> :lol:
 
This is a hardware forum? 😱

I knew coming to work drunk every day would get me in trouble eventually.

<font color=blue>If you don't buy Windows, then the terrorists have already won!</font color=blue> - Microsoft
 
i hope you noticed that he talks about 386, and 486 in there, which is two processors....
the intel 386, and the intel 486. the 486 is faster than the 386, which isnt much, but it would beat it. i think the 386 was 33 mhz, and the 486 was 66 mhz. in fact, i might have a 486dx2 laying around somewhere....not sure..but i might. my family's first computer was the 486dx2...
but i am moving away from the actual arguement....well....yeah....right...AHYUK!

-DAvid

-Live, Learn, then build your own computer!-
 
Looking at AMD's Tbred roadmap indicates that the Tbred is not going to have that much to offer. A 2200+ somewhere in between mid may and end of June, 2400+ can be expected by the end of the year and the top dog is going to be 2600+. If this info is to be believed, AMD is going to have a rough time for the rest of Q2, Q3 and Q4.
Click to expand...

They are coasting towards hammer I think thats plainly obvious.

PS: the top p4 will be 3ghz (according to intel) if the 2100+ = 2.2ghz pe, the 2600+ should be about a 2.6ghz or so, hammer in october, I see no problem personally.

I also think this backs up the claim that Athlon was not designed for extreme high speed as the P4. Also it indicates that AMD, with the optimizations made in XP and process refinements implemented (0.13 micron Class gates), obtained some of the performance that could be expected from a die-shrink, thus leaving only a very modest gain to be obtained with the introduction of Tbred.
Click to expand...

I think you are neglecting the hammer's effect on the roadmaps and are reading way into things which have nothing to do with a die shrink.


Also they claim that the first Athlon XP processor on Thoroughbred core will be marked as 2200+ and will work at 1800MHz. Then in Q3 the rating of Athlon XP (Thoroughbred) will reach 2400+ and by the end of the year the top Athlon XP will be the 2600+ one.
Click to expand...

A: rumor, but regardless, the hammer should be out before the end of the year, so they dont NEED, nor will they want to rush out fast tbreds, its all about the hammer.

we will see if my predictions are accurate when they release some tbreds and we get overclocking results, I highly doubt amd will ever release a tbred running at its cores limit, the athlon core is being retired.

:wink: The Cash Left In My Pocket,The BEST Benchmark :wink:
 
In case you did not know the 386 was the predecessor to the 486 both ran at 33 MHz internally same speed as their FSB.(there were earlier 386's
that ran with slower internal frequency). The DX2 66 ran internally at 66 MHz but still had FSB of 33MHz.Have we not come a long way.If you guys don't remember or were to young
I still stick to my assessment, for everybodies sake help
us with our queries. Ignorant does not equal
stupid.All you people over fourteen please ignore this!



avandonk
 
You must log in or register to reply here.

TRENDING THREADS

Latest posts

Moderators online

Share this page

COMPANY
RESOURCES
FOLLOW
Top Bottom