Every once in a while, we talk about AMD fanatics (whom we call AMDroids).
There is an Intel equivalent to an AMDroid. We call such people Intelverts. We don’t talk much about Intelverts because they hardly write to us at all. We know they exist; we just don’t hear from them.
So when we get a note from someone who is at least a borderline one, in fairness, we thought we ought to share:
(Email text in bold italics, edited for spelling and grammar; my comments in plain):
“I think it’s about time you write down WHY NOT INTEL, because I don’t see
your point here. You tell people “Don’t buy A64 for like a year now. And yet “Uhh, open your
eyes, “cheap” S939 A64 are comming. And uhm, they are in 90nm.
“Yeah, great, now AMD has changed the laws of
physics and their 90nm works just fine. Yep, the smallest chipmaker starts making 90nm and gets it better then
the big 2 before eerrrrright.
“Hmm why is there no Highend 90nm A64? why is AMD making a new HE in
130nm. Ah, maybe they just felt like it, or maybe 90nm A64 will get as hot as
Prescott.”
Gee, you’d think we never wrote a number of articles pointing out that AMD must be having some problems with 90nm since
they’re not putting out high-end 90nm chips? 🙂
Actually, we don’t think 90nm Hammers are ever going to be a radically faster chip than the current 130s. However, AMD uses SOI and Intel does not.
All indications so far indicate that SOI helps the heat problem considerably. The level of improvement may not be blindingly obvious due to the different
architecutures, but with SOI, Hammers run a bit cooler instruction for instruction than Intel’s PIV design. Without, they’d probably melt.
We think SOI will give AMD enough of an advantage at standard cooling levels to give it the edge over any Prescotts. Probably not by a whole lot, but by some.
“Why do you think AMD is selling so few A64? Because most AMDroids got
it, they were {screwed}. They fell in love with that thin girl, laughing about the other one which
had some more on her hips.
“Now this thin cutie has turned into a moving barrel. Some said “well ____ it,
I just hate intel” and the rest said either “I am happy.. no need for a
new one” .. or they start asking “What is a good IntelMoBo?””
That a unique way to describe AMD’s efforts to increase prices. 🙂
Again, we’ve indicated before that we think a sizable chunk of AMD fans will stay on the sidelines, even
find something else to do, but I’ve not detected any surge to Intel, nor expect to. If AMD fans stayed away
from Intel while AMD offered them too little for too much, why would they jump to Intel now when matters are
going to improve very shortly?
AMD is not implementing any new stuff, they made a big leap in putting
the MC into the CPU, a REAL BIG LEAP if you ask me.
He’s being sarcastic, but that’s probably the most productive single hardware change anyone’s made to a CPU in recent memory. If that’s nothing, what would be something?
And i like that one… kinda, except that now i’d need a new CPU / MoBo
/ RAM if i am going to DDR2.
That’s unfortunately the price to be paid for that productive change.
“I don’t get it, really. Intel is just out of the game for you because of
PressHot. Now the CPU can be changed, doesn’t cost as much as a whole new sys . . .”
Well, this presumes you have an Intel motherboard to begin with (a decidedly minority interest in this audience), and even if you do, if it’s a
socket 478 board, you’re going to have a really interesting time keeping something from going wrong
with any big-time overclock.
. . . besides you call your website OVERCLOCKERS.com
Here: 52% OC with a PressHot.
Well, if you look at the links, we’re talking about phase change/dry ice/liquid nitrogen cooling. We’ve indicated in the past that PressHot was an interesting chip for the Mr. Freezes of the overclocking world, but not too many people use phase-change, and nobody uses liquid nitrogen for regular cooling.
Not exactly typical.
I don’t see why I should get an AMD over an Intel today, maybe you can tell me.
Somehow, I get the feeling that no answer would be good enough. 🙂
Nonetheless, for the rest:
There’s no ideal here. It’s like how many view the U.S. presidential election; it’s a choice between bad and worse. Neither CPU family looks likely to give you the increased performance you’d expect from a process shrink, and both face substantial heat and/or other problems running at high speeds.
However, unless 90nm Hammers turn out to be really terrible, they ought to be a bit better than any Prescotts we’re likely to see. Not by a whole lot, but by some.
Given the typical configuration of the average overclocker’s system, we think the average person will get the same or a bit better performance with less hassle with a 90nm Hammer than with a PressHot.
Is it a slam-dunk call? No. Depending on people’s situations, some can rationally make some rational tradeoffs and choose Intel. But heat’s a big factor here, and it’s one most people shouldn’t take lightly.
The fact is, all the calls as of late have been relatively close, and often decided for non-performance reasons. For instance, I’ve had folks write me saying, “Why do you have such a problem with socket 754 systems?” It’s not like there is/going to be some huge performance difference between the two, the main objection was one of instant obsolescence.
Shedding Light
There’s a good way and a bad way to ignore me.
No single recommendation will fit all people and all situations. What one ought to do when seeing someone’s recommendation is to see the reasoning behind it, and see it as a sort of checklist for significant factors to seriously consider. All recommendations have underlying assumptions and priorities.
If your particular situation differs radically from those assumptions and priorities, odds are any recommendation would change also. For instance, if additional-bang-for-the-buck and waiting for a bit more bang is an issue; it makes a big difference if you currently have a Barton system or a Duron running at 900MHz.
Priorities can sometimes rationally differ quite a bit. If you have a Northwood system, and bang-for-the-buck is important, you may get somewhat less bang and somewhat more heating hassle popping in a Prescott, but you’ll also get less financial and installational grief, too.
That’s the good way. The point is to get you to think about the important items before you buy, and knowing why you’re buying what you’re buying.
On the other hand, if your reaction to something I say boils down to “You didn’t choose MY side, make MY choice,” well, you’re not looking for information, you’re looking for someone to tell you how smart you are thinking the way you do.
We don’t do that here. We’re not here to tell you what you want to hear; we’re hear to tell you what you need to hear.
A good way to make a decision consists of accumulating facts to come up with an answer. A bad way to make a decision is coming up with an answer, then looking for facts to back it up. That’s using facts as decoration, not as a foundation.
In many ways, that’s the difference between an AMD/Intel fan and an AMDroid/Intelvert, the difference between a fan and a fanatic. The first looks at facts to know the answer. The second knows the answer, then looks for facts, and if something good doesn’t quite fit, they’ll change it until it does.
While this note is nowhere near as extreme as some I’ve seen; it does show a lot of selective reading and thinking. Every point raised has been addressed in an article, sometimes many times, but that doesn’t register because the answer didn’t come out “right.” It’s pretty clear from the email that the only acceptable answer is going to be Intel, and any major inconvenient facts about Prescott like heat just get tossed aside to help the cause.
That may be fine if you have a 300-watt freezing unit handy; then it’s not a factor at all. It’s not so fine when you’re telling people literally blowing hot air on the problem that it’s no problem.
It’s better to let the 2% or 5% with such units to make any necessary adjustments than to not tell the 90% with a two or three-inch fan about a big problem.
Be the first to comment