QX6700 at NewEgg and TigerDirect

[Idontcare]

Originally posted by: myocardia

Originally posted by: Idontcare
I am buying one at $1500. I would have bought one at $3000. Hell I would have bought it at $4500.

For me I am simply elated that I can buy one at any price.

Why? Because in my business I make money depending on CPU power at my disposal. 8 years ago I was building 12-node Beowulf's with 800MHz Athlons to get stuff done. Now I can put all that horsepower in one box. Priceless. Economies of scale at work.

Don't tell me it is overpriced, I'll make $100k's from one chip. Don't tell me I am a "niche" user. I work in an industry of folks who will make $100k's off of the numbers these chips will grind thru. Just because folks in my industry don't come to AT forum and talk about our life as end users doesn't mean 10's of thousands of us aren't out there.

And I am not so arrogant as to assume my industry is the only one who will see a 4X increase in productivity with this chip. Too many folks on this forum live in one hell of a small fishbowl, its sad to read, mostly, but alas I am off to continue making my fortunes with my "way overpriced" CPU. Whee!

You are a niche user. Will Intel sell millions of these quad-cores to corporations, both large and small? You bet. But, if crunching numbers was making you $100k's, I can assure you that you would own 10's, if not more, of quad-core boxes that have been available for the last ~2 years. I mean, there's a member of these forums who runs a business out of his home, who has two quad-core systems, plus a dual-core system, and he's had them for a year or more. And since a quad-core is going to quadruple your productivity, that means you've got a pretty slow dual-core system now, huh?

edit: If crunching numbers, and only crunching numbers, was my business, I can guarantee you that I'd have an entire fleet of 2P Opteron 185 systems already. Well, assuming they were actually making me $100k's, anyway.

Ha ha, I wish mindlessly crunching numbers was all my business took. It is iterative, crunch numbers, interpret numbers, tweak code, crunch numbers, repeat and repeat. 4 cores is about right, I could do more with more but not if it took longer to have more. This code runs single-thread, multi-instance.

So fast time to finish rules the roost, extra cores and instances gives you more ways to have slightly varying answers (think weather forcast computing techniques) and that helps up to a certain arbitrary threshold related to this human's ability to multi-task.

I have no doubt I could saturate 12-24 cores with enough instances to still make it a value-added proposition, but it would't really help me if those cores were clocked low.

Time will tell which path turns out the most rewarding, who knows I may try them all in parallel just to answer the question for myself.

 

[myocardia]

Originally posted by: Idontcare
Ha ha, I wish mindlessly crunching numbers was all my business took. It is iterative, crunch numbers, interpret numbers, tweak code, crunch numbers, repeat and repeat. 4 cores is about right, I could do more with more but not if it took longer to have more. This code runs single-thread, multi-instance.

So fast time to finish rules the roost, extra cores and instances gives you more ways to have slightly varying answers (think weather forcast computing techniques) and that helps up to a certain arbitrary threshold related to this human's ability to multi-task.

I have no doubt I could saturate 12-24 cores with enough instances to still make it a value-added proposition, but it would't really help me if those cores were clocked low.

Time will tell which path turns out the most rewarding, who knows I may try them all in parallel just to answer the question for myself.

Ahh, I didn't realize it was interactive. And since it's single threaded, speed wins, of course. Multiple instances doesn't really give you "more", it just gives you more variations, which is obviously a good thing, in your business. Makes perfect sense now.

BTW, I understand exactly what you mean about the human factor. I have a hard enough time keeping my new dual-core saturated, on the few times that I've attempted it, with single-threaded apps. Still, I find it hard to believe that there are all that many "strategy backtesting for automated forex trading" users out there. You know, enough of you that the term niche no longer applies.

 

[coldpower27]

Originally posted by: myocardia

Originally posted by: coldpower27
Though I concur in that the Q6600's should be below 700 USD 10 Months after it's introduction as that is only a reduction of less then 20% in order to satisfy that requirement.

Right. What I forgot to say in that last post was that if the price on 65nm Q6600's doesn't go down, the prices of 45nm Q6600's would have to be higher, or there would never be another 65nm Q6600 sold again, ever, after the introduction of the 45nm Q6600. Anytime any company wants to get parts out the door (think Presler), then they have to either raise the price of the new part, or drop the price of the older part. There's really no other way.

Not really, there are still consumers not educated enough to tell the differences, it doesn't matter in the end, the 45nm versions will be like the Willamette to Northwood transition, in that they are the same architecture with more cache.

You also have the possibility of only the low end SKU's going down the pricing charts, with the mid end to high end being phased out entirely, considering Intel only expects 3% of their Desktop Sales to be Quad Core even with the Q6600.

Actually from what I recall when Intel introduced the Presler incarnation of Pentium D they were the same price as the 90nm versions, except for the Pentium D 950 which was a high end SKU, and was more epxensive but the Pentium D 940 was cheaper then the Pentium D 840 funny that. Pentium D 920 and 930 were on identical pricing as the 8xx versions.

When the superior 45nm versions come in I expect them to be the same as the same clockspeed version on 65nm. It's what Intel has done with the past 2 transistions now. Same price as the old version at first, with prices drops on the newer process versions later, with lower end cheaper versions of the older tech versions introduced to move remaining inventory (if need be).

 

[Idontcare]

Well I am not really going to disagree about being called a niche user, but only because there is no such thing as the non-niche user when you get right down to it.

Even gamers are niche users if you are so inclined as to step back far enough when breaking down the financials of the semiconductor industry. I bet Intel (and AMD) sell far more chips for non-gaming laptop users than otherwise.

But for my particular industry, they are holding a contest right now for automated traders (we hone our codes thru exhaustive and extensive back-testing) and the latest contest attracted 258 participants. Expert Advisor Competition

258 sounds like a very very small number, and it is, but it represents maybe 1/1000 or even less of the total number of metatrader users. Most folks who I know who are decent at making money with automated forex traders will not give out their code so freely as is required in such competitions.

You might be surprised how many people are in this particualr industry, and the majority of us can't wait for faster and faster desktop systems.

Alas I am not here to give a sales pitch on forex trading, merely stopped by to decrease the level of negativity I saw being leveled towards a chip I wish was available last year for $10k.