mcornelio
09-27-2006, 02:09 PM
Intel Announces Quad-Core Processor
By Scott M. Fulton, III (%[email protected]), BetaNews
September 26, 2006, 5:48 PM
At an early morning keynote address to open the semi-annual Intel Developers' Forum in San Francisco, CEO Paul Otellini confirmed what hardware industry resources had already known for weeks: A Core 2 Extreme quad-core processor is coming this November for the enthusiast segment of the PC market, with Core 2 Quad processors for the mainstream segment and Quad-Core Xeon server processors soon to follow in the first quarter of next year.
It will be Intel's best opportunity in years to effectively grind competitor AMD's nose into the silicon dirt, if you will, with regard to both price and performance. Last July's introduction of the Core 2 Duo and Core 2 Extreme desktop PC processors, based on Intel's Conroe architecture, saw a dramatic leap in both performance and power efficiency in a category that AMD had dominated for the past few years.
<script language="javascript" src="http://ads.betanews.com/adserve.iframejscript/www.betanews.com/MPU@Top?539064802"></script>
Suddenly, Intel's top-of-the-line models exhibited 33 percent better performance than their previous models, by independent analysis, and 27.5 percent better performance than AMD's top-of-the-line Athlon 64 FX-62.But today, Otellini made an extremely bold claim, which, if disproven by hardware analysts, could put a damper on what would otherwise be considered solid performance gains. He claimed the new Core 2 Extreme quad-core, likely to be given the model number QX6700, will exhibit 70 percent better performance than the company's current X6800.
The likelihood of that delta actually being borne out is low. Overall performance is generally measured by enthusiasts -- the very segment to whom Intel intends to target its first quad-core -- as a mixture or average of rankings in several departments. On the other hand, one or two benchmarks might have to post stellar gains of this degree, if the company is to make its case that a quad-core processor is a better value than two dual-core processors.
"Two dual-core processors" describes the architecture AMD is planning to release in mid-November, perhaps in tandem with the QX6700 announcement. Ironically, AMD might have the tougher case to make; if doubling up two FX-62s, essentially, doesn't produce just about double the performance of a single FX-62, analysts are likely to blame some unseen design deficiency.
In any event, the question now after the close of Day 1 of IDF is, is Intel setting the bar too high?
Here's the situation: The first quad-core enthusiast processor probably should not sell for more than $2,000, given the fact that street prices for the current Core 2 Extreme X6800 have fallen below $1,000. Processor manufacturers tend to choose their "tray prices" (wholesale prices for sales of 1,000-unit lots) based on a general performance curve, whereby the higher a CPU ranks in terms of performance, the greater the value of each performance point versus others on the same curve.
Based on a mathematical model of the current state of Intel CPU prices, using data from reliable sources, we plugged in a hypothetical QX6700 with 70 percent greater performance then an X6800. Given Intel's current curve, the company could conceivably afford to charge the impossible sum of $8,222.51 per unit.
Of course, Intel won't; it would be hard enough for the company to explain why it should charge more than double the X6800's price, let alone more than eight times that much. But suppose Intel were to make as phenomenal a performance gain with its first quad-core as it accomplished last July with Core 2 Extreme - which was, by anyone's measure, a monumental feat.
Based on our model, Intel could produce a processor with about a 33.6 percent performance gain, with a price of three times that of the X6800, while staying on the price/performance curve the company has already exhibited.
By that measure, if the QX6700 shows a performance gain that's more within the line of rational expectations, and yet prices it at about double the X6800's current number -- perhaps a little more -- then reviewers could chock up the new model as another Intel victory.
By Scott M. Fulton, III (%[email protected]), BetaNews
September 26, 2006, 5:48 PM
At an early morning keynote address to open the semi-annual Intel Developers' Forum in San Francisco, CEO Paul Otellini confirmed what hardware industry resources had already known for weeks: A Core 2 Extreme quad-core processor is coming this November for the enthusiast segment of the PC market, with Core 2 Quad processors for the mainstream segment and Quad-Core Xeon server processors soon to follow in the first quarter of next year.
It will be Intel's best opportunity in years to effectively grind competitor AMD's nose into the silicon dirt, if you will, with regard to both price and performance. Last July's introduction of the Core 2 Duo and Core 2 Extreme desktop PC processors, based on Intel's Conroe architecture, saw a dramatic leap in both performance and power efficiency in a category that AMD had dominated for the past few years.
<script language="javascript" src="http://ads.betanews.com/adserve.iframejscript/www.betanews.com/MPU@Top?539064802"></script>
Suddenly, Intel's top-of-the-line models exhibited 33 percent better performance than their previous models, by independent analysis, and 27.5 percent better performance than AMD's top-of-the-line Athlon 64 FX-62.But today, Otellini made an extremely bold claim, which, if disproven by hardware analysts, could put a damper on what would otherwise be considered solid performance gains. He claimed the new Core 2 Extreme quad-core, likely to be given the model number QX6700, will exhibit 70 percent better performance than the company's current X6800.
The likelihood of that delta actually being borne out is low. Overall performance is generally measured by enthusiasts -- the very segment to whom Intel intends to target its first quad-core -- as a mixture or average of rankings in several departments. On the other hand, one or two benchmarks might have to post stellar gains of this degree, if the company is to make its case that a quad-core processor is a better value than two dual-core processors.
"Two dual-core processors" describes the architecture AMD is planning to release in mid-November, perhaps in tandem with the QX6700 announcement. Ironically, AMD might have the tougher case to make; if doubling up two FX-62s, essentially, doesn't produce just about double the performance of a single FX-62, analysts are likely to blame some unseen design deficiency.
In any event, the question now after the close of Day 1 of IDF is, is Intel setting the bar too high?
Here's the situation: The first quad-core enthusiast processor probably should not sell for more than $2,000, given the fact that street prices for the current Core 2 Extreme X6800 have fallen below $1,000. Processor manufacturers tend to choose their "tray prices" (wholesale prices for sales of 1,000-unit lots) based on a general performance curve, whereby the higher a CPU ranks in terms of performance, the greater the value of each performance point versus others on the same curve.
Based on a mathematical model of the current state of Intel CPU prices, using data from reliable sources, we plugged in a hypothetical QX6700 with 70 percent greater performance then an X6800. Given Intel's current curve, the company could conceivably afford to charge the impossible sum of $8,222.51 per unit.
Of course, Intel won't; it would be hard enough for the company to explain why it should charge more than double the X6800's price, let alone more than eight times that much. But suppose Intel were to make as phenomenal a performance gain with its first quad-core as it accomplished last July with Core 2 Extreme - which was, by anyone's measure, a monumental feat.
Based on our model, Intel could produce a processor with about a 33.6 percent performance gain, with a price of three times that of the X6800, while staying on the price/performance curve the company has already exhibited.
By that measure, if the QX6700 shows a performance gain that's more within the line of rational expectations, and yet prices it at about double the X6800's current number -- perhaps a little more -- then reviewers could chock up the new model as another Intel victory.