AMD's new chip can help it gain on Intel

Reuters ^ | August 4, 2002 | Barron's

[JameRetief]

AMD's new chip can help it gain on Intel--Barron's

Last Updated: August 04, 2002 02:49 PM ET

NEW YORK, Aug 4 (Reuters) - Advanced Micro Devices Inc. AMD.N is hailed as having the "next big thing" with its upcoming eighth-generation microprocessor, and this could make the depressed stock a long-term winner, Barron's said.

The Aug. 5 edition of the Wall Street financial weekly cites Fred Hick, publisher of the newsletter High-Tech Strategist, as saying AMD's next line of microprocessors, code-named Hammer, can give AMD a multiyear lead on arch-rival Intel Corp. INTC.O .

AMD's edge could lie in the move by Intel to make its Itanium chips work only with advanced software that capitalizes on its high-speed features, the article said. By contrast, AMD's Hammer PC chips are compatible both with hundreds of thousands of existing software programs as well as new ones.

AMD is set to release its first Hammer chips during the current third-quarter and a "backward compatible" version to run powerful computers known as servers is due out in the first half of 2003, the article said.

This could lead AMD to gain back ground over the next three to four years that it has lost to Intel. With AMD's stock at $7.31 and market capitalization around $2.5 billion (compared with Intel's $110 billion), "AMD might just regain the exalted investment status it enjoyed in the giddy days of 2000," Barron's said.

[JameRetief]

Ping for another short article you might be interested in reading.

[JameRetief]

FYI, Unless the author of this article has some inside info, the soonest that I've heard for Hammer being available would be October.

[4Freedom]

What kind of speed are we talking here?

[JameRetief]

The last that I heard about speeds was in this article.

And it is likely that it will be even faster than that.

[texas booster]

AMD has not officially announced a CPU speed that I can find. I think that this is in line with earlier marketing efforts to utilize a "PR" rating (e.g. 1500), rather than raw CPU Ghz (1.33 Ghz).

The new Athlon Thoroughbreds have roadmaps that start at 2000+ and run up to 2600+ in late 2003. We can expect the Opteron (AKA, Hammer) to start at similar clock settings. AMD's contention is that CPU cycles isn't everything, and that their chips are more "efficient" per cycle than are the P4's cycles.

AMD is also working on a new 666 mhz front-side bus and several other improvements that will make this chips much faster than an equivalent "7th generation" chip.

[Lorenb420]

What kind of speed are we talking here?

I heard the first Hammers will debut at 3400+, which would be about 2.2 Ghz or around there.

[Ernest_at_the_Beach]

Thanks for the ping!

I want one!

Just saw this:

AMD to "gain four years lead" over Intel - Barron's

[Ernest_at_the_Beach]

To find all articles tagged or indexed using tech_index

Click here: tech_index

[general_re]

AMD's contention is that CPU cycles isn't everything, and that their chips are more "efficient" per cycle than are the P4's cycles.

Waaaaaaaahhhh - please tell me you're kidding...

Will we never hear the end of this "megahertz myth" BS?

[4Freedom]

Thanks for all the info.

[texas booster]

Well, I know that I'll need another couple of hundred million cpS to play my solitare, Age of Empires and other examples of productivity software on my system.

In reality, if MS adds any more "features" to Office, we might all need new systems just to run the bloatware that is tossed at us.

[4Freedom]

We're getting closer to that day when we can just about throw our keyboard and mouse away.

Carpal tunnel syndrome will be a thing of the past.

[BlessingInDisguise]

"Will we never hear the end of this "megahertz myth" BS?"

It's not a myth, it's true.

[BlessingInDisguise]

"In reality, if MS adds any more "features" to Office, we might all need new systems just to run the bloatware that is tossed at us."

That's the plan Stan.

[BlessingInDisguise]

That's one of the reasons why I still run Windows 2000.

[Dinsdale]

I've got a friend who has access to a pre-release clawhammer clock locked at 600MHz.

At that speed it outperforms a 2 GHz Pentium 4 (on AMDs benchmark of course).

These puppies are smoking fast.

[Dinsdale]

Will we never hear the end of this "megahertz myth" BS?

Which part of the long vs short pipeline engineering trade-off don't you understand? Do all opcodes execute in the same number of cycles in your conceptual universe?

Clock cycles are definitly not clock cycles.

That being said to achive very high clockspeeds everybodys needs to go to the many stage pipeline solution. Which is why Motorola lost it's clock for clock speed advantage (such as it ever had it, never did on floating point) when it pushed the PowerPC up to 1 GHz. Look at the internals, op codes that used to execute in 3 cycles now take 5 or 6.

[general_re]

Which part of the long vs short pipeline engineering trade-off don't you understand? Do all opcodes execute in the same number of cycles in your conceptual universe?

I understand it just fine. I am, however, still waiting for official SPEC results for the G4 to show me just how badly the G4 smokes the P4/AMD chips...

What's that? No official results? Maybe because the unofficial results floating around make the G4 look like it sucks s*** harder than a shop vac in a septic tank? Hmmmm.... ;)

The bottom line is that nobody cares how "efficient" a processor is, despite Steve Jobs's year-long tapdance - it's not a car, where the mileage means something to the end user. Either the raw performance is there, or it isn't, and it sounds to me like the AMD chips have got it where it counts, unlike the G4, which should be all the selling points they need. Unlike the G4, when it's unleashed on some actual, real-world benches (insofar as that's a meaningful statement), it'll show what it's capable of, and AMD shouldn't have to fall back on silly arguments about how wonderfully "efficient" their chips are...