backtop


Print E-mail del.icio.us 119 comment(s) - last by Justin Case.. on Jul 7 at 1:00 AM


AMD "Barcelona" die shot. (Source: AMD)
AMD bets on low-power, low-clock processors for its next generation debut

AMD announced today what analysts have been dreading for months: the company will launch its next-generation architecture this August, at top-out frequencies of 2.0 GHz.  This next-generation CPU will become the successor of the existing K8 AMD Opteron lineup.

AMD's press release claims, "With planned availability at launch in a range of frequencies up to 2.0 Ghz, AMD expects its native quad-core processors to scale to higher frequencies in Q407 in both standard and SE (Special Edition) versions."

AMD Barcelona roadmaps from February 2007 indicated the company would launch at a top-out frequency of 2.3 GHz, with an eventual ramp to 2.6 GHz on the way. 

Just this past June the company also announced it could demonstrate Barcelona on working systems immediately.  The part the company left out from its announcement was that the demonstration CPUs ran at 1.6 GHz -- still 400 MHz under the estimated launch speed and 700 MHz under the roadmapped top-out frequency.

To add insult to injury, when DailyTech benchmarked the pre-production 1.6 GHz Barcelona, the CPU did not match Intel's 65nm quad-core offering clock-for-clock.  AMD engineers stress to DailyTech that this benchmark was premature, and that final silicon and software will allow for SSE optimizations and better performance. 

AMD's press release claims the Barcelona Opteron will provide a 70% performance increase in certain database applications; and a 40% increase on certain floating point applications.  The company has maintained this conclusion for nearly six months, though the only benchmarks it has provided to back its case up are simulated SPECint and SPECfp benchmarks released in April 2007.

The 2.0 GHz K10 Opteron announcement comes with another twist; the company will launch its energy-efficient models alongside its performance models at launch.  Originally, these low-power Barcelona derivatives were scheduled for a late 2007, possibly early 2008 launch. 

“AMD has prioritized production of our low power and standard power products because our customers and ecosystem demand it, and we firmly believe that the introduction of our native Quad-Core AMD Opteron processor will deliver on the promise of the highest levels of performance-per-watt the industry has ever seen,” added Randy Allen, corporate vice president of AMD's Server and Workstation division.

The upcoming Barcelona processors are drop-in compatible with current-generation Opteron motherboards.


Comments     Threshold


This article is over a month old, voting and posting comments is disabled

Why...
By JTKTR on 6/29/2007 1:01:57 PM , Rating: 2
I have no idea why I'm an AMD fanboy.... See, I bought an X2 when Conroe was king, I'll by a Phenom even if Penryn is king...But I have no idea why! Can someone fill me in as to why I will consistently by slower chips as long as they are made by the hands of an AMD worker?




RE: Why...
By melgross on 6/29/2007 1:44:33 PM , Rating: 5
Because you're not thinking your purchases through clearly.


RE: Why...
By PCDestroyer on 6/29/2007 2:43:31 PM , Rating: 1
They run a lot cooler. They rate the TDP as a max while Intel rates it as typical TDP


RE: Why...
By defter on 6/29/07, Rating: -1
RE: Why...
By omnicronx on 6/29/2007 2:54:30 PM , Rating: 3
no.. your wrong.. intel and amd do measure their tdp rating differently. as the poster said amd is maximum and intel is more of an average.

quote:
To make matters more confusing, both companies measure TDP in a different way, oh joy. AMD measures it as the maximum power that a CPU can possibly draw while Intel calls TDP the maximum you are likely to see sustained. One Intel TDP ~= .85 AMD TDP. To correct for this, AMD bumps up the Intel numbers to max draw.


http://www.theinquirer.net/default.aspx?article=32...

i know its inquirer.. but the point gets accross


RE: Why...
By JackPack on 6/29/07, Rating: -1
RE: Why...
By omnicronx on 6/29/2007 3:10:19 PM , Rating: 5
thats nice.. thanks for no link, or quote.. or anything.
On another note all i said was TDP is not TDP. You only proved my point, but thanks for trying.

just incase you need it.. (which you might)
if 65w tdp for AMD =/ 65w tdp for intel
then TDP =/ TDP.. thanks for coming!


RE: Why...
By JackPack on 6/29/07, Rating: -1
RE: Why...
By Accord99 on 6/29/2007 4:24:28 PM , Rating: 2
Here's one review:

http://www.behardware.com/articles/674-4/intel-p35...

The new G0 stepping looks very impressive, even further improving on the excellent power usage of the B2 stepping C2Ds.


RE: Why...
By kenji4life on 6/30/2007 1:50:56 PM , Rating: 4
The only thing he clearly implied is that Intel rates their TDP's lower than AMD does.

He never made any claim that any AMD processor is superior in TDP.

But I will argue that the AMD method is much more reliable, and responsible.


RE: Why...
By TomZ on 6/30/07, Rating: 0
RE: Why...
By Justin Case on 7/2/2007 10:52:20 AM , Rating: 4
Actually, AMD's method is more "responsible" from the point of view of end users, because it lets people upgrade the CPU (as long as it's within the same "family"), or use it under load 24/7 without worrying if their standard cooling solution will do.

Basing a system's thermal design on "average power dissipation" can spell disaster when the person using that system doesn't run it under "average" conditions. Which is why many OEMs ignore Intel's "average TDP" and do their own testing, to make sure the chips won't fry.

For that matter, most also ignore AMD's TDP, because it's usually 10-15% higher than the CPU will ever dissipate, even under full load. But if you're going to use the manufacturer-supplied numbers, I'd much rather use AMD's than Intel's.


RE: Why...
By Accord99 on 7/3/2007 2:23:12 AM , Rating: 2
quote:
Actually, AMD's method is more "responsible" from the point of view of end users, because it lets people upgrade the CPU (as long as it's within the same "family"), or use it under load 24/7 without worrying if their standard cooling solution will do.

Intel uses family TDPs now too.

quote:
For that matter, most also ignore AMD's TDP, because it's usually 10-15% higher than the CPU will ever dissipate, even under full load. But if you're going to use the manufacturer-supplied numbers, I'd much rather use AMD's than Intel's.

Though power tests have shown there is no significant difference and if anything, it's Intel whose TDP is over-rated.


RE: Why...
By TomZ on 7/3/07, Rating: 0
RE: Why...
By Justin Case on 7/3/2007 3:44:52 PM , Rating: 3
You seem to have missed the point, completely. Hint: the difference lies in the different meaning of the words "maximum" and "typical".


RE: Why...
By TomZ on 7/3/2007 4:41:40 PM , Rating: 1
I don't know where you're getting "typical" from - Intel doesn't rate TDP in terms of typical. They rate it in terms of power dissiptated given the maximum expected computational load. While this is not the same as total worst case power, it is worst case relative to what is expected to be encountered by customers in actual use.

If it was typical (i.e., average) as you say, then Intel's TDP would be useless for the purpose it was designed for, which is clearly not the case.


RE: Why...
By Justin Case on 7/4/2007 12:17:18 AM , Rating: 3
Tell that to the people with dual-core Netburst chips... :P

Apparently the "maximum expected computational load" used to calculate their TDP consisted of running Solitaire.


RE: Why...
By TomZ on 7/4/07, Rating: 0