News
 

Bookmark and Share

(12) 

An unknown server end-user or hardware vendor has apparently attempted to test an engineering sample of sixteen-core AMD Opteron "Interlagos" microprocessors powered by Bulldozer micro-architecture. While the benchmark numbers are unavailable, previously unknown clock-speeds of the novelty transpired.

The unidentified recipient of alleged sixteen-core AMD Opteron "Interlagos" utilized processor marked as AMD Eng[ineering] Sample ZS182045TGG43_28 clocked at 1.80GHz as well as Supermicro H8DGU mainboard, according to a listing at OpenBenchmarking.org web-site. While the benchmark results of the dual-socket 32-core server are unavailable, the list of programs that were run via the Phoronix Test Suite includes a long list of applications and conditions.

It is interesting to note that the unknown hardware vendor or end-user compared to dual-socket 32-core machine against four-socket 48-core AMD Opteron 6168 (12-core, 1900MHz, 6MB L2, 12MB L3, 80W ACP, 115W TDP) server. Potentially, this may indicate efficiency of AMD Bulldozer micro-architecture versus previous-generation microprocessors, but this is not an evidence, though.

While 1.80GHz is hardly a high clock-speed, it does not make a lot of sense to estimate clock-speed potential of Bulldozer in general and Interlagos multi-chip-module (MCM) implementation in particular based on the frequency of the engineering sample. Firstly, thermal design power (TDP) of the new central processing units (CPUs) is unlear; in case the chip belongs to Opteron HE versions (85W TDP), then 1.80GHz for a 16-core chip may not be bad, but for the standard version 1.80GHz does not seem to be very good.

Comparison of Various AMD Opteron and Phenom Microprocessors
Brand Opteron FX Opteron Opteron Opteron Opteron Phenom II
Model Interlagos FX-8130P 6180 SE 6176 6166 HE 4184 1100T
Cores 16 8 12 12 12 6 6
Clock-Speed 1.80GHz n/a 2.50GHz 2.30GHz 1.80GHz 2.80GHz 3.30GHz
L2 Cache 16MB 8MB

6MB

6MB 6MB 6MB 6MB
L3 Cache Up to 16MB Up to 8MB 12MB 12MB 12MB 6MB 6MB
TDP/ACP n/a 125W 140W/105W 115W/80W 85W/65W 95W/75W 125W
Memory 4-ch, DDR3 2-ch, DDR3 4-ch, DDR3 4-ch, DDR3 4-ch, DDR3 2-ch, DDR3 2-ch, DDR3
Black Edition - + - - - - +
Turbo Core n/a + - - - - +
Packaging G34 AM3+ G34 G34 G34 C32 AM3
Process Technology 32nm SOI 32nm SOI 45nm SOI 45nm SOI 45nm SOI 45nm SOI 45nm SOI

AMD did not comment on the news-story.

Tags: AMD, Bulldozer, Interlagos, Opteron, 32nm, Orochi

Discussion

Comments currently: 12
Discussion started: 03/23/11 12:55:01 PM
Latest comment: 03/26/11 12:28:06 PM
Expand all threads | Collapse all threads

[1-9]

1. 
How exactly does an engineering sample represent the frequency of the final release product?
Your title for the article is merely a way to attract attention to the article and the site.
0 0 [Posted by: mamisano  | Date: 03/23/11 12:55:01 PM]
Reply

2. 
Yeah, so this is an engineering sample, with a pre-production BIOS, with who knows is or isn't working. You even admit as much in your article body.

Your title is an absolute joke. Incredibly misleading and its simply just a revenue raiser.
0 0 [Posted by: sanity  | Date: 03/23/11 06:18:20 PM]
Reply

3. 
I agree, it's a shameful title, and this type of stupidity on the part of the author just diminishes the reputation of the site for a few page hits. Long term, it's a mistake.

At any rate, it's pretty common knowledge that the Bulldozer will clock pretty high based on the design. It's clearly designed for higher clock speeds than existing processors, and once they work out the kinks should easily hit higher clock speeds than the current processors they sell, on the same fabrication technology.

Even the first real release won't tell the whole story, and certainly an engineering sample says nothing. The design is clearly for higher clock speeds, and we'll see them eventually. It just might not be right away (or it might be, who knows?).
0 0 [Posted by: TA152H  | Date: 03/23/11 09:19:39 PM]
Reply

4. 
This really means nothing and attracts some attention.
It´s sad to see xbitlabs doing such a meaningless article, but take a look at the other sites, they´re no better.

I think AMD bulldozer wil be good, not espectacular, but very good, and that´s what we need to live in a world that´s not dominated by only one chip company!
0 0 [Posted by: bollux78  | Date: 03/23/11 09:38:10 PM]
Reply

5. 
Long chip is long
0 0 [Posted by: TAViX  | Date: 03/24/11 03:02:26 AM]
Reply

6. 
Would four cores at double the speed consume more power?
0 0 [Posted by: Harry Lloyd  | Date: 03/24/11 05:28:17 AM]
Reply
- collapse thread

 
depends on the voltage increase necessary to achieve those speeds.
so really nothing can be said about that yet until we have more information.
0 0 [Posted by: Countess  | Date: 03/26/11 12:27:43 PM]
Reply

7. 
no turbo-core? ouch.
0 0 [Posted by: rektide  | Date: 03/24/11 10:21:30 AM]
Reply
- collapse thread

 
N/A means no information available in this case.
and that isn't exactly true. on the AMD server guys blog (official) he talked a lot about the turbo in upcoming bulldozer based opterons.
how it was able to do up to 500mhz with all cores fully loaded and even more when fewer are loaded.

clearly there is turbo, and clearly this new implementation will be superior to what Intel has got. the question that remains is how good the cores themselves are, for AMD to take full advantage of that.
0 0 [Posted by: Countess  | Date: 03/26/11 12:25:43 PM]
Reply
 
N/A means no information available in this case.
abd that isn't exactly true. on the AMD server guys blog (official) he talked a lot about the turbo in upcoming bulldozer based opterons.
how it was able to do up to 500mhz with all cores fully loaded and even more when fewer are loaded.

clearly there is turbo, and clearly this new implementation will be superior to what Intel has got. the question that remains is how good the cores themselves are, for AMD to take full advantage of that.
0 0 [Posted by: Countess  | Date: 03/26/11 12:28:06 PM]
Reply

8. 
This is such an exceptionally bad story I dont know where to begin.

1. The 16 core chip is server only and thus uses lower power and is clocked slower.

2. The bios used is not even near final release.

3. The expected speed of the 8 core desktop cpu is around 3.5 Ghz though higher clocks have been mentioned.

4. The desktop cpu will have Turbo 2.0 to overclock cores when others are not in use, this has been mentioned as a 500 Mhz boost.

5. Even with all the limitations of using a server 16 core bulldozer with beta bios it was the fastest cpu tested by Phoronix.

I'm waiting for my email invite to write for Xbit.
0 2 [Posted by: DEADTIME  | Date: 03/24/11 10:36:43 PM]
Reply

9. 
No, not a bad article. Who cares if the title doesn't say engineering sample, the article does. That's how the news works.

Thanks Anton, keep giving me whatever updates you can regarding the Bulldozer architecture. I admit there's not a lot of substance here but it's better than nothing, so good job. I'm sure AMD appreciate you stoking the fire too.

0 0 [Posted by: andresgriego  | Date: 03/25/11 09:44:59 AM]
Reply

[1-9]

Add your Comment




Related news

Latest News

Tuesday, July 15, 2014

6:11 am | Apple Teams Up with IBM to Make iPhone and iPad Ultimate Tools for Businesses and Enterprises. IBM to Sell Business-Optimized iPhone and iPad Devices

Monday, July 14, 2014

6:01 am | IBM to Invest $3 Billion In Research of Next-Gen Chips, Process Technologies. IBM to Fund Development of 7nm and Below Process Technologies, Help to Create Post-Silicon Future

5:58 am | Intel Postpones Launch of High-End “Broadwell-K” Processors to July – September, 2015. High-End Core i “Broadwell” Processors Scheduled to Arrive in Q3 2015

5:50 am | Intel Delays Introduction of Core M “Broadwell” Processors Further. Low-Power Broadwell Chips Due in Late 2014

Wednesday, July 9, 2014

4:04 pm | Intel Readies New Quark “Dublin Bay” Microprocessors. Intel’s “Dublin Bay” Chips Due in 2015