News
 

Bookmark and Share

(14) 

Advanced Micro Devices has notified some of the product review web-sites (1, 2) that it had provided misleading transistor count number for its highly-anticipated eight-core microprocessors based on the Bulldozer micro-architecture. The reasons why AMD originally reported a wrong number of elements for its chip are completely unclear.

Based on the latest information provided by the world's second largest developer of x86 microprocessors on the planet, the eight-core chip code named Orochi - which powers the latest FX "Zambezi" product for desktops as well as Opteron "Valencia" and "Interlagos" products for servers - contains 1.2 billion transistors, not 2.0 billion as reported previously when the central processing unit was formally introduced and majority of its details disclosed.

Transistor count is an important characteristic of microprocessors as it describes their complexity and allows professionals to estimate performance and capabilities of such CPUs. The two billion of transistors was considered as incredibly high number for a chip that could barely outperform AMD's previous generation offering and lost the absolute majority of benchmarks to Intel Corp.'s Core i7-/i5-2000-series "Sandy Bridge" microprocessors released earlier this year and which feature a little less than a billion of transistors in their full-speed versions.

The corrected number of transistors - 1.2 billion - on the one hand shows that AMD has managed to squeeze eight x86 processing cores into a relatively low transistor budget, a rather remarkable achievement. But on the other hand the moderate amount of transistors prove that AMD had to sacrifice a lot to increase the core count and as a consequence its eight-core Bulldozer offerings are slower than Intel's quad-core Sandy Bridge chips.

AMD itself remained extremely tight-lipped over the reasons why it reported the wrong transistor count for its latest CPUs initially after the commercial launch. AMD publicly said at a chip conference that each Bulldozer dual-core CPU module with 2MB unified L2 cache contains 213 million transistors and is 30.9mm2 large. As a result, all four CPU modules with L2 cache within Zambezi/Orochi processor consist of 852 million of transistors and take 123.6mm2 of die space. Assuming that 8MB of L3 cache (6 bits per cell) consist of 405 million of transistors, it leaves very few transistors to various input/output interfaces, dual-channel DDR3 memory controller as well as various logic and routing inside the chip. Moreover, since official 315mm2 Orochi die size remained unchanged, it means that there are some secrets within the chip that AMD decided not to uncover.

The most logical explanation is that someone from AMD made a mistake inside the firm's marketing documents. Some suggest that the company used automatic tools to make the layout of its processors so extensively that it did not know the transistor count for sure. Other believe that the firm had to disable certain blocks of its chips of its chips because of the poor manufacturability. Yet another explanation is that AMD had to build a lot of redundant blocks, or even full cores for redundancy reasons but at least this is not something that can be seen on the Orochi die shots.

Tags: AMD, Bulldozer, Zambezi, 32nm, Globalfoundries

Discussion

Comments currently: 14
Discussion started: 12/07/11 08:29:51 PM
Latest comment: 12/09/11 12:38:10 AM
Expand all threads | Collapse all threads

[1-6]

1. 
So delaying is not the only AMD's specialization.
5 1 [Posted by: jpunk  | Date: 12/07/11 08:30:37 PM]
Reply
- collapse thread

 
Unfortunate True fer few of us still believing DAMN Is Not So Devious for years now.

The whole directors board is so Co-Rrupt for years now so this kind of fiasco is only part of artificial evolution of DAMN Co.

If we could only test our pitchforks on boards lam34sses ... well that would gave, to me at least, some remedy for now.
1 2 [Posted by: OmegaHuman  | Date: 12/07/11 10:58:45 PM]
Reply

2. 
So... I've registered specially to post this comment.

Anton, yet again this article shows that you know what you are talking and are making XBITLABS a great site with lots of information but only useful ones and, most importantly, true ones.

Some readers may ask why am I saying this... well... check this out :

"800 million of transistors - which take up a lot of die space - in an incredibly high number for various I/O, memory, logic, etc."

Source : http://www.xbitlabs.com/n...ins_Bulldozer_Fiasco.html

Guess AMD reads your articles and decided to opt out those problematic 800 million transistors... at least in press statements.
4 2 [Posted by: Herr Spiegellman  | Date: 12/08/11 12:24:38 AM]
Reply
- collapse thread

 
Respect to Anton!
2 1 [Posted by: psycho_mccrazy  | Date: 12/08/11 03:34:14 AM]
Reply

3. 
the world's second largest developer of x86 microprocessors on the planet
Seriously guys, "the worlds" and "on the planet" have the same meaning and unless you believe in intelligent aliens unintelligent enough to produce x86 chips then both the statements are redundant. Should say "the second largest developer of x86 microprocessors" and leave the rest out.

/being a pedant.
1 0 [Posted by: Prosthetic_Head  | Date: 12/08/11 07:27:03 AM]
Reply
- collapse thread

 
wanted to say that too

BTW there are only three x86 manufacturers and VIA doesn't seem to be doing anything so it seems there are only two
0 0 [Posted by: madooo12  | Date: 12/08/11 08:22:25 AM]
Reply

4. 
My wild speculation: 'Trintiy' figures with GPU vs BD without.
The 0.8bn transistor count difference sounds about right.

Deneb 4 core = 758M
Llano 4 core = 1450M

So approx GPU count = 700M, pretty close to the 800M 'missing' transistors. The Trinity GPU is likely to have slightly more transistors than the Llano unit so this sounds about right to me.

I have no basis for this other than the (coincidental) transistor count agreement. So just speculating idly.
0 0 [Posted by: Prosthetic_Head  | Date: 12/08/11 07:37:42 AM]
Reply
- collapse thread

 
trinity isn't based on the current bulldozer zambezi it is based on piledriver

and it won't have 8 cores and i don't think there will be 8 with 4 disabled because the die size must be small (for the socket)
1 0 [Posted by: madooo12  | Date: 12/09/11 12:38:10 AM]
Reply

5. 
"Other believe that the firm had to disable certain blocks of its chips of its chips because of the poor manufacturability."

of its chips mentioned twice
3 0 [Posted by: madooo12  | Date: 12/08/11 08:16:44 AM]
Reply

6. 
I was reading

"Some suggest that the company used automatic tools to make the layout of its documents."

Thought what a great joke on AMD xD
0 0 [Posted by: PsiAmp  | Date: 12/08/11 08:31:55 AM]
Reply

[1-6]

Add your Comment




Related news

Latest News

Thursday, November 6, 2014

6:48 am | LG’s Unique Ultra-Wide Curved 34” Display Finally Hits the Market. LG 34UC97 Available in the U.S. and the U.K.

Wednesday, October 8, 2014

8:52 pm | Lisa Su Appointed as New CEO of Advanced Micro Devices. Rory Read Steps Down, Lisa Su Becomes New CEO of AMD

Thursday, August 28, 2014

12:22 pm | AMD Has No Plans to Reconsider Recommended Prices of Radeon R9 Graphics Cards. AMD Will Not Lower Recommended Prices of Radeon R9 Graphics Solutions

Wednesday, August 27, 2014

9:09 pm | Samsung Begins to Produce 2.13GHz 64GB DDR4 Memory Modules. Samsung Uses TSV DRAMs for 64GB DDR4 RDIMMs

Tuesday, August 26, 2014

6:41 pm | AMD Quietly Reveals Third Iteration of GCN Architecture with Tonga GPU. AMD Unleashes Radeon R9 285 Graphics Cards, Tonga GPU, GCN 1.2 Architecture