News
 

Bookmark and Share

(29) 

The next-generation desktop central processing units (CPUs) from Advanced Micro Devices with Bulldozer micro-architecture are projected to offer around 50% higher performance compared to existing chips from AMD and Intel. While the alleged promises presumably from confidential documents seem very "promising", they should be considered as extremely rough performance estimates.

An AMD document obtained by Donanimhaber web-site claims that an eight-core Bulldozer microprocessor offers 50% higher performance compared to quad-core Intel Core i7-950 with HT microprocessor in games, rendering and multimedia applications. The web-sites does not specify which benchmarks were used. In fact, 50% higher performance versus a quad-core chip does not seem to be bad, especially in games that do not usually take advantage of multi-core chips.

The web-site claims that AMD compares performance of the Core i7-950 with performance of AMD Phenom II X6 1100T chip, which is considerably slower than its rival, according to performance tests conducted by X-bit labs. It is unclear whether AMD showed the model 1100T being slower or faster than the i7-950. Test settings were not revealed. Typically, AMD publishes so-called "simulated benchmark results".

Rumours published earlier this month suggested that AMD plans to return the FX products family with extreme performance. Indirectly, this information confirms very high performance of new microprocessors from AMD.

Engineering samples of code-named Zambezi (desktop Bulldozer) chips became available for AMD's partners for testing in December, 2010. Production candidates should be ready by February '11 and the initial production of the company's first desktop microprocessors powered by the long-awaited Bulldozer micro-architecture is scheduled to start in April next year. Probably, the launch of the chip will occur around the same timeframe.

Eight-core Zambezi features four dual-core Bulldozer modules, each of which is believed to have 2MB of shared level-two cache, that will share 8MB L3 cache. In total, the whole chip will pack in whopping 16MB of SRAM, a 77% increase from the current six-core microprocessors that have 9MB of cache in total. Clock-speeds of the chip may be as high as 3.50GHz or even higher.

The first Zambezi microprocessors to be launched are expected to be eight-core products with 95W and 125W thermal design power as well as 8MB L3 cache. Later in the second quarter of 2011 AMD, according to sources with knowledge of the company's roadmap, will release six-core chips with 8MB L3 cache and four-core products with 4MB cache. All of the processors will feature TurboCore 2.0 technology, dual-channel DDR3 memory controller with up to 1866MHz memory support and will be compatible with AM3+ mainboards.

AMD did not comment on the news-story.

Tags: AMD, Zambezi, Orochi, Valencia, Bulldozer, 32nm

Discussion

Comments currently: 29
Discussion started: 01/14/11 03:56:54 PM
Latest comment: 05/15/11 09:01:32 PM
Expand all threads | Collapse all threads

[1-14]

1. 
Cool lets hope that it turns out as expected and not loaded with problems or short comings. Still rocken a ancient 8250e @ 2.7ghz.
0 0 [Posted by: nforce4max  | Date: 01/14/11 03:56:54 PM]
Reply

2. 
I have so many problems with that title I don't know where I should begin.

Ambiguous about which generation of core i7 is being compared.
- Sandy Bridge just came out, but it turns out they're comparing to the first gen.

Equating the core i7 (either generation) with the Phenom II.
- Even the old core i7 is better than the Phenom II, significantly.
- You mention this in the article, but why does your title contradict your article?

"50% better" turns out to mean 25% worse performance per core
- As you say, maybe their benchmarks are not fully threaded. But as there is no disclosure of what these benchmarks are, I would expect they were cherrypicked to make the product look better; i.e., fully threaded to give advantage to their higher core count.


I feel like you tricked me into reading the article by lying in the title. That is the very reason I quit reading slashdot---and I know I'm not alone.
0 0 [Posted by: vith  | Date: 01/14/11 04:21:11 PM]
Reply
- collapse thread

 
Really? I thought that Nehalem was like 10-20% faster than Core2 and Phenom II, and Sandy Bridge was 0-10% faster than Nehalem. That's not even noticeable, let alone "significantly better"... But maybe you build your PCs to run SuperPi and other Intel-sponsored benchmarks....
0 0 [Posted by:  | Date: 01/14/11 08:49:34 PM]
Reply
 
Is SUperPi sponsored by Intel? I don't think so...
But anyway counting SuperPi is so fun, especially when you check the readme and see, that the work people do now for benchmarks in like 30 seconds, older computers (from SuperPi era) were counting for a day!
0 0 [Posted by: knedle  | Date: 01/15/11 01:32:12 AM]
Reply
 
SuperPi, like many synthetic benchmarks, consistently show Intel being exponentially faster than they are in real life. The disparity is nothing short of amazing. If you recall, a low-end Pentium4 could beat a high-end Athlon64 in SuperPi by a solid margin, even though it couldn't beat the AMD chip at anything else. It couldn't be more of a fraud, considering it consistently favors every generation of Intel chip.
0 0 [Posted by:  | Date: 01/15/11 05:17:45 PM]
Reply
 
Stop using superpi go and use Wprime and see how good Amd processors are because its multithreaded, i dont see i7 performs better in games above 1680x1050 , its only faster at lower resolutions so who plays games anyway in lower resolution these days.
0 0 [Posted by: 3Dkiller  | Date: 01/15/11 02:23:58 PM]
Reply

3. 
I'll believe it when I see real reviews on a launched retail product. Really even if they did meet those performance numbers I'm not sure how exciting it will be because

A) what will the CPU's cost
B) what will the motherboards
C) what hardware prerequisite's will be required for it DDR3 or will DDR2 be optionally support on select motherboards?
0 0 [Posted by: knowom  | Date: 01/14/11 05:41:05 PM]
Reply
- collapse thread

 
DDR3 isn't that expensive anymore, why would you be hung up on being able to use DDR2?

Besides, Intel's motherboards are "antitrust expensive"(tm), considering that all of the motherboard's functionality has been moved to the CPU. I don't understand how Intel's prices keep rising.
0 0 [Posted by:  | Date: 01/14/11 08:52:46 PM]
Reply
 
Actually I've build some low costs PCs and it turned out, that using Intel for them was much cheaper, than using AMD (MB with G31 + Celeron Dual Core + 2GB of DDR2).
I wanted so badly to build them on AMD, but we were on really tight budget, so unfortunately everything went to intel.
0 0 [Posted by: knedle  | Date: 01/15/11 01:34:29 AM]
Reply
 
Well, for that to be the case, you must have been using the cheapest-of-the-cheap crappy generic parts. Still, a Sempron and el cheapo 780g board should still be a comparable price.
0 0 [Posted by:  | Date: 01/15/11 05:20:30 PM]
Reply

4. 
Realy, i hadn't expected that such a reputable site like Xbitlabs have published the text like that.
If you, guys, want to make posts like that, i recommend you should publish it in section "Rumours from bloggers" not "News", if you want to stay reputable, of course.
0 0 [Posted by: Azazel  | Date: 01/15/11 01:47:59 AM]
Reply

5. 
Thanks for the forward-looking article Anton. As with all pre-launch articles, we need to take them with a grain of salt.
0 0 [Posted by: bbo320  | Date: 01/15/11 05:35:56 AM]
Reply

6. 
So this claim AMD is making about Bulldozer should make it even faster in performance than even Sandy Bridge if they are saying that the overall performance is 50% higher then intels first gen Core i series. Yeah, lets see the real world benchmarks first. Everything sounds good on paper until you put them through real world testing.
0 0 [Posted by: SteelCity1981  | Date: 01/15/11 06:32:02 AM]
Reply

7. 
Lol at the comments - when xbit is on it's usual pro intel/nvidia stance you guys are all over it like it's gospel, especially idiots like azazel.

As for your "real world testing", you can see how useful that is if xbit actually thinks the i7 950 and 1100T aren't on par - most other sites would agree they were but not here because the benchmarks are biased heavily in intel's favour.

So in summary, I'm sure Bulldozer will thump anything intel has on most review sites. Not here, or on anandtech, but we all know who the shills are.
0 1 [Posted by: jimbo75  | Date: 01/15/11 06:42:48 AM]
Reply

8. 
that's too much cores...
great for serious work, but not my desktop. my amd dual core desktop is pretty much like a car, rarely used its max power.
0 0 [Posted by: zodiacfml  | Date: 01/15/11 09:32:51 AM]
Reply
- collapse thread

 
We should expect AMD to eventually sell defective/crippled 2-modules/4-core Bulldozer as dual-cores.

Not so sure abour how triple cores might work. One and a half BD module?

If the Q2 release holds, I guesstimate that in about 2 months we will be seeing credible leaks of BD performance.
0 0 [Posted by: BernardP  | Date: 01/15/11 01:44:18 PM]
Reply

9. 
I can't deny feeling excited for return of the FX series, they used to own Intel Processors back in the day :D
0 0 [Posted by: PFX  | Date: 01/15/11 04:04:51 PM]
Reply

10. 
Do we need a new mobo for it?!?
0 0 [Posted by: TAViX  | Date: 01/16/11 04:02:30 AM]
Reply
- collapse thread

 
Yes, a new motherboard with AM3+ Socket will be required for the Bulldozer CPU.

But this new socket will be compatible with current AM3 processors.
0 0 [Posted by: BernardP  | Date: 01/16/11 10:36:37 AM]
Reply

11. 
Good for x86 CPU.Not suprise for me.AMD sometime suprise us all.ARM not achieve even 64 bit CPU procesing.
0 1 [Posted by: Blackcode  | Date: 01/16/11 04:27:10 AM]
Reply

12. 
In Paragraph 1, last sentence, Anton wrote:

"While the promises seem very promising, they should be considered as extremely rough estimates."

Promises are usually "promising"

Not just "rough" but EXTREMELY ROUGH estimates!


MRFS
0 0 [Posted by: MRFS  | Date: 01/16/11 09:32:05 AM]
Reply

13. 
This is great if true but when will AMD release AM3+ Motherboards? I dont mind using an AM3 chip in it for a while.
0 0 [Posted by: dizzystuff  | Date: 01/18/11 04:50:03 PM]
Reply

14. 
If bulldozer series beat by 50% an i7 2600k at 5ghz i will give everyone here 1 million dollar
0 0 [Posted by: srgess  | Date: 02/12/11 08:02:19 AM]
Reply
- collapse thread

 
If you make huge bets based on such little information, I am quite sure you don't have any money.
0 0 [Posted by: DEUS_XM  | Date: 05/15/11 09:01:32 PM]
Reply

[1-14]

Add your Comment




Related news

Latest News

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

Monday, August 25, 2014

6:05 pm | Chinese Inspur to Sell Mission-Critical Servers with AMD Software, Power 8 Processors. IBM to Enter Chinese Big Data Market with the Help from Inspur