World's fastest CPU clocked at 128 gigaflops

17. my god i cant believe people are still crowing over which is better, look, 3 years down the line and PS3's are still getting hammered by XBOX's left right and centre, FAR more unit sales and FAR more game selection coupled with a better online experience and much better value for money and there you have the end of the debate.

BUT on a technical note, the Xbox has much better Graphics processing power, which at the end of the day, is what makes the games shinier. Not a more powerful CPU, its allllll about the GPU and anyone that tries to argue that PS3's is better in this dept is a total moron.

The only plus on the PS3's as far as i'm concerned is the Blu-Ray.

Overall it's down to individual preference at this stage in the game, when you take both consoles pro's and con's into consideration, they are about equal whether the fanboys like that or not.

All I keep hearing from the PS3 sources is that "given time, the PS3 will mature and the power of the machine will really be displayed" in my opinion, it's too late for that, come next year I wouldn't be surprised to see another console launched, and if Sony faf about like they did with the PS3, then they will lose that round too.

16. And they use Cell Processors from the PS3 in the RoadRunner Super Computer so don't even say that the 360 has more advanced technology you idiot

15. sgtheadhole is an idiot...the CPU in the PS# does not compare with the speed of a normal cpu...A Playstation 3 uses technology that is superior to that of the Xbox 360. The Cell processor used in the playstation has 6 cores and your precious Xbox 360 has only 3. I have both an Xbox 360 and a Playstation 3 so before you go anointing the Xbox 360 has the better gaming console with better technology you should get your f***ing facts straight. O and my first 360 crapped out after 3 days and I have never had a problem with my PS3...Thank You Sony

Source:

http://en.wikipedia.org/wiki/Cell_%28microprocessor%29

14. Thats TOTAL ********, the PS3's Cell processor isnt 1Tflop in performance, it only has 1 PowerPC IBM Core @ 3.2Ghz, with 7 "SPE's" each at 3.2Ghz also, and these "SPE's" are the same as unified shader cores you find in graphics cards today, and there is no way that those few things can have 1TeraFlops performance, its all Sony marketing ********, none of the consoles are next gen, the PS3 uses older tech than the xbox 360.

13. This article is incorrect, the Cell Processor is the current fastest microprocessor in the world at 1 TerraFLOPS. With the IBM Roadrunner being the fastest supercomputer running at 1.71 PetaFLOPS using over 12k of them, with 6 thousand AMD Opteron's.

12. Mahatmacoat - have you actually *DONE* any research, or are you simply saying that 'it's a nice little article' and making urbane comments for fun??

Not one of your comments has attempted to add anything of value to the original post - nothing that indicates why it's the 'worlds fastest processor', or why in reality - it's not.

I challenge you to *RESEARCH* the information in the article as well as other similar processors - then write a succinct response that explains why this new CPU is in fact the world's fastest. Much of the information that you'd need would already be in the links I've provided. It shouldn't take more than 10 minutes.

Include references (as I have done) to well-respected sources for your information.

===============

I will be more than happy to acknowledge your well-researched defence of your position - if it is accurate.

===============

Also: digitalworldtokyo - which Mahatmacoat indicates is a 'nice little site with plenty of J-tech insights' below has virtually all posts done by three people: Mark Weitzman, Yoshi Suzuki and J Mark Lytle.

I am wondering if one of these three is mahatacoat, in which case this may be a case of astroturfing.

I have no affiliations with Intel, AMD, NVidia or IBM (I don't work for them, nor have I).

http://en.wikipedia.org/wiki/Astroturfing

11. I guess that means everyone covering this pre-release CPU is wrong except you, then!

10. Some traffic indeed...

1. The CPU listed isn't the world's fastest; the Cell processor would be, at 1000 GFLOPS or more (not 128 GFLOPS)

2. A $300 graphics card such as an ATI 4870 will outperform it by 2x - right now

3. The Fujitsui CPU mentioned doesn't seem to be released

4. Intel's Larrabee isn't released - and it's over 1000 GFLOPS according to some estimates (~8x faster than the Fuji processor). It's due for release in 2009/early 2010. When is the Fuji processor being released?

"Plenty Accurate" would cover at a minimum the above four simple points. The article stub doesn't - potentially misleading the reader.

9. Wow - that's some traffic. Still, I'd say the info in there plus the linked source (before it collapsed under the weight!) is plenty accurate for casual readers. Specialists can get what they need at a specialist site, after all. Let's look forward to this chip when it comes out anyway.

8. This article is 5th on the "Top in all categories" at Digg.com. I have no idea how much traffic this article is receiving; but without actual research people would be completely deceived.

7. Good technical points, funkygibbin. The points made here about the new tech all seem pretty valid for a very early, short news item, regardless. It's not an in-depth thesis, so go easy!

xbitlabs.com (currently offline at the link given) as a source is very thorough. I'm happy to get this snippet here and look elsewhere for more if I need it. We'll see what Fujitsu can do anyway when this thing eventually comes out.

6. Thanks Mahatamacoat... I'm not obsessed; more a case of wanting a little research from journalists into new announcements from vendors - rather than repeating the 'news' that the vendor announces. 10 minutes of research into current CPU tech, GPU tech and yet-to-be-released technology would have made a much better article (with the same sensational title perhaps) - but would have informed the readers far more.

As far as I can tell Fujitsu have only "demonstrated" this technology - made a few wafers and tested them - however the technology is not for sale as yet.

Intel with Larrabee have new technology also in the pipe, and have also performance tested it - however this technology not mentioned. In short the article compares "unreleased" technology to "current" technology from Intel - and finds that that the new technology is faster by a factor of 2.5.

Larrabee (also unreleased, due late 2009) from Intel is faster by a factor of around 20 than current Core i7 technology - but this technology isn't mentioned in the article.

Already released technology from AMD and NVidia is well in advance of the Fujitsu CPU, and is well-respected in scientific circles. Floating point performance is approximately 20x current Core i7 speeds at 1 TFLOP

IBM's Cell processor (also released currently) also performs around 1 TFLOP or more according to Wikipedia.

Journalists have a responsibility to both confirm their stories, as well as try to provide some background and further information on the announcement from a vendor - providing a wholistic view of the landscape - otherwise the 'news' can be misleading.

Journalism with a single unresearched viewpoint really went out years ago - which is why sites such as Toms Hardware and Anandtech - which provide this balanced and well-researched viewpoint - have become well-known sources for excellent technical information.

Cell Processor - http://en.wikipedia.org/wiki/Cell_(microprocessor)

5. Funkygibbin, thanks for sharing http://www.digitalworldtokyo.com - nice little site with plenty of J-tech insights. Definitely not for obsessives like you :-)

4. Not to put too fine a point on it, but the author - "J Mark Lytle" - typically writes about "Shuffling wannabe Terminator robot learns the two-step" and "Gadgets of the month – six of the best in Japan tech", as well as "All your blog are belong to us – software robot to do all that tedious writing".

(these are the last few articles written by him here: http://www.digitalworldtokyo.com/)

This is a terrific achievement by Fujitsu; present high end consumer CPUs such as the Core i7 max out at around 50 GFlops (http://www.intel.com/support/processors/sb/cs-023143.htm).

A graphics card with greater than 300 GFLOP performance can be purchased right now - for about USD$300. With the stream computing SDK that comes as a free download and this CPU/GPU can immediately begin to use these GFLOPS.

The best-practice is to research the options for a given desktop or server application, especially as most servers today (databases, file serving, etc) - actually require huge integer / multicore performance, not floating point. Floating point performance is required more for scientific and graphical/3D purposes.

3. Notes: Radeon 4800 series - up to 1200 GFlops (1.2 TFlops) - http://ati.amd.com/products/Radeonhd4800/index.html

Larrabee performance is sketchy, but rumoured to be in the 240 GFlop range with 24 cores (it's rumoured to be shipping with 32) - http://forum.xcpus.com/graphics-displays/16435-larrabee-details.html

More food for thought: the article doesn't mention if these are single or douple precision FLOPs (there's a huge difference). Without this detail it's likely to be single precision - in which case Larrabee is around 400 GFlops... Perhaps a little more research on the CPU / GPU / "fastest CPU" would be better?

2. Yes, but my graphics card can hit 500 GFlops (it has about 128 on-die CPUs as shader engines); the new Larrabee CPU due out in 2009/2010 should hit 1000-2000 GFlops with 32 on-die Pentium class CPUs...

1. Yeah sure but can it run CRYSIS?