Tags

20nm 28nm 32nm AMD Android Apple ARM Business DRAM Flash Geforce GK104 Google HDD Intel iOS Ivy Bridge Kepler Micron Microsoft NAND Nokia Nvidia Qualcomm Samsung Semiconductor SSD TSMC Windows 8 Xbox


News
 

Bookmark and Share

(60) 
AMD Bulldozer Zambezi 32nm

AMD FX-Series Top-of-the-Range Microprocessor to Cost $320 - Report.

AMD's Eight-Core FX-Series "Bulldozer" Chip to Be Relatively Affordable

[05/20/2011 01:03 PM]
by Anton Shilov

Even though Advanced Micro Devices' next-generation FX-series microprocessors code-named Zambezi and powered by Bulldozer micro-architecture are projected to bring the company back onto the market of high-performance desktop microprocessors, the new chips will be relatively affordable. For example, the highest-performing model will cost $320, according to a media report.

As reported previously, the initial family of FX chips will include four models: two eight-core flavours, one six-core version as well as one quad-core chip. The top-of-the-line FX-8130P and FX-8110 will cost $320 and $290, respectively, six-core FX-6110 central processing unit will be priced at $240, whereas quad-core FX-4110 microprocessor will cost $190, according to a news-story by Donanim Haber web-site.

It is rather noteworthy that AMD will sell its highest-performing microprocessor for $190 - $320. Positioning of the chip in AMD’s documents seen by X-bit labs clearly shows that AMD wanted to address the high-end desktop CPU market. In particular, AMD wanted to compete against Intel Core i7-2600-series microprocessors with its eight-core chip. Intel’s quad-core Core i7-2600K costs $317. It remains to be seen how successful will AMD FX-8100-series be against Intel’s six-core code-named Bloomfield Core i7 family chips.

Among the advantages that are mentioned by the AMD documents that X-bit labs has happened to see are “more overclocked cores”, “more cores dual graphics, OpenCL and GPU” compute and others. By the end of the year the firm expects over 10% of its desktop products to be based on the Bulldozer micro-architecture and be in the AM3+ form-factor.

AMD Orochi design is the company's next-generation processor for high-end desktop (Zambezi) and server (Valencia) markets. The chip will feature eight processing engines, but since it is based on Bulldozer micro-architecture, those cores will be packed into four modules. Every module which will have two independent integer cores (that will share fetch, decode and L2 functionality) with dedicated schedulers, one "Flex FP" floating point unit with two 128-bit FMAC pipes with one FP scheduler. The chip will have shared L3 cache, new dual-channel DDR3 memory controller and will use HyperTransport 3.1 bus. The Zambezi chips will use new AM3+ form-factor and will require brand new platforms.

The Sunnyvale, California-based chip designer plans to introduce AMD 900-series chipsets compatible with Zambezi processors in Q2 2011. The Bulldozer processors, Radeon HD 6000 "Northern Islands" discrete graphics cards and AMD 900-series core-logic sets will power AMD's next-generation enthusiast-class platform code-named Scorpius.

AMD did not comment on the news-story.

Tags: AMD, Bulldozer, Zambezi, 32nm

Discussion

Comments currently: 60
Discussion started: 05/20/11 01:46:01 PM
Latest comment: 09/14/11 01:01:33 AM
Expand all threads | Collapse all threads

[1-18]

1. 
Uhm... Not quite convinced about going that route... The first thing that comes into mind is that BD is good, but not good enough with pricing alone.

I really hope I'm wrong about what I'm interpreting here, but that pricing chart is a lil' awkward for a "successful" processor line up IMO. Maybe I'm just accustomed to see Intel charging the hell out of everyone out there.

Cheers!
1 2 [Posted by: YukaKun  | Date: 05/20/11 01:46:01 PM]
Reply
2. 
I can't wait for the benchmarks.
4 2 [Posted by: FLA  | Date: 05/20/11 03:12:46 PM]
Reply
3. 
The highest performing Bulldozer core priced significantly low and this is a GOOD NEWS! However, if this low cost is relative to it's performance, like the very high price if Intel Core i7 990X with a very high performance, then that means Bulldozer doesn't deliver its expected performance... I hope I was wrong with my guess.
0 4 [Posted by: Numonyx  | Date: 05/20/11 04:05:19 PM]
Reply
- collapse thread

 
Well, it's priced to compete with Core i7-2600K at least, so at the very least it looks like they expect similar performance. If so, at least they closed a disgustingly vast performance gap, if not made something superior.

Honestly, with this new architecture, it's anyones guess as to what the performance comparison will be. My guess is that Bulldozer will shine better in servers where being highly threaded will be a bigger strength. In desktops, I still am curious as to whether or not they addressed the single threaded performance gap.
5 13 [Posted by: GaMEChld  | Date: 05/20/11 04:45:31 PM]
Reply
 
If Bulldozer delivers on threaded performance and has a processing power as Core i7 2600K on desktop, I'm sold. Only thing that would get me buying an Intel chip would be an 8 core version of Sandy Bridge, and that does not seem to happen until 2012 and only on the server side. Six cores will continue to be their top performance chips for premium desktops.
3 6 [Posted by: Filiprino  | Date: 05/20/11 08:40:40 PM]
Reply
 
The interesting thing is that with AMD's experience with the 6-core thuban, 12-core magny-cours, and soon to be 16-core interlagos, maybe AMD could release a 10-core desktop variant when the 8-core sandy bridge comes around. Now I'm just wildly theorizing.

Just out of curiosity, what do you need the 8-threads for that makes you less concerned with the single threaded performance? For my usage, I'm wondering I should stick to the 6-core version of bulldozer if it has more overclocking headroom.
5 12 [Posted by: GaMEChld  | Date: 05/21/11 09:25:44 AM]
Reply
 
Why are you "theorizing" ?
Check the roadmap and then you'll see a 6-10core Komodo planned for 2012.
4 2 [Posted by: Bingle  | Date: 05/23/11 03:12:48 AM]
Reply
 
Ah, touche. I was too lazy to look it up and didn't remember the future map from memory, but thanks for the info.
4 13 [Posted by: GaMEChld  | Date: 05/28/11 10:54:38 PM]
Reply
4. 
If the chip is any good, I would fear that resellers will just take advantage of this and start price gouging customers.
1 4 [Posted by: Kyrono  | Date: 05/20/11 08:21:01 PM]
Reply
- collapse thread

 
only if supply is tight, otherwise others will just undercut them.
1 3 [Posted by: Countess  | Date: 05/21/11 04:49:59 AM]
Reply
5. 
My guess is none of you read benchmarks on this site? Bulldozer has no chance at all of even hoping to touch Sandy Bridge in performance per clock. The areas where BD will be competitive are going to be heavily multi-threaded benchmarks like WPrime Pi calculations, Cinebench and video encoding (although with QuickSync on Z68 chipset, SB will still mop the floor here).

I mean just take a look at these benchmarks:
http://www.xbitlabs.com/a...7-2600k-990x_9.html#sect1

F1 2010, Starcraft 2, Metro 2033 with PhysX. AMD would need to increase performance per clock by about 40-50% just to be competitive.

Another indication that BD will not be as fast as SB per clock is because they are positioning an 8 core processor vs. a 4 core 2600k (Hyper Threading does not equal 2x as many cores!). And also, they are positioning a 6 core processor vs. a 4 core 2500k. So that means there is no way BD will be anywhere close to performance per clock.

The most reasonable estimation would be for them to aim at 1st generation Core i5/i7 performance per clock. Even if they achieve that, it will already be very impressive given the amount of cores they will have to work with.
4 6 [Posted by: BestJinjo  | Date: 05/20/11 10:11:22 PM]
Reply
- collapse thread

 
who cares about the per clock?
bulldozer is designed to clock higher with its 18 stage pipeline (vs 14 voor phenom2 and i7's)
in fact they both increased the pipeline AND increased the IPC that bulldozer can do vs phenom2.

i dont expect the early version to all reach the clocks AMD is expecting to hit eventually. its a new design on a new manufacturing node(32nm) for AMD after all.

also, much of the performance advantage intel had was down to the turbo. well, AMD's turbo is better. higher range with more cores. and it can even works when all cores active in certain loads.
0 3 [Posted by: Countess  | Date: 05/21/11 04:45:31 AM]
Reply
 
Since most programs are not multi-threaded still, performance per clock is what gives SB a significant advantage in modern applications such as games, photoshop, etc. Therefore, performance per clock / watt is of utmost importance unless all you do is video transcoding work or rendering.

Based on the benchmarks I linked, you can see that Phenom II is at about 40-50% slower per clock. Therefore, even if Bulldozer launches at 3.5ghz, it would need to be 50% faster per clock the Phenom II is today to match a Core i7 2600k in most tasks. That's the point.

Also, since Intel has the more mature 32nm process than AMD, it's also very unlikely that AMD processors will have as impressive power consumption at load, or the same overclocking headroom.
1 2 [Posted by: BestJinjo  | Date: 05/21/11 05:37:03 PM]
Reply
 
"performance per clock is what gives SB a significant advantage"
no its performance per clock at a certain clockspeed.
my point was that ipc alone isn't that important without knowing what the clockspeeds will be.

but you are comparing game performance. games are very specific and unpredictable type of load and they like a lot of cache, at a fast speed and low latencies.
the difference in per clock performance for other tasks is closer to 25% or less with the exception of Photoshop (which has always been highly intel optimized)
and i have to disagree with your statement that most programs aren't multi-core optimized yet. almost all programs were performance really matters and its doable to multithread have been optimized to a large extent.

but back to games.
bulldozers l1 is the same latency of sandy bridge but AMD's is larger. the l2 latency is higher (18 vs 11 cycles,) but AMD's is 8 times larger, or 4 times larger per core.
the L3 latency we dont know yet.(sandy bridges is 48 cycles) but given that its now going to run at the same speed as the rest of the CPU, unlike with the phenom2 its latency should see a very large improvement.
and bandwidth on all of them has been improved.
also AMD added advanced pre-fetches between the l1 and 2 and the l2 and 3 for the first time, reducing cache misses significantly.

then there is the much more advanced branch predictor and better decoder with micro-op fusion compared to phenom2.

most of these things aren't THAT important for most types of loads. but they play a large role in game performance and so bulldozer should be a significant step up from phenom2 in terms of gaming performance.
can it match sandy bridge? the first models should be close, and after that it will depend on how fast AMD can ramp up the clock speed.
their novel SOI+HKMG process should help them do that.
1 4 [Posted by: Countess  | Date: 05/22/11 11:02:47 AM]
Reply
 
You already know that Sandy Bridge can overclock to 4.7-4.8ghz with ease. You think Bulldozer will be able to clock higher than that on 32nm? The reason performance per clock is critical is because AMD doesn't have any more overclocking headroom compared to Intel.

If AMD's processors were clocked at 6.0ghz or something like that, then I could see the argument for not caring about IPC.

Also, I still think that most programs don't benefit from more than 4 threads at the moment. This is why we see a $115 Core i3-2100 dual core with HT still beat a Phenom II 955, which is a true quad-core.

http://www.xbitlabs.com/a...i3-2120-2100_4.html#sect1


Therefore, this tells me that IPC for 4 core processors are still more critical at least for the next 2-3 years than having a slower 6 or even an 8 core processor.

Obviously I welcome much more competitive processors from AMD to maintain strong competition, but I'd rather be conservative in my expectations for BD after the hype that Phenom I/II were before they launched (yes, I still remember how much hype there was around them).
0 4 [Posted by: BestJinjo  | Date: 05/23/11 01:17:02 PM]
Reply
 
the phonom2 can hit 4.2ghz on 45nm. (its actually still the current record holder for a overclocked quad-core)
the bulldozer has a longer pipeline, is made on 32nm, and uses SOI+HKMG.
its a new process with a new design so it will take a bit of time but yes they will defiantly clock higher then sandy bridge.

as for threads that's mostly about games not responding well to more threads.
that's what the turbo is for.
1 4 [Posted by: Countess  | Date: 05/25/11 05:31:24 AM]
Reply
 
The benchmarks on your link don't show anything that would suggest how the Bulldozer architecture will perform. Also, you say that AMD is positioning an 8 core CPU vs. a 4 core CPU. You're misleading people by saying that - AMD is positioning their 8 threaded CPU vs. Intel's 8 threaded CPU. Keep in mind that AMD's "8-core" statement is essentially based on the number of ALUs, there are still only 4 sections of L2 cache, 4 fetchers, and 4 decoders. AMD only calls it 8-cores for marketing purposes to emphasize that their SMT is much better than Intel's hyperthreading. I look forward to seeing real benchmarks, because currently hyperthreading is counterproductive to the types of workloads my servers run (we turn HT OFF at the bios level to avoid the degraded performance).
1 3 [Posted by: FlavusSnow  | Date: 05/21/11 09:58:30 AM]
Reply
 
Servers are where this architecture will really shine I feel, I'm not so sure how useful the massive amounts of threads will be for consumers. Developers are still slow to provide multi-threading support.
4 13 [Posted by: GaMEChld  | Date: 05/21/11 01:27:13 PM]
Reply
 
They aren't supposed to show you how a BD architecture will perform. But they show how much faster 1st and 2nd generation Core architecture is over Phenom II.

1st generation Core i7 is about 40-50% faster than a Phenom II X4 based on those Xbitlabs benchmarks. Here is a 2nd source that backs up these numbers:

http://www.pcgameshardwar...ks-with-18-CPUs/Practice/

You can see Core i7 920 @ 3.5ghz is 45% faster than Phenom II X4 @ 3.5ghz. We know that SB is another 15% faster per clock than 1st generation i7 processors are. Therefore, right now SB is at least 50-60% faster than Phenom II is.

Are you expecting BD to be 50-60% faster per clock than Phenom II is? That's not going to happen. We also know AMD processors don't have any overclocking advantages either.

Therefore, AMD's only competitive advantage will come from being faster in multi-threaded benchmarks (video encoding, rendering, etc.). Otherwise, SB will continue to be better for the average consumer (and gamer).

Of course anything is better than Phenom II. So it will beneficial for the marketplace to have a faster AMD and for AMD to be able to sell processors at higher prices (which will boost their profit margins and allow them to invest more $ into R&D for a better BD2).
0 3 [Posted by: BestJinjo  | Date: 05/21/11 07:47:53 PM]
Reply
 
How can you draw the conclusion that BD won't be 50-60% faster per clock? It is a completely new architecture. Based on your earlier post about the 8 "core" AMD vs 4 core Intel you haven't the slightest clue what has changed.

When Conroe was first released, intel improved per-clock speeds by well over 80%+ compared to their Pentium D line, and over 50% per-clock of what AMD was offering at the time.

So to say that it can't happen is just foolish. Whether it will or not we won't know until late june/july, but I think you are jumping to conclusions a bit early.
2 2 [Posted by: 10Ghz  | Date: 05/22/11 11:49:16 PM]
Reply
 
Think about it, it took Intel 2 full generations to go from Core 2 Quad to Sandy bridge to increase performance per clock 40-50%. Intel has the most advanced manufacturing facilities for CPUs in the world, with the most mature 32nm process too. So you are telling me AMD will be able to achieve the same performance, same power efficiency and same overclocking potential, all in just 1 generational jump? Not going to happen.
0 4 [Posted by: BestJinjo  | Date: 05/23/11 09:13:28 AM]
Reply
6. 
Quite honestly all I need is an inexpensive Processor which can beat my Q9450 @3.0 by a good margin and great price. Whichever one gives me much better performance and lower energy bill, compared to my current system, for the least amount of money will get my money period. That includes my next Graphic card, currently HD4890. I'm looking forward to benchmarks and actual market prices once they come out. Reading around and comparing is half the fun!
1 2 [Posted by: Upfront  | Date: 05/20/11 11:53:11 PM]
Reply
7. 
actually its very difficult to compare 8 bulldozer cores with conventional 8 cores. the bulldozer 8 cores arent really 8 physical cores rather its 8 modules which shares the fetch unit etc etc. so actually 8 bulldozer cores are a cross bw 4physical cores with ht & and real 8 cores. so 8 bulldozer cores or modules are actually like equivalent to 5.5-6 actual cores. Considering this, I feel AMD will still fall short of intel in per core per thread per clock performance & they are hoping to make up for it by adding more modules ("cores??" to counter that. That being said 8 bulldozer modules will be decidedly better at handling multi-threaded tasks than 4 SB cores with 4 logical cores from intel. All in all IMHO AMD will most likely come within 10% of the performance of intel's fastest 4 core k -series part with this generation of bulldozer modules. They are likely to give intel a run for their money with their next generation bulldozer parts when ivy bridge comes. Personally I cant wait to see the Athlon 64 vs Pentium 4 HT days again.
0 5 [Posted by: devvfata1ity  | Date: 05/21/11 12:19:28 AM]
Reply
- collapse thread

 
its 4 modules for a 8 core and the cores ARE physical, unlike the HT 'cores' from intel.
the AMD cores just share certain components.
in some workloads that wont have any effect, and in other the effect will be larger.
how exactly we'll have to wait and see.
1 5 [Posted by: Countess  | Date: 05/21/11 04:46:56 AM]
Reply
 
The only shared resources being shared are the Flex-FP units. Integer cores are fully functional and independent.
0 4 [Posted by: Filiprino  | Date: 05/21/11 06:51:49 AM]
Reply
8. 
Will be good if later AMD will develop 4 cores per module, greater than today 2 cores per module.
By that will upset a bit Hyperthreading of Intel in servers.

Is a bit late, but ok. They had to adopt hyperthreading from beginning (6-7 years ago) today they will have been developed in more advanced way the multithreading of Buldozer.
0 2 [Posted by: tbaracu  | Date: 05/21/11 03:29:59 AM]
Reply
9. 
Anything from the other camp is great news! I have been with AMD in the past and now Intel the past few generations purely because they offered the best performance,The downside is the lack of competition means us consumers pay top price for intel processors. So all I can say is go AMD,Hope it's successful for you so it pushes down Intel prices as well!
2 2 [Posted by: ozegamer  | Date: 05/21/11 04:49:51 AM]
Reply
10. 
Always is better more real core than Hyper thread.
0 11 [Posted by: Blackcode  | Date: 05/21/11 07:57:33 AM]
Reply
- collapse thread

 
In AMD's case, not really. they share fetch, decode and L2 functionality, & the fpu unit. When the cpu gets slammed with 256-bit fpu instructions the damm thing will act like a 4 core piecer, slowing the the cpu down. Then add muilti-tasking people that runs 8 or more heavy threaded apps on top of that, the thing will act like a high-end 4 core/ 8 core crippled cpu, because almost if not all the resources will be used. Also amd is also hoping GPGPU computing will give them more performance. In reality right now, there are't millions of apps for ATI gpu's yet. Overall there high-end 4-module is just really 6 cores / 8 core hindered crap.

Look at the low-power AMD E-350 based on bobcat. AMD had high hopes for it. Here's a quote that i found on techreport.com
AMD's bravado last summer in declaring that the Bobcat core would reach "90% of today's mainstream performance in less than half the silicon area" was fun to witness, but it was a bit too optimistic, in our view.


It fell short, period. It couldn't out due a lowly Pentium Extreme 840 cpu.

Wow!!! Its new but still crap, like intel atoms just 20-30% better.

AMD BULLDOZER architecture sucks all the way around. AMD is charging more for a strip-down CPU, like 1 1/2 cores per module, instead of 2 and they are still going to call it 2 cores just "marketing gimmick". Just to sell their hindered CPU's. I can't wait for the benchmarks to be revealed for AMD bulldozer, wait AMD bulls**t.

At least Intel has the balls to include 2 fpu's for each core, and reintroduce HTT back into their cpus, even though th P4 was a disaster. I think they did that, so people that bought intel cpu's with HTT, can multitask better than amd's cheap piecer's, and be that much better than AMD.

Die a slow & pain full death AMD. We still have VIA lol, they're still a joke though lol and still around. Wait i have a via chipset.... Opps lol
13 9 [Posted by: macintoshci  | Date: 05/21/11 01:23:16 PM]
Reply
 
Interesting you bash the E-350 when most everyone else gave it great reviews. I guess you should make yourself into a top rated reviewer, what with all the information and benchmarks you seem to have.
6 14 [Posted by: GaMEChld  | Date: 05/21/11 01:31:40 PM]
Reply
 

To Stop your nonsense:
E-350 is not based on Bulldozer Arq, it's based on Athlon 64/K8 arq.

Destroys Atom, everyone praises them. Good try troll.
8 13 [Posted by: Nintendork  | Date: 05/21/11 05:11:40 PM]
Reply
 
Hey least your right about one thing... The architecture name it is bobcat, and now i will quote from X-bit labs Article: "AMD Brazos Platform and AMD E-350 (Zacate) CPU Review" page 1, paragraph 5, line 2

"The Bobcat is the first new microarchitecture to be embodied in silicon, namely in two processor series codenamed Ontario and Zacate."


So its not base off of anything for all i care AMD FAN BOY. Its a new architecture. If you really want to go where a small glimpse of where it is based off, then from same the article on page 2

"The outcome looks more like AMD’s K6 processor rather than like a modern Phenom."


and that doesn't even prove its from based off of the K6 architecture

So in your case its not a K10, K7 or from your Quote
"Athlon 64/K8"

Its not that either.

Get your facts straight wannabee. Thank you for the correction on the simple architecture name though.... Not lol
12 6 [Posted by: macintoshci  | Date: 05/21/11 06:06:12 PM]
Reply
 
It wasn't a simple mistake on your part. It would be the equivalent of mistaking Atom for Sandy Bridge.
4 14 [Posted by: GaMEChld  | Date: 05/21/11 11:27:18 PM]
Reply
 
It was a simple mistake, on my part.... But with both of you... and i quote
"E-350 is not based on Bulldozer Arq, it's based on Athlon 64/K8 arq."


When you miss by an inch you, you miss by a mile, Nintendork & GaMEChld. You had the huge mistake, & i proved it, with facts to back me up. You had nothing, except your pathetic self....

I'm not dissing your E-350 thing though for that but when you pair an atom & a ion up. While doing 1080P res gaming. The atoms stomps the E-350... And i refer to thee quote & to my quote. To reiterate my point.

Thee Quote
AMD's bravado last summer in declaring that the Bobcat core would reach "90% of today's mainstream performance in less than half the silicon area" was fun to witness, but it was a bit too optimistic, in our view.


My Quote
It fell short, period. It couldn't out due a lowly Pentium Extreme 840 cpu.


My prediction will be that the AMD bulldozer architecture. Is going to fail at any real performance gains...

Go stick it Wannabee
11 7 [Posted by: macintoshci  | Date: 05/22/11 12:04:14 AM]
Reply
 
1080p res gaming on an atom/bobcat, whichever processor comes up on top is meaningless.
2 9 [Posted by: wh3resmycar  | Date: 05/22/11 12:20:23 PM]
Reply
 
It tells me that one of the components in the E-350 is a bottle neck when doing hi-res gaming.

If it is meaningless, how come tech websites and or companies do benchmarks on a broad range of computer hardware??? Why do people make benchmark programs in the first place???
Because its meaningless. Think again. it does matter. Its how educated consumers make an inform decision.

Meaningless or not to you "wh3resmycar". it does matter. It does matter to someone, other than your meaningless "wh3resmycar" self.
12 5 [Posted by: macintoshci  | Date: 05/22/11 05:15:30 PM]
Reply
 
Whoa there champ, it was someone else who was arguing with. So you go stick it you over zealous asshat. All those quotes you referenced and was responding to were from someone else. Keep making little mistakes all in a row buddy, it does wonders for your credibility. First wrong chip names, then responding to wrong user names. Sheesh...
4 13 [Posted by: GaMEChld  | Date: 05/28/11 10:59:30 PM]
Reply
 
Get your facts straight. I was addressing you both, and anyone who reads this stuff. As you can see with my comments, i'm open to correction. Unlike your mean-less self.

wrong chip names


What chip names??? I didn't know a chip had more than one name... lol I know a family of chips has an architecture name, then a chip within a family has it own designated name to separate them out between the family of chips. But a chip that has 1+ names in the same family...

Talk about a DUMBASS....

GaMEChld:wrong chip names, then responding to wrong user names.


You haven't said anything that can back your claims up yet, poorAMDfag.
12 5 [Posted by: macintoshci  | Date: 06/02/11 11:40:58 AM]
Reply
 
Learn how to read, dumb troll. They didn't reach "90% of today's mainstream performance" They failed at that. I'm predicting that AMD bulldozer, wait AMD bulls**t, will be the same bull**** from amd, and it will fail to reach any "REAL" performance gains. Like in the example of the AMD E-350.
12 6 [Posted by: macintoshci  | Date: 05/21/11 11:35:36 PM]
Reply
 
U act like fanboy I mean always better is real core than hyperthread, not only in Intel case ,AMD also.Hyperthread have security issue.
2 14 [Posted by: Blackcode  | Date: 05/21/11 11:38:40 PM]
Reply
 
How so, what does hyperthreading have to do with security holes?
1 11 [Posted by: jumpingjack  | Date: 05/22/11 08:33:24 AM]
Reply
 
http://www.crn.com/news/s...-threading-technology.htm

Security Threat Seen In Intel's Hyper-Threading Technology
0 14 [Posted by: Blackcode  | Date: 05/22/11 08:43:25 AM]
Reply
 
That just shows a one side of the story. here's the link to the actual paper, & plus that news is dated back in 2005, things have progress since than.
http://www.daemonology.net/papers/htt.pdf

Here's the 1st two paragraph's from the paper of "CACHE MISSING FOR FUN AND PROFIT" BY COLIN PERCIVAL
Abstract.

"Abstract. Simultaneous multithreading | put simply, the sharing of the execution resources of a superscalar processor between multiple execution threads | has recently become widespread via its introduction (under the name "Hyper-Threading" ) into Intel Pentium 4 processors. In this implementation, for reasons of efficiency and economy of processor area, the sharing of processor resources between threads extends beyond the execution units; of particular concern is that the threads share access to the memory
caches.

We demonstrate that this shared access to memory caches provides not only an easily used high bandwidth covert channel between threads, but also permits a malicious thread (operating, in theory, with limited privileges) to monitor the execution of another thread, allowing in many cases for theft of cryptographic keys."


Also from Section 7 Subsection 1 on pages 10 & 11. I quote


"7. Recommendations
Our recommendations are as follows:
1. CPU designers should, on all future processors which implement simultaneous multithreading, use cache eviction strategies which re-spect threading and minimize the extent to which one thread can evict data used by another thread. Similarly, "multi-core" processors should either avoid sharing caches between the processor cores or use thread-aware cache eviction strategies."


And if if you look at lines 6 and 7 and i quote


"Similarly, "multi-core" processors should either avoid sharing caches between the processor cores"


That means any multiple-core cpu whether its from AMD or Intel that shares caches resources has this very same security threat that the the old Pentium 4 HT has.

Don't worry though modern processors have Hardware enforcement of "Data Execution Prevention". Also Moderns OS's have or should have software enforcement of DEP.

Next time... Cite a relevant article, or an article that shows all sides of the problem and or story.

Either you guys are pathetic fan-boy wannabes trolls, and don't know anything that you are talking about. Which makes you guys worthless in the first place.

Or you guys, did a shortcut. Took a 3-third party source, like that website article as an example that only showed a small part of a one sided of the story and took it as fact.

Or Both

I'm going for both.....
12 4 [Posted by: macintoshci  | Date: 05/22/11 04:53:40 PM]
Reply
 
I agree... more cores are better, but more cores with Hyper-threading is almost always better, than just more cores.
13 3 [Posted by: macintoshci  | Date: 05/22/11 05:19:44 PM]
Reply
 
but almost double the amount of real cores in just slightly more space (excluding l3 cache) then a core+ht is far still.
0 8 [Posted by: Countess  | Date: 05/23/11 02:21:26 AM]
Reply
 
According to Intel the first implementation only used 5% more die area than the comparable non-hyperthreaded processor, but the performance was 15–30% better.


So if intel used the same exact design or a tweak design of AMD bulldozer, with an HTT unit or units, and both ran at the same speed the intel version will be 15-30% faster or even more because that 15-30% was with the P4's I'm sure the htt unit got upgrades through the years.

Which leads me back to my main point. More cores are better, but more cores with Hyper-threading is almost always better, than just more cores.

This part is for stupid people alike. This is in theory, i know that Intel won't use Amd designs.... lol There would be entangle in legal battles if they did.
12 3 [Posted by: macintoshci  | Date: 05/23/11 08:08:24 AM]
Reply
11. 
Is there any software/OS combination that can exploit 8 cores?
0 3 [Posted by: jb3177  | Date: 05/21/11 10:15:11 AM]
Reply
- collapse thread

 
there are machines with 48cores, so yes 8 cores are supported some some combinations.

as for on the desktop, yet there 2. software today is often made multithreaded, which means it doesnt matter how many cores you have, it will try to use all of them.
0 2 [Posted by: Countess  | Date: 05/23/11 02:23:14 AM]
Reply
12. 
Any video/audio/render/image software can fully utilize them unless youre one those sheep that use single thread itunes.
1 14 [Posted by: Nintendork  | Date: 05/21/11 05:13:03 PM]
Reply
13. 
My guess is that the performance will be on pair with an Core i7-2600K, therefore the similar price. Cannot wait for the realese, maybe Intel will finally cut the price of all I7 procs.
2 1 [Posted by: TAViX  | Date: 05/22/11 10:21:57 AM]
Reply
14. 
If the numbers match up with intel then AMD have a real winner on there hands.. only 2 weeks until you know for sure! when they come out on the 7th on June at E3
2 2 [Posted by: vid_ghost  | Date: 05/22/11 06:18:30 PM]
Reply
- collapse thread

 
I can't wait! finally a release date for bulldozer.... I hope.
12 1 [Posted by: macintoshci  | Date: 05/22/11 08:43:17 PM]
Reply
15. 
Donainim Haber got the story from here: http://news.softpedia.com...ealed-Report-201335.shtml
1 0 [Posted by: djklax  | Date: 05/22/11 11:04:42 PM]
Reply
- collapse thread

 
thank you...
10 0 [Posted by: quad840  | Date: 05/24/11 12:19:50 PM]
Reply
16. 
Its really kind of sad to see just how threatened Intels fan boys are. I think they maybe crossing into the apple fan boy zone where they treat their hardware choice like a religion.
1 15 [Posted by: DEADTIME  | Date: 05/23/11 04:49:34 PM]
Reply
- collapse thread

 
"Another crap from poorAMDfags. :D"
13 5 [Posted by: quad840  | Date: 05/23/11 07:53:22 PM]
Reply
 
I think you just might represent the average Intel fanboys IQ on this board. You set the bar low and came in beneath your expectations.
2 13 [Posted by: DEADTIME  | Date: 05/24/11 01:22:52 AM]
Reply
 
You set the bar low and came in beneath your expectations.

Well thank you... poorAMDfag. You had no bar to begin with. Get one next time!
12 3 [Posted by: quad840  | Date: 05/24/11 12:27:24 PM]
Reply
17. 
Its even sadder to see how AMD fanboys react. Technology delayed to match what will be yesterday's processors in a few months is the salvation AMD fanboys are looking for. You can all brag about "if this matches Intel's stuff then we got a winner" all you want, but face the reality: AMD was ONCE the king in single core processors but has been inferior ever since. The best use only the best and not some catch-up series of processors that look good on paper, err, when actually used yields far less than the paper specs and here comes someone saying- "at least its cheaper than the other guy's stuff" (are you seriously making that statement?).

AMD, give me something that will best Intel's next generation and you will receive an entire department's worth of purchases for that. Until then, we're sticking with what we know works, and won't compromise with these new "old market" devices. As for your enterprise class hardware, you're years behind and losing ground there faster than you're making it elsewhere. Where's the competition?
4 0 [Posted by: thoufand  | Date: 06/15/11 02:16:27 AM]
Reply
18. 
The FX-8150 is $266 right now.
0 2 [Posted by: jewie27  | Date: 09/14/11 01:01:33 AM]
Reply

[1-18]

Add your Comment

[Login] [Forgot password?] [Registration]




Related news

Latest News

Thursday, May 24, 2012

10:00 pm | Microsoft Clarifies Its Exec's Claims: 500 Million Windows 8 Copies in 2013 Are "Potential" Upgrades. Microsoft Retracts Statement Regarding 500 Million Windows 8 Licenses to Be Sold in 2013

8:35 pm | ECS Reveals "NonStop" Mainboard Family with Extended Stability and Reliability. ECS Boosts Stability and Reliability with Premium Components and Rigorous Testing on NonStop Platforms

6:28 pm | AMD Rumoured to Start Production of Next-Gen FX-Chips in Q3. AMD to Start Making FX "Vishera" Chips Next Quarter

11:32 am | UMC Begins to Build Fab to Make 28nm, 20nm and 14nm Chips. UMC Spends $8 Billion on Expansion, Confirms Development of 14nm FinFET Process Technology

Wednesday, May 23, 2012

11:17 pm | OCZ Quietly Introduces 7mm Vertex 3 LP SSDs for Ultrathin Laptops. OCZ Reveals Low-Profile Vertex 3 LP Solid-State Drives

11:01 pm | Toshiba Vows to Release Hybrid Hard Disk Drive in September. Toshiba Confirms Development of Hybrid Hard Disk Drive

10:28 pm | Jury Finds: Google Android Did Not Infringe Oracle's Patents. Google Not Guilty of Infringing Oracle's Patents

8:48 pm | 208 Million "Phablets" Set to Be Sold by 2015 - Analysts. Smartphones with Ultra Large Screens Will Continue Gaining Popularity

8:23 pm | Microsoft Expects 500 Million People to Use Windows 8 Next Year [UPDATED]. Microsoft Projects Breakthrough for Windows 8 Sales Thanks to Compatibility with Different Devices

7:07 pm | Thunderbolt Interface to Receive Green Light from Mainboard Makers in Second Half of 2012 - Rumour. Thunderbolt Interconnection to Make It to Mainstream Systems Later in 2012

6:18 pm | Apple and Samsung Fail to Reach Settlement at Special Meeting. Court-Ordered Meeting Does Not Resolve Apple Vs. Samsung Dispute

12:10 pm | Seagate Set to Take Over LaCie. Seagate and LaCie Agree About Acquisition

more news >>>

Our Sponsors

Crucial.com by Micron Technology, Inc. Kingston Technology