Information

Did you know you could become an X-bit labs fan on Facebook or follow us on Twitter?



Discussion

Discussion on Article:
AMD Needs to Boost Clock-Speed of FX "Bulldozer" Chips - Sources.

Started by: RtFusion | Date 05/30/11 01:11:38 PM
Comments: 85 | Last Comment:  01/03/13 10:13:10 AM

Expand all threads | Collapse all threads

[1-20 | 21-21]

1. 
http://www.anandtech.com/...90fx-and-z68-motherboards

"ust above Llano we will have the long awaited Bulldozer CPU. AMD originally wanted to launch Bulldozer at Computex but performance issues with its B0 and B1 stepping chips pushed back the launch. Now we're looking at a late July launch with B2 silicon, but performance today is a big unknown. Apparently the performance of B1 stepping silicon doesn't look too good."


Hmmmmm.
0 1 [Posted by: RtFusion  | Date: 05/30/11 01:11:38 PM]
Reply
2. 
This reminds me soo much of the first gen Phenoms and the problems they faced with their clock speed limits.
14 2 [Posted by: SteelCity1981  | Date: 05/30/11 02:03:40 PM]
Reply
3. 
"This will slowdown revenue growth of the chip developer and will also hit its reputation"

http://www.xbitlabs.com/n...dge_Processors_Slide.html

No mention of the same thing in the Ivy Bridge delay article only 4 days ago. I wonder why, who's paying your rent Shilov?

The laughable thing is, intel has said Q1 for ivy bridge on loads of ocassions, while AMD only ever said 2011 for BD. Get your shit together Shilov.
14 7 [Posted by: jimbo75  | Date: 05/30/11 02:04:11 PM]
Reply
- collapse thread

 
Firstly, unlike AMD, Intel does and will have competitive processors for the high-end desktop CPU market. Small delay means nothing for their financial results or image among end-users.

Secondly, Bulldozer has been delayed for several times already, unlike Ivy Bridge.

Thirdly, AMD did say that the desktop Zambezi was due in Q2 2011 and server Valencia/Interlagos were set to arrive in Q3 2011.
http://www.xbitlabs.com/n...r_Chip_at_Conference.html

AMD originally expected to start shipments of Bulldozer in 2010, not in the second half of 2011: http://www.xbitlabs.com/n...ples_Due_in_2009_AMD.html
7 6 [Posted by: Anton  | Date: 05/30/11 07:03:05 PM]
Reply
 
Yeah right. AMD has 3 brand new archictectures launching this year and 2 of them are already a great success. Intel does a minor speed bump on nehalem, loses half of their netbook market to bobcat and then messes the IB launch and all you can do is have a go at AMD for one arch that hasn't gone perfectly.

You know what they say, 2 out of 3 aint bad.
11 6 [Posted by: jimbo75  | Date: 05/31/11 04:37:30 AM]
Reply
 
You're really not calling Sandy Bridge a minor speed bump on Nehalem, are you? No one is that stupid. Or are they?

Sandy Bridge is a completely new microarchitecture. It's pipelines are redone from the ground up, and it's got a lot of Pentium 4 in it, not the least of which is the Physical Register File, and a new type of trace cache.

Nehalem is the last of the Pentium Pro line. Sandy Bridge is a new architecture.

Having said that, I like Bobcat, and anyone who thought Bulldozer was going to compete on a thread level with Sandy Bridge was an idiot anyway. It doesn't have to, to be successful. It will be very cost/power effective with integer apps with many threads, but it will not compete in single threaded performance.
6 1 [Posted by: TA152H  | Date: 05/31/11 08:40:02 AM]
Reply
 
What are the three architectures? I know that Bulldozer and Bobcat are the two brand-new micro-architectures that are released this year.

Bulldozer was due in 2009 - 2010 timeframe.
Bobcat was due in 2008 - 2009 timeframe (http://www.xbitlabs.com/n...splay/20070608225159.html).

Llano, an interesting chip design with outdated K10.5+ micro-architecture, was originally scheduled for release in early 2011 (http://www.xbitlabs.com/n...duction_in_Late_2010.html), not mid-2011.

What I really don't know is how to "measure" the success of Bulldozer, Bobcat and Llano, provided that only Bobcat has been released so far commercially?
3 4 [Posted by: Anton  | Date: 05/31/11 12:44:44 PM]
Reply
 
Jesus you are really pedantic aren't you. If you are going to hark back to 2008 and 2009 why not go write a article on how Larrabee is late but "coming soon". Or are you only pedantic when it comes to AMD?

And really don't even try to convince anyone that Llano isn't a far more ambitious step than SB was.
9 4 [Posted by: jimbo75  | Date: 05/31/11 02:15:39 PM]
Reply
 
"while AMD only ever said 2011 for BD. Get your shit together Shilov." jimbo75

The SSE5 specification, which is being made available to the developer community today at http://developer.amd.com/SSE5, will be implemented in products based on AMD’s next-generation “Bulldozer” core, available in 2009.

http://www.amd.com/us/pre...Press_Release_119091.aspx

Seems the ones who have to get their act together is AMD and jimbo...
2 2 [Posted by: Xtractor  | Date: 06/01/11 12:53:43 AM]
Reply
 
When I am accused of bribing and asked to get my "shit straight", there are two options for me:

- ban you and delete your account without any further ado.

- show you that you don't know the facts.

I chose the second option and it looks like you really do not know all the facts surrounding various chips.

Larrabee as we know it is dead. It is officially admitted. Just like AMD's Swift, Montreal, etc.

Llano and Sandy Bridge are comparable. Llano has relatively new GPU core and old CPU cores, whereas Sandy Bridge has old GPU core and new CPU cores.
2 3 [Posted by: Anton  | Date: 06/01/11 02:47:05 AM]
Reply
 
Lol swift and montreal now as well? Why not just keep going back Shilov, then we can talk about Netburst and Itanium. Why not just go all the way back to intels FDIV? Or did you want to stop at 2007 because that suits your agenda more?
9 3 [Posted by: jimbo75  | Date: 06/01/11 04:24:59 AM]
Reply
 
Itanium makes more money than AMD's whole server business. So I guess you mean AMD's server products are a failure as a whole.
2 4 [Posted by: Xtractor  | Date: 06/01/11 05:05:10 AM]
Reply
 
BTW instead of trying to figure it out while we both throw AMD and intels failures at each other, the point is you chose to single out AMD for criticism when both have had many faults over the past few years.

The second point - the financial one - is even more bullshit. Bulldozer would only be worth 10% revenue to AMD in 2011 had it been released this month, so lets say that's down to 5% worse case. Ok they lose $300m, not really because they will just flog more X6's they have in inventory but still, it's a small loss overall.

How much more do you think intel is set to lose with a 3 month delay (at least) on ALL of their products slower crossover to 22nm? It's a lot more than what AMD is set to lose and that's for sure.

Drop your websites negative slant on all things AMD.
8 4 [Posted by: jimbo75  | Date: 06/01/11 04:48:52 AM]
Reply
 
Intel will loose much less with the 3month delay than AMD's 2 year delay of bulldozer caused them or the last 5 years straight in red.

At least Intel still owns it's fabs.
2 3 [Posted by: Xtractor  | Date: 06/01/11 05:07:41 AM]
Reply
 
Yeah that's why it's desperately trying to hawk supply to apple and others. It won't be long before a lot of those fabs are sitting empty.
7 4 [Posted by: jimbo75  | Date: 06/01/11 05:18:49 AM]
Reply
 
What you do not know is that prior to acquisition of ATI, AMD planned to aggressively expand its production capacities and considered making chips for others in order to keep them busy...

BTW. Intel not only plans to make some custom chips for others, but also hopes to expand its product lineup further to address emerging markets like smartphones, tablets and other new form-factors.

I suggest you really should start reading various news-stories at this and other web-sites before trying to argue or criticize.
2 4 [Posted by: Anton  | Date: 06/01/11 05:57:42 AM]
Reply
 
No offence but if I wanted actual news this would be the last place I'd visit. If I wanted to read pro-spintel propaganda I'll make sure to keep you bookmarked though.
8 2 [Posted by: jimbo75  | Date: 06/01/11 06:02:19 AM]
Reply
 
I heard it already so many times that Intel will be destroyed by AMD... and look who got bought up by an investment group.
2 3 [Posted by: Xtractor  | Date: 06/01/11 06:16:51 AM]
Reply
 
When Intel fails to deliver products on time or at all, it receives its amount of criticism.

The problem is that 60 day Ivy Bridge delay from January to March is less important for Intel than another 90 days delay of Bulldozer for AMD.

The reason is obvious. AMD's server business is going down. AMD's most expensive desktop chip today costs $205. The most expensive FX chip was supposed to cost $320. Have you ever heard of profit margins? Do you understand why AMD is not profitable?

I can go back as far as Intel Timna and imply that this was supposed to be the first x86 APU...

How much more do you think intel is set to lose with a 3 month delay (at least) on ALL of their products slower crossover to 22nm.


Enlighten us, please! With actual numbers. Just keep in mind that by the end of 2012 the company will have five fabs producing chips at 22nm, which will speed up transition to 22nm in general.
2 4 [Posted by: Anton  | Date: 06/01/11 05:53:45 AM]
Reply
 
AMD IS profitable FYI, maybe you should check the recent financials?
8 2 [Posted by: jimbo75  | Date: 06/01/11 06:06:21 AM]
Reply
 
Then how about aking a look at the last 5 years as a whole ?
2 3 [Posted by: Xtractor  | Date: 06/01/11 06:22:28 AM]
Reply
 
It shows profitability because it either receives payments from Intel, sells its stock in Globalfoundries or receives other kind of payments.

Otherwise, the company either shows minimum profits or is not profitable.

So, did you count how much money will Intel lose by delaying desktop Ivy Bridge chips by 60 days? Did you count how many quarters will it take Intel to transit its major products to 22nm given the fact that it will have 22nm fabs by late 2012?
2 2 [Posted by: Anton  | Date: 06/01/11 06:32:38 AM]
Reply
 
And what was the reason for those payments from intel? Do you even understand that the reason AMD had such a torrid time was precisely because of the crap intel pulled?

When you are busy lauding intel's profits did you count the $5 billion spent on bribing Dell?

Or the many billions spent on fines?

Or the billions wasted on Larrabee?

Or the $8 billion on McAfee? That will take 4 years to pay off at current levels assuming nothing goes wrong that is.

No I guess not. And no I cant be arsed figuring out how much money intel will lose because of the IB delay, but it's more than what AMD will lose on the BD delay for sure. A lot more.
8 3 [Posted by: jimbo75  | Date: 06/01/11 08:41:10 AM]
Reply
 
Wrong again.

The 2 year delay for bulldozer pushed AMD's whole lineup under 200$ and their server market share under 10%. Intel makes 25% more on Itanium than AMD's whole server business thanks to this massive delay.

The past 5 years of continuous losses had nothing to do with dell intel.
Leaky 65nm, badly executed over hyped K10, not being able to match Conroe performance for years, are the main reasons.
2 3 [Posted by: Xtractor  | Date: 06/01/11 11:14:18 AM]
Reply
 
Intel has spent more on fines and bribes in the past 7 than AMD has lost in its entire 42 year history. As far as wasting cash on useless projects goes, it's far ahead of AMD's cash losses.

That's how "awesome" a company intel is. It wastes more money that it puts to good use.
7 2 [Posted by: jimbo75  | Date: 06/01/11 11:56:28 AM]
Reply
 
Then it's probably just a coincidence most big warehouses still sell only or mainly Intel-based PC's, even in those sectors AMD provides a competitive or even a better alternative?

Also, it costed them a fortune to buy ATI, which is turning into a good business only by now.
2 1 [Posted by: dezz  | Date: 06/01/11 05:35:17 PM]
Reply
 
It is profitable even without those payments, since more than a year.
2 2 [Posted by: dezz  | Date: 06/01/11 05:18:19 PM]
Reply
 
They even (Intel) have said 2011Q4 lately...
2 2 [Posted by: dezz  | Date: 05/31/11 04:34:57 AM]
Reply
 
Really?...Fanboy!
1 1 [Posted by: jonup  | Date: 05/31/11 05:07:30 AM]
Reply
4. 
Bulldozer was scheduled to launch in 2009. So Ivy seems pretty good compared to that.

I really hope after 4 years of waiting bulldozer will blow us away... Another K8 would be great but, the 320$ top end FX price seems quite different from the 1000+$ of the K8 era, so I guess not.
2 3 [Posted by: Xtractor  | Date: 05/30/11 03:15:38 PM]
Reply
- collapse thread

 
Don't we all wish for AMD to be more competitive?

Sandy Bridge has a larger IPC advantage over Phenom II. Add to that Intel's mature 32nm process that allows their 2600k to overclock well into 4.8ghz, and even more competition coming in from SB-E LGA2011 socket, and AMD is in a tough spot. The main draw of AMD's new processors will likely be cheap mobo bundles and those users who absolutely need to use 8 threads for rendering/encoding etc. I can see BD beating SB in multi-threaded tasks.

But even then Ivy Bridge may still beat BD should AMD have any chance at competing with SB. So AMD will still be 2nd best.
15 4 [Posted by: BestJinjo  | Date: 05/30/11 07:06:10 PM]
Reply
 
Well, The FX-51 was "only" $600 or so when it came out, right? AMD desktop chips didn't officially hit the $1K mark until FX-57 or FX60 IIRC. $320 sounds rather sweet by comparison.
1 1 [Posted by: cheeseman  | Date: 05/31/11 06:28:57 AM]
Reply
 
That one wasn't the same chip as the new one. That one used a 45nm process, the new one is for 32nm. Also, there can be substantial changes to the developed microarchitecture.
1 2 [Posted by: dezz  | Date: 05/31/11 05:14:11 PM]
Reply
 
You are just trying to justify the delay.
Instead of 2 years it takes 4 for AMD to release their next uarch...
2 3 [Posted by: Xtractor  | Date: 06/01/11 12:55:42 AM]
Reply
 
Justifying or not justifying, I'm just saying that what they wanted to launch two years ago is not the same thing they will launch by now, and probably this one is as much the better as more the time it needed to develop, not mentioning it's 32nm, which is an advantage in itself.

Do you think Intel have built SB from scratch?
1 2 [Posted by: dezz  | Date: 06/01/11 05:52:10 PM]
Reply
5. 
Toms Hardware said that motherboard companies had B0 samples in an article today. However,it seems know that unknown sources have B1 CPUs too!!

It could all be true or not.
0 2 [Posted by: USER2000  | Date: 05/30/11 03:48:44 PM]
Reply
6. 
This leaked info seems to follow perfectly after Intel said it was delaying its 22NM processors. Again,it could be coincidence.
0 2 [Posted by: USER2000  | Date: 05/30/11 04:02:28 PM]
Reply
7. 
Am I the only one reading this article thinking "What happened to you AMD?" 2.5 ghz? Puh-lease you can argue as much as you want that the ghz race is over but I'll take an Intel 3.0+ any day when AMD needs 40% more clock speed just to give equivalent performance as Intel.
2 4 [Posted by: AnonymousGuy  | Date: 05/30/11 05:05:05 PM]
Reply
- collapse thread

 
500 MHz hasn't been 40% of most CPU's clock speed since PIII. The Phenom II X4 980 performs as well as the Core i7 960 without HyperThreading. With HyperThreading a better match for the X4 980 is probably the i5 760. They're priced about the same and they're both old mid-range products so PLEASE intel fanboys don't come over yelling "haha AMD can barely compete with a two year old Core i5" because you'll all be crying when the 4 GHz STOCK AMD Zambezi FX (Bulldozer) 8150P octo-core CPU comes out.
4 6 [Posted by: elcommenter  | Date: 05/30/11 06:50:57 PM]
Reply
 
show the post
3 7 [Posted by: AnonymousGuy  | Date: 05/30/11 11:02:38 PM]
Reply
 
Are you aware that the Bulldozer is a brand new microarchitecture? So, better not to base your expectations on Phenom II's (which is basically an enhanced Athlon64 with more cores). Yes, they can do it, provided it reaces the indended clockrates, which it should, sooner or later.
5 2 [Posted by: dezz  | Date: 05/31/11 04:43:04 AM]
Reply
 
It's sad/funny how some posts are, like this one. You discredit AMD with this type of post.

Bulldozer IS NOT intended to compete head on with Sandy Bridge in single threaded performance. AMD knew they couldn't win that battle, so made Bulldozer with an eye on performance per watt on threaded applications that did mostly integer.

So, stop whining, because AMD will not beat Sandy Bridge, and it doesn't have to. But, it will perform well enough for most people, and if you've got something needing 8 threads, and AMD 4 module/8 integer core unit is going to offer you a solution perfectly suited for your application. Or even a four thread. Intel can throw in a big four core unit, or hyper thread a 2-core, but the former is bigger/more power hungry, the latter will not perform as well, as the AMD unit.

Absolute performance is NOT the end all. Bulldozer will lose that battle, but will win in other areas, and that's fine for AMD. They win at nothing now in that market.

Also, despite these clock speeds, you should expect the Bulldozer to clock higher than Sandy Bridge based on the design. AMD designed this chip for high clock speeds, and when they work out all the warts, it should clock quite high.
8 2 [Posted by: TA152H  | Date: 05/31/11 08:46:37 AM]
Reply
 
Note that Sandy Bridge got its high single-thread performance out of turbo (f.ex. substantially higher clock when only one core is working hard). A thing the BD also has - with even higher clockrates!
2 2 [Posted by: dezz  | Date: 05/31/11 04:59:32 PM]
Reply
8. 
I love this stuff "I'm not a fanboy your a fanboy.. fanboy" I mean please anyone who is talking up AMD at this current time is a fanboy. They are behind right now in process and in design. Maybe this processor will catch them up in one category but it would take a miracle to overcome the advantage intel has right now. Making billions off of a virtual duapolly with microsoft gave them this lead and they are not a stupid bunch that will waste it away easily.
4 4 [Posted by: Davinchy  | Date: 05/30/11 08:10:10 PM]
Reply
- collapse thread

 
No, not miracle just technology and creativity, moreover courage.
3 3 [Posted by: dezz  | Date: 05/31/11 04:46:01 AM]
Reply
9. 
show the post
2 15 [Posted by: Azazel  | Date: 05/30/11 08:29:15 PM]
Reply
10. 
Unless someone is able to stump up 50 billion dollars for AMD RnD I don't see AMD beating Intel anytime soon even with the new BD architecture... I'd really like to believe AMD could do it again but things aren't looking to good so far
3 1 [Posted by: Athlonite  | Date: 05/30/11 11:42:13 PM]
Reply
11. 
screw it. i was going to wait but I'll just build a rig with SB. there no point in waiting anymore.
5 4 [Posted by: goury  | Date: 05/31/11 12:34:31 AM]
Reply
12. 
If this delay is confirmed, I might just go ahead and build myself that i5-2500K rig that I have been eying for a few weeks now. Was just waiting for the Dozer, but another three months' delay. That is too much for me to take....
5 3 [Posted by: psycho_mccrazy  | Date: 05/31/11 04:24:10 AM]
Reply
13. 
From now on, I'll always support second player to give they alive.
3 1 [Posted by: beautyless  | Date: 05/31/11 05:33:08 AM]
Reply
14. 
http://www.anandtech.com/...79-and-llano-motherboards

Anand says B1 easily hits 3.8 GHz Shilov. I know who I'd believe first - what's it like to be so far out of the loop? Did you and Theo make that 2.5 GHz number up all by yourselves?
13 2 [Posted by: jimbo75  | Date: 05/31/11 07:00:42 PM]
Reply
- collapse thread

 
It seems Toms Hardware said it was B0 samples which were tested not B1 samples. Anandtech and xbitlabs also said B1 samples had problems too.
0 2 [Posted by: USER2000  | Date: 06/01/11 03:31:28 AM]
Reply
 
If they haven't any problems then they should show us the cpus!
3 1 [Posted by: Tester128  | Date: 06/01/11 04:53:59 AM]
Reply
15. 
It's official, they admit it at Computex they had to delay bulldozer one more time...
3 4 [Posted by: Xtractor  | Date: 06/01/11 08:00:22 AM]
Reply
- collapse thread

 
Don't twist their words, they didn't said they "had to".
2 2 [Posted by: dezz  | Date: 06/02/11 04:32:38 PM]
Reply
16. 
http://semiaccurate.com/f...115650&postcount=1484

AMD Europe says BD has no issues and it's a simple strategic decision to hold off on release.
8 3 [Posted by: jimbo75  | Date: 06/01/11 01:05:27 PM]
Reply
- collapse thread

 
show the post
2 5 [Posted by: Anton  | Date: 06/01/11 05:49:27 PM]
Reply
 
You should ask them. Of course AMD doesn't waste time talking to you so you'd be better off making a story up by yourself instead. Ask Theo for help if you get stuck.
14 3 [Posted by: jimbo75  | Date: 06/01/11 06:04:09 PM]
Reply
 
Do you think Llano is for nettops???
Do you think Opteron is for netbook???

Do you really hate AMD that much as it seems?
5 2 [Posted by: dezz  | Date: 06/02/11 04:11:52 PM]
Reply
17. 
I am not so sure that boosting clock speed and power dissipation is what is needed.
To the user, is there so much improvement going from 95% idle to 98% idle?
I think overall performance with multitasking and graphics would be a better selling point.
1 1 [Posted by: jetgraphics  | Date: 06/02/11 07:06:37 AM]
Reply
- collapse thread

 
Then take a look on AMD Llano...
2 2 [Posted by: dezz  | Date: 06/02/11 04:08:45 PM]
Reply
18. 
Intel rules AMD in every possible way. Its so dominant to argue about it begs credulity.
4 3 [Posted by: beck2448  | Date: 06/02/11 09:23:55 AM]
Reply
- collapse thread

 
Really? Who has had the first commercial x86 CPU with an IMC and a point-to-point link to the chipset? And who has followed up years later...? Whose intellectual child x86-64 is (formerly AMD64, hint, hint)? Whose intellectual child the fusion of CPU and GPU is, taking advantage of the hunderds of stream processors, in order to bring GPGPU to everydays use, allowing for an unprecedented boost in performance? Who has a fully functional OpenCL 1.1 driver since last year? And who don't, because the IGP of Sandy Bridge is not up to the task, anyway...?
3 2 [Posted by: dezz  | Date: 06/02/11 04:58:08 PM]
Reply
19. 
Enough of the arguing, it's tiresome to keep rehashing the same apologies over and over. The time has come, put up or shut up, show me the CPU's and I don't mean 3 months from now.
2 0 [Posted by: Rowdy  | Date: 06/03/11 02:31:22 PM]
Reply
20. 
One of things missing in these comments about the "evil" of Intel is this: Business is War. Limiting what you can do to a corporate opponent is even more stifling than the Geneva Conventions on treatment of POWs. POWs live as nothing more than an information resource to be exploited. Sometimes you're gentle, sometimes you're not. Intel could destroy AMD by operating at a loss for a single year... crying foul about that is survival of the lowest. Capitalism is the ultimate extension of mankind; if AMD can't compete by any means, it should go into the furnace of history.
0 0 [Posted by: dac7nco  | Date: 06/17/11 03:18:35 PM]
Reply
- collapse thread

 
Competition breeds better products and lower prices for consumers. If you think it's some kind of boon to mankind for Intel to be the sole X86 supplier, you are a fool. What reason would Intel have had to develop its Core architecture if not for the fact they needed to surpass Athlon? We may very well have been stuck with a Pentium 4 design for quite a while longer. If you get rid of anti-competitive regulation like that, then you HAVE to get rid of patents as well. That's the only way for newcomers to break in and nip at the heels of giants. I'd be fine with no regulation if that included removal of patent laws and copyright, etc.
0 1 [Posted by: GCR117  | Date: 09/14/11 04:01:19 PM]
Reply

[1-20 | 21-21]

Back to the Article

Add your Comment


Our Sponsors

Crucial.com Western Digital G.Skill