Sign in with
Sign up | Sign in
Your question

AMD intros SSE5 instructions

Last response: in CPUs
Share
August 30, 2007 6:27:46 PM

AMD is still gasping for air as they have now announced SSE5, the new instructions they have been talking about for awhile now. They are supposedly releasing the libraries now and I would think they would release the instructions with Shanghai\Montreal.

New Instruction Set “SSE5” Continues Tradition of AMD x86 Innovation, Including 3DNow!, AMD x86-64 Architecture, AMD Virtualization and Light-Weight Profiling Specification

SUNNYVALE, Calif.--AMD (NYSE:AMD) today announced further plans to innovate the x86 architecture by introducing SSE5, a new extension of the x86 instruction set that is designed to allow software developers to simplify code and achieve greater efficiency for the most performance-hungry applications. SSE5 will give developers additional capabilities to help maximize the performance of applications that have daily impact on consumers and enterprises, including high performance computing, multimedia and security applications. By making the SSE5 specification available to developers today, AMD expects to ease the adoption of the new instructions for tool providers and software vendors who develop these performance-intense applications.

“Chip advancements and software improvements go hand-in-hand, to the benefit of consumers and enterprises alike,” said Phil Hester, senior vice president and chief technology officer, AMD. “The impact of our designs are best realized when AMD-based servers, PCs and devices enable software to more effectively solve every-day problems and enhance every-day experiences. By announcing our plans to add SSE5 instructions to the x86 instruction set ─ and by making the specification available today ─ we are enabling open and collaborative software innovation that will bring AMD’s advancements to life for our customers and end-users.”



I guess they're not dead yet.

ALL HAIL THE DUOPOLY!!!
August 30, 2007 7:42:30 PM

Well it is totally inline with their entire marketing strategy! Just keep the fanboys HOPING for something worth buying, keep the investors HOPING they will produce something worth buying!
Oh yes great a new instset to be released in 2009! Well then I think I will wait to upgrade! Forget needing a new system to play bioshock and crysis on NOW (or before christmas), because AMD will release an intel-crusher 5 yeas from now!
I can't believe how much crap we put up with from AMD and their paper statements...they're garbage, doesn't anyone realize this? End of November...it is a LONG wait (I am SO itching to build a new system), but by then we will see. If AMD is still promising good proc's by then, I'll be laughing on my X48 and intel 45nm. If I can get 790 and 3ghz phenom by then I will, but I am not counting on it.
August 30, 2007 8:47:20 PM

Quote:
Keep laughing. I laughed at Intel for 3 YEARS while AMD walked all over P4 and PD.
Truth is AMD has a lot more market share than they did in the late 90's, they arent going anywhere, their plenty happy being back in number 2.



Yes, they have 62% of desktop retail revenue and cost savings should start to raise their income. They also got 44% of the mobile retail market. Two years ago they had barely ANY mobile share. if Intel had announced this it would be the second coming.
Related resources
August 30, 2007 10:01:29 PM

BaronMatrix said:
Yes, they have 62% of desktop retail revenue and cost savings should start to raise their income. They also got 44% of the mobile retail market. Two years ago they had barely ANY mobile share. if Intel had announced this it would be the second coming.


It's fine and dandy to have 62% and 44% of retail market, but when ASPs are so low that they are on the fringe of almost giving them away, what's the point of having such a large retail market share? I'd rather have 20% with very high ASPs, than 50% with super low ASPs.

How much revenue will they show this quarter? They gained retail and section share last quarter, and lost $600M. What will these gains cost them?
August 30, 2007 11:05:39 PM

Superb post nmdante -- I laugh when people look at such a small picture --> oh man they have x% of market share and oh you forgot to mention I make nothing off of it because I have to seel it so cheap to keep the share, ROFL
August 31, 2007 12:15:24 AM

NMDante said:
It's fine and dandy to have 62% and 44% of retail market, but when ASPs are so low that they are on the fringe of almost giving them away, what's the point of having such a large retail market share? I'd rather have 20% with very high ASPs, than 50% with super low ASPs.

How much revenue will they show this quarter? They gained retail and section share last quarter, and lost $600M. What will these gains cost them?


Hector said, "break the monopoly" not make more money than ever. The fact that they had to lower prices while Intel still had A WHOLE LOT of Netburst stock shows.....ehh never mind.

Don't forget that Morgan Stanley is paid off and no cash goes toward that this quarter, at least after a certain accounting day. Why are you so against the duopoly? You complain when they're behind and you complain when they aren't...sounds like a personal problem not having anything to do with the worldwide CPU market.

Your opinion of what's best is not "gospel." I don't pretend to know but I'm typing this on a 4400+(2.3GHz) so I know what AMD is capable of.
August 31, 2007 12:55:51 AM

Yup this must suck real bad, just like all of AMDs other innovations. Intel is faster right now, therefore, they will always be faster. If you have an AMD chip you can't even open web-sites or play any games cause they suck so bad. Let's all get together and tell AMD to quit trying to innovate so Intel can sit back and release and cut all that R&D funds. AMD no one want more efficient programs so just sit back and die so Intel can finally have its much desired monopoly in the PC market.
August 31, 2007 1:04:24 AM

BaronMatrix said:
Hector said, "break the monopoly" not make more money than ever. The fact that they had to lower prices while Intel still had A WHOLE LOT of Netburst stock shows.....ehh never mind..
Allow me to finish that statement.....shows nothing.

Well, Perhaps instead of 'ALL HAILING THE DUOPOLY', we should "ALL HAIL" Hector, since, as you say he "said, "break the monopoly" not make more money than ever" he is clearly so set on 'breaking the monopoly' rather than making money that he is willing to bankrupt AMD to do it.


BaronMatrix said:

Don't forget that Morgan Stanley is paid off and no cash goes toward that this quarter, at least after a certain accounting day. Why are you so against the duopoly? You complain when they're behind and you complain when they aren't...sounds like a personal problem not having anything to do with the worldwide CPU market.


Morgan Stanley was paid off by going to senior market shares....one step away from going to a loan shark. Merely to trade off debt (for the purpose of freeing themselves from the financial restrictions apllied by Morgan Stanely), not clear it. The debt still exists, in the form of maturing Senior Shares, however now it is the stock holders who are going take the loss in the shorts if AMD goes Chapter 11 rather than Morgan Stanley and the Senior Share Holders. No doubt Morgan Stanley was ecstatic about AMD issuing the second set of Senior Shares as that removed their investment from a company which is, for all intensive purposes, rushing towards Chapter 11 at the rate of $600+ million per quarter.

BaronMatrix said:

Your opinion of what's best is not "gospel." I don't pretend to know but I'm typing this on a 4400+(2.3GHz) so I know what AMD is capable of.

Being capable of something, and doing it are two different things. Russia had the 'capability' to send men to the moon but they didnt. The US did.
AMD had the capabilty to steal the majority of the overall marketshare from Intel last year, but they didnt because they choose to keep ASPs (and thus margins) on a superior product high. That created a short term financial gain for them, but in the long term, now that they no longer have the superior product, they must struggle to maintain the marketshare they had by lowering prices so much as to cripple the margins and thus their revenue share, in the process making any gains in marketshare financially ineffectual.
August 31, 2007 2:15:07 AM

BaronMatrix said:
Hector said, "break the monopoly" not make more money than ever. The fact that they had to lower prices while Intel still had A WHOLE LOT of Netburst stock shows.....ehh never mind.

Don't forget that Morgan Stanley is paid off and no cash goes toward that this quarter, at least after a certain accounting day. Why are you so against the duopoly? You complain when they're behind and you complain when they aren't...sounds like a personal problem not having anything to do with the worldwide CPU market.

Your opinion of what's best is not "gospel." I don't pretend to know but I'm typing this on a 4400+(2.3GHz) so I know what AMD is capable of.



So, what you're saying is that AMD doesn't care about making money, just breaking this monopoly? I'm sure their stockholders would think differently. Especially in light of all the senior notes that are all over the place now, which pretty much hoses the average stock holder from making a cent, if the worse happens.

You said a lot of things, but none answered my question -
What good is gaining retail share, when ASPs are so low? At what cost will AMD be willing to risk, in order to gain 2-3% of market share?

You can spout off all you like about "HAILING THIS" or "HAILING THAT", but the bottom line is that AMD is selling their products "dirt cheap" in order to gain a percentage here or there. Along the way, they will continue to bleed red, and all you can reply back is "break the monopoly". So, as turpit has mentioned, Hector is willing to drive this company into the ground for market share, no matter the cost. Unfortunately, it could cost AMD everything - but as long as they "break the monopoly" right?

Gospel? Ha. You're funny. Especially coming from someone who "almost had sex with a woman on the dance floor".

Oh, and by the way, I'm typing this with an E6700 powered system, cause I do know what Intel is capable of. Maybe next week, I'll type with a Quad, can you do that? Oh, wait, that's right...the memory was too expensive for that 4X4 system. Shame.

August 31, 2007 2:26:09 AM

The best predictor of future behavior is relevant past behavior.

I'm a newbie, and even I've heard of BaronMatrix's antics.

I began reading the initial post about AMD and SSE5 instructions with quite a bit of interest, but as soon as I saw who posted it, I stopped reading.

My advise to Baron would be to, somehow, "create a new history" that's not so tainted by past issues, like

1) Chronic bias (AMD in his case, but bias in general can be a deterrent)
2) Frequent "rewrites of past history" when said past history is putting him in a negative context
3) Persistent defense of his actions when they portray him in a negative light, even when he's blind to that negative light.

Bottom line: know when to say "I'll shut up now." After all, do you want to be right or do you want to be happy?
August 31, 2007 2:53:25 AM

Neh, the Article is real, it's on Anandtech website. If AMD can hold out that long, and some how convince Intel to also adopt SSE5, that being the major obstacle, the instruction set should improve x86 computing drastically.

We all know this isn't going to be part of barcelona. We also all already know that, Barc doesn't perform as well as the c2d based business and server procs, at least according to the last benchies I saw. Granted things may have changed by the time they go "gold" with them. In my opinion, Barcelona is what AMD is using to test the waters for the native Quad core, and to work out and issues while at a lower clock speed. It makes me wonder what tricks AMD has up its sleeves when it comes to Phenom. Think they'll use Barcelona to play Intel's hand, and if they got something good going with Phenom, trump it with that. I'd also be interested to see, whether or not they're sneaking SSE5 onto the Phenom core. From what I've read they're saying something about that core being able to pull more instructions per cycle as it is.

Otherwise, I think SSE5 may be stopped before it even gets out of the gate if AMD doesn't find a way to make some real revenue. Either that or may be stalled if someone swoops in and takes them back private after sacking the management.
August 31, 2007 4:14:19 PM

I doubt SSE5 will be snuck into Penom. AMD's lucky enough to have been able to adapt a portion of SSE4 (intel's) to work in the Barcy arch.

Releasing news of SSE5 may either work for or against AMD here. Imagine if Intel's had this information for some time and has embedded their version of SSE5 into Nehalam. That would give Intel a year lead on AMD's SSE5 introduction. As with all instruction sets, it requires actually using the functionality before it becomes a big win (SSE4 is still gaining traction).

This is where we need jack back to explain the full technical details of SSE4 and SSE5. Not everyone knows assembly :) 
August 31, 2007 4:17:02 PM

Didn't the article actually say that SSE5 isn't supposed to be released until 2009!?
No, it is not going into anything anytime soon, and it likely won't actually be released until 2011. This is all marketing.
August 31, 2007 4:53:51 PM

@iceman71 The only other CPU brand, running x86 instructions and still selling, is to my knowledge VIA Technologies that takes the third and probably are happy with that possision as well.
September 1, 2007 2:08:01 AM

Its a good idea by AMD to release additional instructions so early to help widespreaden the adoption of them before cpu even released. Would you prefer buying cpu and get use of all its features after several years only, when you about to upgrade from it, or when its just released? I hope Intel will do the same with new instruction sets.

NMDante said:
It's fine and dandy to have 62% and 44% of retail market, but when ASPs are so low that they are on the fringe of almost giving them away, what's the point of having such a large retail market share? I'd rather have 20% with very high ASPs, than 50% with super low ASPs.


Tell me about the way for AMD to get "20% with very high ASPs". It sounds cool, in theory, too bad its not connected to reality atm, as AMD hadnt any product to pull this off.
a b à CPUs
September 1, 2007 3:13:30 AM

Given that Fusion is scheduled for 2009 it makes sense to me that releasing SSE5 instructions in 2007 is right in line with AMD's long term roadmap. From Wikipedia...

AMD claims SSE5 will provide dramatic performance improvements, particularly in high performance computing (HPC), multimedia and computer security applications, including a 5x performance gain for AES encryption and a 30% performance gain for DCT used to process video streams. said:
AMD claims SSE5 will provide dramatic performance improvements, particularly in high performance computing (HPC), multimedia and computer security applications, including a 5x performance gain for AES encryption and a 30% performance gain for DCT used to process video streams.

What's so special about either Intel or AMD releasing instruction instruction sets before the release of a processor. This is nothing new here. I don't know proc history as well as some, but I'm quite sure Intel released instruction sets far enough in advance for the industry to make use of them.

Don't understand some of the anti-AMD statements made in this thread.
Especially since no noise was made about SSE4a being released wioth
Barcelona. Unless it's because Baron Matrix seeded this thread or
because it's AMD relateds news.

Seems pretty cool to me... :D 
September 1, 2007 3:34:30 AM

chunkymonster said:
Given that Fusion is scheduled for 2009 it makes sense to me that releasing SSE5 instructions in 2007 is right in line with AMD's long term roadmap. From Wikipedia...


What's so special about either Intel or AMD releasing instruction instruction sets before the release of a processor. This is nothing new here. I don't know proc history as well as some, but I'm quite sure Intel released instruction sets far enough in advance for the industry to make use of them.

Don't understand some of the anti-AMD statements made in this thread.
Especially since no noise was made about SSE4a being released wioth
Barcelona. Unless it's because Baron Matrix seeded this thread or
because it's AMD relateds news.

Seems pretty cool to me... :D 


+1 *applauds*
September 1, 2007 4:09:04 AM

Harrisson said:
Its a good idea by AMD to release additional instructions so early to help widespreaden the adoption of them before cpu even released. Would you prefer buying cpu and get use of all its features after several years only, when you about to upgrade from it, or when its just released? I hope Intel will do the same with new instruction sets.



Tell me about the way for AMD to get "20% with very high ASPs". It sounds cool, in theory, too bad its not connected to reality atm, as AMD hadnt any product to pull this off.


Um, they did do it, when the K8 series was the faster CPU. During H106 when the now $68 (US) X2 3800 was selling for over $300 (US).

After the past 2 years, the question isnt if AMD could do it, since they had, pushing their way towards 1/3 of the overall marketshare at high (relatively) ASPs, but can they do it again? The fluxuations in marketshare and revenue share show that the market is shifting based on value and performance as backed by production capacity, not any 'backhanded' dealing as many want to believe. AMD had both performance and value leadership, allowing them to eat away at Intels share lead at an admirable rate, but they failed to capitalize. The real question is: is AMDs current situation their own doing, or was the current situation ineveitable. Many people (mysef included) are of the opinion that AMD became complacent with Intel and Intel's Netburst 'ineptitude', expecting Intel to continue bumbling along indefinately. This is believable considering AMD's descision to secure a large loan as well as spend a significant portion of their available funds to purchase ATI. The purchase of ATI seemingly contradicts the complacency theory, but only to an extent. AMDs purchase of ATI indicates a move towards marketing a hetrogeneous platform or line of platform machines along with the known plans for integration of CPU and GPU (purportedly for mobile) and was a brilliant strategy; forward thinking and not in any way complacent. Unfortunaelty K8 was an older design and AMD was not prepared to compete with C2D on a one to one basis. This is no longer a question, but hindsight. So it is quite possible that AMD was complacent in expecting the "netburst" market trend to continue and as such focused too far into future, in the process devoting all the resources they could towards that end, as a result depriving themselves of near term agility. Regardless of whether AMD was concious of the threat posed by Intel or not, the purchase of ATI, while brilliant for future product developement, hurt them badly in terms of near term financial liquidity and thus R&D. With less money to throw at the R&D of K10 to speed its introduction, AMD has allowed the perfromance imbalance to exist long enough that their only recourse was to lower prices, resulting in the reduced margins they are now suffering. They really had no choice, but if Barcelona outperforms C2D and Penryn by the same margin C2D out performs K8, AMD will once again be in a position to hold ASPs high and increase their margins while simultaniously taking market share away from Intel. If not AMD will continue bleeding until the are dry.
September 1, 2007 4:34:51 AM

turpit said:
Um, they did do it, when the K8 series was the faster CPU. During H106 when the now $68 (US) X2 3800 was selling for over $300 (US).

They could have higher profit before C2D, NOT after its release with very affordable prices, so the question stays to nmdante, how does he suggests for AMD to get "20% with very high ASPs" in this period before Barce is introduced. I heard this "claim" too many times and would like to hear ANY decent reasoning :) 

turpit said:
They really had no choice, but if Barcelona outperforms C2D and Penryn by the same margin C2D out performs K8, AMD will once again be in a position to hold ASPs high and increase their margins while simultaniously taking market share away from Intel. If not AMD will continue bleeding until the are dry.

Thats what I'm talking about, AMD cornered themselves and had no choice but to sell loads of cpu's for cheap to keep affloat till they have product they could have higher margins from. Not sure if Barce can turn the tables arround, probably it will be just good enough to keep them up till Fusion and Buldozer kicks in. One of the sweet spots could be R600 based FireGL proffesional series, it looks good against Quadro and it will give some profits if they will sell plenty of them.
September 1, 2007 4:56:27 AM

I like Turtles.

But seriously I want to hear what the greatgrapeape has to say on this...he seems to be quite learned.
September 1, 2007 5:24:27 AM

Harrisson said:
They could have higher profit before C2D, NOT after its release with very affordable prices, so the question stays to nmdante, how does he suggests for AMD to get "20% with very high ASPs" in this period before Barce is introduced. I heard this "claim" too many times and would like to hear ANY decent reasoning :) 

Thats what I'm talking about, AMD cornered themselves and had no choice but to sell loads of cpu's for cheap to keep affloat till they have product they could have higher margins from. Not sure if Barce can turn the tables arround, probably it will be just good enough to keep them up till Fusion and Buldozer kicks in. One of the sweet spots could be R600 based FireGL proffesional series, it looks good against Quadro and it will give some profits if they will sell plenty of them.


Sorry, but I think there is some misunderstanding on my post. I never said AMD can achieve 20% higher ASPs. I said that I, personally, would rather have a higher ASP, over higher market share. I didn't mean to make it sound like AMD could've kept their high ASP when Core 2 was released.

BM claimed that getting 42% and 62% of the retail market share was a great thing, but I questioned how good it was, when ASPs were so low, that the razor thin margins to get those gains, could potentially be another quarterly loss. That's when I said that "I'd rather have 20% of the market with very high ASPs, than 50% at super low ASPs."

So, I apologize if I made it sound like AMD could've achieved a higher ASP with their products, when Core 2 launched.
September 1, 2007 6:03:29 AM

Harrisson said:
They could have higher profit before C2D, NOT after its release with very affordable prices, so the question stays to nmdante, how does he suggests for AMD to get "20% with very high ASPs" in this period before Barce is introduced. I heard this "claim" too many times and would like to hear ANY decent reasoning :) 


Thats what I'm talking about, AMD cornered themselves and had no choice but to sell loads of cpu's for cheap to keep affloat till they have product they could have higher margins from. Not sure if Barce can turn the tables arround, probably it will be just good enough to keep them up till Fusion and Buldozer kicks in. One of the sweet spots could be R600 based FireGL proffesional series, it looks good against Quadro and it will give some profits if they will sell plenty of them.


I understand Dantes point. Remember too, mometum carried AMD for the better part of 2 quarters after the introduction of C2D, allowing them to keep ASPs relatively high, with retailers charging over $500 for the (then) non existant X2 5000s. At that time, they still had the publics distrust of Intel working in their favor, but now, well after C2D has been proven not to be another netburst, IMO they have no choice but to keep prices lower than Intels. Binning also plays a huge role in AMDs fate. If AMD is unable to bin large numbers of their better performing chips (say 5400 and up ) as some people suspect, then they are forced to live off the margins of the lower chips. They get so many chips from a wafer, regardless of whether they bin out as 4xxx, 5xxx or 6xxxs. The 6xxxs pull the higher ASP, where AMD can maximize their margins but if they can only get a 5% yield of 6xxx's per wafer, that significantly reduces what margin they have left. In addition to the obvious loss of income due to lower ASPs, the rumours of both 65nm production problems and heat issues on 90nm forcing low end bins certainly seems more plausable given AMDs 1/2% revenue gain for the 4.7% recoup they experianced last quarter.

Whether barcelona can save AMD or not remains to be seen, but AMDs situation now appears more desperate than ever before. Yes, AMD is no stranger to operating in the red, operating with low available funds and at low margins, but at losses $600 mil per quarter for the past 3 quarters, they are rapidly running out of cash. Because of that, going private is no longer an option as they have insufficent funds. They have issued senior bonds not once but twice this year, and describing their attempts at PR and marketing the past 13 months as disasterous is not an exagerration. IMO selling entry level machines is the best strategy they can follow right now, but it is finite. A retreating tactic that trades profit for time. But sooner or later they must release Barcelona, and it must deliver. It must deliver strongly enough to allow AMD to set the prices rather than react to Intels pricing. If not, I just dont beleive they will survive long enough to get either Fusion or Buldozer to market.
September 2, 2007 3:06:33 AM

NMDante said:
Sorry, but I think there is some misunderstanding on my post. I never said AMD can achieve 20% higher ASPs. I said that I, personally, would rather have a higher ASP, over higher market share. I didn't mean to make it sound like AMD could've kept their high ASP when Core 2 was released.

BM claimed that getting 42% and 62% of the retail market share was a great thing, but I questioned how good it was, when ASPs were so low, that the razor thin margins to get those gains, could potentially be another quarterly loss. That's when I said that "I'd rather have 20% of the market with very high ASPs, than 50% at super low ASPs."

So, I apologize if I made it sound like AMD could've achieved a higher ASP with their products, when Core 2 launched.

From the context it seemed you were talking about how AMD "could have done it better with lower market share and very high ASPs". If you didnt meant that, then your comment is a bit out of place, I would prefer both high market share and high ASP, what does it have to do with the topic? ;) 
September 2, 2007 3:15:36 AM

turpit said:

Whether barcelona can save AMD or not remains to be seen, but AMDs situation now appears more desperate than ever before. Yes, AMD is no stranger to operating in the red, operating with low available funds and at low margins, but at losses $600 mil per quarter for the past 3 quarters, they are rapidly running out of cash. Because of that, going private is no longer an option as they have insufficent funds. They have issued senior bonds not once but twice this year, and describing their attempts at PR and marketing the past 13 months as disasterous is not an exagerration. IMO selling entry level machines is the best strategy they can follow right now, but it is finite. A retreating tactic that trades profit for time. But sooner or later they must release Barcelona, and it must deliver. It must deliver strongly enough to allow AMD to set the prices rather than react to Intels pricing. If not, I just dont beleive they will survive long enough to get either Fusion or Buldozer to market.

AMD wont go bankrupt IMO, if K10 and R670/R700 will fail, shares will reach bottom and someone will buy them out. There were plenty of rumours this year and its not unlikely, AMD have promising products in near and long term to be attractive target for the rich sharks.
September 2, 2007 3:39:05 AM

Thats a possibility, but for an entity considering buying out AMD, it might very well be cheaper to allow AMD to go Chap11, then pick up the pieces afterwords. A Fed subsidy is also a possibility. Taking out another loan to buy their own shares is a possibility, but extremely unlikely due to their current finacial state and the existing senior bonds.

There are many possibilities, time and Barcelona's performance will tell what will happen
September 2, 2007 4:33:14 AM

turpit said:
Thats a possibility, but for an entity considering buying out AMD, it might very well be cheaper to allow AMD to go Chap11,
There are many possibilities, time and Barcelona's performance will tell what will happen

anyone remember when the inquirer reported AMD was buying ATi then there was nothing for about 3-4months then bang AMD purchased ATi.
I think if Bacrolona is a flop Samsung or AMAT will buy em dirt cheap.
September 2, 2007 5:05:23 AM

Any thing is possible. Who knows, Barcelona may even outperform C2D by 400% and Intel may go BK as Sharikou predicts. Not liekly, but possible.
September 2, 2007 6:28:44 AM

BaronMatrix said:
AMD is still gasping for air as they have now announced SSE5, the new instructions they have been talking about for awhile now. They are supposedly releasing the libraries now and I would think they would release the instructions with Shanghai\Montreal.

New Instruction Set “SSE5” Continues Tradition of AMD x86 Innovation, Including 3DNow!, AMD x86-64 Architecture, AMD Virtualization and Light-Weight Profiling Specification

SUNNYVALE, Calif.--AMD (NYSE:AMD) today announced further plans to innovate the x86 architecture by introducing SSE5, a new extension of the x86 instruction set that is designed to allow software developers to simplify code and achieve greater efficiency for the most performance-hungry applications. SSE5 will give developers additional capabilities to help maximize the performance of applications that have daily impact on consumers and enterprises, including high performance computing, multimedia and security applications. By making the SSE5 specification available to developers today, AMD expects to ease the adoption of the new instructions for tool providers and software vendors who develop these performance-intense applications.

“Chip advancements and software improvements go hand-in-hand, to the benefit of consumers and enterprises alike,” said Phil Hester, senior vice president and chief technology officer, AMD. “The impact of our designs are best realized when AMD-based servers, PCs and devices enable software to more effectively solve every-day problems and enhance every-day experiences. By announcing our plans to add SSE5 instructions to the x86 instruction set ─ and by making the specification available today ─ we are enabling open and collaborative software innovation that will bring AMD’s advancements to life for our customers and end-users.”



I guess they're not dead yet.

ALL HAIL THE DUOPOLY!!!


Maybe AMD should just go out and build a chip from the ground up instead of licensing the ISA x86 from Intel... But, until they do, they can write all the extensions they like and the other licensees can review, and vote on, whether to adopt them or not. Oh, forgot to mention... Intel owns the instruction set... So unless this set of instructions is a great benefit (to the owner) then it will not be approved by the standards body.
September 2, 2007 5:21:08 PM

Harrisson said:
From the context it seemed you were talking about how AMD "could have done it better with lower market share and very high ASPs". If you didnt meant that, then your comment is a bit out of place, I would prefer both high market share and high ASP, what does it have to do with the topic? ;) 


Well, I was referring to BM's insistence that higher retail market share is great, but considering the low ASPs for each unit sold, how good is it?

That's when I said it would've been better to have low retail market but with higher ASPs. When AMD was flying high with K8, they had about 21-23% of the overall market, but probably low retail market share - with high ASPs, and they flourished. So, higher retail market means little without a decent ASP, which right now, AMD does not have.

Of course every company would prefer high ASPs, along with high market, but when a company strives for market share at any costs, ASPs are the first to usually drop, and they are usually the hardest to bring back up.

That's what I was talking about.
!