Does AMD has some future?

Page 6 - Seeking answers? Join the Tom's Hardware community: where nearly two million members share solutions and discuss the latest tech.
Toggle sidebar Toggle sidebar
Status
Not open for further replies.
http://www.intel.com/content/www/us/en/processors/xeon/xeon-phi-detail.html
http://www.amd.com/en-us/innovations/software-technologies/processors-for-business/compute-cores

let me just copy this..

Although no commercially successful general-purpose computer hardware has used a dataflow architecture, it has been successfully implemented in specialized hardware such as in digital signal processing, network routing, graphics processing, telemetry, and more recently in data warehousing. It is also very relevant in many software architectures today including database engine designs and parallel computing frameworks.

your totally wrong that parallelism is not a real thing and that its not being used this very second..
 


Specialized hardware is very different from general purpose systems (which HSA is aimed at, for example). I won't deepen this argument, since I don't have enough information about how "popular" those systems are.

And honoring your name, the idea of parallelism is like saying that Quantum computing will replace everything next year or something like that. Paradigm shifts (specially in general purpose systems) takes years and years to happen.

Cheers!
 


You can make specialized DSP's for specialized tasks, and get a lot more performance then a general purpose CPU can. At a really large scale, that's all GPUs are.

As far as design, as long as you follow a LOAD/STORE architecture, there's only so much you can do. We've extracted as much efficiency as we can out of individual instructions and pipeline enhancements, we've done about as much as we can by adding cache to hide slow memory access speeds, we've done away with external clocks to sync the system. We're down to putting multiple serial processors together, and frankly, there's only so much you can do with that method. We're basically at the point where the current design of CPUs has hit a brick wall.
 


Not the whole industry: Intel, IBM, Fujitsu, NEC and others are pushing alternatives against GPGPUs.
 


I don't know why they joined. What I know is the following:

I know the details of the Heterogeneous System Architecture and I am far from impressed because there are better alternatives.

I know that a bunch of big players joining means nothing. Giants players such as HP, IBM, Dell, Sun Microsystems, Compaq, Silicon Graphics... (the list is much more impressive than HSA members) joined Intel on the Itanium architecture, but it was a sound fiasco

http://www.pcmag.com/article2/0,2817,2339629,00.asp
 


Yea, but you could say that at almost anypoint. Tubes hit a brick wall. Single cores hit a brick wall. Single pattern litho hit a brick wall. Planar Silicon hit a brick wall.

How about a software that can define it's own instruction set extensions, via an in ASIC FPGA? Or a parallel load RAM to push into known states? It's really not hard to imagine improvements, but the thing is that they have to make sense economically. Desktop class CPU's will become mature and commoditized at some point in the next decade or so, not because they can't be made better, but because investment won't make ROI. For people who want a ton of processing power it's not like there's not new stuff going on, quantum computing et al.
 
Samsung won't buy AMD, because there's nothing there that's attractive to Samsung. They'll all in on ARM, they can't get the X86 license via a sale, and frankly, AMD's GPU unit isn't as attractive as it used to be. There's zero benefit for Samsung to buy AMD.
 
From an Investor POV if you are holding AMD, liquidate your position before you get caught holding the bag. Unfortunately many technology companies miss the boat in a rapidly changing environment. They simply fail to change and shift directions. Many can be named like Kodak, Memorex, Xerox, all Print Media and the list goes on.

Who would have ever thought Blu-Ray disc would be antiquated in such a short period of time? Next Flex did.
Who would have ever thought book and software stores would ever be antiquated. Amazon did.
Who would have ever thought Ma Bell would be antiquated? Verizon and ATT did. They dumped the wireline for wireless.

What will be the next big fall? CATV will be the next giant to be slain by Google and ATT with Gigabit fiber to the house. Will make your 50 mb/s CATV service look like dial up.
 


They've fallen under a dollar before. And if the bottom drops out, it wouldn't take long for investors to bail.

Then again, I brought them last time they fell under a buck, and made 4/1 on my money when I cashed out.
 


The part reporting the troubles of AMD and the lack of future is precise.

The part reporting why Samsung would buy AMD is difficult to accept.

AMD's experience doing CPUs means nothing. Looks at Apple and Nvidia, without the same experience both have first class CPUs in their market, whereas AMD better design in the same market (Jaguar/Puma) is not competitive. If Samsung want to do a CPU they only need to hire a group of good engineers (like Apple and Nvidia did) instead purchasing an entire company with debts like AMD.

The same about GPUs. Samsung only needs to hire engineers. In fact this is what Samsung did. Samsung GPU is expected to arrive this year. Again AMD is irrelevant.

Samsung canceled its server plans. And the last partner to chose would be AMD. AMD has the record of losing the x86 server market and now are losing the ARM server market as well: people is purchasing APM and Cavium real products and avoiding AMD's Seattle hype. Again AMD is irrelevant.

Why would Samsung want a x86 license (even assuming it is transferable/purchasable) when x86 has no future?
 


This sounds as that famous quote from former AMD head: "I can't fail"

He did fail and "laid the foundation for a fundamentally inefficient capital structure that AMD never recovered from."
 
I think that unless AMD has a winner architecture with Zen, they'd be better off becoming ATi, just spend all the R&D on graphics and make ARM + Radeon combos to sell in tablets and phones.

They're really good at graphics, they should stick to it and not let their losing CPUs take a bite of the graphics investment.
 


Can I ask you a question? What is the purpose of this thread? Not to mention the other thread. Lots of tech companies that were once mighty are dead or dying. SGI, Novell, SUN, XEROX, H.P. Wang. The list is huge All you do is bash AMD. You say the same thing over and over. What will you do if they die? Will you go away? Does it make you feel smart? Like you could run AMD better than the people in charge there? This thread should be locked due to the premise that there is nothing constructive about it at all. It has not one redeeming quality.
 


Their GPUs are losing market share month after month. Moreover their GPUs are completely irrelevant in the GPGPU market, where their market share must be 5% or so.

If they want to sell GPUs for tablets and phones now then why sold their mobile graphics IP to Qualcomm?

http://www.informationweek.com/desktop/qualcomm-buys-amds-mobile-graphics-division/d/d-id/1075767?
 


For one if you don't feel the need to have a thread that can talk about future products from a company then perhaps you should get off of it.

For example people like to talk about the future of some sport team however no one really questions why this is happening.

Some people look at businesses as a sport and yes those companies matter to that you mentioned and i guess there is people who are more knowledgeable to discuses those areas are at some other forum. BUT this forum is to discuses Amd's future and this forum was made when the other one closes.

Obviously people want to talk about it if its here and even you since you clicked on it unless you were trying to put down others for reading it.

Edit

Also if you feel someone is bashing a certain company when all they are doing is providing evidence then possibly you have a bias yourself. For example if you read a gtx 970 review and think for even a second a 290X is better in terms of performance per watt in gaming you would be mistaken and possibly bias.

It's like getting mad at reviewers for telling their audience how it is if you don't like it then fine but others like to read about the information.
 


1.) Blu-Ray is not antiquated...it is now what DVD was 10 years ago. I also believe you reference NetFlix, which is having their own problems. Not aware of any "NextFlex", but you can buy Sony stock currently, since they hold the patent to Blu-Ray, you may even make money on that "iffy gamble".

2.) Good thing companies like Barnes & Noble are going the Amazon route...their stock is not what it once was, but far from a penny stock.

3.) Ma Bell = AT&T, they are still around too...

4.) Also, most cable companies offer comparable speeds to Google Fiber for less money now. Unless you live in BFE.

 

BluRay is dead and replaced with Streaming and Cloud Service. No reason to buy a disc today, Block Buster and Video rental stores are relics like record stores. No reason to buy movies or software today on CD/DVD/Blu-Ray because you can download or stream all of it from home.



Around yes, relevant no.



How old are you? Bell and ATT were broke up in 1982 33 years ago. The 7 Baby bells are Land Line twisted copper aka POTS which has been dead for years and replaced by VOIP and Wireless which are carried by CATV and ATT, Verizon and other wireless/internet providers. I work in telecom for 35 years. I was laid off from Southwestern Bell in 1985. Bell switching and head quarter offices are ghost towns. Twisted copper is DEAD. The 7 Baby Bells missed their opportunity in the late 90's and early 2000's when they failed to implement FIOS to the home. All that dark fiber is now being snatched up by GOOGLE and ATT. Fact is ATT and Bell have no ties and are estranged siblings, Judge Green broke up that monopoly 33 years ago. .



How is 50 mb/s CATV comparable to 1 gb/s Ethernet from Google and ATT?

 
You can say that about the baby bells, but it seems to me when I was in college, they were talking about the breakup, and how that ATT and various others were actually buying back up the companies, thus starting to effectively reconstitute the old monopolies. Not in the same form of course, but just interesting how things come back around.
 


Although completly irrelevant to this thread i agree with you 100%.
 
To some extent yes, bu tonly litle peices of the money makers Bell had to sell of to remain a float.

Example all 7 Baby Bells use to have cellular, no more they sold that to ATT. Bell also use to have carriers (fiber and microwave radio between cities for long distance, internet, ect.... ATT bought that up along with all the fiber optic routes. All the baby Bells have left is the Public Switched Telephone Network (PSTN) and twisted copper to the public and associated DSL. The PSTN and DSL market is dying. Who wants a twisted copper line today with dial tone and DSL? No one.

Remember Nortel aka Northern Telecom? They were the largest switch manufacture and a huge company. They do not exist, gone in 2005. Nortel switching went to Erricson, and the cellular CDMA/LTE/4G went to Lucent. The rest was divided up and sold in chunks. PSTN is a dead technology giving way to VOIP. LTE/4G is wireless VOIP. Even CDMA and GSM networks will be gone soon like Analog cellular because they are a PSTN technology. The Baby Bells are 8-Track tapes.

 
Status
Not open for further replies.

TRENDING THREADS

Latest posts

Moderators online

Share this page

COMPANY
RESOURCES
FOLLOW
Top Bottom