Overblog Follow this blog
Administration Create my blog
June 28 2015 1 28 /06 /June /2015 16:20

After AMD last week has presented the revised graphics cards of the R3x0-series and this week the all-new FuryX graphics-card, equipped with revolutionary components, the 2015 line-up is almost complete. And since the latest AMD graphics-cards have caused significant fuss in the hardware-scene, to which I briefly took position before the release, so I'm making a short assessment of the new products here after their release.

Let's get straight to the content:

Brief summary to the AMD Radeon R3x0 Line-up

  • In the presentations run-up there has been a fierce debate, whether the R300 Radeons should be considered a „rebrand“ (= pure renaming, in the specific case of the R200 chip) or a „refresh“ (with small changes on the chips).
    I am turning out more detail below, in my opinion it is something like the „smallest possible refresh“.

  • The chip design is basically identical to its predecessors, but the new cards are a tad faster and tend to be at bit more energy-efficient.

  • Furthermore, some third-party manufacturers have used the opportunity (of a new products) to improve the PCBs (printed circuit boards) and especially the cooling solutions.

  • The technical differences to its predecessors are little.
    The competitiveness towards Nvidia is slightly improved, but basically the same:

    • Nvidia scores with a somewhat higher computational-performance and especially with energy-efficiency.

    • AMD scores with better-value-for-money and the larger number of cards in various performance-segments.

All in all AMD makes a slightly better deal than before. The improvements are only little, so an upgrade from its R200 predecessors makes sense only in exceptional cases.

The new AMD Radeon-Fury family

An eagerly awaited highlight, particularly by AMD-fans, are the new graphics-cards top-models "Fury". These are intended to be the counterpart to NVIDIAs top-models "Titan" and Geforce-980(Ti).

 

After evaluating several reviews and user-ratings, I'd give the Fury-X school-mark 2 (in Austria 1 = very good, 5 = insufficient).

It's doing well in the tests, but is not able to reach outstanding scores so far. In terms of gaming-performance Fury’s benchmarks are in between Nvidia's „normal“ Geforce-980 and the Ti-version, mostly closer to the Ti-version.

 

Particularly noteworthy with the Furys is, that they are carriers of forward-looking technology:

They are the 1st graphics-cards with revolutionary, extremely fast HBM-memory. The top-model „Fury-X“ also has a water cooling system. Air-cooled versions, the "normal" „Fury“ and the particularly compact „Fury-Nano“, are going to follow.

AMDs changes for the Radeon-R300-series in detail

A heavily discussed question in recent weeks was, what exactly distinguishes the R3x0 GPUs from its predecessors. This issue still, even after the launch is controversial and it has cost me some research-work to clarify it. The best listing of the changes is provided by German site golem.de:

  • An improved manufacturing-process

  • micro-code updates

  • a revised power-management

  • support for newer, better memory chips.

Infos in English about the changes can be found at Anandtech and Hexus.

So now new, very efficient GDDR5 chips by Hynix are supported, which allow 6 GBs transfer-rates even at low 1,35V supply-voltage.

In addition, some third-party manufacturers have improved even their PCBs (circuit/printing-boards) and cooling-solutions.

 

Just given the controversies, it is however incomprehensible why AMD itself doesn't communicate comprehensive and clear information on the 300-series improvements.

 

3DCenter.org, another great German hardware-site, estimates that AMD's improvements beyond higher chip- and memory-clocks offer a 1-3% performance increase. So really a very small surplus, but nevertheless a surplus.

My conclusion to AMDs R3x0-Radeons

Usually I am only dealing with certain products that detailed, if I'm going to buy them for myself – what I'm not doing in this case. After having extensively dealt with AMD anyway for my marketing-case-study, of which I just published an English summary recently, so this "launch assessment" of the new Radeons is almost a „by-product“.

And just from the overall-view I can understand AMD's strategy spending it's GPU-lineup only a small-update – Fury nominally does not belong to the 300-series.

 

Sure, many users are disappointed by the little progress. However, even all-new chips based on the current GCN-1.2/1.3 architecture would not have performed much better. Thus the small and cheap solution for now makes sense.

 

Finally the user gets a minor update of the products, which are definitely more tempting than their predecessors at the same price-points. NVIDIA however also makes really good graphics card and so I'm already sure that there are going to be GPUs of both manufacturers in my next purchase-advice.

Share this post

Repost 0
Published by critical-constructive - in computers
write a comment
June 22 2015 2 22 /06 /June /2015 23:00

7 months ago on 23rd November 2014 I published my big marketing-study in German. As both a fan of AMD as well as a PC-freak I wanted to publish my comprehensive knowledge about computers and about AMDs current situation. And AMDs situation currently is definitely a difficult one.

 

As AMD is an American company and most of the audience is outside the German-speaking countries, it's really important to present my thoughts also in English. Well, several months later than originally intended by me: Here it is, my “AMD-Marketing-Summary-2014.pdf”.

What does my paper deal with

  1. First with a comprehensive analysis for each of the 4Ps in marketing & about economic and strategic aspects of the company’s policy in the last couple of years.

  2. With strategic as well as with operational solution-approaches, what & how AMD could do better.

  3. A closer look on AMDs corporate responsibility – something very important to me as a “Greenie” as well as for companys in the long term.

  4. New and not contained in my German paper is a wrap-up of new developments at AMD from December 2014 until today. And there have been made several remarkable changes since then.

What are the most important insights of my study:

  • AMDs basic-strategy, breaking out the 2-front-long-time-static-warfare with Intel & Nvidia by seeking for new product-fields basically is a 100% right decision, but AMD has overshot this target.
    As a result the company has neglected traditional product-fields, most of all the Opteron & FX-CPU-families too much, which resulted in big economic losses.
     

  • This has to be changed and AMD has set the course for an Opteron & FX rebirth in 2016.
     

  • AMDs products have weaknesses, especially against some Intel CPUs. But they have also their strengths and are still in the 1st half of 2015 rather competitive.
    Nevertheless many of them they need to be improved or replaced by successors.

     

  • AMD has significant weaknesses especially in distribution, but also on communication.
    Their products quite seldom make it into third-party-manufacturers devices and so didn't sell well – even when they were better suitable for some devices, than competitors products.
    Also in communication with stakeholders there have been shortcomings, which in the meantime seem to have been addressed quite well.

     

  • AMDs just introduced R300 graphics-cards offer little improvement over their R200-predecessor an perform reasonable against Nvidia Geforce GTX-900-series as well.
     

  • All we know from the revolutionary ZEN-CPU-architecture looks very promising. If AMD is able to keep their promises, they should do well also against Intel’s Core-i7 & Xeons in late 2016.
     

  • But that’s exactly my main point of criticism:
    Letting the Vishera-FX-CPUs silicon unchanged in the markets for 4 years absolutely is a no-go for some reasons:

    • Communication-wise: The FX-series is AMDs “top-dog” in the product-range. All products & the whole brand is very much measured by the FXs performance.

    • Product-wise: AMD's FX-CPUs top-models in terms of application-performance (but not at games) are competitive with Intel's current Haswell-Core-i5 CPUs.
      But with the introduction of the Skylake-generation in autumn 2015 this is going to change.

    • Price-wise: After Skylakes launch AMD either will have to accept sales-declines when maintaining the current prices or to accept margin-declines when bringing down prices for maintaining sales-numbers.

    • And there are also other reasons like long-term effect for affordable prices with future products or less trust by customers.

As ZEN-based CPUs are said to arrive in Q4 2016 this is too long for not being competitive against Intel's upcoming Core-i5s. So in my eyes it's an absolute key-products for AMD introducing deeply revised Vishera-FX-CPUs, that I call “Vishera-2.0” and for which I have presented specific measures to stay competitive.

But I will make a separate blog-entry about Vishera-2.0 in the next weeks telling you details.

Final words

AMD is in a really difficult situation right now. But there are also several very promising opportunities. And many products i.e. it's new graphic-cards or it's new notebook APUs (codename “Carrizo”) are competitive.

If AMD sets the right measures I see good opportunities returning to the road of success.

 

But read all the details in my paper. And I hope, you enjoy my thoughts in this copious summary of my big marketing-study!

 

P.S.: For everybody interested in more details & understanding German, please see my original documents "AMD-Marketing-Analyse-2014.pdf" and "AMD-Vishera-2.0.pdf"

Share this post

Repost 0
Published by critical-constructive - in computers
write a comment
June 13 2015 7 13 /06 /June /2015 14:40

The introduction of the new AMD R300-series graphics-cards is imminent. Telling it exactly this should happen on 16th June 2015.

 

After it had seemed to be clear, that there will be no all-new graphics-chips, with the exception of the new and revolutionary HBM-memory-equipped „Fury“ top-models, a little „shitstorm“ brake upon the R300 product-line: Many Users are disappointed because in their eyes the new graphics cards seem to use just "old" chips, that have simply been renamed.

 

So folks, I am also a hardware-fan and I am also interested what's happening there. But quite apart from the fact that there are more important things in life – for example the 2nd anniversary of the incredible revelations of Edward Snowden, for my part there are following things to say:

I believe that it's highly premature, when users and in particular also the specialised press assess the new cards, although these have not been introduced yet.

How useful is it, to form an opinion on the base of rumours or at the best on information-snippets? Shouldn't you wait until there is comprehensive information available and only evaluate then?

 

But I also think, that some of the media is to blame here as well: Some of them deal really messy with the terms "rebrand(ing)" and "refresh" and use both as interchangeable.

But the fact is, that only a "rebrand" applies the same chips unchanged, just given a new name.

While a "refresh" on the other hand uses the same technical-basis, which however has undergone improvement.

 

And there is indication that the existing chips for use in the R300-series have been slightly revised.

And there are some „screws“ AMD could tighten up:

  • A new chip-stepping, in which the chips are slightly higher clocked and more energy-efficient.

  • The use of new, more energy-efficient GDDR5 memory.

  • Possible further changes not on the graphics-chip, but at the cards underpinning.

 

However, I am also very curious to see what AMD actually delivers in „real world“.

A pure re-branding, I of course would also see very negative.

But if there is a slightly increased computational performance and/or a somewhat lower energy-consumption were realized, then I would consider the upcoming GPU-generation as positive.

 

Ah, yes, not to forget another note from me as a marketing-guy:

AMD absolutely should defuse the users-debate, because they are afraid the R300-series will be only a pure rebrand, by officially clarifying, whether this series is a „rebranding“ or a „refresh“.

And for the chip-specifications in detail, the user can easily wait a few days!

Share this post

Repost 0
Published by critical-constructive - in computers
write a comment
December 4 2014 5 04 /12 /December /2014 19:30

AMD is in a difficult situation but also has big possibilitiesI have been dealing with AMD a couple of times in my blogs, in my German one as well as in my English one.
Up to now I mostly have put my focus on technical aspects of the US semi-conductor company.

In late November this has changed:
After around 2 months of hard work I have finished 2 big case-studies about AMD on my German-language blog:

  1. Marketing case-study about AMD 2014
  2. Case-study about Vishera-2.0 FX-CPUs

I am not going to translate these big documents into English, but I am going to translate the Abstract and the conclusions of all chapters of both documents.
It's on my schedule in late December 2014, hopefully I am able to keep it.

Right now I want to hint all German-speaking visitors on my papers.
In addition I also want to summarize the most important key-points right now:

  • AMD has been involved in a 2-front long-term static-warfare with Intel and Nvidia. As usually there is nothing to win in static-warfares you have to break out of this situation.
  • So AMD has changed it's strategy reducing it's focus on traditional PC- and server-processors and has made enormous efforts to get it's feet into new market-segments.
  • Tough this basically is the absolutely right step, AMD has overshot the target:
    The company has lost much more ground in the traditional market-segments than it has been able to gain in new areas.
  • As a result AMD should modify it's strategy and decrease efforts in new product-categories and re-increase efforts in some traditional product-lines.
  • Especially the FX-CPUs are a key product. They are the top-models in the company's line-up, are often sold directly to end-customers and to a high degree stand for the whole company's expertise
  • Also from the technical point of view the current FX-processors are not bad, but they won't remain competitive against Intel's 2015 CPUs.
  • So AMD should deeply revise them and introduce Vishera-2.0 as soon as possible.

I really hope to give a broad public but most of all AMD's management an understanding of how important these measures are.
And I also hope, that the loyal AMD fans are going to make pressure and to demand the strategy-adaptation towards more traditional products insistently.

Share this post

Repost 0
Published by critical-constructive - in computers
write a comment
October 2 2014 5 02 /10 /October /2014 22:15

It's hard for me writing English blog-entries frequently as this language isn't my mother tongue and it takes me much more time realizing entries in solid quality than in German. But this entry is quite a compulsory one.

After having criticised AMD for supposingly not introducing revised FX processors respectively supposing, that they are going to raise the processors TDP (the power class), I have to make a statement to their new FX-CPUs, codename "Vishera" (1.0).

 

In late August 2014 the Californian semiconductor-company has introduced 3 new 4-module FX-processors:

  • FX-8370 – 4.0 / 4.3 GHz, 125 watt TDP
  • FX-8370E – 3.3 / 4.3 GHz, 95 watt TDP
  • FX-8320E – 3.2 / 4.0 GHz, 95 watt TDP

So the 8370's TDP stays the same as it's predecessor's 8350.

In addition 2 new, more energy-efficient 4-module CPUs with 95 watt are now available as well.

 

So how do I judge the new products:

  • Definitely a step in the right direction and so positive. Especially the 95-watt versions are something new to most markets and for sure something useful. Now for users more efficient CPUs with good applications performance are available.
    In general energy-efficiency has slightly increased, also for the 125 watt 8370.
    So, you could call theses CPUs Vishera 1.5.
  • On the other hand, the silicon still is exactly the same since October 2012. And because of that more improvements have not been realized. And because of this for many PC-buyers, the difference limen is too small, and they will not go for new products.
  • It's also unsatisfying, that no new 2- or 3-module processors have been introduced.
  • Nevertheless it's positive to hear news from AMDs FX-processors and to see little improvement

As mentioned a couple of weeks ago, I can't understand AMD for not having revised the piledriver-cores used in the current FX-CPUs.

Spending them just the same revisions as for the Richland APUs 1 ½ years ago, would slightly increase the processing-performance und significantly improve the performance per watt.

 

Still the current Vishera 1.0 / 1.5 processors are competitive even against Haswell Core i3 and i5 in terms of application performance and performance per Euro / Dollar. But this could change with Intel's upcoming Broadwell CPU-range.

As AMD wants to stay in midrange-CPU market-segment and also considers a comeback in high-end segment 2016, it were really important, offering improved products. On one hand to stay competitive against Intel also in 2015. But even more than that for pleasing AMDs loyal CPU-fans, who are lusting for – if not completely new, excavator-based – significantly updated products.

 

However: I still like AMD, especially for their APUs (I have a A8-3870K Llano myself) and still do recommend their products in my purchase guidances. My guidances up to now are only available in German language, but I plan to publish them in English as well.

Share this post

Repost 0
Published by critical-constructive - in computers
write a comment
August 15 2014 6 15 /08 /August /2014 20:00

AMD seems to introduce no FX-CPUs in the foreseeable futureUS CPU-manufacturer AMD has something really extra-ordinary in semi-conductors business: A loyal fanbase, you can even say a base of hardcore-fans of the company.

 

As long as AMD processors are at least reasonable competitive to Intel, the top-dog among the CPU-manufacturers, the fans buy AMDs products.

 

But the company obviously doesn't value their loyal fans attitude adequately: AMD hasn't introduced any update to the Piledriver-cores of it's current Vishera FX-CPUs since October 2012, when they made their debut.

Though extending the Vishera product-range in the first half of 2013 with higher clocked and higher power-consuming models – no new stepping has been introduced.

 

What makes AMDs decision very opaque is the fact, that in summer 2013 the Richland-APUs, using the same Piledriver CPU-cores, have received a new stepping making them significantly more energy-efficient.

It seems to be likely, that AMD does not place value to "classical" CPUs at the moment, but on APUs for notebooks & mobile devices as well as on semi-custom chips used i.e. in Sony's "Play Station 4" and Microsoft's "Xbox One".

 

To the news in June 2014, that energy-waster CPUs FX-9590 and FX-9370, both with an incredible high TDP of 220 watt, are being offered with water-coolers ex-factory, according to postings in web-forums the AMD fans reception mostly was unsympathetic because not getting more energy-efficient new CPUs.

The fans, especially the fans of "classical CPUs" feel neglected if not kidded.

 

A positive signal is the release of the FX-8300 also as a retail-version (you can buy separately). The OEM version of this CPU has been released in December 2012.

There are also rumours for a FX-8370, a higher clocked FX-8350. Though yet no clear data is available for this processor. But as there is not a single sign of AMD make a new stepping, it's very likely, that such a processor will have a higher energy-consumption.

My guess is, that the TDP will increase from 125W (8350) to 140W (8370).

 

So all signs point to AMD not realizing a new stepping for it's FX-series, not making a Vishera-2.0 - although promised making enhancements to it in 2012. This were sad. Though I don't need a new processor myself right now, I totally understand every AMD fan as well as every CPU-buyer being fed up for not getting the best possible product.

Though the hardware improvements were limited to the current Visheras especially the performance-per-watt would improve appreciably. Vishera 2.0 were more competitive towards Intel products and would give AMD positive public-relations through a significant higher attention by the press.

 

And a FX-8450, a FX 8420 and a FX-6400, as described in detail in my blog-entry from January 2014, with slightly increased clock rates and reduced power-consumption are just appealing.

 

"Hope dies at last" is the translation of a famous German proverb.

AMD really should not turn it's back to traditional CPUs but should reward it's loyal fans with not even completely new but significantly updated products. Last but not least AMD would benefit itself most making this step.

 

 

Update 18th of August 2014:
Today I surprisingly have found new information to upcoming AMD Vishera FX-CPUs.
According to Hexus.net in autumn 2014 3 new FX-8xxx processors (with 8 integer-cores) are going to be introduced.
It that becomes true - it were just great :-)

Share this post

Repost 0
Published by critical-constructive - in computers
write a comment
January 17 2014 6 17 /01 /January /2014 21:35

AMD-FX-Logo-Vector.jpgAs written in my last blog entry, so it is fixed that AMD towards spring 2014 revises its Opteron server-processors.

However it is unclear so far whether also an update for the FX desktop-CPUs will be left over.

 

For me and probably for all loyal AMD fans however is clear that there must come one.

  • Intel has launched its new HASWELL processors on the market already in mid-2013, AMD's counterattack is pending; additionally Intel wants to present it's next generation of CPUs, codenamed "Broadwell", in autumn 2014.
  • From all we know about the revised Piledriver-2.0 cores, already used in the Richland-APUs, they can significantly improve the Performance-per-Watt - which is absolutely worth to be improved for the current FX-CPUs
  • The AMD fans really are lusting for new products;
    AMD has - despite significant deficiencies in energy efficiency and computing-performance - a loyal fanbase, that seeks for new products and deserves them
  • For marketing reasons, it is highly useful to revise the FX-product-range:
    Higher processing-power or energy-efficiency and thus more competitiveness compared to Intel, with this lots of media-coverage and thus public attention would be sure for AMD

The Vishera-2.0 lineup could look like this:

  • FX-4400: 2 modules, TDP 95 Watt, clock 4,0 / 4,1 / 4,2 GHz
  • FX-6400: 3 modules, TDP 95 Watts, clock 3.8 / 3.9 / 4.0 / 4.2 GHz
  • FX-8420: 4 modules, TDP 95 Watts, clock 3.5 / 3.7 / 3.9 / 4.1 GHz
  • FX-8450: 4 modules, TDP 110 watts, clock 4.0 / 4.1 / 4.2 / 4.3 GHz

3 (instead of 2 so far) turbo-levels, analogous to the Richland APUs, are also likely for the Vishera 2.0 FX-CPUs.

 

In any case AMD should focus on energy efficiency, Intel already has done so and other, upcoming AMD processors will pay high attention to the efficiency as well.

 

For AMD 2014 it's also not just about to introduce an evolution-level of FX-processors for factual reasons but also to keep it's own word regarding the implementation of it's fixed-route.

Finally AMD in 2012 has prepared the plan, introducing a revision of the FX-series *) into the market in 2013.

 

 

*) "Vishera” will also carry over into 2013 but with some enhancements over the 2012 version.

Share this post

Repost 0
Published by critical-constructive - in computers
write a comment
January 17 2014 6 17 /01 /January /2014 21:30

AMD-Opteron-Logo-Vector.jpgOver 1 year ago, AMD has introduced its current FX-series-processors, codenamed "Vishera". The Piledriver-Cores used by the Visheras 2013 should have been revised as well as at the Trinity APUs. They haven't been so far however.

 

This has been incomprehensible for experts, as the Piledriver-2.0-cores in the Richland-APUs are already in use since mid-2013.

But now AMD's strategy seems to crystallize:

At least, Opteron Server-CPUs, that are closely related with the FX-CPUs, new models with the code name "Warsaw" will be introduced around spring 2014.

The Warsaw-Opterons work with revamped Piledriver-Cores, so Piledriver 2.0 (or maybe even 2.5).

 

Supposed background for this seems to be that AMD wants to deploy the new SteamrollerB-Cores only in the Kaveri APUs, but not in the Opteron- or FX-CPUs. At the "normal" CPUs, this generation obviously should be skipped and from 2015 new models with the then up-to-date Excavator-Cores should be presented. The 2015 processors probably will be no pure CPUs, but APUs with integrated graphics-units.

 

This strategy is generally very reasonable from my point of view.

Since even though the SteamrollerB-Cores IPC (= instructions per cycle, performance per MHz) are approximately 15% higher, as those of their Piledriver-Predecessors, they are also lower clocked. The CPUs computational-power should rise only around 5-7% (Remark: According to the first reviews it has only risen around 1.5%).

And the expenses introducing a completely new processor are a multiple of what the costs revising an existing one.

 

For the Kaveri-APUs the situation looks differently: They also have integrated graphics-units - which approximately get 15% more processing-power on the floor than their Richland-Predecessors.

 

In this respect I endorse AMD's plan under 1 condition:

That AMD puts the revised cores not only into the Warsaw-Opterons for the server market, but also into the FX-CPUs for the desktop.

Anything else would be an outrage for the AMD fans!

Share this post

Repost 0
Published by critical-constructive - in computers
write a comment
March 9 2012 6 09 /03 /March /2012 20:48

In January 2011 AMD has revealed the mid-range models of its newly designed Radeon HD 7000-series.

Compared to its predecessors, the 6000-series graphics cards, the current generation is manufactured not only in 28nm instead of 40nm design, also the chip architecture is completely new. Accordingly the expectations were high that the Radeon HD 7000 cards will bring a big performance gain.

 

In the meantime pretty much the entire press has tested the 7750 and 7770 HD Radeons and has come to a unanimous verdict:

The new AMD mid-range-cards deliver the same performance with greatly improved efficiency.
And: Almost all of them interpret this insofar as unpleasant as gamers - in their opinion - expect significantly higher performance gains.

 

Well, quite possibly they may be true. Many hardcore gamers just want to put as much power as possible into their computers. No matter how much energy these parts are consuming, sometimes even no matter whether they really need this power.

 

AMD with its two mid-range 7000 Radeons is going the road of efficiency-increase, even though with different tactics for one of each:

·        The Radeon HD 7750 is opted according to the economic minimization-principle - the same output of computing power while reducing the energy-consumption compared to predecessors

·        The Radeon HD 7770 AMD is opted according to the economic maximization-principle – get out the maximum of computing power at the same energy-consumption

 

The Radeon HD 7750 delivers a minimally slower performance in games, especially in GPGPU applications a minimally faster performance than the Radeon HD 6770. On average both of them deliver virtually the same performance.
Compared to its direct predecessor HD 6750 it's clearly superior in all applications.
But there is one crucial difference: The 7750s maximum power-consumption is 61 watt - compared to 86 watt of the 6750 and even 108 watt of the 6770.
This is around a 53.5% (!) better performance per watt.

 

The Radeon HD 7770 in turn with 108 watt has the same maximum power consumption as its direct predecessor HD 6770. Performance-wise, its playing in the same league as the higher placed 6000-models Radeon HD 6790 and HD 6850 – in gaming closer to the 6790 in GPGPU applications closer to the 6850.
At the same power consumption level, the HD 7770 offers the performance of a 1 ½ class’s higher ranked model of the previous generation. You could say that it’s a kind of fictitious Radeon HD 6830 in permanent power-saving mode.

 

Anyway in terms of power-consumption for graphic-cards I have an explicit view:

Graphics cards with power consumption of at least 150 watt in terms of resource-efficiency are borderline, cards with a power consumption of 200 watt or more for most purposes are intolerable.

At resolutions up to 1600 x 900 pixels AMD's Radeon HD 7770 really is an excellent choice, offering really powerful graphics performance. On 22-inch computer-screens you simply don’t need more.

To me AMD with its efficiency strategy is on the absolutely right way, from an ecological point of view even on the only possible way. Thus eco-sexy.

Share this post

Repost 0
Published by critical-constructive - in computers
write a comment
February 26 2012 1 26 /02 /February /2012 20:21

In „AMD stagnating“ I have made a critical analysis of AMDs status quo in early 2012.

Today I want to show ways how AMD can get out of this mess. Especially I'm going to focus on AMDs processor-division, their graphics-division is already well on the way.

 

At the entry-range with the C-and E-series APUs AMD is on track. Indeed the introduction of their 2nd generation APUs in 28nm manufacturing-process was canceled in early 2012 due to technical problems. Nevertheless the current generation is still fully competitive with Intel’s Atom processors and is selling well.

 

The A-series medium-range APUs not only have got mostly positive reviews by the specialists press but are also very popular among buyers. Given to AMDs production-troubles too popular.

 

Even the FX processors are well received by the AMD fans. Their sales numbers are solid – though they are performing pretty mediocre.

Their architecture is completely new and still yet fully designed for Windows 8, which probably will be published in autumn 2012. However it will take until 2013, when the upcoming Windows will be significantly spread.
Up to then, it is necessary bridging 1 year.

 

And this bridge AMD can build with 2 products:

1.      A further evolution of the Llano APUs

2.      The introduction of Phenom III processors, based on Llano

 

Under Windows 7 the A-Series APUs and the old Phenom II processors are still delivering appropriate performance. So what’s more obvious than furthermore expediting this decisive bridge-technology in 2012?

To put it simply the Llanos are Phenom II processors shrinked from 45nm to 32nm, that have an integrated graphical processor instead a L3-cache.

 

At the Llanos, troubles in the manufacturing-process have occurred. That’s why their yield-rate is low and up to now AMD has not been able to significantly improve it.
The main reason is believed to be that there are problems integrating the graphics-unit in the processor (http://www.brightsideofnews.com/news/2011/8/30/amd-bulldozer-yields-are-fine2c-fusion-llano-apus-are-not-gpu-is-the-issue.aspx).

 

Nevertheless, a slight increase in terms of clock-speed together with the introduction of a turbo-mode to the 100-watt class Llano models such as the A6-3670 and the A8-3870K must be possible.

As the new flagship, AMD should introduce an A8-3950K with at least 3.1, better 3.2 GHz base-clock and a 300 MHz higher turbo mode (i.e. 3.4 to 3.5 GHz)..
Such a processor would be almost on the same level like the fastest four-core Phenom II CPUs that have expired in 2011. In addition they would also offer a powerful graphics-unit.

 

And especially for gamers a Phenom III is indispensable for AMD.

The Phenom III basically were a Llano, but instead the integrated graphics unit 8 MB of L3 cache would be fitted on the processor.
In late 2011, numerous benchmark-tests have proven that still now AMDs fastest processors for gamers often is the Phenom II with 4 cores.
The quickest 4-core-processor, the Phenom II X4 980 has beaten the FX-8150 with 8 cores as well as the Phenom II X6 1100 with 6-cores in around 50 % of all gaming-tests.

 

The Phenom II X4 980 offers 4 processor-cores at a clock frequency of 3.7 GHz.
A Phenom X4 1000 III with 4 cores and a clock frequency of 4.0 GHz and maybe a Phenom III X4 975 with 3.6 GHz would be the perfect complement to AMDs product-lineup.
In nearly all games the Phenom X4 III would crown itself to AMDs fastest current CPU and would also be able keeping up with Intel's Sandy-Bridge CPUs well.

 

For multi-threaded applications, e.g. data-compression or video-editing, the FX processors already are strong on the road.

 

However I am afraid, that AMD is not going to seize the idea of a Phenom III. The course is still set for further FX-models.

But for a third evolution-step of Llano, I am at least a little bit optimistic. This step would be very important for AMD for coming well through the year 2012.

 

By all means AMD is in dire need for sales hits.

Finally I furthermore want to use computers with processors from the likeable Californians.

Share this post

Repost 0
Published by critical-constructive - in computers
write a comment