...
AMD has for some time been a member of BAPCo, an industry organization that promotes, among other things, a benchmark known as SYSmark. In the past year or so AMD, with openness and transparency, has tried to explain why we believe this benchmark is misleading with respect to today’s commonplace applications − about a year ago I published a blog designed to explore this. If you work for a company that believes in transparency and integrity – and I do – then you have to take a stand and speak up when something is wrong.

BAPCo’s response to this blog was a threat to expel AMD from the consortium.

The heart of our complaint is this: the SYSmark benchmark is not only comprised of unrepresentative workloads (workloads that ignore the importance of heterogeneous computing and, frankly, favor our competitor’s designs), but it actually generates misleading results that can lead to very poor purchasing decisions, causing governments worldwide to historically overspend somewhere in the area of approximately $8B!
...
... our good intentions were met with an outcome that we believe does a disservice to the industry and our customers. We weren’t able to effect positive change within BAPCo, and the resulting benchmark continues to distort workload performance and offers even less transparency to end users. Once again, BAPCo chose to ignore the opportunity to promote openness and transparency.

While SM2012 is marketed as rating performance using 18 applications and 390 measurements, the reality is that only 7 applications and less than 10 percent of the total measurements dominate the overall score. So a small class of operations across the entire benchmark influences the overall score.
In fact, a relatively large proportion of the SM2012 score is based on system performance rated during optical character recognition (OCR) and file compression activities − things an average user will rarely if ever do.
And SM2012 doesn’t represent the evolution of computer processing and how that evolution is influencing average users’ experience. SM2012 focuses only on the serial processing performance of the CPU, and virtually ignores the parallel processing performance of the GPU. In particular, SM2012 scores do not take into account GPU-accelerated applications that are widely used in today’s business environments.

There are more things that AMD objects to in SM2012, like the excessive wall clock time consumed by its installation and execution. But this explanation will hopefully help you understand why, ultimately, we couldn’t look in the other direction.
...

So how can AMD stay in BAPCo? Simply put, we can’t. We have resigned from BAPCo and asked that our name and logo be removed from marketing materials promoting SM2012.

We will have our own bulldozer

"Bulldozer is going to disappoint people because we did not get the resources to build a great CPU, and it's not that we needed billions of dollars to make it a leader. We needed investment in people, tools and technology."
As every competitive spirit, one of our sources went on to confirm the quote above and expressed concern that "Bulldozer’s client performance is not going to be enough to deliver growth in the enterprise and in my view, that is the biggest driver behind our current strategy."

When asked about core performance, surprising information was that a Bulldozer core versus the existing cores in Llano will result in minimal improvements overall. Our sources went on to say that the launch of Llano clearly shows what is the current and future strategy - downplay CPU performance every chance you get. Everything has to revolve around the GPU.

Comments continued "As soon as you have analyzed the Bulldozer core, you're quickly going to find out that it disappoints at SYSmark. It won't matter, because by that time I am sure the current marketing strategy will be changed again when we have our new CEO in place."

The source argues that "we could not compete with Kite Refresh in '07, Puma and Congo in '08... neither could we compete with Tigris in '09 and we're still challenged with Nile and Danube."

2007: Kite Refresh - Tyler CPU, M690T Chipset
2008: Puma - Griffin CPU, RS780M Chipset
2009: Congo - Conesus CPU, RS780M Chipset; Tigris - Caspian CPU, RS880M Chipset
2010: Nile - Geneva CPU, RS880 Chipset; Danube - Champlain CPU, RS880 Chipset


Our source is dismayed at the longevity of the chipsets and alleged lack of innovation as well as lack of will to invest in removing the bugs which plagued consumers such as long standing issues with USB 2.0 performance.

The author of these lines believes that the key reason that pushed AMD into regression was the failed merger between NVIDIA and AMD in 2005. This lead to a panic acquisition of ATI Technologies, for which AMD paid a grand total of 5.9 billion dollars (5.4B plus 350 plus 120 million).

This emptied the coffers and engineering team was left without resources to go and fight against the rising Intel threat. AMD's reaction after the departure of Henri Richard was appointing Nigel Dessau, seasoned executive with IBM and SUN Microsystems behind him. Nigel was faced with an impossible task of competing against Intel with blanks, and the only choice left was to deploy the Fear, Uncertainty, Doubt i.e. FUD strategy patented by his previous company, centennial Big Blue. A shift occurred and all engineers were left was not making a discussion that would go along the lines of "ok, how do we make our architecture better?", "how do we work to regain performance leadership?". According to one of our highly positioned sources, the culture switched to "how do we discredit benchmarks and skew the numbers?"