AMD Reorganizes

[Phynaz]

And Randy "40%" Allen is gone.

Text

 

[brxndxn]

Sweet..

I want processor/gpu "raid 5" and parity.. using uberCrossfire X.. Imagine.. dropping more graphics cards in to get faster cpu AND graphics!

Then, you could even take older cpu/graphics cards and assign them more mundane tasks because Microsoft will have a really awesome OS that lets you do that by then..

 

[Idontcare]

AMD really did seem to be the nexus of bad managers/ment there for a while. Possibly even surpassing auto-industry quality, but thankfully not in duration.

I've got really high expectations of Dirk, I can't think of anyone in the industry more capable of getting AMD turned around (outside of Andy Grove perhaps).

 

[ShawnD1]

Originally posted by: brxndxn
Sweet..

I want processor/gpu "raid 5" and parity.. using uberCrossfire X.. Imagine.. dropping more graphics cards in to get faster cpu AND graphics!

Then, you could even take older cpu/graphics cards and assign them more mundane tasks because Microsoft will have a really awesome OS that lets you do that by then..

http://en.wikipedia.org/wiki/DirectX_11#Direct3D_11

Microsoft unveiled Direct3D 11 at the Gamefest 08 event in Seattle, with the major scheduled features including GPGPU support, tessellation[11][12] support, and improved multi-threading support to assist video game developers in developing games that better utilize multi-core processors. Direct3D 11 will run on Windows Vista, Windows 7, and all future Windows operating systems. Parts of the new API such as multi-threaded resource handling can be supported on Direct3D 9/10/10.1-class hardware. Hardware tessellation and Shader Model 5.0 will require Direct3D 11 supporting hardware. [13] Microsoft has since released the Direct3D 11 Technical Preview. [14]. DirectX 11 is scheduled to be released July 2009.

IDC, does merging the CPU and GPU teams sound like a good idea? I don't know the details of how to design a processor, but this sounds like one of those things where merging two high quality yet different products creatives one crappy product.

 

[soonerproud]

Originally posted by: Idontcare
AMD really did seem to be the nexus of bad managers/ment there for a while. Possibly even surpassing auto-industry quality, but thankfully not in duration.

I've got really high expectations of Dirk, I can't think of anyone in the industry more capable of getting AMD turned around (outside of Andy Grove perhaps).

Lets hope they pull it off, though it looks grim for them now. I like the competition and I like having real choices in processors.

 

[Idontcare]

Originally posted by: ShawnD1
IDC, does merging the CPU and GPU teams sound like a good idea? I don't know the details of how to design a processor, but this sounds like one of those things where merging two high quality yet different products creatives one crappy product.

I view it as absolutely necessary.

As for the outcome, if we look at it pragmatically as a simple 2x2 orthogonal matrix where the inputs are "GPU" and "CPU" and the outcomes are "Good" or "Bad" we realize that while having a "good" cpu is still 50/50 and having a "good" GPU is also 50/50 the chances of having a great (good GPU and good CPU) fusion product have been reduced to 1 in 4.

This is of course flawed analysis but the point of the exercise is that it speaks to the elevated complexity of creating a product that needs to kill two birds with one stone...should it fail in either category the end result could be disastrous for the aggregate.

AMD is in a situation where they basically have to die trying, otherwise they stand serious risk of dying whilst not trying. By dying I don't mean bankruptcy but rather I mean what they stand for market-wise and position will die, Cyrix/Via style, and they will be relegated to a niche processor provider same as Via (if atom doesn't kill them too).

 

[VirtualLarry]

And here I thought that their execution with respect to the Phenom II and the HD4xxx GPUs was pretty good overall.

 

[soonerproud]

Originally posted by: VirtualLarry
And here I thought that their execution with respect to the Phenom II and the HD4xxx GPUs was pretty good overall.

It was but it should have happened at least a year sooner. They are going to get a good 6 months before the i5 is out, then they are back in the same spot they were with the X2's and C2D's.

 

[ShawnD1]

Originally posted by: Idontcare
AMD is in a situation where they basically have to die trying, otherwise they stand serious risk of dying whilst not trying. By dying I don't mean bankruptcy but rather I mean what they stand for market-wise and position will die, Cyrix/Via style, and they will be relegated to a niche processor provider same as Via (if atom doesn't kill them too).

There's nothing wrong with having a second-rate product (look how successful Walmart is), but there's a big problem with a lack of vendor support. Even when the Athlon was destroying the Pentium 4 in test after test, you couldn't buy an AMD based Dell or HP computer. That was several years ago, and the problem still exists in different ways. If you go to the Apple store, you'll notice that all Apple products are Intel/Nvidia.

It's nice to see AMD has significant support from Walmart, but and important market they don't have any footing in is the office workstation market. Most computers I've seen at past jobs were either Dell or IBM. All Dell workstation computers only use Intel processors. IBM's website is too complicated to figure out, but I'm going to assume they're mostly Intel since I've never actually seen an AMD based IBM workstation, although their servers have a lot of AMD options.

Merging teams doesn't really address this problem. Making a great product doesn't help if the product will not be purchased by some of the largest vendors.

 

[Martimus]

Originally posted by: Idontcare
AMD really did seem to be the nexus of bad managers/ment there for a while. Possibly even surpassing auto-industry quality, but thankfully not in duration.

I've got really high expectations of Dirk, I can't think of anyone in the industry more capable of getting AMD turned around (outside of Andy Grove perhaps).

The worst thing for the auto industry was that Chrysler came back after almost being dead. What that did was make auto makers realize they can make more money by becoming asset light like Chrysler did, and in the process the quality took a huge nose dive as they didn't truly understand yet how to control quality of so many suppliers. That has since been fixed, but it is hard to undo the damage of the cars they sold in the meantime that had unacceptable quality. The funny thing is that GM probably has the best quality of any car company in the world, but yet you hear that they should increase quality to compete with foreign markets. (I know this because I worked for a auto supplier that sold parts to GM, Fiat, Ford, Opel, and Volkswagen (at least they sold the products that I helped developed to these companies) And GM had by far the strictest quality requirements (followed by Ford surprisingly).

 

[Dark4ng3l]

The problem in the x86 cpu market is that there is basically no room for differentiation. If you have better cpu's then you can make better low power cpu for notebooks, a more powerful cpu, cheaper and more powerfull budget cpu's etc. Performance is everything more or less. AMD if trying to differentiate their products by going with a lot more cores and integrating the video to the cpu. In the end even if AMD had a faster product unless the difference is large(>20%) they won't be able to make large market share gains.

 

[Idontcare]

Originally posted by: Martimus

Originally posted by: Idontcare
AMD really did seem to be the nexus of bad managers/ment there for a while. Possibly even surpassing auto-industry quality, but thankfully not in duration.

I've got really high expectations of Dirk, I can't think of anyone in the industry more capable of getting AMD turned around (outside of Andy Grove perhaps).

The worst thing for the auto industry was that Chrysler came back after almost being dead. What that did was make auto makers realize they can make more money by becoming asset light like Chrysler did, and in the process the quality took a huge nose dive as they didn't truly understand yet how to control quality of so many suppliers. That has since been fixed, but it is hard to undo the damage of the cars they sold in the meantime that had unacceptable quality. The funny thing is that GM probably has the best quality of any car company in the world, but yet you hear that they should increase quality to compete with foreign markets. (I know this because I worked for a auto supplier that sold parts to GM, Fiat, Ford, Opel, and Volkswagen (at least they sold the products that I helped developed to these companies) And GM had by far the strictest quality requirements (followed by Ford surprisingly).

As this continues to draw suitable parallels to AMD's product situation, I'll say that what struck me with the big-three and the quality "image" is it appeared to me (as a consumer) that they expected the quality "gap" to evaporate via some mechanism in which the quality was supposed to be self-evident to the consumer.

They relied on the same discredited marketing dept to convince consumers the quality had been addressed, but after having used the same approach to sell crappy cars who was about to sink $30k into believing them this time? So in the absence of credible marketing team what else was the consumer exposed to which would have convinced them the quality gap had been addressed (or even reversed)?

I think AMD has similar challenges ahead of them, the Randy Allen's of the business did some damage to AMD that won't simply be undone by AMD releasing a superior product and having the same PR team tell everyone so.