AMD: What happened?

[Blitzvogel]

AMD got complacent over their performance leads in 2005 and 2006. Then Core 2 Duo came out while they had been buying ATi, and AMD has barely had it's head above water since then.

 

[beginner99]

The way I see it is that AMD made a judgment call 4-5 years ago and assumed that software progression would keep up with and even out-pace an ever-upward-scaling core count in the computing world that would try to feed the software that favored moar coars over IPC. It was a piss-poor call.

Depends. BD was meant for servers and software in server world scales pretty good (RDBMS, virtualization,...). The issue is if your 8 cores are not faster than intels 4 and you use more power -> not attractive buy. Then if you have per core licensing of software...your great idea is actually pretty bad.

 

[mrcool63]

to put it simply, when you can get an 8-core for the price of a 4-core it will be inferior to the 4-core in anthing that does not use all its cores... so for the rest there is intel
you may argue about the ipc, or the shared resources but in the end the above statement is all that matters..

 

[blckgrffn]

I'd also say that BD was built on Intel 32nm and SB was built on GF 32nm, the conversation might be different

AMD (and everyone else?) is being out classed and outspent by Intel in more ways than one and its important to keep that in mind.

 

[amenx]

Intel is building a 14nm fab at the cost of $5 billion. Thats almost what AMD as a whole is worth today. Thats what the problem is. Peanuts vs billions.

 

[frozentundra123456]

to put it simply, when you can get an 8-core for the price of a 4-core it will be inferior to the 4-core in anthing that does not use all its cores... so for the rest there is intel
you may argue about the ipc, or the shared resources but in the end the above statement is all that matters..

I would put it a little differently. I dont really care about how many "cores" a processer has, or whether Bulldozer is really 8 core or 4 core or something in between.

What matters is performance per dollar and performance per watt. Right now Intel wins on both counts vs Bulldozer in all but a few heavily multithreaded cases. Doesnt matter why, really, just the bottom line of how good is the performance.

 

[Dravic]

It can happen, look at Apple. At one time they were nearly out of business but now are making money hand and foot. I dont particularly like their products, business practices, and they are grossly overpriced. But they did turn things around. Not saying it can happen often, but it did happen.

I guess what I am saying is that I sympathize with AMD to an extent because of the lower amount of resources they have relative to intel, but they have made some bad decisons and executed poorly as well, and cant blame that on a lack of R and D.

Apple didn't make it because of their core products, they almost went out of business trying to fight MS and the entire PC industry on its own[os, hardware, and app development]. It took a major cash deal with MS to put Office on the Macs officially that saved Apple. It wasn't until a new take [ipod] on an existing consumer device [mp3's] that apple started to turn around.


Even with apple's success now they are in a tough spot long term. There is no way 1 company can out innovate a whole industry by itself. Android as %50+ of the smartphone market because it's open, and android tablets will be > %50 of the market by the end of they year. You'd think they would have learned their lesson the first time they tried to take on the hole industry.

But they have sooooo much cash they should be able to figure something out.


AMD got there $2B cash infusion from Intel for their practices, and now we will see how they fair. AMD may pull something out every once in a while that makes them competitive, but i wouldnt expect it every new product launch with the kind of resource disadvantage they have.