AMD is currently shipping its first server processors based on x86 “Bulldozer” architecture.
Initial production of the 16-core x86 processor, codenamed “Interlagos,” apparently kicked off in August – with many of the chips now headed to large custom supercomputer installations.
According to Charlie Demerjian of SemiAccurate, Bulldozer’s official launch was slightly delayed due to a number of issues, including yield and 32nm wafer allocation priorities at GlobalFoundries.
“Bulldozer is built on the same process as Llano, 32SHP, so the yield problems are going to hit that chip at least as hard,” wrote Demerjian.
“Llano has a large portion of it’s area taken up by GPU shaders, so you can salvage a lot of lower binned dies because of that, and more by going to 3 or 2 cores. Assuming any single module can be fused off in Bulldozer, it could have a very high resistance to defects as well.”
Demerjian also noted that Bulldozer is a larger die compared to Llano – meaning yields will noticeably decline.
“Time will tell, but the constant slipping by a week or two here or there is a pretty good sign that this is the case… When Bulldozer is released in volume, what SKUs are offered, and their availability should give a lot of clues about yields and bin splits.”
In addition, explained Demerjian, GlobalFoundries is (currently) only capable of running a finite amount of 32nm wafers. As such, it is up to AMD to allocate the wafers for the most profitable chips.
“[For example], if X * A is bigger than Y * B, make more Llanos. If X * A is smaller, make more Bulldozers… [Sure], yield improvements, customer bitching, and many other things come in to play to influence the split, but the gross calculation is simple. You run the one that makes the most money in the end,” he added.
Note: The launch date for the desktop version of Bulldozer (Zambezi) remains unclear, but industry sources believe it could debut sometime in October.