Bulldozer Lack of Fine Tuning?

Why performance of the long-awaited Bulldozer was below expectations is not only because it was late, but because AMD had adopted design techniques that did not allow it to tweak performance. At least that is according to Cliff A. Maier, an AMD engineer who left the company several years ago.

The management decided to abandon practice of hand-crafting various performance-critical parts of its chips and rely completely on automatic tools. While usage of tools that automatically implement certain technologies into silicon speeds up the design process, they cannot ensure maximum performance and efficiency.

Apparently, automatically-generated designs are 20% bigger and 20% slower than hand-crafted designs, which results in increased transistor count, die space, cost and power efficiency.

"I had been in charge of our design flow in the years before I left, and I had tested these tools by asking the companies who sold them to design blocks (adders, multipliers, etc.) using their tools. I let them take as long as they wanted. They always came back to me with designs that were 20% bigger, and 20% slower than our hand-crafted designs, and which suffered from electro-migration and other problems," the former AMD engineer said.

AMD publicly said that each Bulldozer dual-core CPU module with 2MB unified L2 cache contains 213 million transistors and is 30.9mm2 large. By contrast, die size of one processing engine of Llano processor (11-layer 32nm SOI, K10.5+ micro-architecture) is 9.69mm2 (without L2 cache).

As a result, all four CPU modules with L2 cache within Zambezi/Orochi processor consist of 852 million of transistors and take 123.6mm2 of die space. Assuming that 8MB of L3 cache (6 bits per cell) consist of 405 million of transistors, it leaves around whopping 800 million of transistors to various input/output interfaces, dual-channel DDR3 memory controller as well as various logic and routing inside the chip.

800 million of transistors - which take up a lot of die space - is an incredibly high number for various I/O, memory, logic, etc. For example, Intel's Core i-series "Sandy Bridge" quad-core chip with integrated graphics consists of 995 million.

As a consequence of inefficient design and relatively low performance, AMD has to sell its eight-core FX series processors (315mm2 die size) for up to $245 in 1000-unit quantities. By contrast, Intel sells hand-crafted Core i-series "Sandy Bridge" quad-core chips (216mm2 die size) for up to $317 in 1000-unit quantities. Given the fact that both microprocessors are made using 32nm process technology [and thus have comparable per-transistor/per square mm die cost], the Intel one carries much better profit margin than AMD's microprocessor.

For more information, read here:
Insideris : AMD Spreads Propaganda, Ex-Employee Speaks Out
Mr. Maier posting on MacRumor Forum

Bulldozer for Desktop is Finally Out

Zambezi is the fourth-longest river in Africa (the one with the beautiful Victoria Falls), and also the codename for the new FX series of "Bulldozer" desktop enthusiast CPUs from AMD. After months of speculation and spin about the launch date and performance numbers, the new processors have finally hit the online shops. Yes, we saw the launch today, and the long awaited Bulldozer is finally out.


That new AMD processors based on Bulldozer microarchitecture are one of the most highly anticipated products not only of the year, but at least of the last five years. There are several reasons for that as well as for the fact that AMD products have so many fans. Remember the times when AMD processors were better than Intel ones in all aspects? Some users like AMD products for the balanced combination of price and performance that they have to offer. And some may have been carried away by the passion, with which they talked about the advantages of the new microarchitecture they have been working on. All this combined with the years of waiting for the new Bulldozer processor generation produced high interest and excitement.

Here we bring you some reviews of the top of the line FX-8150:
Tech Report
X-Bit Labs
Guru3D
Bit-Tech
AnandTech