This site may earn chapter commissions from the links on this page. Terms of employ.

A new class-action lawsuit against AMD is arguing that the visitor engaged in fraud and deceptive marketing when it claimed that its Bulldozer processors had eight cores. The suit, Dickey v. Advanced Micro Devices Inc, argues that AMD's Bulldozer (and presumably Piledriver, Kaveri, Carrizo, and so on) all effectively incorporate just half the number of cores AMD claims they exercise, and that the company'southward decision to combine certain aspects of the core into a module and share resources constitutes a deceptive practice.

If the writeup at Legalnewsline.com is remotely accurate, the lawsuit is utterly without technical merit. The accommodate supposedly alleges that because Bulldozer shares certain core resources, the cores can no longer work independently and the chip is no longer capable of performing eight instructions simultaneously. If that's the hook Dickey is hoping to hang his lawsuit on, he picked a bad one. While it's true that AMD shared core resources within Bulldozer, the chip doesn't piece of work the way Dickey alleges it does.

Diving into 'Dozer

Four years ago, we published an investigation into how Bulldozer handled multi-threading scenarios and what kind of penalty the chip took when running two threads on the aforementioned module. Our investigations showed that Bulldozer took roughly a 20% performance striking when scaling up to viii cores compared to the scaling we'd look to come across from a conventionally designed processor.

Cinebench scaling when using four threads.

Cinebench scaling when using iv threads. Clock speed was locked to 3.6GHz.

The graph above shows Bulldozer functioning when Cinebench has been told to use only 4 threads in every example. Nosotros and then controlled where the threads executed. The 4M / 4C test means we locked 1 thread to each module. The 4M / 8C test means we specified four threads, only let them run on any cores. The 2M/4C examination means we locked all four threads to run on a single module. The results show that the 4M / 4C way is conspicuously faster than the 2M/4C mode — virtually 18% faster.

Bulldozer's total operation with all eight cores enabled, withal, was even so higher than the 4M / 4C way, as shown below:

Cinebench performance when all eight threads ran on eight cores. CPU clock speed was locked to 3.6GHz.

Cinebench operation when all eight threads ran on eight cores. Clock speed was locked to 3.6GHz.

An eight-core Bulldozer was capable of roughly the same performance every bit a six-cadre Thuban. The performance striking from combining and sharing cores, in other words, was nowhere nigh the fifty% that Dickey alleges.

Since Bulldozer shipped, AMD has made small-scale improvements to the CPU's overall efficiency and functioning. Kaveri cut the penalty for multi-threading in half, from ~20% to ten% compared with typical core scaling. If AMD hadn't been forced to lower clock speeds to recoup for its 28nm manufacturing procedure, Kaveri would've outperformed Richland across the board. The 10-20% penalty for multi-threading compared to conventional core designs is nowhere nigh the halving that Dickey alleges.

At present, it'southward absolutely true that the Bulldozer family unit of products has had much lower single-thread performance than either previous AMD CPUs (in many cases) or Intel fries (in virtually all cases). Simply this lawsuit doesn't appear to argue that AMD mismarketed its CPUs considering single-threaded performance was weaker than expected, but considering multi-threaded scaling was critically harmed by the decision to share various aspects of the underlying architecture. Weak single-threaded performance and high power consumption created a state of affairs in which BD could neither hit its target clock frequencies nor its IPC targets. Critically, these issues do not disappear when the CPU is run in one-thread per module mode.

Dickey's lawsuit is incorrect on other areas of fact likewise. Bulldozer does share a single FPU cake per work unit, but consumer workloads are rarely FPU-heavy. Each CPU module does comprise  the eight integer pipelines y'all'd wait in a typical dual-cadre conventional chip (four ALU + 4 AGU per module). Dickey refers to Bulldozer as being unable to "perform eight calculations simultaneously," only this is imprecise, inexact language that does not reflect the complexity of how a CPU executes code. Bulldozer is absolutely capable of executing eight threads simultaneously, and executing eight threads on an viii-core FX-8150 is faster than running that same bit in a four-thread, four-module mode. Bulldozer tin can decode xvi instructions per clock (not eight) and it tin can keep far more than eight instructions in flying simultaneously.

Courts can't define cores

This lawsuit essentially asks a court to define what a core is and how companies should count them. As annoying as it is to meet vendors occasionally abuse core counts in the name of dubious marketing strategies, asking a courtroom to make declarations about relative performance between companies is a cure far worse than the illness. From big iron enterprise markets to mobile devices, companies deploy vastly unlike architectures to solve different types of problems. An eight-cadre, Cortex-A7-based, mobile SoC is a very different animal from an eight-core big.Little Cortex-A57 / Cortex-A53 configuration. That chip is very different from an Oracle M7 or the SPARC T5. The T5 doesn't pack the per-cadre performance of Intel'due south xviii-core Xeons, or IBM's Power8.

CPU-Scaling

Equally this chart makes clear, CPU IPC and clock speed vary widely, even within Intel products over time.

Bulldozer may accept performed more like a quad-core chip from Intel, but that doesn't mean it actually was a quad-cadre chip. The performance benefits from running the CPU in a quad-core-equivalent configuration weren't well-nigh large enough to make that merits. The argument might stand if AMD had marketed BD every bit having groovy floating-bespeak functioning, only the company's disclosures and briefings all clearly stated that BD would have only 4 floating-indicate units. Anyone buying the arrangement for FPU work would have known that long before hardware shipped.

AMD has, in a very real sense, been thoroughly punished for the CPU it brought to marketplace in 2022 — and this lawsuit makes claims that don't concord upwardly to technical scrutiny.