New PC Power Rules Only Apply to Idle, as Alienware Squawks Alone

New PC Power Rules Only Apply to Idle, as Alienware Squawks Alone

When we covered the news that Alienware will no longer sell most of its desktops in California, Colorado, Hawaii, Oregon, Vermont, and the State of Washington, we noted that there had to be more to the story. Alienware might be loudly declaring that it can’t sell in these states due to power efficiency requirements, but nobody else is. In fact, the word from other manufacturers is that nobody considers these requirements much of a problem.

It turns out, there’s a crucial bit of context Alienware didn’t disclose: The new power requirements that went into effect on July 1 apply to idle power consumption, not load. The sharp constraints on total PC power consumption do not apply when the system is being used for things. They apply when it isn’t. All of the standards and requirements, including the ones that went into effect back in 2018, explicitly target idle power consumption, and nobody but Alienware seems particularly chuffed about it.

This actually explains one of the strangest things about the entire affair. The function that the California Energy Commission used to calculate projected GPU power consumption is based entirely on memory bandwidth and allocates virtually no electricity to GPUs as memory bandwidth scales up. Under the published formula, GPU power consumption is allowed to scale as follows:

New PC Power Rules Only Apply to Idle, as Alienware Squawks Alone

A GPU with 3.2TB/s of memory bandwidth is therefore allowed 1.91x the power consumption of a GPU with 25GB/s of memory bandwidth. This is ludicrous when considered as a target for a loaded GPU. It makes more sense as a target for an idle card. I’m not sure why the commission chose to base its formula on memory bandwidth, specifically, but the formula is intended to calculate the power consumption of a GPU when it is not in use.

There appears to be little concern as to whether these goals are achievable. Intel launched a program to help motherboard manufacturers make certain they came in on-target with 400-series chipsets several years ago. There are no regulations in place regarding the maximum potential power consumption of PCs under active load, OEMs do not appear to be restricted from selling systems that consume large amounts of electricity in any fashion, and this entire affair is an attempt to push manufacturers to waste less power when the computer isn’t doing things, not restrict machines from their fullest potential when performing useful work. According to this Intel slide, Alienware systems that include GPUs with frame buffers of more than 600GB/s aren’t subject to this restriction at all.

New PC Power Rules Only Apply to Idle, as Alienware Squawks Alone

It looks increasingly as if Alienware was caught by a set of regulations they didn’t prepare for and is attempting to foist the blame on to regulations passed five years ago rather than acknowledging its own failure to comply with well-known upcoming regulations. It was always the plan to phase these requirements in over time, and the Tier 2 compliance date of 2021 was discussed in this 2016 Reuters story. Alienware has known this was coming.

The actual regulations don’t do a great job spelling out that these requirements only apply to idle power, but a recent fact sheet on the new rules states: “The Energy Commission recognizes four different non-active operational modes: short-idle, long-idle, sleep and off-modes. Some computers consume 50 watts of electricity in these idle modes. The Title 20 standards are cost-effective, technically feasible and limit the amount of electricity computers and monitors are allowed to consume when not in active mode.”

Long and short idle refer to times when the PC is performing operational background tasks and spending most of its time in the C6/C7 low-power states, with no ongoing processor tasks. Sleep and off are self-explanatory. The guidelines as published are an attempt to reduce waste electricity that PCs expend without performing useful work and not any kind of effort to choke off performance.

There is currently no sign that any manufacturer other than Alienware is having any problems meeting these standards. There is no indication of upset from any of the companies we cover on a regular basis. We don’t know exactly why Alienware’s systems don’t pass muster, but it does not appear to be for any reason related to maximum power consumption — only idle. For now, this does not appear to be an industry problem.

Continue reading

Apple’s M1 Continues to Impress in Cinebench R23, Affinity Photo
Apple’s M1 Continues to Impress in Cinebench R23, Affinity Photo

New Cinebench R23 benchmarks paint AMD in a more competitive light against the M1, but Apple's SoC still acquits itself impressively. The Affinity Photo benchmark, however, is a major M1 win.

Why Apple’s M1 Chip Threatens Intel and AMD
Why Apple’s M1 Chip Threatens Intel and AMD

Intel's own history suggests it and AMD should take Apple's new M1 SoC very seriously.

Nvidia Unveils ‘Grace’ Deep-Learning CPU for Supercomputing Applications
Nvidia Unveils ‘Grace’ Deep-Learning CPU for Supercomputing Applications

Nvidia is already capitalizing on its ARM acquisition with a massively powerful new CPU-plus-GPU combination that it claims will speed up the training of large machine-learning models by a factor of 10.

The Xbox Series S Is Handicapped by Its Storage Capacity
The Xbox Series S Is Handicapped by Its Storage Capacity

The Xbox Series S has been favorably received, for the most part, but the console's low base storage makes the Xbox Series X a better value for a lot of people.