News: 0001490654

  ARM Give a man a fire and he's warm for a day, but set fire to him and he's warm for the rest of his life (Terry Pratchett, Jingo)

Intel Efficiency Latency Control "ELC" Feature Slated For Linux 6.12

([Intel] 109 Minutes Ago Intel Uncore Efficiency Latency Control)


Last month I wrote about Intel Linux engineers working on [1]a new Efficiency Latency Control feature for their uncore driver. This ELC option allows for adjusting the behavior of the Intel uncore for efficiency versus latency characteristics. Those Intel ELC patches to the TPMI uncore driver are now queued up for merging with the upcoming Linux 6.12 cycle.

The Intel Efficiency Latency Control allows fine-tuning SoCs -- with a particula emphasis on Xeon processors -- to achieve greater energy efficiency if so desired. The ELC tunables can be controlled under Linux with some new sysfs options exposed by these patches. The ELC Linux patches sum up the capability as:

"In the realm of high-performance computing, particularly with Xeon processors, managing uncore frequency is an important aspect of system optimization. Traditionally, the uncore frequency is ramped up rapidly in high load scenarios. While this strategy achieves low latency, which is crucial for time-sensitive computations, it does not necessarily yield the best performance per watt, —a key metric for energy efficiency and operational cost savings.

The Efficiency vs. Latency Control (ELC) feature allows user to influence the uncore frequency scaling algorithm. Hardware monitors the average CPU utilization across all cores at regular intervals. If the average CPU utilization is below a user defined threshold (elc_low_threshold_percent), the user defined uncore frequency floor frequency will be used (elc_floor_freq_khz), minimizing latency. Similarly in high load scenario where the CPU utilization goes above the high threshold value (elc_high_threshold_percent) instead of jumping to maximum uncore frequency, uncore frequency is increased in 100MHz steps until the power limit is reached."

The patches adding this Efficiency Latency Control feature to the Intel uncore driver code have been queued into [2]platform-drivers-x86.git's for-next branch . With it now in the "for-next" code ahead of the upcoming Linux 6.12 merge window, it should be part of this next kernel cycle barring any last minute issues from coming to light.



[1] https://www.phoronix.com/news/Intel-Uncore-Efficiency-ELC

[2] https://git.kernel.org/pub/scm/linux/kernel/git/pdx86/platform-drivers-x86.git/commit/?h=for-next&id=bb516dc79c4a6334d5ef6bdbd0d262cf8be9db8e



phoronix

SEMINAR ANNOUNCEMENT

Title: Are Frogs Turing Compatible?
Speaker: Don "The Lion" Knuth

ABSTRACT
Several researchers at the University of Louisiana have been studying
the computing power of various amphibians, frogs in particular. The problem
of frog computability has become a critical issue that ranges across all areas
of computer science. It has been shown that anything computable by an amphi-
bian community in a fixed-size pond is computable by a frog in the same-size
pond -- that is to say, frogs are Pond-space complete. We will show that
there is a log-space, polywog-time reduction from any Turing machine program
to a frog. We will suggest these represent a proper subset of frog-computable
functions.
This is not just a let's-see-how-far-those-frogs-can-jump seminar.
This is only for hardcore amphibian-computation people and their colleagues.
Refreshments will be served. Music will be played.