みらい


( futuristic ・ whisper )

Minimalist Async Evaluation Framework for R

Lightweight parallel code execution and distributed computing.

mirai() returns a ‘mirai’ object immediately. Designed for simplicity, a ‘mirai’ evaluates an R expression asynchronously, on local or network resources, resolving automatically upon completion.

State of the art networking and concurrency via nanonext offers reliable and efficient scheduling over fast inter-process communications or TCP/IP secured by TLS.

mirai パッケージを試してみたところ、かなり速くて驚きました

Installation

Install the fast channel release (v0.13.1) from R-releases:

install.packages("mirai", repos = "https://r-releases.r-universe.dev")

Or the slow channel release (v0.12.1) from CRAN:

Or the latest development build from the author’s R-universe:

install.packages("mirai", repos = "https://shikokuchuo.r-universe.dev")

Quick Start

Use mirai() to evaluate an expression asynchronously in a separate, clean R process.

A ‘mirai’ object is returned immediately.

library(mirai)

m <- mirai(
  {
    res <- rnorm(x) + y ^ 2
    res / rev(res)
  },
  x = 10,
  y = runif(1)
)

m
#> < mirai | $data >

Above, all specified name = value pairs are passed through to the ‘mirai’.

The ‘mirai’ yields an ‘unresolved’ logical NA whilst the async operation is ongoing.

m$data
#> 'unresolved' logi NA

To check whether a mirai has resolved:

unresolved(m)
#> [1] FALSE

Upon completion, the ‘mirai’ resolves automatically to the evaluated result.

m$data
#>  [1]  -1.7516360  -2.2421768   1.8608206  -0.0805829  -0.6224308  -1.6066042
#>  [7] -12.4095806   0.5373973  -0.4459952  -0.5708949

Alternatively, explicitly call and wait for the result using call_mirai().

call_mirai(m)$data
#>  [1]  -1.7516360  -2.2421768   1.8608206  -0.0805829  -0.6224308  -1.6066042
#>  [7] -12.4095806   0.5373973  -0.4459952  -0.5708949

Daemons

Daemons are persistent background processes created to receive ‘mirai’ requests.

They may be deployed for:

Local parallel processing; or

Remote network distributed computing.

Launchers allow daemons to be started both on the local machine and across the network via SSH etc.

Secure TLS connections can be automatically-configured on-the-fly for remote daemon connections.

Refer to the {mirai} vignette for full package functionality. This may be accessed within R by:

vignette("mirai", package = "mirai")

Integrations

The following core integrations are documented, with usage examples in the linked vignettes:

{parallel} - provides an alternative communications backend for R, implementing a low-level feature request by R-Core at R Project Sprint 2023.

{promises} - ‘mirai’ may be used interchangeably with ‘promises’ by using the promise pipe %...>% or the as.promise() method.

{plumber} - serves as an asynchronous / distributed backend, scaling applications via the use of promises.

{shiny} - serves as an asynchronous / distributed backend, plugging directly into the reactive framework without the need for promises.

{torch} - the custom serialization interface allows tensors and complex objects such as models and optimizers to be used seamlessly across parallel processes.

Powering Crew and Targets High Performance Computing

{targets}, a Make-like pipeline tool for statistics and data science, has integrated and adopted {crew} as its default high-performance computing backend.

{crew} is a distributed worker-launcher extending {mirai} to different distributed computing platforms, from traditional clusters to cloud services.

{crew.cluster} enables mirai-based workflows on traditional high-performance computing clusters using LFS, PBS/TORQUE, SGE and SLURM.

{crew.aws.batch} extends {mirai} to cloud computing using AWS Batch.

Thanks

We would like to thank in particular:

Will Landau, for being instrumental in shaping development of the package, from initiating the original request for persistent daemons, through to orchestrating robustness testing for the high performance computing requirements of {crew} and {targets}.

Henrik Bengtsson, for valuable and incisive insights leading to the interface accepting broader usage patterns.

Luke Tierney, R Core, for discussion on R’s implementation of L’Ecuyer-CMRG streams, used to ensure statistical independence in parallel processing.

Daniel Falbel, for discussion around an efficient solution to serialization and transmission of {torch} tensors.

Joe Cheng, for discussion around the Shiny integration, and authoring {promises}, for which this package provides a method.

◈ mirai R package: https://shikokuchuo.net/mirai/

mirai is listed in CRAN Task View:
- High Performance Computing: https://cran.r-project.org/view=HighPerformanceComputing

◈ nanonext R package: https://shikokuchuo.net/nanonext/

NNG website: https://nng.nanomsg.org/

Please note that this project is released with a Contributor Code of Conduct. By participating in this project you agree to abide by its terms.