Grid Computing @ Firepear
Grid
Published: 2020-10-25

We believe in the power of science and a knowledge-based approach to life in general. So we’re proud to be a part of several community grid computing projects which work to solve open problems in medicine, environmental science, and mathematics.

Our contributions

We’re currently active in World Community Grid, Rosetta@Home, Einstein@Home, GPUGrid, and Asteroids@Home. You can check out our detailed stats across all projects.

WCG stats

Updated every 10 years of CPU time. A WU is a “workunit”; a packet of computational tasks sent by the project and returned when complete.

Subproject Team CPU Time Team WUs Subproject status
OpenPandemics 16 years+ 59_000+ Active
Africa Rainfall Project 5 years+ 3_000+ Active
Microbiome Immunity Project 33 years+ 202_000+ Active
Smash Childhood Cancer 4 years+ 47_000+ Active
Help Stop TB 186 days+ 424 Active
Mapping Cancer Markers 96 years+ 285_000+ Active
Beta Testing 67 days+ 688 Intermittent
OpenZika 7y 290d 04:23:01 65_778 Complete (2016-05 to 2019-12)
FightAIDS@Home 2 10y 352d 12:24:47 30_648 Complete (2015-09 to 2020-10)
Outsmart Ebola Together 3y 101d 15:01:33 9_312 Complete (2014-12 to 2018-12)
180 years+ 705_000+

Other project stats

Most projects don’t provide rich statistics, so we generate some numbers from our local logs.

Project Team Compute Time Team WUs Team status
Einstein (astrophysics) 177 days+ 10_000+ Active
GPUGrid (bioinformatics) 349 days+ 4_000+ Active
Asteroids (astronomy) 324d 16:40:28 4_418 Inactive
Rosetta (bioinformatics) 13y 118d 07:54:14 15_428 Inactive

Some compute times are so low compared to other projects because those are GPU projects. BOINC counts 1 second of wall-clock time as 1 second of compute time, regardless of how many hundreds (or thousands!) of GPU cores are working on a task. Meanwhile, for CPU projects, 1 second of time is counted for each core/thread, and cumulative runtime goes up much faster as a result.

Our projects

  • We develop the Homefarm cluster orchestration suite
  • We created the Greenhouse multi-machine chassis project
  • For testing Homefarm on ARM CPUs, we created Greenroom

Our fleet

These machines run 24/7/365, crunching workunits.

Machine CPU Cores RAM GPU
node01 Ryzen 3900X 12c24t 32GB GTX 750 Ti
node02 Ryzen 3900X 12c24t 32GB GTX 1650
node03 Ryzen 3900X 12c24t 16GB
node04 Ryzen 3900X 12c24t 16GB
node05 Ryzen 2700 08c16t 16GB
node06 Ryzen 2700 08c16t 32GB GTX 1650
nodepi01 RPi 4 B 04c04t 4GB
nodepi02 RPi 4 B 04c04t 4GB
nodepi03 RPi 4 B 04c04t 4GB
nodepi04 RPi 4 B 04c04t 4GB
80c144t 160GB

And we try to maximize the goodness by turning upgraded parts into new machines that we give to people who can use them to do even more good work.

Team news

For more in-depth information on computing milestones and plans, check out our most recent updates.