Pico is an algorithm that quickly computes the CMB scalar and tensor power spectra, the matter transfer function and the WMAP 3 year likelihood. It is intended to accelerate parameter estimation codes. More generally, Pico allows using massively parallel computing resources, including distributed computing projects such as Cosmology@Home, to speed up the slow steps in inherently sequential calculations.

Pico is described in detail in W. A. Fendt and B. D. Wandelt, astro-ph/0606709. Improvements to the algorithm as well as results for a nonflat model are presented in astro-ph/0712.0194. Go to the manual for instructions on using Pico with CosmoMC and CAMB.The source code for Pico, including the training code, along with sets of regression coefficients can be found in the download section.

• Red line uses CAMB and WMAP software (1.5 weeks).
• Green line uses Pico and WMAP software (4 hours).
• Blue line uses only Pico (1.5 hours).
• The 2D plots show the 68% and 99% contours.

---

• Compute the CMB power spectrum, matter transfer function and WMAP likelihood in a few milliseconds.
• Sufficient accuracy for parameter estimation from Planck data.
• Training generally requires 1 hour or less.
• Training is linear and does not rely on proprietary codes.
• Incorporates into parameter estimation codes, such as CosmoMC.

• Red line is the mean error over the test set.
• Blue line is the 99% error bars over the test set.
• Power spectrum error is plotted in units of cosmic standard deviation.
• Matter transfer function error is plotted as percent error.

---