Notes. The first column shows the name of the method. The second column indicates when the code is using a prior on the SZ cluster shape. The uperscript ^a indicate that the detection did not use a shape prior but that the computation of the SZ flux did. The third column gives the performance in terms of CPU hours needed to complete the analysis. The fourth and fifth column show whether the analysis was made using all-sky maps or projected patches, their area in square degrees and the number of patches. Methods ILC1 and ILC2 work with full sky maps for producing an SZ map (by ILC on needlet coefficients) and then work with 504 small patches for cluster detection by matched filtering (ILC1) or using SExtractor (ILC2). The sixth and seventh columns provide information about any specific method used for subtracting point sources (PS) and Galactic foregrounds (FG). Note that both the MMF and the ILC methods have a built-in way for subtracting both point sources and diffuse foregrounds, by treating them as additional noise (of astrophysical origin) correlated across the channels. Note, also, that the study is made only on clusters at  |b| > 20 degrees Galactic latitude for all methods. The eighth column summarizes the main characteritics of each algorithm in terms of yield at 90% purity and photometric accuracy.