In this report, we summarize the results obtained from comparing the performance of six map-making packages for PACS photometer data. The codes which participated in this exercise are, in alphabetical order: JScanam, MADmap, SANEPIC, Scanamorphos, Tamasis and Unimap. All these codes are publicly available and are designed to correct for noise in the data, while preserving emission on all angular scales. To test the performance of these codes, we considered a combination of simulated and real data, and a set of five different metrics, namely: 1. power spectrum estimation; 2. difference matrix; 3. point source photometry (for both bright and faint sources); 4. comparison with ancillary data (IRAS and Spitzer/MIPS); 5. analysis of the noise. The analysis evidences that all the codes preserve emission on large angular scales and that they are able to handle the PACS data with a high level of photometric accuracy. Finally, we point out that differences in the generated maps are typically small yet noticeable.