A laser microparticle detector used in conjunction with continuous sample melting provided a more than 1500 m long record of particle concentration and size distribution of dust in the NGRIP2 ice core at high resolution, covering continuously the depth range 1404.7–2930.4 m (Ruth, 2002). The data set of dust concentrations has been made publicly available in low resolution at earlier occasions, but we here provide the full record of dust particle counts in millimetre resolution. The data are supplied on a depth scale, and were used as such in the construction of the GICC05 time scale (Rasmussen et al., 2006). The data cover approximately 10,280–107,600 years before 2000 CE (denoted b2k) on the GICC05modelext age model. The melting system is common for the dust data presented here and several other records of impurities documented thoroughly by Erhardt et al. (2022). The measurement technique is described in detail in Ruth (2002). Data were previously released in 1.65 m depth resolution in connection with the work of Ruth et al. (2003), archived at https://doi.org/10.1594/PANGAEA.107444, and in 0.55 m resolution with the work of Ruth et al. (2007) archived at https://doi.org/10.1594/PANGAEA.587836. Please see these sources for a full description of the measurement techniques. The data file released here is in millimeter resolution, but the shortest features that can be safely detected in the data have a peak-to-peak separation of 1-2 centimeters (Erhardt et al., 2022). There could be uncertainties of a similar magnitude in the depth registration, arising e.g. from difficulties in precise co-registration between the different measurements subsystems. This means that the observed relative phasing between dust and soluble impurity records on annual and sub-annual scales (even from the same core and melter setup) should be interpreted with extreme caution. Furthermore, around core breaks there is an increased risk of contamination and imperfect depth assignments.