First results of the ALICE-HLT fast clusterfinder preprocessor

The track reconstruction of the ALICE TPC in the High-Level-Trigger (HLT) computing farm is a two stage process. First the clusters of the charge distributions in the TPC are calculated which are then used as the seeds for the tracking algorithm. In the HLT the first stage, the cluster finding, can be done in software or in hardware. The hardware solution is implemented in an FPGA which is located directly in the data path from the TPC to the HLT. It processes the raw data in real-time and stores it in the main memory of the HLT computing farm for further processing by the tracker. For each of the more than 550.000 channels of the TPC a gain correction factor is applied. Furthermore the center of gravity of the charge clusters and the deviation are calculated. Depending on the multiplicity of the charged particles in the TPC a speed-up of factor 10-50 with respect to the software algorithm can be obtained thus allowing the HLT to analyze events online at maximum TPC rate. First results with real data obtained in the first collisions at LHC are shown.