Purpose: We evaluated the performance of a 3D diode device ArcCHECK (Sun Nuclear) for IMRT patient plan verification and compared the results with our current IMRT QA 2D diode array device MapCHECK (Sun Nuclear). Methods: The ArcCHECK QA device consists of 1386 diodes of 0.8×0.8 mm2 size geometrically arranged in 21 adjacent rings spaced 1 cm apart. An acrylic insert with an ionization chamber holder was plugged into ArcCHECK to form a 26.6 cm diameter cylindrical phantom. A compact ion chamber was placed inside the holder to measure dose. Seven IMRT patient plans generated using Pinnacle 9.0 were evaluated for this study, including 3 Head & Neck, 3 prostate and one whole pelvis. Actual beam angles were used for ArcCHECK QA plans, but set to 0 degrees for MapCHECK. Gamma criteria of 3%/3mm were used for analysis. Results: Results using our gamma criteria fall between 90.4% to 96.8% for ArcCHECK and 96% to 100% for MapCHECK. For 3 Head & Neck plans ArcCHECK achieved 96% or above pass rate using our gamma criteria, very similar to our MapCHECK result. 90.4% of ArcCHECK measurements passed the gamma analysis for the whole pelvis while 100% of MapCHECK measurements passed. The gamma analysis for three prostate plans was 91.5%, 95% and 96.7% using Arc Check measurements and 98.2%, 96% and 99.2% using MapCHECK. Doses between ion chamber measurement and calculation were in good agreement (0.89%∼2.44%) for all 7 plans. Conclusions: ArcCHECK achieved above 90% pass rate using gamma criteria of 3%/3mm. ArcCHECK measurements produced very similar results to MapCHECK for most plans. Using ArcCHECK in the clinic for IMRT QA can mimic patient dose delivery and examine gantry angle and couch transmission effects.