https://community.intel.com/t5/Intel-oneAPI-Math-Kernel-Library/BKM-to-init-matrix-for-FFT-transform/m-p/914768#M12468 I have large arrays that I need to do FFT transform many times. The Vtune shows the memset as the hottest spot. I used memset to zero init all array values before I assign. Not all matrix elements got assigned values. It is more like a sparse matrix. If I don't use memset and use for loop instead, the program runs even slower. How do I init the array more efficiently? Thanks. Tue, 05 Jan 2010 20:07:16 GMT Zhu_W_Intel 2010-01-05T20:07:16Z https://community.intel.com/t5/Intel-oneAPI-Math-Kernel-Library/BKM-to-init-matrix-for-FFT-transform/m-p/914768#M12468 I have large arrays that I need to do FFT transform many times. The Vtune shows the memset as the hottest spot. I used memset to zero init all array values before I assign. Not all matrix elements got assigned values. It is more like a sparse matrix. If I don't use memset and use for loop instead, the program runs even slower. How do I init the array more efficiently? Thanks. Tue, 05 Jan 2010 20:07:16 GMT https://community.intel.com/t5/Intel-oneAPI-Math-Kernel-Library/BKM-to-init-matrix-for-FFT-transform/m-p/914768#M12468 Zhu_W_Intel 2010-01-05T20:07:16Z