Hi,
Somebody know if there are complex float matrix inversion and QR-decomposion algorithms optimized for the C6678?
This thread has been locked.
If you have a related question, please click the "Ask a related question" button in the top right corner. The newly created question will be automatically linked to this question.
Hi,
Somebody know if there are complex float matrix inversion and QR-decomposion algorithms optimized for the C6678?
QR decomposition solution work is under progress and will be available sometime in July/Aug. Whether you are also interested in Cholesky?
Regards,
Asheesh
Hi,
I'm also interested in LU decomposition.
The source matrix is not hermitian (but it is symmetric), so I think Cholesky is not applicable.
LU decomposition will also be available in the July/Aug time time frame.
Regards
Asheesh
Hi Asheesh,
We need urgent the matrix inversion routine LU based.
Could you tell me which will be the planed processing time for 128 and 512 matrix inversion?
Thanks
Carlos
Asheesh Bhardwaj said:LU decomposition will also be available in the July/Aug time time frame.
Any news about the availability of the LU and QR decomposition?
Any anticipation about the performance?
Regards
Hi,
sorry to be insistent, any news about the C6678 optimized LU decomposition release date?
I reach my deadline to decide whether or not wait for it or try my own implementation.
Alberto,
Please download the latest DSPLIB 3_4_0_0 release from the http://www.ti.com/tool/sprc265 link.
Regards,
Asheesh
Hi,
Please, could somebody tell me if the functions for LU decomposition and matrix inverse in DSPLIB 3_4_0_0 are capable of perform the calculation for multicores working in parallel?
If yes, could you somebody tell me which is the time efficiency improvement?. Do you have a benchmark according with the quantity of cores used in parallel? (obviously depending of how many cores are performing the processing)?
Thanks.
Carlos
Carlos,
The functions for LU decomposition and matrix inverse are based on single core like any other DSPlib kernel.
Refer the MCSDK - HPC release for running the examples in multicore environment which you can use for developing the code for your application.
http://software-dl.ti.com/sdoemb/sdoemb_public_sw/mcsdk_hpc/03_00_01_04/index_FDS.html
Regards
Asheesh