Part Number: TMS320F28379D
Tool/software: Code Composer Studio
Hi,
Previously I was using CPU1 for my complete current implementation where I was running short of data memory for which I merged the GSXRAM sectors in CPU1 linker file when running the code from FLASH as shown in the snap shot below.
Now for some parallel computations, i am trying to utilize CLA1 with CPU1 for which I am using the CLA linker file "2837xD_FLASH_CLA_lnk_cpu" regarding which I have some queries.
1) Does the program use only LSXRAM sectors while program execution using CPU1 and CLA1? Can the GSxRAM sectors also be used and merged when running out of memory?
2) When i will be using CLA linker file, which parts of the LSXRAM and GSXRAM can be merged to accommodate data memory code sections (which in my case is taking a lot of space)? and how?
Just to describe my case briefly, i am declaring an array of 400 float variables which will be exchanged between CLA and CPU using "pragma" declarations. On the CPU, i am trying to accommodate around 5000 to 7000 float variables. Previously using just CPU1, by merging GSXRAM sectors, i was able to fit the program sections both in program and data memory. Please guide how can this be done using both CPU and the CLA.
