I seem to recall Keith mentioning that there will or might be a new chip coming out in the C3x line. Any word on that? Is this in the works, or just in the pre-planning stage... or just a wish list?

Rick Collins

r...@arius.com

Arius - A Signal Processing Solutions Company

Specializing in DSP and FPGA design http://www.arius.com

4 King Ave 301-682-7772 Voice

Frederick, MD 21701-3110 301-682-7666 FAX

Hi

The problem here is that the destination in the ADDF3 part of the instruction is limited to either R2 or R3 (the MPYF3 accepts R0,R1 as destinations).

MPYF3 *AR1++,*AR2++%,R0 ;R0=(AR1)*(AR2) || ADDF3 R0,R2,R2 ;in parallel with accumulation=>R1 Best regards Keith Larson
male_uk00 wrote:

I have set up circular buffers for both the left and right channels for a simple band pass FIR filter i get an error when i uses the parralell multiply and accumalate instruction in my code undefined or invalid op code code is below any idears whats wrong?? .include "c3xmmrs.asm" ;include memory map .start ".data",0x809800 ;start address of data .data HN .float -0.000705659 .float -0.003488302 .float -0.008788407 .float -0.014505267 .float -0.015254676 .float -0.006203771 .float 0.011028171 S0_run .word 0xEBC0000 ;Set global control register for 32bit LRCIN .set 0x80A000 ;LRCIN holds the address of the CPLD HN_ADDR .word HN ;starting address of HN array XN_ADDR_L .word XN_L ;starting address of XN_L buffer XN_ADDR_R .word XN_R ;starting address of XN_R buffer XB_ADDR_L .word XN_L+LENGTH-1 ;last (bottom) address of XN_L XB_ADDR_R .word XN_R+LENGTH-1 ;last (bottom) address of XN_R LENGTH .set 32 ;size of circular buffer .brstart "XN_BUFFER_L",32 ;align a buffer of size 32 XN_L .sect "XN_BUFFER_L" ;buffer section of XN .loop LENGTH ;loop length times .float 0 ;init all XN values to zero .endloop ;end of loop .brstart "XN_BUFFER_R",32 ;align a buffer of size 32 XN_R .sect "XN_BUFFER_R" ;buffer section of XN .loop LENGTH ;loop length times .float 0 ;init all XN values to zero .endloop .entry BEGIN ; start of code BEGIN ldp @stack ; Set stack space ldi @stack,SP ; ldp T0_ctrl ; Use kernel data page and stack ldp HN_ADDR ;init to data page 128 ldi @XB_ADDR_L,AR2 ;AR1=bottom address XN_L buffer ldi @HN_ADDR,AR1 ;AR2=starting address of HN array ldi @XB_ADDR_R,AR4 ;AR4=bottom address XN_R buffer ldi LENGTH,BK ;BK2, size of circular buffer ldi @S0_rdata,R0 ;load value at address of SP0 receive data into R1 direct lsh 12,R0 ;left shift for sign extention FLOAT R0 ;convert interger to floating point STF R0,*AR2++% ;store value at bottom of XN_L buffer LDF 0,R0 ;init R0=0 LDF 0,R1 ;init R1=0 RPTS LENGTH-1 ;execute next 2 instr 0-31 times MPYF3 *AR1++,*AR2++%,R0 ;R0=(AR1)*(AR2) || ADDF3 R0,R1,R1 ;in parallel with accumulation=>R1 ADDF R0,R1 ;last accumulation => R1 FIX R1,R0 ash -12,R0 ;right shift keeping the sign bit sti R0,@S0_xdata

I have set up circular buffers for both the left and right channels for a simple band pass FIR filter i get an error when i uses the parralell multiply and accumalate instruction in my code undefined or invalid op code code is below any idears whats wrong?? .include "c3xmmrs.asm" ;include memory map .start ".data",0x809800 ;start address of data .data HN .float -0.000705659 .float -0.003488302 .float -0.008788407 .float -0.014505267 .float -0.015254676 .float -0.006203771 .float 0.011028171 S0_run .word 0xEBC0000 ;Set global control register for 32bit LRCIN .set 0x80A000 ;LRCIN holds the address of the CPLD HN_ADDR .word HN ;starting address of HN array XN_ADDR_L .word XN_L ;starting address of XN_L buffer XN_ADDR_R .word XN_R ;starting address of XN_R buffer XB_ADDR_L .word XN_L+LENGTH-1 ;last (bottom) address of XN_L XB_ADDR_R .word XN_R+LENGTH-1 ;last (bottom) address of XN_R LENGTH .set 32 ;size of circular buffer .brstart "XN_BUFFER_L",32 ;align a buffer of size 32 XN_L .sect "XN_BUFFER_L" ;buffer section of XN .loop LENGTH ;loop length times .float 0 ;init all XN values to zero .endloop ;end of loop .brstart "XN_BUFFER_R",32 ;align a buffer of size 32 XN_R .sect "XN_BUFFER_R" ;buffer section of XN .loop LENGTH ;loop length times .float 0 ;init all XN values to zero .endloop .entry BEGIN ; start of code BEGIN ldp @stack ; Set stack space ldi @stack,SP ; ldp T0_ctrl ; Use kernel data page and stack ldp HN_ADDR ;init to data page 128 ldi @XB_ADDR_L,AR2 ;AR1=bottom address XN_L buffer ldi @HN_ADDR,AR1 ;AR2=starting address of HN array ldi @XB_ADDR_R,AR4 ;AR4=bottom address XN_R buffer ldi LENGTH,BK ;BK2, size of circular buffer ldi @S0_rdata,R0 ;load value at address of SP0 receive data into R1 direct lsh 12,R0 ;left shift for sign extention FLOAT R0 ;convert interger to floating point STF R0,*AR2++% ;store value at bottom of XN_L buffer LDF 0,R0 ;init R0=0 LDF 0,R1 ;init R1=0 RPTS LENGTH-1 ;execute next 2 instr 0-31 times MPYF3 *AR1++,*AR2++%,R0 ;R0=(AR1)*(AR2) || ADDF3 R0,R1,R1 ;in parallel with accumulation=>R1 ADDF R0,R1 ;last accumulation => R1 FIX R1,R0 ash -12,R0 ;right shift keeping the sign bit sti R0,@S0_xdata