Hi, I'm having problems accessing SDRAM in a reasonable time. I read on the periperal guide that consequent access to SDRAM should happen in bursts of 4 words. I understood they are talking about 32-bit words. Is this correct? Because I tried to look at the signals (Chip enable, write enable, read enable) with the oscilloscope, but I cannot see any burst! I wrote simple assembler code to write 16 words to SDRAM, and I observe 16 memory accesses (expected 16/4=4). Each memory access takes 1 EMIF clock cycle. The time between each access is about 10 EMIF clock cycles. I don't understand what the dsp is doing during tha time. Has anybody else experienced problems like this with memory access? Am I doing anything wrong? TIA Federica Legger |
|
SDRAM memory access with 6711 DSK
Federica,
The CPU does not know that there will be 15 more
write accesses to
If you can instead write the data to SRAM address
space and then program
I guess in your example, SDRAM space is non-cacheable
or else the CPU
Subrangshu
fedelegger wrote: Hi,
|