Closed hyunjongL closed 1 year ago
Yes, try the "--max-proc" command line option. For example, "--max-proc 32". This is not officially supported, so it's possible you'll run into problems. Let me know if that happens.
If I synthesize 2 models for 32 processors each, it will each have configurations for 32 processors. Can the two be allocated for processors 0-31 and 32-63? They would have different row/column count registers and will have settings for themselves. Problems I can imagine is when the aggregator tries to fetch data from all 64 processors, or when the processors would want run in sync.
Correct, unfortunately this is not supported in the general sense.
Thanks!! I think I got a somewhat understanding. I just need to process on how the aggregator works.
I know the synthesis in the current form does not support this, however, is there a way to run two pass for less processors by setting the registers? (I am comfortable with the registers on the accelerator, so you can explain using them)