[FPGA-Bitstream] note limitations
This commit is contained in:
parent
ffac5a66e1
commit
a7e87b9432
|
@ -223,6 +223,11 @@ class FabricBitstream {
|
|||
* 101x1 -> 10101 -> 21
|
||||
* - bit-x number: which encodes the 'x' bits into a number. For example,
|
||||
* 101x1 -> 00010 -> 2
|
||||
*
|
||||
* TODO: There is a limitation here, when the length of address vector is more than 64,
|
||||
* A size_t number overflows (cannot represent any binary number > 64 bit).
|
||||
* Such thing can entirely happen even in a medium sized FPGA.
|
||||
* A solution can be use multiple size_t to fit. But clearly, we should not use vector in vector, which causes large memory overhead!
|
||||
*/
|
||||
vtr::vector<FabricBitId, size_t> bit_address_1bits_;
|
||||
vtr::vector<FabricBitId, size_t> bit_address_xbits_;
|
||||
|
|
Loading…
Reference in New Issue