You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
When testing the ZfaF/ZfaD extensions in arch-verif, there seemed to be a mismatch with ImperasDV at PCE=0x8001fc90, running this elf file, on both the rv32gc as well as the rv64gc:
Upon looking at the waveforms simulating on the rv32gc spec, it appears as though the root cause of the mismatch occurs earlier, when testing the froundnx instruction. At PCE=8001ccf8, the first instance of the froundnx instruction being executed, the fflags get set to 0xX1, meaning that the NX flag is being raised as is expected, but the NV flag is being set to a floating value.
The text was updated successfully, but these errors were encountered:
When testing the ZfaF/ZfaD extensions in arch-verif, there seemed to be a mismatch with ImperasDV at PCE=0x8001fc90, running this elf file, on both the rv32gc as well as the rv64gc:
ZfaFDBug.tar.gz
Upon looking at the waveforms simulating on the rv32gc spec, it appears as though the root cause of the mismatch occurs earlier, when testing the
froundnx
instruction. At PCE=8001ccf8, the first instance of thefroundnx
instruction being executed, the fflags get set to 0xX1, meaning that the NX flag is being raised as is expected, but the NV flag is being set to a floating value.The text was updated successfully, but these errors were encountered: