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
cycle_idx,/* diff between curr clock cycle and start clock cycle of the current asmop */
));
(asmop,false)
}
// if the next asmop is the first in the list and is at a greater than current clock cycle
// or if the current op is not a part of any asmop, return None.
else{
(None,false)
}
}
It should be simplified to point the asmop_idx to the right decorator, regardless of the direction.
The final implementation should be as simple/succinct as possible.
The outcome of this issue should probably simplify the DebugInfo struct to not have a complex relationship between decorators and operations. Right now, we have the operations, that will increase one per clock tick, and the decorators, that will be a tuple (clock, AsmOp). One alternative is to have a wrapper struct that will bind all of the three together with any eventual metadata we might need.
The debugger expects a max number of cycles related to a given instruction, and that is computed manually during the iteration cycles. Instead, it should probably be precomputed when the debug info is generated - the iterator should just consume the information that was compute, not have logic of its own.
The text was updated successfully, but these errors were encountered:
This code is not prepared for backward step as it assumes the clock always increases in order to reposition the decorator index
miden-vm/processor/src/debug.rs
Lines 67 to 122 in 159a04f
It should be simplified to point the
asmop_idx
to the right decorator, regardless of the direction.The final implementation should be as simple/succinct as possible.
The outcome of this issue should probably simplify the
DebugInfo
struct to not have a complex relationship between decorators and operations. Right now, we have the operations, that will increase one per clock tick, and the decorators, that will be a tuple(clock, AsmOp)
. One alternative is to have a wrapper struct that will bind all of the three together with any eventual metadata we might need.The debugger expects a max number of cycles related to a given instruction, and that is computed manually during the iteration cycles. Instead, it should probably be precomputed when the debug info is generated - the iterator should just consume the information that was compute, not have logic of its own.
The text was updated successfully, but these errors were encountered: