10 level logic in VHDL (for simulation only)

[SSL]

Hi,

I'm taking a class in numerical numerical methods and we have an assignment on floating-point arithmetic.

I thought it might be fun to implement this in VHDL.

But the assignment require us to use FP numbers with base10 instead for of base2

Is there any way to implement logic in VHDL with 10 logic levels instead the normal 2?

I don't care if it will never synthesize, but for simulation I can't see any reasons for it not being possible? I just can't find any types or methods that will support 10 level logic - is it possible? If, how?

[Bassman59]

Hi,

I'm taking a class in numerical numerical methods and we have an assignment on floating-point arithmetic.

I thought it might be fun to implement this in VHDL.

But the assignment require us to use FP numbers with base10 instead for of base2

Is there any way to implement logic in VHDL with 10 logic levels instead the normal 2?

I don't care if it will never synthesize, but for simulation I can't see any reasons for it not being possible? I just can't find any types or methods that will support 10 level logic - is it possible? If, how?

VHDL has long supported the real type. It's generally not synthesizable but for simulation purposes it's used all the time.

[SiliconWizard]

Well, I guess the OP still wants to implement FP arithmetic "from scratch", and not use a FP type? (Otherwise I wouldn't see the point of this assignment? )

The OP's not completely full clear, but from what I understand, they want to implement FP in base 10? I don't really see the problem. Doing this even with very low-level logic. The OP may not have heard of BCD.

In very short: encode each digit with 4 bits (only using the 0-9 values obviously). The rest is up to you.

[SSL]

Well, I guess the OP still wants to implement FP arithmetic "from scratch", and not use a FP type? (Otherwise I wouldn't see the point of this assignment? )

Correct. It must be implemented from scratch so a FP type is not an option.

The OP's not completely full clear, but from what I understand, they want to implement FP in base 10? I don't really see the problem. Doing this even with very low-level logic. The OP may not have heard of BCD.

In very short: encode each digit with 4 bits (only using the 0-9 values obviously). The rest is up to you.

I think you understood the assignment correctly.

I have heard of BCD and used it previously. But only to print my internal binary values in a more human friendly form on a 7-segment or some other type of display.

Using BCD to work with base10 numbers in base2 in this context haven't crossed my mind.

Thank you very much!