64 Bit Integers

[Rick C]

I know the standard requires 32 bit signed integer range, but does not restrict vendors from supplying larger ranges. However extending the range of "integer" may break existing code. Is there any consideration to a new data type which would be larger integers? Then the usages would be independent and not subject to breaking existing code.

I think the same thing applies to the REAL type. It's just that many languages allow simple indication of the size of integer and real you wish to work with. Why is this so hard to figure out in HDL?

[KJ]

I know the standard requires 32 bit signed integer range, but does not restrict vendors from supplying larger ranges. However extending the range of "integer" may break existing code. Is there any consideration to a new data type which would be larger integers? Then the usages would be independent and not subject to breaking existing code.

Larger integers were added to VHDL-2019

Kevin Jennings

[KJ]

Post by KJ

I know the standard requires 32 bit signed integer range, but does not restrict vendors from supplying larger ranges. However extending the range of "integer" may break existing code. Is there any consideration to a new data type which would be larger integers? Then the usages would be independent and not subject to breaking existing code.

Larger integers were added to VHDL-2019

Specifically, VHDL-2019 defines...
The range of INTEGER is implementation dependent and shall include the range –(2^63) to (2^63)–1 inclusive.

Kevin Jennings

[Rick C]

Post by KJ

Post by KJ

I know the standard requires 32 bit signed integer range, but does not restrict vendors from supplying larger ranges. However extending the range of "integer" may break existing code. Is there any consideration to a new data type which would be larger integers? Then the usages would be independent and not subject to breaking existing code.

Larger integers were added to VHDL-2019

Specifically, VHDL-2019 defines...
The range of INTEGER is implementation dependent and shall include the range –(2^63) to (2^63)–1 inclusive.
Kevin Jennings

Ok, that's great. Now we just need to wait for the various vendors to embrace the new standard.

Thanks, I'm not sure how I was not aware this standard had been approved. I'm sure it was mentioned here. I see in searches it covers some new aspects of interfaces. I should read up on that to see if it is useful for me.