floating point

[mmt1]

Hi all.
I want to have an integer to floating point converter. On the other
hand I want a floating point generator to convert some numbers from
integer to floating point . so I can use them as elements of the
"primitive table" in verilog code. Have any one help me ? how I can
get such tools? I also want to implement the 16 bit floating point
adder and multiplier in ASIC. So I must know the operation of them.
Please help me. Thanks.
Regards, mehr.

 

[Uncle Noah]

mmt1 wrote:

Hi all.
I want to have an integer to floating point converter. On the other
hand I want a floating point generator to convert some numbers from
integer to floating point . so I can use them as elements of the
"primitive table" in verilog code. Have any one help me ? how I can
get such tools? I also want to implement the 16 bit floating point
adder and multiplier in ASIC. So I must know the operation of them.
Please help me. Thanks.
Regards, mehr.

During FPU development at Opencores, I recall Rudi Usselmann coding
Verilog cores that did exactly this conversion. These files are still
accessible in the opencores mailing lists, dated circa June-August
2000.
http://www.opencores.org/forums.cgi/cores/

Also, check out the (complete) Open FPU core
http://www.opencores.org/projects.cgi/web/fpu/overview
You will also need SoftFloat to run their test pattern generator (C
program). (easy to find in Google).

I have a question: 16-bit FP arithmetic is not defined in IEEE-754, so
what is your approach? Have you ever heard of the OpenEXR tool? I think
they use sign=1-bit, exponent=5-bits, mantissa=10-bits.

cheers

Nikolaos Kavvadias