5 replies to this topic

[Robin]

HI:
I have built a project using Templates: One-process testbench(floating). in CoDeveloper, it is all ok!
But, when i compile it in Quartus II, it encounted an error, here is the error message:

Error (10482): VHDL error at impulse_float.vhd(1201): object "altfp_convert" is used but not declared

How to solve it??


Regards,
Robin

[RalphBodenner]

Hi Robin,

Have you included the impulse_float.vhd file in your Quartus project? CoDeveloper would have exported it to an "impulse_lib" subdirectory of the Quartus project.

Which version of Quartus are you using?

Regards,
Ralph

[Robin]

Hi Robin,

Have you included the impulse_float.vhd file in your Quartus project? CoDeveloper would have exported it to an "impulse_lib" subdirectory of the Quartus project.

Which version of Quartus are you using?

Regards,
Ralph

Thanks.
I use Quartus II 7.2.
In the folder called "impulse_lib", there are not only impulse_float.vhd, but also many VHDLs about float, like:fp_div_lut1.vhd or fp_div.vhd, etc....
I have included all of them in my Quartus project.
I use notepad watching the code of these VHDLs, but all of them is a messy code, not vhdl code, except impulse_float.vhd. is that right??
but yesterday, I delete the code about altfp_convert in impulse_float.vhd, it works! no error!

Another question: dose Altera floating-point library support double precision types?

Regards,
Robin

[RalphBodenner]

I'm glad to hear it's working now. I believe the altfp_convert cores are not supported in Quartus II 7.2--only in 8.0 or later. Therefore, type conversion to/from floating-point types is only supported with Quartus 8.

The "messy code" you saw is due to the fact that those files are encrypted by Altera, for use only with a valid Quartus license.

The Altera floating-point library does support double precision. Please make sure the "Allow double-precision types and operators" option is checked in the Impulse C project options.

Regards,
Ralph

[Robin]

The Altera floating-point library does support double precision. Please make sure the "Allow double-precision types and operators" option is checked in the Impulse C project options.

Regards,
Ralph

But, when I use double, I encounted an error:
Error (10395): VHDL Conditional Signal Assignment error at HelloWorld_mem_comp.vhd(331): conditional waveforms must have same number of elements

I just simply replaced float to double, In float situation, it works well!
Is it my Quartus II version(7.2) that caused this error ??

Regards,
Robin

[RalphBodenner]

Hi Robin,

Could you send the complete Impulse C project and source code files that demonstrate the problem to support@impulsec.com? We can then try to reproduce and diagnose the problem.

Thanks,
Ralph