Floating point limitation for C6747

Doug Kim

Hi ,

I am using C6747 on OMAP137. I had ran my algorithm realtime mode using double precision successfully, but I just change double to float variable, it has problem. I have some question about this issue

1. What is floating point limitation for C6747?

2. Could you reply maximum number and minim number for float variable?

for example, -1.0 x 10^-?? ~ 1.0 x 10^??

Thank you

Doug

over 12 years ago

0 Sivaraj Kuppuraj over 12 years ago

TI__Mastermind 35645 points

Hi Doug,

Thanks for your post.

There are standard single and double precision formats mentioned in the C674x DSP CPU and Instruction Set, in which please refer Section 3.3.1 & 3.3.2 as below:

http://www.ti.com/lit/ug/sprufe8b/sprufe8b.pdf

Thanks & regards,

Sivaraj K

-------------------------------------------------------------------------------------------------------

Please click the Verify Answer button on this post if it answers your question.
--------------------------------------------------------------------------------------------------------

0 Sivaraj Kuppuraj over 12 years ago

TI__Mastermind 35645 points

Hi Doug,

In the most common format (ANSI/IEEE Std. 754-1985), the largest and smallest numbers are ±3.4×10³⁸ and 1.2x10^-38, respectively.

Thanks & regards,

Sivaraj K

-------------------------------------------------------------------------------------------------------

Please click the Verify Answer button on this post if it answers your question.
--------------------------------------------------------------------------------------------------------

0 Doug Kim over 12 years ago in reply to Sivaraj Kuppuraj

Intellectual 300 points

I saw in the document,

Normalized: -1s ×2(e - 127) ×1.f 0<e<255

Denormalized (Subnormal): -1s × 2-126 × 0.f e = 0; f is nonzero

but, I cannot understand these number 3.4x10^38 and 1.2x10-38 ?

Doug

Processors

Processors forum

Floating point limitation for C6747