Multiple precision modes (FP) ????

parhelia

Newcomer
Why are the multiple precision modes discussed in all recent reviews(12-bit integet, 16-bit floating point, and 32-bit floating point) ?
Why wasn't this point discussed about DX8.1 level cards? Don't they have a FP?
Sorry if I look like a noob, I'm just confused.
 
No, DX8 level cards don't have FP pixel shader precision. The only cards with different PS precisions are the GeForceFX series.
 
parhelia said:
Why are the multiple precision modes discussed in all recent reviews(12-bit integet, 16-bit floating point, and 32-bit floating point) ?
Why wasn't this point discussed about DX8.1 level cards? Don't they have a FP?
Sorry if I look like a noob, I'm just confused.
Cards have had multiple precisions for some time now.

DX9 introduced the 16-bit and 32-bit floating point formats. Current DX9 hardware supports these storage formats. ATI's does all calculations at 24-bit FP. nVidia does its calculations at the precision the developer chooses (12-bit int, 16-bit FP, or 32-bit FP).
 
"We don't have at tower sir. Only a bridge."

"No tower? Dammit! Why does nobody tell me these things!"
 
Dio said:
"We don't have at tower sir. Only a bridge."

"No tower? Dammit! Why does nobody tell me these things!"

"Ok, here's what we're going to do. I want you to dig a trench all the way round the building. Fill it with gasoline."

"I was Over Unger, and Unger was Under Dunn."

"Over Macho Grande?"
"No, I don't think I'll ever be over Macho Grande."

"Looks like I picked the wrong week to give up sniffing glue..."

Off topic? Me? Naaaah...
 
Back
Top