Help FPU converting real8 -> real4

flat assembler
Message board for the users of flat assembler.

Index > Main > Help FPU converting real8 -> real4

Author

Thread

hopcode

Joined: 04 Mar 2008
Posts: 563
Location: Germany

hopcode 22 Nov 2009, 14:50

Hallo, is there a way to improve this code where
myvalue is a dqword and in eax should be the (truncated) dword as result ?:

Code:

 lea edx,[myvalue]
  fldz
  fld qword[edx]
  xor eax,eax
  ftst
  fstsw ax
  fwait
  sahf              
  ja @f   ;<-- should be in addition a jp .error_handler implemented ?
  fldz
@@:
  fist dword[edx]
  xchg eax,[edx]   ; <--- is this a read after write ?

Thanks,
hopcode

Last edited by hopcode on 23 Nov 2009, 00:51; edited 1 time in total

22 Nov 2009, 14:50

asmfan

Joined: 11 Aug 2006
Posts: 392
Location: Russian

asmfan 22 Nov 2009, 15:12

cvtsd2ss - sse2

22 Nov 2009, 15:12

hopcode

Joined: 04 Mar 2008
Posts: 563
Location: Germany

hopcode 22 Nov 2009, 15:23

asmfan wrote:

cvtsd2ss - sse2

Oh.. yes. I have looked up right now shortly in Intel docs...
CVTSD2SS—Convert Scalar Double-Precision FP Value to Scalar Single-
Precision FP Value
At a glance, it should be a wrapping around the fpu instructions...
Anyway useful and slim.
F2 0F 5A /r CVTSD2SS xmm1,xmm2/m64

Thanks

BTW:
Where can i find a sse good tutorial ?

22 Nov 2009, 15:23

chaoscode

Joined: 21 Nov 2006
Posts: 64

chaoscode 23 Nov 2009, 01:05

you don't need a tutorial for SSE.
you need SSE every time if you have to work on a lot of Floatingpoint Data.
and to implement an algorithm. Then is the only thing you need the docs, which SSE-Instructions exists and how they work.

23 Nov 2009, 01:05

< Last Thread | Next Thread >

Forum Rules:

You cannot post new topics in this forum
You cannot reply to topics in this forum
You cannot edit your posts in this forum
You cannot delete your posts in this forum
You cannot vote in polls in this forum
You cannot attach files in this forum
You can download files in this forum