Re: filling big array of double

From: "Aleksandr Sharahov" <alsha-on-mail333.com>
Date: Tue, 31 May 2005 18:05:53 +0400

One more ;-)

procedure FillDouble(var Dest; Count: integer; Value: double); register;
asm
neg edx
jge @ret
shl edx, 3
sub eax, edx
movq mm0, [ebp+8]
test edx, 8
jz @loop
movq [eax+edx], mm0
add edx, 8
jge @ret
@loop:
movq [eax+edx], mm0
movq [eax+edx+8], mm0
add edx, 16
jl @loop
@ret:
emms
end;

.

Follow-Ups:
- Re: filling big array of double
  - From: Henri Gourvest

References:
- filling big array of double
  - From: Henri Gourvest

Prev by Date: Re: filling big array of double
Next by Date: Re: filling big array of double
Previous by thread: Re: filling big array of double
Next by thread: Re: filling big array of double
Index(es):
- Date
- Thread

Relevant Pages

Re: filling big array of double
... asm neg edx jge @ret shl edx, 3 sub eax, edx movq mm0, test edx, 8 jz @loop movq, mm0 add edx, 8 jge @ret @loop: ...
(borland.public.delphi.language.basm)
Re: OOPS: 2.6.16-rc6 cpufreq_conservative
... testl %eax, %eax ... movq _cpu_pda, %rdx ... movq $per_cpu__runqueues, %rax ... cmpl $9, %edx ...
(Linux-Kernel)
Re: Pros and cons of assumed-shape arrays
... just the instructions inside the add loop. ... addl -116, %edx ... instructions off %edx the compiler can take advantage of that. ...
(comp.lang.fortran)
Re: Larkin, Power BASIC cannot be THAT good:
... I got SSE2 code for the inner loop when compiling the C ... xorl %edx, %edx ... xorl %eax, %eax ... int s; ...
(sci.electronics.design)
RE: 2.6.29 acpi regression: acpi_ex_extract_from_field -- div by zero
... Subject: 2.6.29 acpi regression: acpi_ex_extract_from_field -- div by zero ... Division before the loop is apparently OK. ... xorl %edx, %edx # ... More majordomo info at http://vger.kernel.org/majordomo-info.html ...
(Linux-Kernel)