[9fans] 6c CMP reg, $0 optimization

[9fans] 6c CMP reg, $0 optimization

[cinap_lenrek]

i changed the compiler to remove unneeded compare with
zero instructions when the previous instruction already
sets the zero flag in the peephole optimizaiton pass.

http://felloff.net/usr/cinap_lenrek/6cpeep.diff

this removes compares for zero/non zero tests only.
it only looks at the previous non-nop instruction
to see if it sets our compare value register.

it seems to work fine so far. our pc64 kernel got
arround 8K smaller, ghostscript got 5K smaller.

i'm not a compiler expert tho, so i'd like to hear if
there are any issues with this approach or if there are
better ways to do this.

--
cinap

Re: [9fans] 6c CMP reg, $0 optimization

[erik quanstrom]

> it seems to work fine so far. our pc64 kernel got
> arround 8K smaller, ghostscript got 5K smaller.

it would be nice if ghostscript worked much at all on
64-bit machines.  :-)

- erik

Re: [9fans] 6c CMP reg, $0 optimization

[cinap_lenrek]

what? i'm having no issues with ghostscript on amd64.
do you have issues with some specific documents? the
ones in /sys/doc work fine for me.

--
cinap

Re: [9fans] 6c CMP reg, $0 optimization

[erik quanstrom]

On Sat Mar 29 00:08:06 EDT 2014, cinap_lenrek@felloff.net wrote:
> what? i'm having no issues with ghostscript on amd64.
> do you have issues with some specific documents? the
> ones in /sys/doc work fine for me.

any intel datasheet, for example this one works on arm, mips

- erik
---

; page 82574.pdf	# google 82574 datasheet

reading through pdf...
Ghostscript Error:
Error: /undefined in --setcolor--
Operand stack:
   --dict:7/7(L)--   (Order)   0   0   0   --nostringval--   --nostringval--   cs_prepare_color   --dict:13/20(G)--   --nostringval--
Execution stack:
   %interp_exit   .runexec2   --nostringval--   --nostringval--   --nostringval--   2   %stopped_push   --nostringval--   --nostringval--   --nostringval--   false   1   %stopped_push   1   3   %oparray_pop   1   3   %oparray_pop   1   3   %oparray_pop   1   3   %oparray_pop   .runexec2   --nostringval--   --nostringval--   --nostringval--   2   %stopped_push   --nostringval--   --nostringval--   --nostringval--   --nostringval--   --nostringval--   %array_continue   --nostringval--   false   1   %stopped_push   --nostringval--   %loop_continue   --nostringval--   --nostringval--   --nostringvafatal: readimage error
l--   %array_continue   --nostringval--   --nostringval--   6   8   %oparray_pop
Dictionary stack:
   --dict:1127/1686(ro)(G)--   --dict:2/20(G)--   --dict:78/200(L)--   --dict:105/127(ro)(G)--   --dict:258/347(ro)(G)--   --dict:21/24(L)--   --dict:4/6(L)--   --dict:25/32(L)--
Current allocation mode is local

Re: [9fans] 6c CMP reg, $0 optimization

[Charles Forsyth]

[-- Attachment #1: Type: text/plain, Size: 184 bytes --]

On 29 March 2014 03:56, <cinap_lenrek@felloff.net> wrote:

>
> http://felloff.net/usr/cinap_lenrek/6cpeep.diff


at a glance that looks fine to me; i'll merge it in to my copy.

[-- Attachment #2: Type: text/html, Size: 563 bytes --]