On Tue, Dec 22, 2009 at 3:46 PM, Laurent Desnogues
wrote:
> On Tue, Dec 22, 2009 at 1:02 AM, Richard Henderson wrote:
>> On 12/21/2009 03:08 PM, Laurent Desnogues wrote:
>>>
>>> If you wanted to use movcond, you'd have to make
>>> cond + move a special case...
>>
>> You'd certainly want the ARM f
On Tue, Dec 22, 2009 at 1:02 AM, Richard Henderson wrote:
> On 12/21/2009 03:08 PM, Laurent Desnogues wrote:
>>
>> If you wanted to use movcond, you'd have to make
>> cond + move a special case...
>
> You'd certainly want the ARM front-end to use movcond more often than that.
> For instance:
>
>
On Mon, Dec 21, 2009 at 12:28:03PM -0800, Richard Henderson wrote:
> On 12/21/2009 01:13 AM, Laurent Desnogues wrote:
>> The question for the generalized movcond is how useful is it?
>> Which front-ends would need it and would the cost to generate
>> code for it on some (most?) back-ends be amortiz
On 12/21/2009 03:08 PM, Laurent Desnogues wrote:
If you wanted to use movcond, you'd have to make
cond + move a special case...
You'd certainly want the ARM front-end to use movcond more often than
that. For instance:
addeq r1,r2,r3
-->
add_i32 tmp,r2,r3
movcond_i32 r1,ZF,0,tmp,r1,eq
On 12/21/2009 02:21 PM, Laurent Desnogues wrote:
As far as I know these CPU's don't need the full movcond but
only the variant with vtrue.
I know that. And I looked into TCG very closely to figure out how to
implement just that. Except then I have to modify TCG to special-case
movcond to kn
On Mon, Dec 21, 2009 at 11:50 PM, Richard Henderson wrote:
[...]
>> Even if movcond was quick to generate
>> host code, for instance for ARM, you'd have to explicitly detect
>> conditional moves
>
> One of us is confused. Why would I have to explicitly detect conditional
> moves?
Most ARM instru
On Mon, Dec 21, 2009 at 9:28 PM, Richard Henderson wrote:
> On 12/21/2009 01:13 AM, Laurent Desnogues wrote:
>>
>> The question for the generalized movcond is how useful is it?
>> Which front-ends would need it and would the cost to generate
>> code for it on some (most?) back-ends be amortized?
>
On 12/21/2009 01:13 AM, Laurent Desnogues wrote:
The question for the generalized movcond is how useful is it?
Which front-ends would need it and would the cost to generate
code for it on some (most?) back-ends be amortized?
... Any front end that has a conditional move instruction?
Sparcv9, Mi
On Sun, Dec 20, 2009 at 06:00:48PM -0800, Richard Henderson wrote:
> On 12/20/2009 02:57 PM, Paul Brook wrote:
>> On Saturday 19 December 2009, Richard Henderson wrote:
>>> Changes from round 3:
>>>
>>> * Drop movcond for now.
>>> * Only use movzbl and not xor in setcond.
>>
>> I'm still catchi
On Mon, Dec 21, 2009 at 3:00 AM, Richard Henderson wrote:
[...]
> I *am* convinced that to remove either VTRUE or VFALSE as arguments to the
> movcond primitive (implementing dest = (cond ? vtrue : dest) would do too
> much violence to both the liveness analysis and the register allocator
> within
On 12/20/2009 02:57 PM, Paul Brook wrote:
On Saturday 19 December 2009, Richard Henderson wrote:
Changes from round 3:
* Drop movcond for now.
* Only use movzbl and not xor in setcond.
I'm still catching up on mail backlog from this thread, but I'm concerned that
we're exposing setcond to
On Saturday 19 December 2009, Richard Henderson wrote:
> Changes from round 3:
>
> * Drop movcond for now.
> * Only use movzbl and not xor in setcond.
I'm still catching up on mail backlog from this thread, but I'm concerned that
we're exposing setcond to the target translation code if we're p
On Sat, Dec 19, 2009 at 11:43 PM, malc wrote:
> On Sun, 20 Dec 2009, Aurelien Jarno wrote:
>
>> On Sat, Dec 19, 2009 at 10:52:52AM -0800, Richard Henderson wrote:
>> > Changes from round 3:
>> >
>> > * Drop movcond for now.
>> > * Only use movzbl and not xor in setcond.
>> >
>> >
>>
>> Thanks fo
On Sun, 20 Dec 2009, Aurelien Jarno wrote:
> On Sat, Dec 19, 2009 at 10:52:52AM -0800, Richard Henderson wrote:
> > Changes from round 3:
> >
> > * Drop movcond for now.
> > * Only use movzbl and not xor in setcond.
> >
> >
>
> Thanks for the update, it looks like ready, I only have cosmetic
On Sat, Dec 19, 2009 at 10:52:52AM -0800, Richard Henderson wrote:
> Changes from round 3:
>
> * Drop movcond for now.
> * Only use movzbl and not xor in setcond.
>
>
Thanks for the update, it looks like ready, I only have cosmetic or
minor comments. See my comments in the individual patches
Changes from round 3:
* Drop movcond for now.
* Only use movzbl and not xor in setcond.
r~
Richard Henderson (5):
tcg: Generic support for conditional set
tcg-x86_64: Implement setcond.
tcg-i386: Implement small forward branches.
tcg: Add tcg_invert_cond.
tcg-i386: Implement setcon
16 matches
Mail list logo
