Have you (with your suggested fix?)
Sun
On Mon, Jul 11, 2011 at 10:32 AM, Gang Yu <yugang...@gmail.com> wrote:
> The unrelated-stuff make the compiler build broken. So, we have to fix it.
>
> Thanks
> Gang
>
>
> On Mon, Jul 11, 2011 at 10:31 AM, Sun Chan <sun.c...@gmail.com> wrote:
>>
>> does this implementation make sense for you folks?
>> Sun
>>
>> On Mon, Jul 11, 2011 at 10:28 AM, Gang Yu <yugang...@gmail.com> wrote:
>> > Sun:
>> >
>> > We have a check on this (Is_Target_SSE41()), it is defined in
>> > osprey/common/com/X8664/config_targ.h:
>> >
>> > 182#define Is_Target_SSE41() (Target_SSE41 == TRUE)
>> >
>> > We believe it is unrelated to other targets, so still should we define
>> > this
>> > Is_Target_SSE41() for other targets?
>> >
>> > Thanks
>> >
>> > Gang
>> >
>> >
>> > On Sat, Jul 9, 2011 at 7:03 AM, Sun Chan <sun.c...@gmail.com> wrote:
>> >>
>> >> i meant the #ifdef be done inside the function Is_Target_SSE4()
>> >> Sun
>> >>
>> >> On Fri, Jul 8, 2011 at 5:28 PM, Gang Yu <yugang...@gmail.com> wrote:
>> >> > Thanks, you are right. The LNO_Iter_threshold is set based on target.
>> >> > So,
>> >> >
>> >> > Index: osprey/common/com/config_lno.cxx
>> >> > ===================================================================
>> >> > --- osprey/common/com/config_lno.cxx (revision 3682)
>> >> > +++ osprey/common/com/config_lno.cxx (working copy)
>> >> > @@ -1232,7 +1232,9 @@
>> >> > the flag is set based on target. Otherwise use user-specified
>> >> > value.
>> >> > */
>> >> > if(LNO_Iter_threshold == 1) {
>> >> > +#ifdef TARG_X8664
>> >> > LNO_Iter_threshold = (Is_Target_SSE41())? 8 : 0;
>> >> > +#endif
>> >> > }
>> >> > }
>> >> >
>> >> > Index: osprey/be/cg/whirl2ops.cxx
>> >> > ==============================
>> >> > =====================================
>> >> > --- osprey/be/cg/whirl2ops.cxx (revision 3681)
>> >> > +++ osprey/cg/whirl2ops.cxx (working copy)
>> >> > @@ -3166,6 +3166,7 @@
>> >> > WN *compare;
>> >> > VARIANT variant;
>> >> >
>> >> > +#ifdef TARG_X8664
>> >> > if (opcode == OPC_V16I1V16I1SELECT) {
>> >> > TN* op1 = Expand_Expr(WN_kid0(select), select, NULL);
>> >> > TN* op2 = Expand_Expr(WN_kid1(select), select, NULL);
>> >> > @@ -3177,6 +3178,7 @@
>> >> > Expand_Select(result, op1, op2, op3, MTYPE_V16I1, FALSE,
>> >> > &New_OPs);
>> >> > //FALSE passed as dummy arg
>> >> > return result;
>> >> > }
>> >> > +#endif
>> >> >
>> >> >
>> >> > Gang
>> >> >
>> >> >
>> >> > On Fri, Jul 8, 2011 at 5:15 PM, Sun Chan <sun.c...@gmail.com> wrote:
>> >> >>
>> >> >> the #ifdef can be done inside Is_target_SSE...
>> >> >> Sun
>> >> >>
>> >> >> On Fri, Jul 8, 2011 at 1:24 AM, Gang Yu <yugang...@gmail.com> wrote:
>> >> >> > Hi,
>> >> >> >
>> >> >> > SL target builds fail today due to the LNO check-in v3681, the
>> >> >> > build
>> >> >> > fail
>> >> >> > comes from the patch:
>> >> >> >
>> >> >> > Index: config_lno.cxx
>> >> >> >
>> >> >> > ===================================================================
>> >> >> > --- config_lno.cxx (revision 3643)
>> >> >> > +++ config_lno.cxx (revision 3681)
>> >> >> > @@ -1224,5 +1227,12 @@
>> >> >> > Mhd_Options.L[i].TLB_Miss_Penalty;
>> >> >> > }
>> >> >> > }
>> >> >> > +
>> >> >> > + /* Value of 1 for LNO_Iter_threshold is interpreted as default
>> >> >> > in
>> >> >> > which
>> >> >> > case
>> >> >> > + the flag is set based on target. Otherwise use
>> >> >> > user-specified
>> >> >> > value.
>> >> >> > + */
>> >> >> > + if(LNO_Iter_threshold == 1) {
>> >> >> > + LNO_Iter_threshold = (Is_Target_SSE41())? 8 : 0;
>> >> >> > + }
>> >> >> > }
>> >> >> > Index: /home/yugang/trunk/trunk/osprey/be/cg/whirl2ops.cxx
>> >> >> >
>> >> >> > ===================================================================
>> >> >> > --- /home/yugang/trunk/trunk/osprey/be/cg/whirl2ops.cxx (revision
>> >> >> > 3666)
>> >> >> > +++ /home/yugang/trunk/trunk/osprey/be/cg/whirl2ops.cxx (revision
>> >> >> > 3681)
>> >> >> > @@ -3166,9 +3166,18 @@
>> >> >> > WN *compare;
>> >> >> > VARIANT variant;
>> >> >> >
>> >> >> > + if (opcode == OPC_V16I1V16I1SELECT) {
>> >> >> > + TN* op1 = Expand_Expr(WN_kid0(select), select, NULL);
>> >> >> > + TN* op2 = Expand_Expr(WN_kid1(select), select, NULL);
>> >> >> > + TN* op3 = Expand_Expr(WN_kid2(select), select, NULL);
>> >> >> >
>> >> >> > + if (result == NULL)
>> >> >> > + result = Allocate_Result_TN (select, NULL);
>> >> >> > +
>> >> >> > + Expand_Select(result, op1, op2, op3, MTYPE_V16I1, FALSE,
>> >> >> > &New_OPs);
>> >> >> > //FALSE passed as dummy arg
>> >> >> > + return result;
>> >> >> > + }
>> >> >> >
>> >> >> > the "Is_Target_SSE41()", "OPC_V16I1V16I1SELECT" and "MTYPE_V16I1"
>> >> >> > are
>> >> >> > specific to X86 targets, these patches will cause other targets
>> >> >> > fail,
>> >> >> >
>> >> >> > The suggest patche is:
>> >> >> > Index: config_lno.cxx
>> >> >> >
>> >> >> > ===================================================================
>> >> >> > --- config_lno.cxx (revision 3681)
>> >> >> > +++ config_lno.cxx (working copy)
>> >> >> > @@ -1228,11 +1228,13 @@
>> >> >> > }
>> >> >> > }
>> >> >> >
>> >> >> > +#ifdef TARG_X8664
>> >> >> > /* Value of 1 for LNO_Iter_threshold is interpreted as default
>> >> >> > in
>> >> >> > which
>> >> >> > case
>> >> >> > the flag is set based on target. Otherwise use
>> >> >> > user-specified
>> >> >> > value.
>> >> >> > */
>> >> >> > if(LNO_Iter_threshold == 1) {
>> >> >> > LNO_Iter_threshold = (Is_Target_SSE41())? 8 : 0;
>> >> >> > }
>> >> >> > +#endif
>> >> >> > }
>> >> >> >
>> >> >> > Index: ../../be/cg/whirl2ops.cxx
>> >> >> >
>> >> >> > ===================================================================
>> >> >> > --- ../../be/cg/whirl2ops.cxx (revision 3681)
>> >> >> > +++ ../../be/cg/whirl2ops.cxx (working copy)
>> >> >> > @@ -3166,6 +3166,7 @@
>> >> >> > WN *compare;
>> >> >> > VARIANT variant;
>> >> >> >
>> >> >> > +#ifdef TARG_X8664
>> >> >> > if (opcode == OPC_V16I1V16I1SELECT) {
>> >> >> > TN* op1 = Expand_Expr(WN_kid0(select), select, NULL);
>> >> >> > TN* op2 = Expand_Expr(WN_kid1(select), select, NULL);
>> >> >> > @@ -3177,6 +3178,7 @@
>> >> >> > Expand_Select(result, op1, op2, op3, MTYPE_V16I1, FALSE,
>> >> >> > &New_OPs);
>> >> >> > //FALSE passed as dummy arg
>> >> >> > return result;
>> >> >> > }
>> >> >> > +#endif
>> >> >> >
>> >> >> >
>> >> >> >
>> >> >> > Would a gatekeeper help a review?
>> >> >> >
>> >> >> > Thanks
>> >> >> >
>> >> >> >
>> >> >> > Gang
>> >> >> >
>> >> >> >
>> >> >> >
>> >> >> >
>> >> >> > ------------------------------------------------------------------------------
>> >> >> > All of the data generated in your IT infrastructure is seriously
>> >> >> > valuable.
>> >> >> > Why? It contains a definitive record of application performance,
>> >> >> > security
>> >> >> > threats, fraudulent activity, and more. Splunk takes this data and
>> >> >> > makes
>> >> >> > sense of it. IT sense. And common sense.
>> >> >> > http://p.sf.net/sfu/splunk-d2d-c2
>> >> >> > _______________________________________________
>> >> >> > Open64-devel mailing list
>> >> >> > Open64-devel@lists.sourceforge.net
>> >> >> > https://lists.sourceforge.net/lists/listinfo/open64-devel
>> >> >> >
>> >> >> >
>> >> >
>> >> >
>> >
>> >
>
>
------------------------------------------------------------------------------
All of the data generated in your IT infrastructure is seriously valuable.
Why? It contains a definitive record of application performance, security
threats, fraudulent activity, and more. Splunk takes this data and makes
sense of it. IT sense. And common sense.
http://p.sf.net/sfu/splunk-d2d-c2
_______________________________________________
Open64-devel mailing list
Open64-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/open64-devel
