* RFC PRE-ing REFERENCE expressions
@ 2009-10-30 15:51 Rahul Kharche
2009-10-30 17:48 ` Richard Guenther
0 siblings, 1 reply; 4+ messages in thread
From: Rahul Kharche @ 2009-10-30 15:51 UTC (permalink / raw)
To: rguenth; +Cc: gcc, sdkteam-gnu
Hi Richi,
Following up your suggestion on PR41488, I am continuing to test with
loop header copy before FRE in GCC 4.4.1. One regression I am trying
to tackle is from the test case below.
typedef struct {
int degree;
int c[(256)];
} swbcht_Polynomial;
typedef struct {
int m;
int n;
const swbcht_Polynomial *primPoly;
unsigned short alpha[(1 << (13))];
} swbcht_GF;
typedef struct {
swbcht_GF gf;
} swbcht_Handle;
static const swbcht_Polynomial _primPoly13;
void _ComputeGFLookupTables (swbcht_Handle *bchH, int m)
{
int i, j;
swbcht_GF *gf = &bchH->gf;
gf->m = m;
gf->n = (1 << m) - 1;
gf->primPoly = &_primPoly13;
for (i = 0; i < gf->m; i++) {
gf->alpha[gf->m] ^= (gf->primPoly->c[i] * gf->alpha[i]);
}
}
The dump after CH - FRE looks like
_ComputeGFLookupTables (struct swbcht_Handle * bchH, int m)
{
int i;
short unsigned int D.1241;
short int D.1240;
short int D.1239;
short unsigned int D.1238;
short unsigned int D.1237;
short unsigned int D.1236;
const int D.1235;
const struct swbcht_Polynomial * D.1233;
short int D.1232;
short unsigned int D.1231;
int D.1230;
int D.1229;
int D.1228;
<bb 2>:
bchH_2(D)->gf.m = m_4(D);
D.1228_5 = 1 << m_4(D);
D.1229_6 = D.1228_5 + -1;
bchH_2(D)->gf.n = D.1229_6;
bchH_2(D)->gf.primPoly = &_primPoly13;
if (m_4(D) > 0)
goto <bb 5>;
else
goto <bb 6>;
<bb 6>:
goto <bb 4>;
<bb 5>:
<bb 3>:
# i_35 = PHI <0(5), i_23(7)>
D.1230_10 = bchH_2(D)->gf.m;
D.1231_11 = bchH_2(D)->gf.alpha[D.1230_10];
D.1232_12 = (short int) D.1231_11;
D.1233_13 = bchH_2(D)->gf.primPoly;
D.1235_15 = D.1233_13->c[i_35];
D.1236_16 = (short unsigned int) D.1235_15;
D.1237_18 = bchH_2(D)->gf.alpha[i_35];
D.1238_19 = D.1236_16 * D.1237_18;
D.1239_20 = (short int) D.1238_19;
D.1240_21 = D.1239_20 ^ D.1232_12;
D.1241_22 = (short unsigned int) D.1240_21;
bchH_2(D)->gf.alpha[D.1230_10] = D.1241_22;
i_23 = i_35 + 1;
D.1230_8 = bchH_2(D)->gf.m;
if (i_23 < D.1230_8)
goto <bb 7>;
else
goto <bb 8>;
<bb 7>:
goto <bb 3>;
<bb 8>:
<bb 4>:
return;
}
1. Why does FRE miss eliminating expression bchH_2(D)->gf.primPoly in
bb 3 with &_primPoly13. This is normally the case if we ran FRE then
CH.
Further PRE identifies bchH_2(D)->gf.primPoly as a partially redundant
expression and hoists it into predecessor bb 7 and we get
<bb 3>:
# i_35 = PHI <0(5), i_23(7)>
# prephitmp.25_49 = PHI <m_4(D)(5), D.1230_8(7)>
# prephitmp.27_51 = PHI <&_primPoly13(5), pretmp.26_50(7)>
D.1230_10 = prephitmp.25_49;
D.1231_11 = bchH_2(D)->gf.alpha[D.1230_10];
D.1232_12 = (short int) D.1231_11;
D.1233_13 = prephitmp.27_51;
D.1235_15 = D.1233_13->c[i_35];
D.1236_16 = (short unsigned int) D.1235_15;
D.1237_18 = bchH_2(D)->gf.alpha[i_35];
D.1238_19 = D.1236_16 * D.1237_18;
D.1239_20 = (short int) D.1238_19;
D.1240_21 = D.1239_20 ^ D.1232_12;
D.1241_22 = (short unsigned int) D.1240_21;
bchH_2(D)->gf.alpha[D.1230_10] = D.1241_22;
i_23 = i_35 + 1;
D.1230_8 = bchH_2(D)->gf.m;
if (D.1230_8 > i_23)
goto <bb 7>;
else
goto <bb 8>;
<bb 7>:
pretmp.26_50 = bchH_2(D)->gf.primPoly;
goto <bb 3>;
Clearly prephitmp.27_51 will make a redundant induction variable.
Stepping through the insert_into_preds_of_block in tree-ssa-pre.c
I can see the value numbers for expression bchH_2(D)->gf.primPoly
available through bb 3 and through bb 2 are different.
2. Is this because alias analysis cannot determine bchH_2(D)->gf
has a unique target?
Many Thanks,
Rahul
^ permalink raw reply [flat|nested] 4+ messages in thread
* Re: RFC PRE-ing REFERENCE expressions
2009-10-30 15:51 RFC PRE-ing REFERENCE expressions Rahul Kharche
@ 2009-10-30 17:48 ` Richard Guenther
2009-11-02 14:59 ` Rahul Kharche
0 siblings, 1 reply; 4+ messages in thread
From: Richard Guenther @ 2009-10-30 17:48 UTC (permalink / raw)
To: Rahul Kharche; +Cc: rguenth, gcc, sdkteam-gnu
On Fri, Oct 30, 2009 at 3:22 PM, Rahul Kharche <rahul@icerasemi.com> wrote:
> Hi Richi,
>
> Following up your suggestion on PR41488, I am continuing to test with
> loop header copy before FRE in GCC 4.4.1. One regression I am trying
> to tackle is from the test case below.
>
> typedef struct {
> int degree;
> int c[(256)];
> } swbcht_Polynomial;
>
>
> typedef struct {
> int m;
> int n;
> const swbcht_Polynomial *primPoly;
> unsigned short alpha[(1 << (13))];
> } swbcht_GF;
>
>
> typedef struct {
> swbcht_GF gf;
> } swbcht_Handle;
>
>
> static const swbcht_Polynomial _primPoly13;
>
> void _ComputeGFLookupTables (swbcht_Handle *bchH, int m)
> {
> int i, j;
>
> swbcht_GF *gf = &bchH->gf;
>
> gf->m = m;
> gf->n = (1 << m) - 1;
>
> gf->primPoly = &_primPoly13;
>
> for (i = 0; i < gf->m; i++) {
> gf->alpha[gf->m] ^= (gf->primPoly->c[i] * gf->alpha[i]);
> }
> }
>
>
> The dump after CH - FRE looks like
>
> _ComputeGFLookupTables (struct swbcht_Handle * bchH, int m)
> {
> int i;
> short unsigned int D.1241;
> short int D.1240;
> short int D.1239;
> short unsigned int D.1238;
> short unsigned int D.1237;
> short unsigned int D.1236;
> const int D.1235;
> const struct swbcht_Polynomial * D.1233;
> short int D.1232;
> short unsigned int D.1231;
> int D.1230;
> int D.1229;
> int D.1228;
>
> <bb 2>:
> bchH_2(D)->gf.m = m_4(D);
> D.1228_5 = 1 << m_4(D);
> D.1229_6 = D.1228_5 + -1;
> bchH_2(D)->gf.n = D.1229_6;
> bchH_2(D)->gf.primPoly = &_primPoly13;
> if (m_4(D) > 0)
> goto <bb 5>;
> else
> goto <bb 6>;
>
> <bb 6>:
> goto <bb 4>;
>
> <bb 5>:
>
> <bb 3>:
> # i_35 = PHI <0(5), i_23(7)>
> D.1230_10 = bchH_2(D)->gf.m;
> D.1231_11 = bchH_2(D)->gf.alpha[D.1230_10];
> D.1232_12 = (short int) D.1231_11;
> D.1233_13 = bchH_2(D)->gf.primPoly;
> D.1235_15 = D.1233_13->c[i_35];
> D.1236_16 = (short unsigned int) D.1235_15;
> D.1237_18 = bchH_2(D)->gf.alpha[i_35];
> D.1238_19 = D.1236_16 * D.1237_18;
> D.1239_20 = (short int) D.1238_19;
> D.1240_21 = D.1239_20 ^ D.1232_12;
> D.1241_22 = (short unsigned int) D.1240_21;
> bchH_2(D)->gf.alpha[D.1230_10] = D.1241_22;
> i_23 = i_35 + 1;
> D.1230_8 = bchH_2(D)->gf.m;
> if (i_23 < D.1230_8)
> goto <bb 7>;
> else
> goto <bb 8>;
>
> <bb 7>:
> goto <bb 3>;
>
> <bb 8>:
>
> <bb 4>:
> return;
>
> }
>
> 1. Why does FRE miss eliminating expression bchH_2(D)->gf.primPoly in
> bb 3 with &_primPoly13. This is normally the case if we ran FRE then
> CH.
>
> Further PRE identifies bchH_2(D)->gf.primPoly as a partially redundant
> expression and hoists it into predecessor bb 7 and we get
>
> <bb 3>:
> # i_35 = PHI <0(5), i_23(7)>
> # prephitmp.25_49 = PHI <m_4(D)(5), D.1230_8(7)>
> # prephitmp.27_51 = PHI <&_primPoly13(5), pretmp.26_50(7)>
> D.1230_10 = prephitmp.25_49;
> D.1231_11 = bchH_2(D)->gf.alpha[D.1230_10];
> D.1232_12 = (short int) D.1231_11;
> D.1233_13 = prephitmp.27_51;
> D.1235_15 = D.1233_13->c[i_35];
> D.1236_16 = (short unsigned int) D.1235_15;
> D.1237_18 = bchH_2(D)->gf.alpha[i_35];
> D.1238_19 = D.1236_16 * D.1237_18;
> D.1239_20 = (short int) D.1238_19;
> D.1240_21 = D.1239_20 ^ D.1232_12;
> D.1241_22 = (short unsigned int) D.1240_21;
> bchH_2(D)->gf.alpha[D.1230_10] = D.1241_22;
> i_23 = i_35 + 1;
> D.1230_8 = bchH_2(D)->gf.m;
> if (D.1230_8 > i_23)
> goto <bb 7>;
> else
> goto <bb 8>;
>
> <bb 7>:
> pretmp.26_50 = bchH_2(D)->gf.primPoly;
> goto <bb 3>;
>
>
> Clearly prephitmp.27_51 will make a redundant induction variable.
> Stepping through the insert_into_preds_of_block in tree-ssa-pre.c
> I can see the value numbers for expression bchH_2(D)->gf.primPoly
> available through bb 3 and through bb 2 are different.
>
> 2. Is this because alias analysis cannot determine bchH_2(D)->gf
> has a unique target?
You should move to GCC 4.5 - the alias-oracle in GCC 4.4 is too weak
to discover all this and virtual operands are not helpful because these
are all indirect accesses through pointers without points-to information.
Richard.
>
> Many Thanks,
> Rahul
>
>
^ permalink raw reply [flat|nested] 4+ messages in thread
* RE: RFC PRE-ing REFERENCE expressions
2009-10-30 17:48 ` Richard Guenther
@ 2009-11-02 14:59 ` Rahul Kharche
2009-11-02 15:01 ` Richard Guenther
0 siblings, 1 reply; 4+ messages in thread
From: Rahul Kharche @ 2009-11-02 14:59 UTC (permalink / raw)
To: Richard Guenther; +Cc: rguenth, gcc, sdkteam-gnu
Hi Richard,
We would like to hang on to 4.4 for a little while. I was hoping I could
grab only the alias analysis improvements from the trunk. Do you suspect
this would be troublesome?
Cheers,
Rahul
-----Original Message-----
From: Richard Guenther [mailto:richard.guenther@gmail.com]
Sent: 30 October 2009 14:50
To: Rahul Kharche
Cc: rguenth@gcc.gnu.org; gcc@gcc.gnu.org; sdkteam-gnu
Subject: Re: RFC PRE-ing REFERENCE expressions
On Fri, Oct 30, 2009 at 3:22 PM, Rahul Kharche <rahul@icerasemi.com> wrote:
> Hi Richi,
>
> Following up your suggestion on PR41488, I am continuing to test with
> loop header copy before FRE in GCC 4.4.1. One regression I am trying
> to tackle is from the test case below.
>
> typedef struct {
> int degree;
> int c[(256)];
> } swbcht_Polynomial;
>
>
> typedef struct {
> int m;
> int n;
> const swbcht_Polynomial *primPoly;
> unsigned short alpha[(1 << (13))];
> } swbcht_GF;
>
>
> typedef struct {
> swbcht_GF gf;
> } swbcht_Handle;
>
>
> static const swbcht_Polynomial _primPoly13;
>
> void _ComputeGFLookupTables (swbcht_Handle *bchH, int m)
> {
> int i, j;
>
> swbcht_GF *gf = &bchH->gf;
>
> gf->m = m;
> gf->n = (1 << m) - 1;
>
> gf->primPoly = &_primPoly13;
>
> for (i = 0; i < gf->m; i++) {
> gf->alpha[gf->m] ^= (gf->primPoly->c[i] * gf->alpha[i]);
> }
> }
>
>
> The dump after CH - FRE looks like
>
> _ComputeGFLookupTables (struct swbcht_Handle * bchH, int m)
> {
> int i;
> short unsigned int D.1241;
> short int D.1240;
> short int D.1239;
> short unsigned int D.1238;
> short unsigned int D.1237;
> short unsigned int D.1236;
> const int D.1235;
> const struct swbcht_Polynomial * D.1233;
> short int D.1232;
> short unsigned int D.1231;
> int D.1230;
> int D.1229;
> int D.1228;
>
> <bb 2>:
> bchH_2(D)->gf.m = m_4(D);
> D.1228_5 = 1 << m_4(D);
> D.1229_6 = D.1228_5 + -1;
> bchH_2(D)->gf.n = D.1229_6;
> bchH_2(D)->gf.primPoly = &_primPoly13;
> if (m_4(D) > 0)
> goto <bb 5>;
> else
> goto <bb 6>;
>
> <bb 6>:
> goto <bb 4>;
>
> <bb 5>:
>
> <bb 3>:
> # i_35 = PHI <0(5), i_23(7)>
> D.1230_10 = bchH_2(D)->gf.m;
> D.1231_11 = bchH_2(D)->gf.alpha[D.1230_10];
> D.1232_12 = (short int) D.1231_11;
> D.1233_13 = bchH_2(D)->gf.primPoly;
> D.1235_15 = D.1233_13->c[i_35];
> D.1236_16 = (short unsigned int) D.1235_15;
> D.1237_18 = bchH_2(D)->gf.alpha[i_35];
> D.1238_19 = D.1236_16 * D.1237_18;
> D.1239_20 = (short int) D.1238_19;
> D.1240_21 = D.1239_20 ^ D.1232_12;
> D.1241_22 = (short unsigned int) D.1240_21;
> bchH_2(D)->gf.alpha[D.1230_10] = D.1241_22;
> i_23 = i_35 + 1;
> D.1230_8 = bchH_2(D)->gf.m;
> if (i_23 < D.1230_8)
> goto <bb 7>;
> else
> goto <bb 8>;
>
> <bb 7>:
> goto <bb 3>;
>
> <bb 8>:
>
> <bb 4>:
> return;
>
> }
>
> 1. Why does FRE miss eliminating expression bchH_2(D)->gf.primPoly in
> bb 3 with &_primPoly13. This is normally the case if we ran FRE then
> CH.
>
> Further PRE identifies bchH_2(D)->gf.primPoly as a partially redundant
> expression and hoists it into predecessor bb 7 and we get
>
> <bb 3>:
> # i_35 = PHI <0(5), i_23(7)>
> # prephitmp.25_49 = PHI <m_4(D)(5), D.1230_8(7)>
> # prephitmp.27_51 = PHI <&_primPoly13(5), pretmp.26_50(7)>
> D.1230_10 = prephitmp.25_49;
> D.1231_11 = bchH_2(D)->gf.alpha[D.1230_10];
> D.1232_12 = (short int) D.1231_11;
> D.1233_13 = prephitmp.27_51;
> D.1235_15 = D.1233_13->c[i_35];
> D.1236_16 = (short unsigned int) D.1235_15;
> D.1237_18 = bchH_2(D)->gf.alpha[i_35];
> D.1238_19 = D.1236_16 * D.1237_18;
> D.1239_20 = (short int) D.1238_19;
> D.1240_21 = D.1239_20 ^ D.1232_12;
> D.1241_22 = (short unsigned int) D.1240_21;
> bchH_2(D)->gf.alpha[D.1230_10] = D.1241_22;
> i_23 = i_35 + 1;
> D.1230_8 = bchH_2(D)->gf.m;
> if (D.1230_8 > i_23)
> goto <bb 7>;
> else
> goto <bb 8>;
>
> <bb 7>:
> pretmp.26_50 = bchH_2(D)->gf.primPoly;
> goto <bb 3>;
>
>
> Clearly prephitmp.27_51 will make a redundant induction variable.
> Stepping through the insert_into_preds_of_block in tree-ssa-pre.c
> I can see the value numbers for expression bchH_2(D)->gf.primPoly
> available through bb 3 and through bb 2 are different.
>
> 2. Is this because alias analysis cannot determine bchH_2(D)->gf
> has a unique target?
You should move to GCC 4.5 - the alias-oracle in GCC 4.4 is too weak
to discover all this and virtual operands are not helpful because these
are all indirect accesses through pointers without points-to information.
Richard.
>
> Many Thanks,
> Rahul
>
>
^ permalink raw reply [flat|nested] 4+ messages in thread
* RE: RFC PRE-ing REFERENCE expressions
2009-11-02 14:59 ` Rahul Kharche
@ 2009-11-02 15:01 ` Richard Guenther
0 siblings, 0 replies; 4+ messages in thread
From: Richard Guenther @ 2009-11-02 15:01 UTC (permalink / raw)
To: Rahul Kharche; +Cc: Richard Guenther, rguenth, gcc, sdkteam-gnu
[-- Attachment #1: Type: TEXT/PLAIN, Size: 5191 bytes --]
On Mon, 2 Nov 2009, Rahul Kharche wrote:
> Hi Richard,
>
> We would like to hang on to 4.4 for a little while. I was hoping I could
> grab only the alias analysis improvements from the trunk. Do you suspect
> this would be troublesome?
Uh, definitely - it's not easy to grab them in isolation.
Richard.
> Cheers,
> Rahul
>
> -----Original Message-----
> From: Richard Guenther [mailto:richard.guenther@gmail.com]
> Sent: 30 October 2009 14:50
> To: Rahul Kharche
> Cc: rguenth@gcc.gnu.org; gcc@gcc.gnu.org; sdkteam-gnu
> Subject: Re: RFC PRE-ing REFERENCE expressions
>
> On Fri, Oct 30, 2009 at 3:22 PM, Rahul Kharche <rahul@icerasemi.com> wrote:
> > Hi Richi,
> >
> > Following up your suggestion on PR41488, I am continuing to test with
> > loop header copy before FRE in GCC 4.4.1. One regression I am trying
> > to tackle is from the test case below.
> >
> > typedef struct {
> > Â int degree;
> > Â int c[(256)];
> > } swbcht_Polynomial;
> >
> >
> > typedef struct {
> > Â int m;
> > Â int n;
> > Â const swbcht_Polynomial *primPoly;
> > Â unsigned short alpha[(1 << (13))];
> > } swbcht_GF;
> >
> >
> > typedef struct {
> > Â swbcht_GF gf;
> > } swbcht_Handle;
> >
> >
> > static const swbcht_Polynomial _primPoly13;
> >
> > void _ComputeGFLookupTables (swbcht_Handle *bchH, int m)
> > {
> > Â int i, j;
> >
> > Â swbcht_GF *gf = &bchH->gf;
> >
> > Â gf->m = m;
> > Â gf->n = (1 << m) - 1;
> >
> > Â gf->primPoly = &_primPoly13;
> >
> > Â for (i = 0; i < gf->m; i++) {
> > Â Â gf->alpha[gf->m] ^= (gf->primPoly->c[i] * gf->alpha[i]);
> > Â }
> > }
> >
> >
> > The dump after CH - FRE looks like
> >
> > _ComputeGFLookupTables (struct swbcht_Handle * bchH, int m)
> > {
> > Â int i;
> > Â short unsigned int D.1241;
> > Â short int D.1240;
> > Â short int D.1239;
> > Â short unsigned int D.1238;
> > Â short unsigned int D.1237;
> > Â short unsigned int D.1236;
> > Â const int D.1235;
> > Â const struct swbcht_Polynomial * D.1233;
> > Â short int D.1232;
> > Â short unsigned int D.1231;
> > Â int D.1230;
> > Â int D.1229;
> > Â int D.1228;
> >
> > <bb 2>:
> > Â bchH_2(D)->gf.m = m_4(D);
> > Â D.1228_5 = 1 << m_4(D);
> > Â D.1229_6 = D.1228_5 + -1;
> > Â bchH_2(D)->gf.n = D.1229_6;
> > Â bchH_2(D)->gf.primPoly = &_primPoly13;
> > Â if (m_4(D) > 0)
> > Â Â goto <bb 5>;
> > Â else
> > Â Â goto <bb 6>;
> >
> > <bb 6>:
> > Â goto <bb 4>;
> >
> > <bb 5>:
> >
> > <bb 3>:
> > Â # i_35 = PHI <0(5), i_23(7)>
> > Â D.1230_10 = bchH_2(D)->gf.m;
> > Â D.1231_11 = bchH_2(D)->gf.alpha[D.1230_10];
> > Â D.1232_12 = (short int) D.1231_11;
> > Â D.1233_13 = bchH_2(D)->gf.primPoly;
> > Â D.1235_15 = D.1233_13->c[i_35];
> > Â D.1236_16 = (short unsigned int) D.1235_15;
> > Â D.1237_18 = bchH_2(D)->gf.alpha[i_35];
> > Â D.1238_19 = D.1236_16 * D.1237_18;
> > Â D.1239_20 = (short int) D.1238_19;
> > Â D.1240_21 = D.1239_20 ^ D.1232_12;
> > Â D.1241_22 = (short unsigned int) D.1240_21;
> > Â bchH_2(D)->gf.alpha[D.1230_10] = D.1241_22;
> > Â i_23 = i_35 + 1;
> > Â D.1230_8 = bchH_2(D)->gf.m;
> > Â if (i_23 < D.1230_8)
> > Â Â goto <bb 7>;
> > Â else
> > Â Â goto <bb 8>;
> >
> > <bb 7>:
> > Â goto <bb 3>;
> >
> > <bb 8>:
> >
> > <bb 4>:
> > Â return;
> >
> > }
> >
> > 1. Why does FRE miss eliminating expression bchH_2(D)->gf.primPoly in
> > bb 3 with &_primPoly13. This is normally the case if we ran FRE then
> > CH.
> >
> > Further PRE identifies bchH_2(D)->gf.primPoly as a partially redundant
> > expression and hoists it into predecessor bb 7 and we get
> >
> > <bb 3>:
> > Â # i_35 = PHI <0(5), i_23(7)>
> > Â # prephitmp.25_49 = PHI <m_4(D)(5), D.1230_8(7)>
> > Â # prephitmp.27_51 = PHI <&_primPoly13(5), pretmp.26_50(7)>
> > Â D.1230_10 = prephitmp.25_49;
> > Â D.1231_11 = bchH_2(D)->gf.alpha[D.1230_10];
> > Â D.1232_12 = (short int) D.1231_11;
> > Â D.1233_13 = prephitmp.27_51;
> > Â D.1235_15 = D.1233_13->c[i_35];
> > Â D.1236_16 = (short unsigned int) D.1235_15;
> > Â D.1237_18 = bchH_2(D)->gf.alpha[i_35];
> > Â D.1238_19 = D.1236_16 * D.1237_18;
> > Â D.1239_20 = (short int) D.1238_19;
> > Â D.1240_21 = D.1239_20 ^ D.1232_12;
> > Â D.1241_22 = (short unsigned int) D.1240_21;
> > Â bchH_2(D)->gf.alpha[D.1230_10] = D.1241_22;
> > Â i_23 = i_35 + 1;
> > Â D.1230_8 = bchH_2(D)->gf.m;
> > Â if (D.1230_8 > i_23)
> > Â Â goto <bb 7>;
> > Â else
> > Â Â goto <bb 8>;
> >
> > <bb 7>:
> > Â pretmp.26_50 = bchH_2(D)->gf.primPoly;
> > Â goto <bb 3>;
> >
> >
> > Clearly prephitmp.27_51 will make a redundant induction variable.
> > Stepping through the insert_into_preds_of_block in tree-ssa-pre.c
> > I can see the value numbers for expression bchH_2(D)->gf.primPoly
> > available through bb 3 and through bb 2 are different.
> >
> > 2. Is this because alias analysis cannot determine bchH_2(D)->gf
> > has a unique target?
>
> You should move to GCC 4.5 - the alias-oracle in GCC 4.4 is too weak
> to discover all this and virtual operands are not helpful because these
> are all indirect accesses through pointers without points-to information.
>
> Richard.
>
> >
> > Many Thanks,
> > Rahul
> >
> >
>
>
--
Richard Guenther <rguenther@suse.de>
Novell / SUSE Labs
SUSE LINUX Products GmbH - Nuernberg - AG Nuernberg - HRB 16746 - GF: Markus Rex
^ permalink raw reply [flat|nested] 4+ messages in thread
end of thread, other threads:[~2009-11-02 15:01 UTC | newest]
Thread overview: 4+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2009-10-30 15:51 RFC PRE-ing REFERENCE expressions Rahul Kharche
2009-10-30 17:48 ` Richard Guenther
2009-11-02 14:59 ` Rahul Kharche
2009-11-02 15:01 ` Richard Guenther
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).