Commit Graph

51 Commits

Author SHA1 Message Date
Chris Lattner
d998938459 Implement Regression/CodeGen/PowerPC/bswap-load-store.ll by folding bswaps
into i16/i32 load/stores.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@29089 91177308-0d34-0410-b5e6-96231b3b80d8
2006-07-10 20:56:58 +00:00
Chris Lattner
c703a8fbf8 Make PPC call lowering more aggressive, making the isel matching code simple
enough to be autogenerated.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@28354 91177308-0d34-0410-b5e6-96231b3b80d8
2006-05-17 19:00:46 +00:00
Chris Lattner
abde460d4f Instead of implementing LowerCallTo directly, let the default impl produce an
ISD::CALL node, then custom lower that.  This means that we only have to handle
LEGAL call operands/results, not every possible type.  This allows us to
simplify the call code, shrinking it by about 1/3.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@28339 91177308-0d34-0410-b5e6-96231b3b80d8
2006-05-16 22:56:08 +00:00
Chris Lattner
8ab5fe574a Switch the PPC backend over to using FORMAL_ARGUMENTS for formal argument
handling.  This makes the lower argument code significantly simpler (we
only need to handle legal argument types).

Incidentally, this also implements support for vector argument registers,
so long as they are not on the stack.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@28331 91177308-0d34-0410-b5e6-96231b3b80d8
2006-05-16 18:18:50 +00:00
Chris Lattner
90564f26d1 Implement an important entry from README_ALTIVEC:
If an altivec predicate compare is used immediately by a branch, don't
use a (serializing) MFCR instruction to read the CR6 register, which requires
a compare to get it back to CR's.  Instead, just branch on CR6 directly. :)

For example, for:
void foo2(vector float *A, vector float *B) {
  if (!vec_any_eq(*A, *B))
    *B = (vector float){0,0,0,0};
}

We now generate:

_foo2:
        mfspr r2, 256
        oris r5, r2, 12288
        mtspr 256, r5
        lvx v2, 0, r4
        lvx v3, 0, r3
        vcmpeqfp. v2, v3, v2
        bne cr6, LBB1_2 ; UnifiedReturnBlock
LBB1_1: ; cond_true
        vxor v2, v2, v2
        stvx v2, 0, r4
        mtspr 256, r2
        blr
LBB1_2: ; UnifiedReturnBlock
        mtspr 256, r2
        blr

instead of:

_foo2:
        mfspr r2, 256
        oris r5, r2, 12288
        mtspr 256, r5
        lvx v2, 0, r4
        lvx v3, 0, r3
        vcmpeqfp. v2, v3, v2
        mfcr r3, 2
        rlwinm r3, r3, 27, 31, 31
        cmpwi cr0, r3, 0
        beq cr0, LBB1_2 ; UnifiedReturnBlock
LBB1_1: ; cond_true
        vxor v2, v2, v2
        stvx v2, 0, r4
        mtspr 256, r2
        blr
LBB1_2: ; UnifiedReturnBlock
        mtspr 256, r2
        blr

This implements CodeGen/PowerPC/vec_br_cmp.ll.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@27804 91177308-0d34-0410-b5e6-96231b3b80d8
2006-04-18 17:59:36 +00:00
Chris Lattner
e87192a854 Rename get_VSPLI_elt -> get_VSPLTI_elt
Canonicalize BUILD_VECTOR's that match VSPLTI's into a single type for each
form, eliminating a bunch of Pat patterns in the .td file and allowing us to
CSE stuff more aggressively.  This implements
PowerPC/buildvec_canonicalize.ll:VSPLTI


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@27614 91177308-0d34-0410-b5e6-96231b3b80d8
2006-04-12 17:37:20 +00:00
Chris Lattner
140a58f9df Change the interface to the predicate that determines if vsplti* can be used.
No functionality changes.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@27536 91177308-0d34-0410-b5e6-96231b3b80d8
2006-04-08 06:46:53 +00:00
Chris Lattner
f24380e78e Match vpku[hw]um(x,x).
Convert vsldoi(x,x) to work the same way other (x,x) cases work.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@27467 91177308-0d34-0410-b5e6-96231b3b80d8
2006-04-06 22:28:36 +00:00
Chris Lattner
caad163496 Add support for matching vmrg(x,x) patterns
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@27463 91177308-0d34-0410-b5e6-96231b3b80d8
2006-04-06 22:02:42 +00:00
Chris Lattner
116cc48e30 Pattern match vmrg* instructions, which are now lowered by the CFE into shuffles.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@27457 91177308-0d34-0410-b5e6-96231b3b80d8
2006-04-06 21:11:54 +00:00
Chris Lattner
d0608e191f Support pattern matching vsldoi(x,y) and vsldoi(x,x), which allows the f.e. to
lower it and LLVM to have one fewer intrinsic.  This implements
CodeGen/PowerPC/vec_shuffle.ll


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@27450 91177308-0d34-0410-b5e6-96231b3b80d8
2006-04-06 18:26:28 +00:00
Chris Lattner
ddb739e5ea Compile the vpkuhum/vpkuwum intrinsics into vpkuhum/vpkuwum instead of into
vperm with a perm mask lvx'd from the constant pool.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@27448 91177308-0d34-0410-b5e6-96231b3b80d8
2006-04-06 17:23:16 +00:00
Chris Lattner
7ff7e67458 Ask legalize to promote all vector shuffles to be v16i8 instead of having to
handle all 4 PPC vector types.   This simplifies the matching code and allows
us to eliminate a bunch of patterns.  This also adds cases we were missing,
such as CodeGen/PowerPC/vec_splat.ll:splat_h.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@27400 91177308-0d34-0410-b5e6-96231b3b80d8
2006-04-04 17:25:31 +00:00
Chris Lattner
bbe77de450 Inform the dag combiner that the predicate compares only return a low bit.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@27359 91177308-0d34-0410-b5e6-96231b3b80d8
2006-04-02 06:26:07 +00:00
Chris Lattner
a17b1557ad Lower vector compares to VCMP nodes, just like we lower vector comparison
predicates to VCMPo nodes.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@27285 91177308-0d34-0410-b5e6-96231b3b80d8
2006-03-31 05:13:27 +00:00
Chris Lattner
7f20b13518 Use normal lvx for scalar_to_vector instead of lve*x. They do the exact
same thing and we have a dag node for the former.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@27205 91177308-0d34-0410-b5e6-96231b3b80d8
2006-03-28 01:43:22 +00:00
Chris Lattner
6d92caddc4 Codegen vector predicate compares.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@27151 91177308-0d34-0410-b5e6-96231b3b80d8
2006-03-26 10:06:40 +00:00
Evan Cheng
5b6a01b59c Remove PPC:isZeroVector, use ISD::isBuildVectorAllZeros instead
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@27149 91177308-0d34-0410-b5e6-96231b3b80d8
2006-03-26 09:52:32 +00:00
Chris Lattner
9c61dcf1aa Codegen things like:
<int -1, int -1, int -1, int -1>
and
 <int 65537, int 65537, int 65537, int 65537>

Using things like:
  vspltisb v0, -1
and:
  vspltish v0, 1

instead of using constant pool loads.

This implements CodeGen/PowerPC/vec_splat.ll:splat_imm_i{32|16}.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@27106 91177308-0d34-0410-b5e6-96231b3b80d8
2006-03-25 06:12:06 +00:00
Chris Lattner
64b3a08bc6 add support for using vxor to build zero vectors. This implements
Regression/CodeGen/PowerPC/vec_zero.ll


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@27059 91177308-0d34-0410-b5e6-96231b3b80d8
2006-03-24 07:48:08 +00:00
Chris Lattner
ecfe55e65b When possible, custom lower 32-bit SINT_TO_FP to this:
_foo2:
        extsw r2, r3
        std r2, -8(r1)
        lfd f0, -8(r1)
        fcfid f0, f0
        frsp f1, f0
        blr

instead of this:

_foo2:
        lis r2, ha16(LCPI2_0)
        lis r4, 17200
        xoris r3, r3, 32768
        stw r3, -4(r1)
        stw r4, -8(r1)
        lfs f0, lo16(LCPI2_0)(r2)
        lfd f1, -8(r1)
        fsub f0, f1, f0
        frsp f1, f0
        blr

This speeds up Misc/pi from 2.44s->2.09s with LLC and from 3.01->2.18s
with llcbeta (16.7% and 38.1% respectively).


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@26943 91177308-0d34-0410-b5e6-96231b3b80d8
2006-03-22 05:30:33 +00:00
Chris Lattner
ef819f8fbb fix duplicate definition errors
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@26896 91177308-0d34-0410-b5e6-96231b3b80d8
2006-03-20 06:33:01 +00:00
Chris Lattner
3c0f9cc90c Check in some intermediate code that adds a skeleton for matching vsplt*
instructions


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@26894 91177308-0d34-0410-b5e6-96231b3b80d8
2006-03-20 06:15:45 +00:00
Chris Lattner
f1d0b2beda Custom lower arbitrary VECTOR_SHUFFLE's to VPERM.
TODO: leave specific ones as VECTOR_SHUFFLE's and turn them into specialized
operations like vsplt*


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@26887 91177308-0d34-0410-b5e6-96231b3b80d8
2006-03-20 01:53:53 +00:00
Chris Lattner
b2177b9405 Custom lower SCALAR_TO_VECTOR into lve*x.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@26868 91177308-0d34-0410-b5e6-96231b3b80d8
2006-03-19 06:55:52 +00:00
Evan Cheng
c4c6257c1a Added getTargetLowering() to TargetMachine. Refactored targets to support this.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@26742 91177308-0d34-0410-b5e6-96231b3b80d8
2006-03-13 23:20:37 +00:00
Chris Lattner
5126984b1d Compile this:
void foo(float a, int *b) { *b = a; }

to this:

_foo:
        fctiwz f0, f1
        stfiwx f0, 0, r4
        blr

instead of this:

_foo:
        fctiwz f0, f1
        stfd f0, -8(r1)
        lwz r2, -4(r1)
        stw r2, 0(r4)
        blr

This implements CodeGen/PowerPC/stfiwx.ll, and also incidentally does the
right thing for GCC bugzilla 26505.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@26447 91177308-0d34-0410-b5e6-96231b3b80d8
2006-03-01 05:50:56 +00:00
Chris Lattner
8c13d0a573 Use a target-specific dag-combine to implement CodeGen/PowerPC/fp-int-fp.ll.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@26445 91177308-0d34-0410-b5e6-96231b3b80d8
2006-03-01 04:57:39 +00:00
Chris Lattner
1efa40f6a4 split register class handling from explicit physreg handling.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@26308 91177308-0d34-0410-b5e6-96231b3b80d8
2006-02-22 00:56:39 +00:00
Chris Lattner
4217ca8dc1 Updates to match change of getRegForInlineAsmConstraint prototype
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@26305 91177308-0d34-0410-b5e6-96231b3b80d8
2006-02-21 23:11:00 +00:00
Chris Lattner
ad3bc8d8ad Implement getConstraintType for PPC.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@26042 91177308-0d34-0410-b5e6-96231b3b80d8
2006-02-07 20:16:30 +00:00
Chris Lattner
763317de1b Add the simple PPC integer constraints
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@26027 91177308-0d34-0410-b5e6-96231b3b80d8
2006-02-07 00:47:13 +00:00
Chris Lattner
ddc787dfdc add info about the inline asm register constraints for PPC
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@25853 91177308-0d34-0410-b5e6-96231b3b80d8
2006-01-31 19:20:21 +00:00
Chris Lattner
281b55ebec Use PPCISD::CALL instead of ISD::CALL
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@25717 91177308-0d34-0410-b5e6-96231b3b80d8
2006-01-27 23:34:02 +00:00
Chris Lattner
bba534dc66 Make llvm.frame/returnaddr not crash on ppc
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@25710 91177308-0d34-0410-b5e6-96231b3b80d8
2006-01-27 22:25:06 +00:00
Nate Begeman
ee625573b5 Remove TLI.LowerReturnTo, and just let targets custom lower ISD::RET for
the same functionality.  This addresses another piece of bug 680.  Next,
on to fixing Alpha VAARG, which I broke last time.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@25696 91177308-0d34-0410-b5e6-96231b3b80d8
2006-01-27 21:09:22 +00:00
Nate Begeman
acc398c195 First part of bug 680:
Remove TLI.LowerVA* and replace it with SDNodes that are lowered the same
way as everything else.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@25606 91177308-0d34-0410-b5e6-96231b3b80d8
2006-01-25 18:21:52 +00:00
Chris Lattner
da6d20f0c1 Give PPCISD:: nodes legible names in dumps.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@25166 91177308-0d34-0410-b5e6-96231b3b80d8
2006-01-09 23:52:17 +00:00
Nate Begeman
9e4dd9dfc9 Pattern-match return. Includes gross hack!
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@24874 91177308-0d34-0410-b5e6-96231b3b80d8
2005-12-20 00:26:01 +00:00
Nate Begeman
993aeb2ed9 Prepare support for AltiVec multiply, divide, and sqrt.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@24700 91177308-0d34-0410-b5e6-96231b3b80d8
2005-12-13 22:55:22 +00:00
Chris Lattner
4172b10ca1 Use new PPC-specific nodes to represent shifts which require the 6-bit
amount handling that PPC provides.  These are generated by the lowering code
and prevents the dag combiner from assuming (rightfully) that the shifts
don't only look at 5 bits.  This fixes a miscompilation of crafty with
the new front-end.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@24615 91177308-0d34-0410-b5e6-96231b3b80d8
2005-12-06 02:10:38 +00:00
Chris Lattner
860e8862c1 Add an initial hack at legalizing GlobalAddress into the appropriate nodes
on Darwin to remove smarts from the isel.  This is currently disabled by
default (uncomment setOperationAction(ISD::GlobalAddress to enable it).
tblgen needs to become smarter about tglobaladdr nodes and bigger patterns
needed to be added to the .td file.  However, we can currently emit stuff like
this:  :)

        li r2, lo16(L_x$non_lazy_ptr)
        lis r3, ha16(L_x$non_lazy_ptr)
        lwzx r2, r3, r2

The obvious improvements will follow.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@24390 91177308-0d34-0410-b5e6-96231b3b80d8
2005-11-17 07:30:41 +00:00
Nate Begeman
4a95945fa5 Add the ability to lower return instructions to TargetLowering. This
allows us to lower legal return types to something else, to meet ABI
requirements (such as that i64 be returned in two i32 regs on Darwin/ppc).


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@23802 91177308-0d34-0410-b5e6-96231b3b80d8
2005-10-18 23:23:37 +00:00
Nate Begeman
21e463b2bf More PPC32 -> PPC changes, as well as merging some classes that were
redundant after the change.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@23759 91177308-0d34-0410-b5e6-96231b3b80d8
2005-10-16 05:39:50 +00:00
Chris Lattner
2668959b88 Rename PowerPC*.h to PPC*.h
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@23743 91177308-0d34-0410-b5e6-96231b3b80d8
2005-10-14 23:51:18 +00:00
Nate Begeman
c09eeec0eb Implement i64<->fp using the fctidz/fcfid instructions on PowerPC when we
are allowed to generate 64-bit-only PowerPC instructions for 32 bit hosts,
such as the PowerPC 970.

This speeds up 189.lucas from 81.99 to 32.64 seconds.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@23250 91177308-0d34-0410-b5e6-96231b3b80d8
2005-09-06 22:03:27 +00:00
Chris Lattner
f76053269e Move FCTIWZ handling out of the instruction selectors and into legalization,
getting them out of the business of making stack slots.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@23180 91177308-0d34-0410-b5e6-96231b3b80d8
2005-08-31 21:09:52 +00:00
Chris Lattner
8a2d3ca7df implement SELECT_CC fully for the DAG->DAG isel!
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@23101 91177308-0d34-0410-b5e6-96231b3b80d8
2005-08-26 21:23:58 +00:00
Chris Lattner
0bbea95433 Make fsel emission work with both the pattern and dag-dag selectors, by
giving it a non-instruction opcode.  The dag->dag selector used to not
select the operands of the fsel, because it thought that whole tree was
already selected.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@23091 91177308-0d34-0410-b5e6-96231b3b80d8
2005-08-26 20:25:03 +00:00
Chris Lattner
e4bc9ea0a5 add initial support for converting select_cc -> fsel in the legalizer
instead of in the backend.  This currently handles fsel cases with registers,
but doesn't have the 0.0 and -0.0 optimization enabled yet.

Once this is finished, special hack for fp immediates can go away.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@23075 91177308-0d34-0410-b5e6-96231b3b80d8
2005-08-26 00:52:45 +00:00