maxjhandsome
/
llvm-project
forked from OSchip/llvm-project
1
0
Fork 0
Code Issues Pull Requests Packages Releases Wiki Activity
23,320 Commits 23 Branches 0 Tags 2.2 GiB
Commit Graph

23320 Commits

Author SHA1 Message Date
Chris Lattner dc2f135f5c Make EmitNode take a SDNode instead of a NodeInfo* 
llvm-svn: 26687
 2006-03-10 07:28:36 +00:00
Chris Lattner b9d8fa0342 Move the VRBase field from NodeInfo to being a separate, explicit, map. 
llvm-svn: 26686
 2006-03-10 07:25:12 +00:00
Chris Lattner 0d443dc91d Store VRBase in a map, not in NodeInfo. 
llvm-svn: 26685
 2006-03-10 07:24:45 +00:00
Chris Lattner c48cfba44b no need to build groups anymore 
llvm-svn: 26684
 2006-03-10 07:15:58 +00:00
Chris Lattner 6f82fe8106 Create SUnits directly from the SelectionDAG. 
llvm-svn: 26683
 2006-03-10 07:13:32 +00:00
Chris Lattner 2f8c7c3d55 Push PrepareNodeInfo/IdentifyGroups down the inheritance hierarchy 
llvm-svn: 26682
 2006-03-10 06:34:51 +00:00
Chris Lattner 243f7dc0a0 make some methods protected instead of private 
llvm-svn: 26681
 2006-03-10 06:30:11 +00:00
Chris Lattner 349e9ddccc Teach the latency scheduler some new tricks. In particular, to break ties, 
keep track of a sense of "mobility", i.e. how many other nodes scheduling one
node will free up.  For something like this:

float testadd(float *X, float *Y, float *Z, float *W, float *V) {
  return (*X+*Y)*(*Z+*W)+*V;
}

For example, this makes us schedule *X then *Y, not *X then *Z.  The former
allows us to issue the add, the later only lets us issue other loads.

This turns the above code from this:

_testadd:
        lfs f0, 0(r3)
        lfs f1, 0(r6)
        lfs f2, 0(r4)
        lfs f3, 0(r5)
        fadds f0, f0, f2
        fadds f1, f3, f1
        lfs f2, 0(r7)
        fmadds f1, f0, f1, f2
        blr

into this:

_testadd:
        lfs f0, 0(r6)
        lfs f1, 0(r5)
        fadds f0, f1, f0
        lfs f1, 0(r4)
        lfs f2, 0(r3)
        fadds f1, f2, f1
        lfs f2, 0(r7)
        fmadds f1, f1, f0, f2
        blr

llvm-svn: 26680
 2006-03-10 05:51:05 +00:00
Jeff Cohen 22841aaadb Put intrinsics.gen in its proper place. 
llvm-svn: 26679
 2006-03-10 04:36:01 +00:00
Chris Lattner 25e2556b71 add an aggregate method for reinserting scheduled nodes, add a callback for 
priority impls that want to be notified when a node is scheduled

llvm-svn: 26678
 2006-03-10 04:32:49 +00:00
Chris Lattner 5353e35ac8 Fix an incorrect intrinsic description 
llvm-svn: 26677
 2006-03-10 04:17:06 +00:00
Jeff Cohen 6ce97687f7 Fix VC++ build breakage. 
llvm-svn: 26676
 2006-03-10 03:57:45 +00:00
Chris Lattner dfea29e2c5 Adding an intrinsic is simpler still. 
llvm-svn: 26674
 2006-03-09 22:38:42 +00:00
Chris Lattner 60f6833376 use autogenerated side-effect information 
llvm-svn: 26673
 2006-03-09 22:38:10 +00:00
Chris Lattner e3c2db3955 generate side-effect info 
llvm-svn: 26672
 2006-03-09 22:37:52 +00:00
Chris Lattner 02d2a6bf66 Simpler still 
llvm-svn: 26671
 2006-03-09 22:32:16 +00:00
Chris Lattner bb40a9fa58 Use autogenerated mod/ref info for intrinsics. 
llvm-svn: 26670
 2006-03-09 22:31:29 +00:00
Chris Lattner 06c7300bf1 Parse mod/ref properties, autogen mod/ref information 
llvm-svn: 26669
 2006-03-09 22:30:49 +00:00
Chris Lattner a0a2e0b25e Intrinsic adding is a little bit simpler now 
llvm-svn: 26668
 2006-03-09 22:07:39 +00:00
Chris Lattner 591693f3bf Use the autogenerated intrinsic verifier 
llvm-svn: 26667
 2006-03-09 22:06:04 +00:00
Chris Lattner 6efe863a70 parse intrinsic types 
autogenerate an intrinsic verifier

llvm-svn: 26666
 2006-03-09 22:05:04 +00:00
Evan Cheng 306c13a8fb Add option -enable-x86-lsr to enable x86 loop strength reduction pass. 
llvm-svn: 26665
 2006-03-09 21:51:28 +00:00
Chris Lattner ff4d4ee163 Use the function name matcher autogenerated from the .td file. 
llvm-svn: 26664
 2006-03-09 20:35:01 +00:00
Chris Lattner 6d8104efd2 autogenerate the function name recognizer 
llvm-svn: 26663
 2006-03-09 20:34:19 +00:00
Chris Lattner 285c95d7eb This rule also depends on tblgen 
llvm-svn: 26662
 2006-03-09 20:29:41 +00:00
Chris Lattner f136299635 add a note 
llvm-svn: 26661
 2006-03-09 20:13:21 +00:00
Chris Lattner 9f57681474 use the enum list autogen'd from Intrinsics.td 
llvm-svn: 26660
 2006-03-09 20:03:31 +00:00
Chris Lattner 213209a248 remove dbg_declare, it's not used yet. 
llvm-svn: 26659
 2006-03-09 20:02:42 +00:00
Chris Lattner ea0aa2868f silly case insensitive file systems... 
llvm-svn: 26658
 2006-03-09 19:55:06 +00:00
Chris Lattner 83e6e19936 Build intrinsics.gen from intrinsics.td 
llvm-svn: 26657
 2006-03-09 19:53:27 +00:00
Evan Cheng bee0b5b83e Add a test case for (store (op (load ..) ..) ..) folding. 
llvm-svn: 26656
 2006-03-09 19:04:30 +00:00
Chris Lattner f133c26a0b Update these tests (which use autoupgrade) to run constprop and check 
that the file parses.

llvm-svn: 26655
 2006-03-09 18:43:07 +00:00
Chris Lattner ab8d2ba046 autoupgrade memcpy/memmove/memset with signed counts. 
CVS: ----------------------------------------------------------------------

llvm-svn: 26654
 2006-03-09 18:42:10 +00:00
Andrew Lenharth 43e569c95f these are copies too 
llvm-svn: 26653
 2006-03-09 18:18:51 +00:00
Chris Lattner 7e7dccd3ab remove some now-dead code 
llvm-svn: 26652
 2006-03-09 18:07:49 +00:00
Andrew Lenharth 70236fc12f fcopysign for mixed mode 
llvm-svn: 26651
 2006-03-09 17:56:33 +00:00
Chris Lattner f0fc6bbaf3 back out my previous hack 
llvm-svn: 26650
 2006-03-09 17:48:46 +00:00
Andrew Lenharth ebfd94fa1d relax fcopysign 
llvm-svn: 26649
 2006-03-09 17:47:22 +00:00
Andrew Lenharth 4a87e7d9a3 alpha and llvm have different oppinions on which arg is the sign bit 
llvm-svn: 26647
 2006-03-09 17:41:50 +00:00
Chris Lattner c6c9e65301 remove temporary option 
llvm-svn: 26646
 2006-03-09 17:31:22 +00:00
Jim Laskey 93d079d757 Forgot this on last check in. 
llvm-svn: 26645
 2006-03-09 17:30:53 +00:00
Andrew Lenharth 16b96d2cb4 Alpha Scheduling classes 
llvm-svn: 26643
 2006-03-09 17:16:45 +00:00
Chris Lattner c44be2c4ef temporary hack to get the build working again, apparently a header 
commit was forgotten

llvm-svn: 26642
 2006-03-09 17:11:08 +00:00
Andrew Lenharth ed7a293b44 fcopysign and get rid of dsnode cruft. custom PA runtimes make this better in some senses 
llvm-svn: 26641
 2006-03-09 14:58:25 +00:00
Andrew Lenharth b8a06a7c6c fcopysign support 
llvm-svn: 26640
 2006-03-09 14:57:36 +00:00
Jim Laskey 4b0e1d9e6c Move bit field endianness to backend. 
llvm-svn: 26639
 2006-03-09 13:28:47 +00:00
Evan Cheng c47620b5d8 Temporary hack to enable more (store (op (load ...))) folding. This makes 
it possible when a TokenFactor is between the load and store. But is still
missing some cases due to ordering issue.

llvm-svn: 26638
 2006-03-09 08:19:11 +00:00
Chris Lattner d17d77aa1d yes yes, enabled debug output is bad 
llvm-svn: 26637
 2006-03-09 07:39:25 +00:00
Chris Lattner 6398c13128 switch the t-d scheduler to use a really dumb and trivial critical path 
latency priority function.

llvm-svn: 26636
 2006-03-09 07:38:27 +00:00
Chris Lattner d4130375c0 Pull latency information for target instructions out of the latency tables. :) 
Only enable this with -use-sched-latencies, I'll enable it by default with a
clean nightly tester run tonight.

PPC is the only target that provides latency info currently.

llvm-svn: 26634
 2006-03-09 07:15:18 +00:00