aboutsummaryrefslogtreecommitdiffstats
path: root/test/CodeGen
Commit message (Collapse)AuthorAgeFilesLines
* Add nounwind.Evan Cheng2010-04-171-6/+6
| | | | git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@101613 91177308-0d34-0410-b5e6-96231b3b80d8
* Add test case for machine-sink on critical edgesJakob Stoklund Olesen2010-04-151-0/+58
| | | | git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@101416 91177308-0d34-0410-b5e6-96231b3b80d8
* Use default lowering of DYNAMIC_STACKALLOC. As far as I can tell, ARM isle ↵Evan Cheng2010-04-151-8/+8
| | | | | | is doing the right thing and codegen looks correct for both Thumb and Thumb2. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@101410 91177308-0d34-0410-b5e6-96231b3b80d8
* Fix PR6847. RegScavenger should ignore DebugValues.Jakob Stoklund Olesen2010-04-151-0/+26
| | | | git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@101392 91177308-0d34-0410-b5e6-96231b3b80d8
* ARM SelectDYN_ALLOC should emit a copy from SP rather than referencing SP ↵Evan Cheng2010-04-151-0/+23
| | | | | | | | | directly. In cases where there are two dyn_alloc in the same BB it would have caused the old SP value to be reused and badness ensues. rdar://7493908 llvm is generating poor code for dynamic alloca, I'll fix that later. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@101383 91177308-0d34-0410-b5e6-96231b3b80d8
* enhance the load/store narrowing optimization to handle aChris Lattner2010-04-151-0/+20
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | tokenfactor in between the load/store. This allows us to optimize test7 into: _test7: ## @test7 ## BB#0: ## %entry movl (%rdx), %eax ## kill: SIL<def> ESI<kill> movb %sil, 5(%rdi) ret instead of: _test7: ## @test7 ## BB#0: ## %entry movl 4(%esp), %ecx movl $-65281, %eax ## imm = 0xFFFFFFFFFFFF00FF andl 4(%ecx), %eax movzbl 8(%esp), %edx shll $8, %edx addl %eax, %edx movl 12(%esp), %eax movl (%eax), %eax movl %edx, 4(%ecx) ret git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@101355 91177308-0d34-0410-b5e6-96231b3b80d8
* teach codegen to turn trunc(zextload) into load when possible.Chris Lattner2010-04-151-13/+39
| | | | | | | | | This doesn't occur much at all, it only seems to formed in the case when the trunc optimization kicks in due to phase ordering. In that case it is saves a few bytes on x86-32. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@101350 91177308-0d34-0410-b5e6-96231b3b80d8
* add a simple dag combine to replace trivial shl+lshr withChris Lattner2010-04-151-1/+1
| | | | | | | and. This happens with the store->load narrowing stuff. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@101348 91177308-0d34-0410-b5e6-96231b3b80d8
* Implement rdar://7860110 (also in target/readme.txt) narrowingChris Lattner2010-04-151-0/+81
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | a load/or/and/store sequence into a narrower store when it is safe. Daniel tells me that clang will start producing this sort of thing with bitfields, and this does trigger a few dozen times on 176.gcc produced by llvm-gcc even now. This compiles code like CodeGen/X86/2009-05-28-DAGCombineCrash.ll into: movl %eax, 36(%rdi) instead of: movl $4294967295, %eax ## imm = 0xFFFFFFFF andq 32(%rdi), %rax shlq $32, %rcx addq %rax, %rcx movq %rcx, 32(%rdi) and each of the testcases into a single store. Each of them used to compile into craziness like this: _test4: movl $65535, %eax ## imm = 0xFFFF andl (%rdi), %eax shll $16, %esi addl %eax, %esi movl %esi, (%rdi) ret git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@101343 91177308-0d34-0410-b5e6-96231b3b80d8
* further tweak this to do something useful.Chris Lattner2010-04-151-6/+4
| | | | git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@101341 91177308-0d34-0410-b5e6-96231b3b80d8
* remove undef control flow.Chris Lattner2010-04-151-22/+1
| | | | git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@101340 91177308-0d34-0410-b5e6-96231b3b80d8
* Remove unneeded types from test.Jakob Stoklund Olesen2010-04-141-69/+3
| | | | git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@101286 91177308-0d34-0410-b5e6-96231b3b80d8
* Don't custom lower bit converts to ARM VMOVDRRD or VMOVDRR when the operandBob Wilson2010-04-141-0/+16
| | | | | | | | does not have a legal type. The legalizer does not know how to handle those nodes. Radar 7854640. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@101282 91177308-0d34-0410-b5e6-96231b3b80d8
* Add test for post-ra machine licm.Evan Cheng2010-04-131-7/+53
| | | | git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@101182 91177308-0d34-0410-b5e6-96231b3b80d8
* Handle a v2f64 formal parameter that is split between registers and memoryBob Wilson2010-04-131-0/+7
| | | | | | | such that the entire second half is in memory. Radar 7855014. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@101181 91177308-0d34-0410-b5e6-96231b3b80d8
* Fix test on non-x86 hosts.Evan Cheng2010-04-131-1/+1
| | | | git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@101163 91177308-0d34-0410-b5e6-96231b3b80d8
* Re-apply 101075 and fix it properly. Just reuse the debug info of the branch ↵Evan Cheng2010-04-132-0/+81
| | | | | | instruction being optimized. There is no need to --I which can deref off start of the BB. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@101162 91177308-0d34-0410-b5e6-96231b3b80d8
* Temporarily revert r101075, it's causing invalid iterator assertionsEric Christopher2010-04-131-39/+0
| | | | | | | in a nightly tester. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@101158 91177308-0d34-0410-b5e6-96231b3b80d8
* add llvm codegen support for -ffunction-sections and -fdata-sections,Chris Lattner2010-04-131-0/+10
| | | | | | | patch by Sylvere Teissier! git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@101106 91177308-0d34-0410-b5e6-96231b3b80d8
* Use .set expression for x86 pic jump table reference to reduce assembly ↵Evan Cheng2010-04-121-2/+7
| | | | | | relocation. rdar://7738756 git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@101085 91177308-0d34-0410-b5e6-96231b3b80d8
* Third time's a charm...Bill Wendling2010-04-121-1/+1
| | | | git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@101081 91177308-0d34-0410-b5e6-96231b3b80d8
* Genericize the label test.Bill Wendling2010-04-121-1/+1
| | | | git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@101079 91177308-0d34-0410-b5e6-96231b3b80d8
* Correct test to test what I mean it to test.Bill Wendling2010-04-121-1/+2
| | | | git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@101077 91177308-0d34-0410-b5e6-96231b3b80d8
* Micro-optimization:Bill Wendling2010-04-121-0/+38
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | If we have this situation: jCC L1 jmp L2 L1: ... L2: ... We can get a small performance boost by emitting this instead: jnCC L2 L1: ... L2: ... This testcase shows an example of this: float func(float x, float y) { double product = (double)x * y; if (product == 0.0) return product; return product - 1.0; } git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@101075 91177308-0d34-0410-b5e6-96231b3b80d8
* Enable post regalloc machine licm by default.Evan Cheng2010-04-121-0/+140
| | | | git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@101023 91177308-0d34-0410-b5e6-96231b3b80d8
* Make sure this test tests something.Benjamin Kramer2010-04-091-3/+7
| | | | git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@100879 91177308-0d34-0410-b5e6-96231b3b80d8
* Add a testcase for svn r100568.Bob Wilson2010-04-091-0/+23
| | | | git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@100876 91177308-0d34-0410-b5e6-96231b3b80d8
* "On SPU, variables in the .bss section that are allocated with the .lcomm ↵Chris Lattner2010-04-091-0/+2
| | | | | | | | | | | | | directive are not aligned on 16 byte boundaries. This causes misaligned loads, as the generated assembly assumes this "default" alignment. this patch disables .lcomm in favour of '.local .comm' Patch by Kalle Raisklia! git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@100875 91177308-0d34-0410-b5e6-96231b3b80d8
* Merge a few fast-isel tests.Dan Gohman2010-04-093-23/+21
| | | | git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@100860 91177308-0d34-0410-b5e6-96231b3b80d8
* Coalescer should not delete copy instructions whose defs are partially dead. ↵Evan Cheng2010-04-081-0/+26
| | | | | | | | | e.g. %RDI<def,dead> = MOV64rr %RAX<kill>, %EDI<imp-def> git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@100804 91177308-0d34-0410-b5e6-96231b3b80d8
* Avoid using f64 to lower memcpy from constant string. It's cheaper to use ↵Evan Cheng2010-04-082-12/+55
| | | | | | i32 store of immediates. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@100751 91177308-0d34-0410-b5e6-96231b3b80d8
* When expanding expressions which are using post-inc mode for multiple loops,Dan Gohman2010-04-081-0/+27
| | | | | | | ensure that the expansion is dominated by the increments of those loops. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@100748 91177308-0d34-0410-b5e6-96231b3b80d8
* add newlines at the end of files.Chris Lattner2010-04-0717-17/+17
| | | | git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@100705 91177308-0d34-0410-b5e6-96231b3b80d8
* Generalize IVUsers to track arbitrary expressions rather than expressionsDan Gohman2010-04-071-0/+277
| | | | | | | | | | | | | | | | explicitly split into stride-and-offset pairs. Also, add the ability to track multiple post-increment loops on the same expression. This refines the concept of "normalizing" SCEV expressions used for to post-increment uses, and introduces a dedicated utility routine for normalizing and denormalizing expressions. This fixes the expansion of expressions which are post-increment users of more than one loop at a time. More broadly, this takes LSR another step closer to being able to reason about more than one loop at a time. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@100699 91177308-0d34-0410-b5e6-96231b3b80d8
* Split big test into multiple directories to cater toDale Johannesen2010-04-0712-0/+397
| | | | | | | | those who don't build all targets. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@100688 91177308-0d34-0410-b5e6-96231b3b80d8
* this has a pr!Chris Lattner2010-04-071-0/+1
| | | | git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@100637 91177308-0d34-0410-b5e6-96231b3b80d8
* fix a latent bug my inline asm stuff exposed: Chris Lattner2010-04-071-0/+16
| | | | | | | MachineOperand::isIdenticalTo wasn't handling metadata operands. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@100636 91177308-0d34-0410-b5e6-96231b3b80d8
* Remove XFAIL for vg_leak as the leaks are fixed by 100601.Sanjiv Gupta2010-04-071-1/+0
| | | | git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@100612 91177308-0d34-0410-b5e6-96231b3b80d8
* Don't try to collapse DomainValues onto an incompatible SSE domain.Jakob Stoklund Olesen2010-04-061-0/+130
| | | | | | This fixes the Bullet regression on i386/nocona. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@100553 91177308-0d34-0410-b5e6-96231b3b80d8
* Add nounwind.Evan Cheng2010-04-051-2/+2
| | | | git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@100482 91177308-0d34-0410-b5e6-96231b3b80d8
* Don't do code sinking on unreachable blocks. It's unprofitable and hazardous.Dan Gohman2010-04-051-0/+30
| | | | git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@100455 91177308-0d34-0410-b5e6-96231b3b80d8
* resolve a fixme.Chris Lattner2010-04-041-2/+2
| | | | git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@100346 91177308-0d34-0410-b5e6-96231b3b80d8
* Correctly lower memset / memcpy of undef. It should be a nop. PR6767.Evan Cheng2010-04-022-5/+15
| | | | git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@100208 91177308-0d34-0410-b5e6-96231b3b80d8
* Revert the recent alignment changes. They're broken for -Os because,Dan Gohman2010-04-022-2/+2
| | | | | | | | in particular, they end up aligning strings at 16-byte boundaries, and there's no way for GlobalOpt to check OptForSize. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@100172 91177308-0d34-0410-b5e6-96231b3b80d8
* After trivial coalescing, the MI being visited may have become a copy. Avoid ↵Evan Cheng2010-04-021-0/+70
| | | | | | | | | adding it to CSE hash table since copies aren't being considered for CSE and they may be deleted. rdar://7819990 git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@100170 91177308-0d34-0410-b5e6-96231b3b80d8
* Remove this initializer so that the optimizer doesn't convertDan Gohman2010-04-021-1/+1
| | | | | | | unaligned loads into aligned loads. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@100166 91177308-0d34-0410-b5e6-96231b3b80d8
* Update this test for the new preferred alignment heuristics.Dan Gohman2010-04-021-1/+1
| | | | git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@100165 91177308-0d34-0410-b5e6-96231b3b80d8
* In 64-bit mode, use i64 to lower memcpy / memset instead of f64.Evan Cheng2010-04-012-9/+25
| | | | git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@100137 91177308-0d34-0410-b5e6-96231b3b80d8
* - Avoid using floating point stores to implement memset unless the value is ↵Evan Cheng2010-04-011-2/+10
| | | | | | | | | zero. - Do not try to infer GV alignment unless its type is sized. It's not possible to infer alignment if it has opaque type. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@100118 91177308-0d34-0410-b5e6-96231b3b80d8
* Add -mcpu to memcpy / memset tests to ensure they behave the same on all ↵Evan Cheng2010-04-013-9/+8
| | | | | | hosts / targets. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@100101 91177308-0d34-0410-b5e6-96231b3b80d8
generated by cgit v1.1 at 2025-11-10 17:58:15 +0000