benchmark/test/state_assembly_test.cc

69 lines
1.8 KiB
C++
Raw Normal View History

#include <benchmark/benchmark.h>
#ifdef __clang__
#pragma clang diagnostic ignored "-Wreturn-type"
#endif
// clang-format off
extern "C" {
extern int ExternInt;
benchmark::State& GetState();
void Fn();
}
// clang-format on
using benchmark::State;
// CHECK-LABEL: test_for_auto_loop:
extern "C" int test_for_auto_loop() {
State& S = GetState();
int x = 42;
// CHECK: [[CALL:call(q)*]] _ZN9benchmark5State16StartKeepRunningEv
// CHECK-NEXT: testq %rbx, %rbx
// CHECK-NEXT: je [[LOOP_END:.*]]
for (auto _ : S) {
// CHECK: .L[[LOOP_HEAD:[a-zA-Z0-9_]+]]:
// CHECK-GNU-NEXT: subq $1, %rbx
Iteration counts should be `uint64_t` globally. (#817) This is a shameless rip-off of https://github.com/google/benchmark/pull/646 I did promise to look into why that proposed PR was producing so much worse assembly, and so i finally did. The reason is - that diff changes `size_t` (unsigned) to `int64_t` (signed). There is this nice little `assert`: https://github.com/google/benchmark/blob/7a1c37028359ca9d386d719a6ad527743cf1b753/include/benchmark/benchmark.h#L744 It ensures that we didn't magically decide to advance our iterator when we should have finished benchmarking. When `cached_` was unsigned, the `assert` was `cached_ UGT 0`. But we only ever get to that `assert` if `cached_ NE 0`, and naturally if `cached_` is not `0`, then it is bigger than `0`, so the `assert` is tautological, and gets folded away. But now that `cached_` became signed, the assert became `cached_ SGT 0`. And we still only know that `cached_ NE 0`, so the assert can't be optimized out, or at least it doesn't currently. Regardless of whether or not that is a bug in itself, that particular diff would have regressed the normal 64-bit systems, by halving the maximal iteration space (since we go from unsigned counter to signed one, of the same bit-width), which seems like a bug. And just so it happens, fixing *this* bug, fixes the other bug. This produces fully (bit-by-bit) identical state_assembly_test.s The filecheck change is actually needed regardless of this patch, else this test does not pass for me even without this diff.
2019-05-13 17:33:11 +08:00
// CHECK-CLANG-NEXT: {{(addq \$1, %rax|incq %rax|addq \$-1, %rbx)}}
// CHECK-NEXT: jne .L[[LOOP_HEAD]]
benchmark::DoNotOptimize(x);
}
// CHECK: [[LOOP_END]]:
// CHECK: [[CALL]] _ZN9benchmark5State17FinishKeepRunningEv
// CHECK: movl $101, %eax
// CHECK: ret
return 101;
}
// CHECK-LABEL: test_while_loop:
extern "C" int test_while_loop() {
State& S = GetState();
int x = 42;
// CHECK: j{{(e|mp)}} .L[[LOOP_HEADER:[a-zA-Z0-9_]+]]
// CHECK-NEXT: .L[[LOOP_BODY:[a-zA-Z0-9_]+]]:
while (S.KeepRunning()) {
// CHECK-GNU-NEXT: subq $1, %[[IREG:[a-z]+]]
// CHECK-CLANG-NEXT: {{(addq \$-1,|decq)}} %[[IREG:[a-z]+]]
// CHECK: movq %[[IREG]], [[DEST:.*]]
benchmark::DoNotOptimize(x);
}
// CHECK-DAG: movq [[DEST]], %[[IREG]]
// CHECK-DAG: testq %[[IREG]], %[[IREG]]
// CHECK-DAG: jne .L[[LOOP_BODY]]
// CHECK-DAG: .L[[LOOP_HEADER]]:
// CHECK: cmpb $0
// CHECK-NEXT: jne .L[[LOOP_END:[a-zA-Z0-9_]+]]
// CHECK: [[CALL:call(q)*]] _ZN9benchmark5State16StartKeepRunningEv
// CHECK: .L[[LOOP_END]]:
// CHECK: [[CALL]] _ZN9benchmark5State17FinishKeepRunningEv
// CHECK: movl $101, %eax
// CHECK: ret
return 101;
}