This PR sets up the new integrated test/bench suite. It then migrates all benchmarks and some related tests to the new suite. There's also some documentation and some linting. For now, a lot of the old tests are left alone so this PR doesn't become even larger than it already is. Eventually, all tests should be migrated to the new suite though so there isn't a confusing mix of two systems.
38 lines
1 KiB
Text
38 lines
1 KiB
Text
import Std.Tactic.BVDecide
|
|
|
|
/-
|
|
Verify: https://en.wikipedia.org/wiki/Lehmer_random_number_generator#Sample_C99_code
|
|
|
|
uint32_t lcg_parkmiller(uint32_t *state)
|
|
{
|
|
return *state = (uint64_t)*state * 48271 % 0x7fffffff;
|
|
}
|
|
|
|
vs
|
|
|
|
uint32_t lcg_parkmiller(uint32_t *state)
|
|
{
|
|
uint64_t product = (uint64_t)*state * 48271;
|
|
uint32_t x = (product & 0x7fffffff) + (product >> 31);
|
|
|
|
x = (x & 0x7fffffff) + (x >> 31);
|
|
return *state = x;
|
|
}
|
|
-/
|
|
|
|
def naive (state : BitVec 32) : BitVec 32 :=
|
|
(((state.zeroExtend 64) * 48271) % 0x7fffffff).extractLsb 31 0
|
|
|
|
def opt (state : BitVec 32) : BitVec 32 :=
|
|
let product := (state.zeroExtend 64) * 48271
|
|
let x := ((product &&& 0x7fffffff) + (product >>> 31)).extractLsb 31 0
|
|
let x := (x &&& 0x7fffffff) + (x >>> 31)
|
|
x
|
|
|
|
--set_option trace.Meta.Tactic.bv true in
|
|
--set_option trace.Meta.Tactic.sat true in
|
|
--set_option trace.profiler true in
|
|
theorem lehmer_correct (state : BitVec 32) (h1 : state > 0) (h2 : state < 0x7fffffff) :
|
|
naive state = opt state := by
|
|
unfold naive opt
|
|
bv_decide (config := { timeout := 120 })
|