修复了大量存在的问题，增加了假鸭子类型等等机制

2026-06-25 14:49:46 +08:00
parent 19f2787db0
commit d88d11b646
827 changed files with 32617 additions and 18316 deletions
--- a/Test/SimdTest/App/main.py
+++ b/Test/SimdTest/App/main.py
@@ -6,6 +6,7 @@ import stdlib
 from stdlib import calloc
 import t, c
 import mpool
+import testcheck


 # 全局内存池
@@ -17,28 +18,13 @@ def clock() -> t.CLong | t.CExtern: pass

 CLOCK_PER_SEC: t.CDefine = 1000

-test_passed: t.CInt = 0
-test_failed: t.CInt = 0
-
-def check(name: str, condition: t.CInt, detail: str):
-    global test_passed, test_failed
-    if condition:
-        test_passed += 1
-        printf("  [PASS] %s\n", name)
-    else:
-        test_failed += 1
-        printf("  [FAIL] %s  -- %s\n", name, detail if detail else "")
-
-def section(title: str):
-    printf("\n=== %s ===\n", title)
-

 # ============================================================
 # 1. AVX2 double4 正确性测试
 # ============================================================

 def test_avx2_add_sub_mul_div():
-    section("AVX2 add/sub/mul/div 4d")
+    testcheck.section("AVX2 add/sub/mul/div 4d")
    a: numpy.ndarray | t.CPtr = numpy.zeros(pool, 4)
    a.data[0] = t.CDouble(1.0)
    a.data[1] = t.CDouble(2.0)
@@ -55,27 +41,27 @@ def test_avx2_add_sub_mul_div():

    # add
    vipersimd.simd_add4d(a.data, b.data, out.data)
-    check("add[0]==6", out.data[0] == t.CDouble(6.0), "")
-    check("add[1]==8", out.data[1] == t.CDouble(8.0), "")
-    check("add[2]==10", out.data[2] == t.CDouble(10.0), "")
-    check("add[3]==12", out.data[3] == t.CDouble(12.0), "")
+    testcheck.check(out.data[0] == t.CDouble(6.0), "add[0]==6", "add[0]==6")
+    testcheck.check(out.data[1] == t.CDouble(8.0), "add[1]==8", "add[1]==8")
+    testcheck.check(out.data[2] == t.CDouble(10.0), "add[2]==10", "add[2]==10")
+    testcheck.check(out.data[3] == t.CDouble(12.0), "add[3]==12", "add[3]==12")

    # sub
    vipersimd.simd_sub4d(a.data, b.data, out.data)
-    check("sub[0]==-4", out.data[0] == t.CDouble(-4.0), "")
-    check("sub[3]==-4", out.data[3] == t.CDouble(-4.0), "")
+    testcheck.check(out.data[0] == t.CDouble(-4.0), "sub[0]==-4", "sub[0]==-4")
+    testcheck.check(out.data[3] == t.CDouble(-4.0), "sub[3]==-4", "sub[3]==-4")

    # mul
    vipersimd.simd_mul4d(a.data, b.data, out.data)
-    check("mul[0]==5", out.data[0] == t.CDouble(5.0), "")
-    check("mul[1]==12", out.data[1] == t.CDouble(12.0), "")
-    check("mul[2]==21", out.data[2] == t.CDouble(21.0), "")
-    check("mul[3]==32", out.data[3] == t.CDouble(32.0), "")
+    testcheck.check(out.data[0] == t.CDouble(5.0), "mul[0]==5", "mul[0]==5")
+    testcheck.check(out.data[1] == t.CDouble(12.0), "mul[1]==12", "mul[1]==12")
+    testcheck.check(out.data[2] == t.CDouble(21.0), "mul[2]==21", "mul[2]==21")
+    testcheck.check(out.data[3] == t.CDouble(32.0), "mul[3]==32", "mul[3]==32")

    # div
    vipersimd.simd_div4d(b.data, a.data, out.data)
-    check("div[0]==5", out.data[0] == t.CDouble(5.0), "")
-    check("div[1]==3", out.data[1] == t.CDouble(3.0), "")
+    testcheck.check(out.data[0] == t.CDouble(5.0), "div[0]==5", "div[0]==5")
+    testcheck.check(out.data[1] == t.CDouble(3.0), "div[1]==3", "div[1]==3")

    a.delete()
    b.delete()
@@ -83,7 +69,7 @@ def test_avx2_add_sub_mul_div():


 def test_avx2_neg_abs_sqrt():
-    section("AVX2 neg/abs/sqrt 4d")
+    testcheck.section("AVX2 neg/abs/sqrt 4d")
    a: numpy.ndarray | t.CPtr = numpy.zeros(pool, 4)
    a.data[0] = t.CDouble(1.0)
    a.data[1] = t.CDouble(-4.0)
@@ -94,31 +80,31 @@ def test_avx2_neg_abs_sqrt():

    # neg
    vipersimd.simd_neg4d(a.data, out.data)
-    check("neg[0]==-1", out.data[0] == t.CDouble(-1.0), "")
-    check("neg[1]==4", out.data[1] == t.CDouble(4.0), "")
-    check("neg[2]==-9", out.data[2] == t.CDouble(-9.0), "")
-    check("neg[3]==16", out.data[3] == t.CDouble(16.0), "")
+    testcheck.check(out.data[0] == t.CDouble(-1.0), "neg[0]==-1", "neg[0]==-1")
+    testcheck.check(out.data[1] == t.CDouble(4.0), "neg[1]==4", "neg[1]==4")
+    testcheck.check(out.data[2] == t.CDouble(-9.0), "neg[2]==-9", "neg[2]==-9")
+    testcheck.check(out.data[3] == t.CDouble(16.0), "neg[3]==16", "neg[3]==16")

    # abs
    vipersimd.simd_abs4d(a.data, out.data)
-    check("abs[0]==1", out.data[0] == t.CDouble(1.0), "")
-    check("abs[1]==4", out.data[1] == t.CDouble(4.0), "")
-    check("abs[2]==9", out.data[2] == t.CDouble(9.0), "")
-    check("abs[3]==16", out.data[3] == t.CDouble(16.0), "")
+    testcheck.check(out.data[0] == t.CDouble(1.0), "abs[0]==1", "abs[0]==1")
+    testcheck.check(out.data[1] == t.CDouble(4.0), "abs[1]==4", "abs[1]==4")
+    testcheck.check(out.data[2] == t.CDouble(9.0), "abs[2]==9", "abs[2]==9")
+    testcheck.check(out.data[3] == t.CDouble(16.0), "abs[3]==16", "abs[3]==16")

    # sqrt
    vipersimd.simd_sqrt4d(out.data, out.data)
-    check("sqrt[0]==1", out.data[0] == t.CDouble(1.0), "")
-    check("sqrt[1]==2", out.data[1] == t.CDouble(2.0), "")
-    check("sqrt[2]==3", out.data[2] == t.CDouble(3.0), "")
-    check("sqrt[3]==4", out.data[3] == t.CDouble(4.0), "")
+    testcheck.check(out.data[0] == t.CDouble(1.0), "sqrt[0]==1", "sqrt[0]==1")
+    testcheck.check(out.data[1] == t.CDouble(2.0), "sqrt[1]==2", "sqrt[1]==2")
+    testcheck.check(out.data[2] == t.CDouble(3.0), "sqrt[2]==3", "sqrt[2]==3")
+    testcheck.check(out.data[3] == t.CDouble(4.0), "sqrt[3]==4", "sqrt[3]==4")

    a.delete()
    out.delete()


 def test_avx2_min_max():
-    section("AVX2 min/max 4d")
+    testcheck.section("AVX2 min/max 4d")
    a: numpy.ndarray | t.CPtr = numpy.zeros(pool, 4)
    a.data[0] = t.CDouble(1.0)
    a.data[1] = t.CDouble(8.0)
@@ -134,16 +120,16 @@ def test_avx2_min_max():
    out: numpy.ndarray | t.CPtr = numpy.zeros(pool, 4)

    vipersimd.simd_min4d(a.data, b.data, out.data)
-    check("min[0]==1", out.data[0] == t.CDouble(1.0), "")
-    check("min[1]==2", out.data[1] == t.CDouble(2.0), "")
-    check("min[2]==3", out.data[2] == t.CDouble(3.0), "")
-    check("min[3]==4", out.data[3] == t.CDouble(4.0), "")
+    testcheck.check(out.data[0] == t.CDouble(1.0), "min[0]==1", "min[0]==1")
+    testcheck.check(out.data[1] == t.CDouble(2.0), "min[1]==2", "min[1]==2")
+    testcheck.check(out.data[2] == t.CDouble(3.0), "min[2]==3", "min[2]==3")
+    testcheck.check(out.data[3] == t.CDouble(4.0), "min[3]==4", "min[3]==4")

    vipersimd.simd_max4d(a.data, b.data, out.data)
-    check("max[0]==5", out.data[0] == t.CDouble(5.0), "")
-    check("max[1]==8", out.data[1] == t.CDouble(8.0), "")
-    check("max[2]==7", out.data[2] == t.CDouble(7.0), "")
-    check("max[3]==6", out.data[3] == t.CDouble(6.0), "")
+    testcheck.check(out.data[0] == t.CDouble(5.0), "max[0]==5", "max[0]==5")
+    testcheck.check(out.data[1] == t.CDouble(8.0), "max[1]==8", "max[1]==8")
+    testcheck.check(out.data[2] == t.CDouble(7.0), "max[2]==7", "max[2]==7")
+    testcheck.check(out.data[3] == t.CDouble(6.0), "max[3]==6", "max[3]==6")

    a.delete()
    b.delete()
@@ -151,7 +137,7 @@ def test_avx2_min_max():


 def test_avx2_scalar_ops():
-    section("AVX2 scalar broadcast ops")
+    testcheck.section("AVX2 scalar broadcast ops")
    a: numpy.ndarray | t.CPtr = numpy.zeros(pool, 4)
    a.data[0] = t.CDouble(1.0)
    a.data[1] = t.CDouble(2.0)
@@ -164,13 +150,13 @@ def test_avx2_scalar_ops():
    out: numpy.ndarray | t.CPtr = numpy.zeros(pool, 4)

    vipersimd.simd_mul_scalar4d(a.data, s.data, out.data)
-    check("mul_scalar[0]==10", out.data[0] == t.CDouble(10.0), "")
-    check("mul_scalar[1]==20", out.data[1] == t.CDouble(20.0), "")
-    check("mul_scalar[3]==40", out.data[3] == t.CDouble(40.0), "")
+    testcheck.check(out.data[0] == t.CDouble(10.0), "mul_scalar[0]==10", "mul_scalar[0]==10")
+    testcheck.check(out.data[1] == t.CDouble(20.0), "mul_scalar[1]==20", "mul_scalar[1]==20")
+    testcheck.check(out.data[3] == t.CDouble(40.0), "mul_scalar[3]==40", "mul_scalar[3]==40")

    vipersimd.simd_add_scalar4d(a.data, s.data, out.data)
-    check("add_scalar[0]==11", out.data[0] == t.CDouble(11.0), "")
-    check("add_scalar[3]==14", out.data[3] == t.CDouble(14.0), "")
+    testcheck.check(out.data[0] == t.CDouble(11.0), "add_scalar[0]==11", "add_scalar[0]==11")
+    testcheck.check(out.data[3] == t.CDouble(14.0), "add_scalar[3]==14", "add_scalar[3]==14")

    a.delete()
    s.delete()
@@ -178,7 +164,7 @@ def test_avx2_scalar_ops():


 def test_avx2_hsum_dot():
-    section("AVX2 hsum/dot 4d")
+    testcheck.section("AVX2 hsum/dot 4d")
    a: numpy.ndarray | t.CPtr = numpy.zeros(pool, 4)
    a.data[0] = t.CDouble(1.0)
    a.data[1] = t.CDouble(2.0)
@@ -194,10 +180,10 @@ def test_avx2_hsum_dot():
    out: numpy.ndarray | t.CPtr = numpy.zeros(pool, 1)

    vipersimd.simd_hsum4d(a.data, out.data)
-    check("hsum==10", out.data[0] == t.CDouble(10.0), "")
+    testcheck.check(out.data[0] == t.CDouble(10.0), "hsum==10", "hsum==10")

    vipersimd.simd_dot4d(a.data, b.data, out.data)
-    check("dot==70", out.data[0] == t.CDouble(70.0), "")
+    testcheck.check(out.data[0] == t.CDouble(70.0), "dot==70", "dot==70")

    a.delete()
    b.delete()
@@ -205,7 +191,7 @@ def test_avx2_hsum_dot():


 def test_avx2_fma():
-    section("AVX2 FMA 4d")
+    testcheck.section("AVX2 FMA 4d")
    a: numpy.ndarray | t.CPtr = numpy.zeros(pool, 4)
    a.data[0] = t.CDouble(1.0)
    a.data[1] = t.CDouble(2.0)
@@ -223,10 +209,10 @@ def test_avx2_fma():
    out: numpy.ndarray | t.CPtr = numpy.zeros(pool, 4)

    vipersimd.simd_fma4d(a.data, b.data, c_val.data, out.data)
-    check("fma[0]==6", out.data[0] == t.CDouble(6.0), "")
-    check("fma[1]==13", out.data[1] == t.CDouble(13.0), "")
-    check("fma[2]==22", out.data[2] == t.CDouble(22.0), "")
-    check("fma[3]==33", out.data[3] == t.CDouble(33.0), "")
+    testcheck.check(out.data[0] == t.CDouble(6.0), "fma[0]==6", "fma[0]==6")
+    testcheck.check(out.data[1] == t.CDouble(13.0), "fma[1]==13", "fma[1]==13")
+    testcheck.check(out.data[2] == t.CDouble(22.0), "fma[2]==22", "fma[2]==22")
+    testcheck.check(out.data[3] == t.CDouble(33.0), "fma[3]==33", "fma[3]==33")

    a.delete()
    b.delete()
@@ -239,7 +225,7 @@ def test_avx2_fma():
 # ============================================================

 def test_batch_array_ops():
-    section("Batch array operations")
+    testcheck.section("Batch array operations")
    n: t.CSizeT = 16
    a: numpy.ndarray | t.CPtr = numpy.arange(pool, t.CDouble(1.0), t.CDouble(17.0), t.CDouble(1.0))
    b: numpy.ndarray | t.CPtr = numpy.full(pool, 16, t.CDouble(2.0))
@@ -247,13 +233,13 @@ def test_batch_array_ops():

    # add array
    vipersimd.simd_add_array(a.data, b.data, out.data, n)
-    check("add_arr[0]==3", out.data[0] == t.CDouble(3.0), "")
-    check("add_arr[15]==18", out.data[15] == t.CDouble(18.0), "")
+    testcheck.check(out.data[0] == t.CDouble(3.0), "add_arr[0]==3", "add_arr[0]==3")
+    testcheck.check(out.data[15] == t.CDouble(18.0), "add_arr[15]==18", "add_arr[15]==18")

    # mul array
    vipersimd.simd_mul_array(a.data, b.data, out.data, n)
-    check("mul_arr[0]==2", out.data[0] == t.CDouble(2.0), "")
-    check("mul_arr[3]==8", out.data[3] == t.CDouble(8.0), "")
+    testcheck.check(out.data[0] == t.CDouble(2.0), "mul_arr[0]==2", "mul_arr[0]==2")
+    testcheck.check(out.data[3] == t.CDouble(8.0), "mul_arr[3]==8", "mul_arr[3]==8")

    # sqrt array
    c: numpy.ndarray | t.CPtr = numpy.zeros(pool, 4)
@@ -263,8 +249,8 @@ def test_batch_array_ops():
    c.data[3] = t.CDouble(25.0)
    out4: numpy.ndarray | t.CPtr = numpy.zeros(pool, 4)
    vipersimd.simd_sqrt_array(c.data, out4.data, t.CSizeT(4))
-    check("sqrt_arr[0]==2", out4.data[0] == t.CDouble(2.0), "")
-    check("sqrt_arr[3]==5", out4.data[3] == t.CDouble(5.0), "")
+    testcheck.check(out4.data[0] == t.CDouble(2.0), "sqrt_arr[0]==2", "sqrt_arr[0]==2")
+    testcheck.check(out4.data[3] == t.CDouble(5.0), "sqrt_arr[3]==5", "sqrt_arr[3]==5")

    # abs array
    d: numpy.ndarray | t.CPtr = numpy.zeros(pool, 4)
@@ -273,9 +259,9 @@ def test_batch_array_ops():
    d.data[2] = t.CDouble(3.0)
    d.data[3] = t.CDouble(-4.0)
    vipersimd.simd_abs_array(d.data, out4.data, t.CSizeT(4))
-    check("abs_arr[0]==1", out4.data[0] == t.CDouble(1.0), "")
-    check("abs_arr[1]==2", out4.data[1] == t.CDouble(2.0), "")
-    check("abs_arr[3]==4", out4.data[3] == t.CDouble(4.0), "")
+    testcheck.check(out4.data[0] == t.CDouble(1.0), "abs_arr[0]==1", "abs_arr[0]==1")
+    testcheck.check(out4.data[1] == t.CDouble(2.0), "abs_arr[1]==2", "abs_arr[1]==2")
+    testcheck.check(out4.data[3] == t.CDouble(4.0), "abs_arr[3]==4", "abs_arr[3]==4")

    # dot array
    e: numpy.ndarray | t.CPtr = numpy.zeros(pool, 4)
@@ -286,7 +272,7 @@ def test_batch_array_ops():
    f: numpy.ndarray | t.CPtr = numpy.ones(pool, 4)
    out1: numpy.ndarray | t.CPtr = numpy.zeros(pool, 1)
    vipersimd.simd_dot_array(e.data, f.data, t.CSizeT(4), out1.data)
-    check("dot_arr==10", out1.data[0] == t.CDouble(10.0), "")
+    testcheck.check(out1.data[0] == t.CDouble(10.0), "dot_arr==10", "dot_arr==10")

    a.delete()
    b.delete()
@@ -304,71 +290,71 @@ def test_batch_array_ops():
 # ============================================================

 def test_llvmir_scalar():
-    section("LLVMIR precise scalar ops")
+    testcheck.section("LLVMIR precise scalar ops")

    # sqrt
    r: t.CDouble = vipersimd.simd_sqrt(t.CDouble(4.0))
-    check("sqrt(4)==2", r == t.CDouble(2.0), "")
+    testcheck.check(r == t.CDouble(2.0), "sqrt(4)==2", "sqrt(4)==2")

    # fabs
    r = vipersimd.simd_fabs(t.CDouble(-3.5))
-    check("fabs(-3.5)==3.5", r == t.CDouble(3.5), "")
+    testcheck.check(r == t.CDouble(3.5), "fabs(-3.5)==3.5", "fabs(-3.5)==3.5")

    # floor
    r = vipersimd.simd_floor(t.CDouble(3.7))
-    check("floor(3.7)==3", r == t.CDouble(3.0), "")
+    testcheck.check(r == t.CDouble(3.0), "floor(3.7)==3", "floor(3.7)==3")

    # ceil
    r = vipersimd.simd_ceil(t.CDouble(3.2))
-    check("ceil(3.2)==4", r == t.CDouble(4.0), "")
+    testcheck.check(r == t.CDouble(4.0), "ceil(3.2)==4", "ceil(3.2)==4")

    # round
    r = vipersimd.simd_round(t.CDouble(3.5))
-    check("round(3.5)==4", r == t.CDouble(4.0), "")
+    testcheck.check(r == t.CDouble(4.0), "round(3.5)==4", "round(3.5)==4")

    # trunc
    r = vipersimd.simd_trunc(t.CDouble(3.9))
-    check("trunc(3.9)==3", r == t.CDouble(3.0), "")
+    testcheck.check(r == t.CDouble(3.0), "trunc(3.9)==3", "trunc(3.9)==3")

    # fma: 2*3+1 = 7
    r = vipersimd.simd_fma(t.CDouble(2.0), t.CDouble(3.0), t.CDouble(1.0))
-    check("fma(2,3,1)==7", r == t.CDouble(7.0), "")
+    testcheck.check(r == t.CDouble(7.0), "fma(2,3,1)==7", "fma(2,3,1)==7")

    # copysign
    r = vipersimd.simd_copysign(t.CDouble(3.0), t.CDouble(-1.0))
-    check("copysign(3,-1)==-3", r == t.CDouble(-3.0), "")
+    testcheck.check(r == t.CDouble(-3.0), "copysign(3,-1)==-3", "copysign(3,-1)==-3")

    # neg
    r = vipersimd.simd_neg(t.CDouble(1.5))
-    check("neg(1.5)==-1.5", r == t.CDouble(-1.5), "")
+    testcheck.check(r == t.CDouble(-1.5), "neg(1.5)==-1.5", "neg(1.5)==-1.5")

    # minnum
    r = vipersimd.simd_minnum(t.CDouble(2.0), t.CDouble(3.0))
-    check("minnum(2,3)==2", r == t.CDouble(2.0), "")
+    testcheck.check(r == t.CDouble(2.0), "minnum(2,3)==2", "minnum(2,3)==2")

    # maxnum
    r = vipersimd.simd_maxnum(t.CDouble(2.0), t.CDouble(3.0))
-    check("maxnum(2,3)==3", r == t.CDouble(3.0), "")
+    testcheck.check(r == t.CDouble(3.0), "maxnum(2,3)==3", "maxnum(2,3)==3")

    # exp
    r = vipersimd.simd_exp(t.CDouble(1.0))
-    check("exp(1)~=2.718", r > t.CDouble(2.71) and r < t.CDouble(2.72), "")
+    testcheck.check(r > t.CDouble(2.71) and r < t.CDouble(2.72), "exp(1)~=2.718", "exp(1)~=2.718")

    # log
    r = vipersimd.simd_log(t.CDouble(2.718281828459045))
-    check("log(e)~=1", r > t.CDouble(0.999) and r < t.CDouble(1.001), "")
+    testcheck.check(r > t.CDouble(0.999) and r < t.CDouble(1.001), "log(e)~=1", "log(e)~=1")

    # sin
    r = vipersimd.simd_sin(t.CDouble(0.0))
-    check("sin(0)~=0", r > t.CDouble(-0.001) and r < t.CDouble(0.001), "")
+    testcheck.check(r > t.CDouble(-0.001) and r < t.CDouble(0.001), "sin(0)~=0", "sin(0)~=0")

    # cos
    r = vipersimd.simd_cos(t.CDouble(0.0))
-    check("cos(0)~=1", r > t.CDouble(0.999) and r < t.CDouble(1.001), "")
+    testcheck.check(r > t.CDouble(0.999) and r < t.CDouble(1.001), "cos(0)~=1", "cos(0)~=1")

    # pow
    r = vipersimd.simd_pow(t.CDouble(2.0), t.CDouble(10.0))
-    check("pow(2,10)==1024", r == t.CDouble(1024.0), "")
+    testcheck.check(r == t.CDouble(1024.0), "pow(2,10)==1024", "pow(2,10)==1024")


 # ============================================================
@@ -405,7 +391,7 @@ def scalar_sqrt_array(a: t.CPtr, out: t.CPtr, n: t.CSizeT):
        i += 1

 def benchmark_simd_vs_scalar():
-    section("Benchmark: SIMD vs Scalar")
+    testcheck.section("Benchmark: SIMD vs Scalar")

    a: numpy.ndarray | t.CPtr = numpy.full(pool, N_ELEM, t.CDouble(1.5))
    b: numpy.ndarray | t.CPtr = numpy.full(pool, N_ELEM, t.CDouble(2.5))
@@ -484,7 +470,7 @@ def benchmark_simd_vs_scalar():
        printf("  speedup=N/A\n")

    # 验证结果正确性
-    check("bench result[0] valid", out.data[0] > t.CDouble(0.0), "")
+    testcheck.check(out.data[0] > t.CDouble(0.0), "bench result[0] valid", "bench result[0] valid")

    a.delete()
    b.delete()
@@ -501,7 +487,7 @@ def main() -> t.CInt:
    arena = calloc(4096 * 1024, 1)  # 4MB arena
    pool = mpool.MPool(arena, 4096 * 1024, 0)  # bump allocator

-    printf("=== ViperSIMD Test Suite ===\n")
+    testcheck.begin("ViperSIMD Test Suite")

    test_avx2_add_sub_mul_div()
    test_avx2_neg_abs_sqrt()
@@ -513,5 +499,4 @@ def main() -> t.CInt:
    test_llvmir_scalar()
    benchmark_simd_vs_scalar()

-    printf("\n=== Results: %d passed, %d failed ===\n", test_passed, test_failed)
-    return 0
+    return testcheck.end()
--- a/Test/SimdTest/output/0991af0d22263577.deps.json
+++ b/Test/SimdTest/output/0991af0d22263577.deps.json
@@ -0,0 +1 @@
+{"mpool": "0991af0d22263577", "numpy.__init__": "1787ce01c2ed0493", "__init__": "1787ce01c2ed0493", "numpy": "1787ce01c2ed0493", "vipermath": "3f7c5e78d8652535", "stdint": "49bd8cba55979f76", "stdlib": "6c2029b306556c00", "stdio": "73edbcf76e32d00b", "testcheck": "9dbecd0942a39782", "main": "b44ac2a895a4072d", "string": "b5f1dc665c52a1bd", "viperio": "c9f4be41ca1cc2b4", "vipersimd": "fe99c83946298690"}
--- a/Test/SimdTest/output/1787ce01c2ed0493.deps.json
+++ b/Test/SimdTest/output/1787ce01c2ed0493.deps.json
@@ -0,0 +1 @@
+{"mpool": "0991af0d22263577", "numpy.__init__": "1787ce01c2ed0493", "__init__": "1787ce01c2ed0493", "numpy": "1787ce01c2ed0493", "vipermath": "3f7c5e78d8652535", "stdint": "49bd8cba55979f76", "stdlib": "6c2029b306556c00", "stdio": "73edbcf76e32d00b", "testcheck": "9dbecd0942a39782", "main": "b44ac2a895a4072d", "string": "b5f1dc665c52a1bd", "viperio": "c9f4be41ca1cc2b4", "vipersimd": "fe99c83946298690"}
--- a/Test/SimdTest/output/1b666d438f4d301e.deps.json
+++ b/Test/SimdTest/output/1b666d438f4d301e.deps.json
@@ -1 +0,0 @@
-{"vipermath": "3f7c5e78d8652535", "stdint": "56cdd754a8a09347", "mpool": "68c4fe4b12c908e3", "string": "6aee24fdefa3cbc0", "stdlib": "6c2029b306556c00", "stdio": "73edbcf76e32d00b", "numpy.__init__": "c3eb91093118e1e1", "__init__": "c3eb91093118e1e1", "numpy": "c3eb91093118e1e1", "viperio": "c9f4be41ca1cc2b4", "vipersimd": "fe99c83946298690"}
--- a/Test/SimdTest/output/3f7c5e78d8652535.deps.json
+++ b/Test/SimdTest/output/3f7c5e78d8652535.deps.json
@@ -1 +1 @@
-{"main": "1b666d438f4d301e", "vipermath": "3f7c5e78d8652535", "stdint": "56cdd754a8a09347", "mpool": "68c4fe4b12c908e3", "string": "6aee24fdefa3cbc0", "stdlib": "6c2029b306556c00", "stdio": "73edbcf76e32d00b", "numpy.__init__": "c3eb91093118e1e1", "__init__": "c3eb91093118e1e1", "numpy": "c3eb91093118e1e1", "viperio": "c9f4be41ca1cc2b4", "vipersimd": "fe99c83946298690"}
+{"mpool": "0991af0d22263577", "numpy.__init__": "1787ce01c2ed0493", "__init__": "1787ce01c2ed0493", "numpy": "1787ce01c2ed0493", "vipermath": "3f7c5e78d8652535", "stdint": "49bd8cba55979f76", "stdlib": "6c2029b306556c00", "stdio": "73edbcf76e32d00b", "testcheck": "9dbecd0942a39782", "main": "b44ac2a895a4072d", "string": "b5f1dc665c52a1bd", "viperio": "c9f4be41ca1cc2b4", "vipersimd": "fe99c83946298690"}
--- a/Test/SimdTest/output/68c4fe4b12c908e3.deps.json
+++ b/Test/SimdTest/output/68c4fe4b12c908e3.deps.json
@@ -1 +0,0 @@
-{"main": "1b666d438f4d301e", "vipermath": "3f7c5e78d8652535", "stdint": "56cdd754a8a09347", "mpool": "68c4fe4b12c908e3", "string": "6aee24fdefa3cbc0", "stdlib": "6c2029b306556c00", "stdio": "73edbcf76e32d00b", "numpy.__init__": "c3eb91093118e1e1", "__init__": "c3eb91093118e1e1", "numpy": "c3eb91093118e1e1", "viperio": "c9f4be41ca1cc2b4", "vipersimd": "fe99c83946298690"}
--- a/Test/SimdTest/output/6aee24fdefa3cbc0.deps.json
+++ b/Test/SimdTest/output/6aee24fdefa3cbc0.deps.json
@@ -1 +0,0 @@
-{"main": "1b666d438f4d301e", "vipermath": "3f7c5e78d8652535", "stdint": "56cdd754a8a09347", "mpool": "68c4fe4b12c908e3", "string": "6aee24fdefa3cbc0", "stdlib": "6c2029b306556c00", "stdio": "73edbcf76e32d00b", "numpy.__init__": "c3eb91093118e1e1", "__init__": "c3eb91093118e1e1", "numpy": "c3eb91093118e1e1", "viperio": "c9f4be41ca1cc2b4", "vipersimd": "fe99c83946298690"}
--- a/Test/SimdTest/output/9dbecd0942a39782.deps.json
+++ b/Test/SimdTest/output/9dbecd0942a39782.deps.json
@@ -0,0 +1 @@
+{"mpool": "0991af0d22263577", "numpy.__init__": "1787ce01c2ed0493", "__init__": "1787ce01c2ed0493", "numpy": "1787ce01c2ed0493", "vipermath": "3f7c5e78d8652535", "stdint": "49bd8cba55979f76", "stdlib": "6c2029b306556c00", "stdio": "73edbcf76e32d00b", "testcheck": "9dbecd0942a39782", "main": "b44ac2a895a4072d", "string": "b5f1dc665c52a1bd", "viperio": "c9f4be41ca1cc2b4", "vipersimd": "fe99c83946298690"}
--- a/Test/SimdTest/output/b44ac2a895a4072d.deps.json
+++ b/Test/SimdTest/output/b44ac2a895a4072d.deps.json
@@ -0,0 +1 @@
+{"mpool": "0991af0d22263577", "numpy.__init__": "1787ce01c2ed0493", "__init__": "1787ce01c2ed0493", "numpy": "1787ce01c2ed0493", "vipermath": "3f7c5e78d8652535", "stdint": "49bd8cba55979f76", "stdlib": "6c2029b306556c00", "stdio": "73edbcf76e32d00b", "testcheck": "9dbecd0942a39782", "string": "b5f1dc665c52a1bd", "viperio": "c9f4be41ca1cc2b4", "vipersimd": "fe99c83946298690"}
--- a/Test/SimdTest/output/b5f1dc665c52a1bd.deps.json
+++ b/Test/SimdTest/output/b5f1dc665c52a1bd.deps.json
@@ -0,0 +1 @@
+{"mpool": "0991af0d22263577", "numpy.__init__": "1787ce01c2ed0493", "__init__": "1787ce01c2ed0493", "numpy": "1787ce01c2ed0493", "vipermath": "3f7c5e78d8652535", "stdint": "49bd8cba55979f76", "stdlib": "6c2029b306556c00", "stdio": "73edbcf76e32d00b", "testcheck": "9dbecd0942a39782", "main": "b44ac2a895a4072d", "string": "b5f1dc665c52a1bd", "viperio": "c9f4be41ca1cc2b4", "vipersimd": "fe99c83946298690"}
--- a/Test/SimdTest/output/c3eb91093118e1e1.deps.json
+++ b/Test/SimdTest/output/c3eb91093118e1e1.deps.json
@@ -1 +0,0 @@
-{"main": "1b666d438f4d301e", "vipermath": "3f7c5e78d8652535", "stdint": "56cdd754a8a09347", "mpool": "68c4fe4b12c908e3", "string": "6aee24fdefa3cbc0", "stdlib": "6c2029b306556c00", "stdio": "73edbcf76e32d00b", "numpy.__init__": "c3eb91093118e1e1", "__init__": "c3eb91093118e1e1", "numpy": "c3eb91093118e1e1", "viperio": "c9f4be41ca1cc2b4", "vipersimd": "fe99c83946298690"}
--- a/Test/SimdTest/output/c9f4be41ca1cc2b4.deps.json
+++ b/Test/SimdTest/output/c9f4be41ca1cc2b4.deps.json
@@ -1 +1 @@
-{"main": "1b666d438f4d301e", "vipermath": "3f7c5e78d8652535", "stdint": "56cdd754a8a09347", "mpool": "68c4fe4b12c908e3", "string": "6aee24fdefa3cbc0", "stdlib": "6c2029b306556c00", "stdio": "73edbcf76e32d00b", "numpy.__init__": "c3eb91093118e1e1", "__init__": "c3eb91093118e1e1", "numpy": "c3eb91093118e1e1", "viperio": "c9f4be41ca1cc2b4", "vipersimd": "fe99c83946298690"}
+{"mpool": "0991af0d22263577", "numpy.__init__": "1787ce01c2ed0493", "__init__": "1787ce01c2ed0493", "numpy": "1787ce01c2ed0493", "vipermath": "3f7c5e78d8652535", "stdint": "49bd8cba55979f76", "stdlib": "6c2029b306556c00", "stdio": "73edbcf76e32d00b", "testcheck": "9dbecd0942a39782", "main": "b44ac2a895a4072d", "string": "b5f1dc665c52a1bd", "viperio": "c9f4be41ca1cc2b4", "vipersimd": "fe99c83946298690"}
--- a/Test/SimdTest/output/fe99c83946298690.deps.json
+++ b/Test/SimdTest/output/fe99c83946298690.deps.json
@@ -1 +1 @@
-{"main": "1b666d438f4d301e", "vipermath": "3f7c5e78d8652535", "stdint": "56cdd754a8a09347", "mpool": "68c4fe4b12c908e3", "string": "6aee24fdefa3cbc0", "stdlib": "6c2029b306556c00", "stdio": "73edbcf76e32d00b", "numpy.__init__": "c3eb91093118e1e1", "__init__": "c3eb91093118e1e1", "numpy": "c3eb91093118e1e1", "viperio": "c9f4be41ca1cc2b4", "vipersimd": "fe99c83946298690"}
+{"mpool": "0991af0d22263577", "numpy.__init__": "1787ce01c2ed0493", "__init__": "1787ce01c2ed0493", "numpy": "1787ce01c2ed0493", "vipermath": "3f7c5e78d8652535", "stdint": "49bd8cba55979f76", "stdlib": "6c2029b306556c00", "stdio": "73edbcf76e32d00b", "testcheck": "9dbecd0942a39782", "main": "b44ac2a895a4072d", "string": "b5f1dc665c52a1bd", "viperio": "c9f4be41ca1cc2b4", "vipersimd": "fe99c83946298690"}
--- a/Test/SimdTest/temp/0991af0d22263577.pyi
+++ b/Test/SimdTest/temp/0991af0d22263577.pyi
@@ -11,7 +11,7 @@ import string

 MPOOL_ALIGN: t.CDefine = 8
 MPOOL_TYPE_SLAB: t.CDefine = 0
-MPOOL_TYPE_BUMP: t.CDefine = 2
+MPOOL_TYPE_BUMP: t.CDefine = 1

 def _align_up(val: t.CSizeT, align: t.CSizeT) -> t.CSizeT: pass

--- a/Test/SimdTest/temp/1787ce01c2ed0493.pyi
+++ b/Test/SimdTest/temp/1787ce01c2ed0493.pyi
@@ -24,8 +24,8 @@ e:  t.CDefine = 2.71828182845904523536
@t.Object
 class ndarray:
    data: t.CDouble | t.CPtr
-    shape: list[t.CSizeT, MAX_NDIM]
-    strides: list[t.CSizeT, MAX_NDIM]
+    shape: t.CArray[t.CSizeT, MAX_NDIM]
+    strides: t.CArray[t.CSizeT, MAX_NDIM]
    ndim: t.CInt
    size: t.CSizeT
    owns_data: t.CInt
--- a/Test/SimdTest/temp/49bd8cba55979f76.pyi
+++ b/Test/SimdTest/temp/49bd8cba55979f76.pyi
@@ -64,4 +64,12 @@ CHAR16: t.CTypedef = t.CChar16T
 CHAR32: t.CTypedef = t.CChar32T
 CHAR8PTR: t.CTypedef = t.CChar8T | t.CPtr
 CHAR16PTR: t.CTypedef = t.CChar16T | t.CPtr
-CHAR32PTR: t.CTypedef = t.CChar32T | t.CPtr
+CHAR32PTR: t.CTypedef = t.CChar32T | t.CPtr
+i8: t.CTypedef = t.CInt8T
+i16: t.CTypedef = t.CInt16T
+i32: t.CTypedef = t.CInt32T
+i64: t.CTypedef = t.CInt64T
+u8: t.CTypedef = t.CUInt8T
+u16: t.CTypedef = t.CUInt16T
+u32: t.CTypedef = t.CUInt32T
+u64: t.CTypedef = t.CUInt64T
--- a/Test/SimdTest/temp/9dbecd0942a39782.pyi
+++ b/Test/SimdTest/temp/9dbecd0942a39782.pyi
@@ -0,0 +1,25 @@
+"""
+Auto-generated Python stub file from testcheck.py
+Module: testcheck
+"""
+
+
+import t, c
+import stdio
+
+_pass_count: t.CExtern | t.CInt
+_fail_count: t.CExtern | t.CInt
+
+def begin(name: str) -> t.CInt: pass
+
+def section(name: str) -> t.CInt: pass
+
+def ok(msg: str) -> t.CInt: pass
+
+def fail(msg: str) -> t.CInt: pass
+
+def check(cond: t.CInt, ok_msg: str, fail_msg: str) -> t.CInt: pass
+
+def info(msg: str) -> t.CInt: pass
+
+def end() -> t.CInt: pass
--- a/Test/SimdTest/temp/_sha1_map.txt
+++ b/Test/SimdTest/temp/_sha1_map.txt
@@ -1,10 +1,11 @@
-1b666d438f4d301e:main.py
+0991af0d22263577:includes/mpool.py
+1787ce01c2ed0493:includes/numpy\__init__.py
 3f7c5e78d8652535:includes/vipermath.py
-56cdd754a8a09347:includes/stdint.py
-68c4fe4b12c908e3:includes/mpool.py
-6aee24fdefa3cbc0:includes/string.py
+49bd8cba55979f76:includes/stdint.py
 6c2029b306556c00:includes/stdlib.py
 73edbcf76e32d00b:includes/stdio.py
-c3eb91093118e1e1:includes/numpy\__init__.py
+9dbecd0942a39782:includes/testcheck.py
+b44ac2a895a4072d:main.py
+b5f1dc665c52a1bd:includes/string.py
 c9f4be41ca1cc2b4:includes/viperio.py
 fe99c83946298690:includes/vipersimd.py
--- a/Test/SimdTest/temp/b44ac2a895a4072d.pyi
+++ b/Test/SimdTest/temp/b44ac2a895a4072d.pyi
@@ -12,6 +12,7 @@ import stdlib
 from stdlib import calloc
 import t, c
 import mpool
+import testcheck

 arena: t.CExtern | t.CUInt8T | t.CPtr
 pool: t.CExtern | mpool.MPool | t.CPtr
@@ -20,12 +21,6 @@ def clock() -> t.CLong | t.CExtern: pass


 CLOCK_PER_SEC: t.CDefine = 1000
-test_passed: t.CExtern | t.CInt
-test_failed: t.CExtern | t.CInt
-
-def check(name: str, condition: t.CInt, detail: str) -> t.CInt: pass
-
-def section(title: str) -> t.CInt: pass

 def test_avx2_add_sub_mul_div() -> t.CInt: pass

--- a/Test/SimdTest/temp/b5f1dc665c52a1bd.pyi
+++ b/Test/SimdTest/temp/b5f1dc665c52a1bd.pyi
@@ -41,4 +41,4 @@ def atoll(src: str) -> t.CInt64T: pass

 def atof(src: str) -> t.CDouble: pass

-def split(s: str, delim: str, result: list[str]) -> int: pass
+def split(s: str, delim: str, result: t.CArray[str]) -> int: pass
				`@@ -0,0 +1 @@`
				`{"mpool": "0991af0d22263577", "numpy.__init__": "1787ce01c2ed0493", "__init__": "1787ce01c2ed0493", "numpy": "1787ce01c2ed0493", "vipermath": "3f7c5e78d8652535", "stdint": "49bd8cba55979f76", "stdlib": "6c2029b306556c00", "stdio": "73edbcf76e32d00b", "testcheck": "9dbecd0942a39782", "main": "b44ac2a895a4072d", "string": "b5f1dc665c52a1bd", "viperio": "c9f4be41ca1cc2b4", "vipersimd": "fe99c83946298690"}`
				`@@ -1 +0,0 @@`
				`{"vipermath": "3f7c5e78d8652535", "stdint": "56cdd754a8a09347", "mpool": "68c4fe4b12c908e3", "string": "6aee24fdefa3cbc0", "stdlib": "6c2029b306556c00", "stdio": "73edbcf76e32d00b", "numpy.__init__": "c3eb91093118e1e1", "__init__": "c3eb91093118e1e1", "numpy": "c3eb91093118e1e1", "viperio": "c9f4be41ca1cc2b4", "vipersimd": "fe99c83946298690"}`
				`@@ -1 +0,0 @@`
				`{"main": "1b666d438f4d301e", "vipermath": "3f7c5e78d8652535", "stdint": "56cdd754a8a09347", "mpool": "68c4fe4b12c908e3", "string": "6aee24fdefa3cbc0", "stdlib": "6c2029b306556c00", "stdio": "73edbcf76e32d00b", "numpy.__init__": "c3eb91093118e1e1", "__init__": "c3eb91093118e1e1", "numpy": "c3eb91093118e1e1", "viperio": "c9f4be41ca1cc2b4", "vipersimd": "fe99c83946298690"}`