Update test vectors

test/data_type/test_mx_fp4.cpp

@@ -326,18 +326,21 @@ TEST(MXFP4, DeviceScaledConvert)
 
 __host__ __device__ float vec16_generator(ck::index_t i)
 {
-    return (i < 8 ? -1.0 : 1.0) * powf(2.0f, i % 8);
+    return type_convert<float>(f4_t(i & 0b00001111));
 }
 
 __host__ __device__ float vec32_generator(ck::index_t i)
 {
     if(i < 16)
     {
-        return vec16_generator(i % 16);
+        return vec16_generator(
+            i); // all positive values, then all negative values in ascending order
     }
     else
     {
-        return 1.5f * vec16_generator(i % 16);
+        return type_convert<float>(ck::NumericLimits<f4_t>::Max()) -
+               vec16_generator(
+                   i); // all positive values, then all negative values in descending order
     }
 }
 
@@ -393,10 +396,12 @@ TEST(MXFP4, DeviceF32x32ToF4x32ScaledConvert)
     device_completed.FromDevice(&completed);
     device_out.FromDevice(out.data());
 
-    auto i = 0;
+    auto i      = 0;
+    auto scale2 = e8m0_bexp_t(2.0f);
 
     ck::static_for<0, N, 1>{}([&](auto ii) {
-        EXPECT_EQ(out[i++], vec32_generator(ii) / 2.0f) << "ii: " << ii << std::endl;
+        EXPECT_EQ(out[i++], vec32_generator(ii) / type_convert<float>(scale2))
+            << "ii: " << ii << std::endl;
     });
 
     EXPECT_EQ(N, completed);
@@ -455,10 +460,12 @@ TEST(MXFP4, DeviceF32x32ToF4x32ScaledConvertSR)
     device_completed.FromDevice(&completed);
     device_out.FromDevice(out.data());
 
-    auto i = 0;
+    auto i      = 0;
+    auto scale2 = e8m0_bexp_t(2.0f);
 
     ck::static_for<0, N, 1>{}([&](auto ii) {
-        EXPECT_EQ(out[i++], vec32_generator(ii) / 8.0f) << "ii: " << ii << std::endl;
+        EXPECT_EQ(out[i++], vec32_generator(ii) / type_convert<float>(scale2))
+            << "ii: " << ii << std::endl;
     });
 
     EXPECT_EQ(N, completed);
@@ -481,14 +488,14 @@ __global__ void test_mx_f32x32_device_scaled_convert(float* p_test, uint64_t* p_
         return;
     }
 
-    auto scale2 = e8m0_bexp_t(4.0f);
+    auto scale2 = e8m0_bexp_t(2.0f);
 
     f4x32_t f4x32{};
     float32_t float32{};
     ck::static_for<0, N / 2, 1>{}([&](auto ii) {
         f4x32.AsType<f4x2_pk_t>()(ck::Number<ii>{})
-            .pack(type_convert<f4_t>(vec32_generator(2 * ii) / 16.0f),
-                  type_convert<f4_t>(vec32_generator(2 * ii + 1) / 16.0f));
+            .pack(type_convert<f4_t>(vec32_generator(2 * ii) / type_convert<float>(scale2)),
+                  type_convert<f4_t>(vec32_generator(2 * ii + 1) / type_convert<float>(scale2)));
     });
 
     float32 = scaled_type_convert<float32_t>(scale2, f4x32);
@@ -515,10 +522,12 @@ TEST(MXFP4, DeviceF4x32ToF32x32ScaledConvert)
     device_completed.FromDevice(&completed);
     device_out.FromDevice(out.data());
 
-    auto i = 0;
+    auto i      = 0;
+    auto scale2 = e8m0_bexp_t(2.0f);
 
     ck::static_for<0, N, 1>{}([&](auto ii) {
-        EXPECT_EQ(out[i++], vec32_generator(ii) / 4.0f) << "ii: " << ii << std::endl;
+        EXPECT_EQ(out[i++], vec32_generator(ii) / type_convert<float>(scale2))
+            << "ii: " << ii << std::endl;
     });
 
     EXPECT_EQ(N, completed);

test/data_type/test_mx_fp4_repro.cpp

@@ -85,15 +85,9 @@ TEST(MXFP4, FP4ToFP32)
     std::vector<float> out(2, -1.0f);
 
     DeviceMem device_out(2 * sizeof(float));
-    // DeviceMem device_completed(sizeof(uint64_t));
-
-    // device_out.SetValue(-21.0f);
-    // device_completed.SetValue(-21.0f);
 
     run_test_mx_fp4_to_fp32<<<1, 1>>>(static_cast<float*>(device_out.GetDeviceBuffer()));
 
-    // uint64_t completed = 0;
-    // device_completed.FromDevice(&completed);
     device_out.FromDevice(out.data());
 
     // f4x2 -> f32x2
@@ -106,12 +100,9 @@ TEST(MXFP4, FP32ToFP4RNE)
     std::vector<float> out(2, -1.0f);
 
     DeviceMem device_out(2 * sizeof(float));
-    // DeviceMem device_completed(sizeof(uint64_t));
 
     run_test_mx_fp32_to_fp4_rne<<<1, 1>>>(static_cast<float*>(device_out.GetDeviceBuffer()));
 
-    // uint64_t completed = 0;
-    // device_completed.FromDevice(&completed);
     device_out.FromDevice(out.data());
 
     // f32x2 -> f4x2
@@ -125,12 +116,9 @@ TEST(MXFP4, FP32ToFP4SR)
     std::vector<float> out(2, -1.0f);
 
     DeviceMem device_out(2 * sizeof(float));
-    // DeviceMem device_completed(sizeof(uint64_t));
 
     run_test_mx_fp32_to_fp4_sr<<<1, 1>>>(static_cast<float*>(device_out.GetDeviceBuffer()));
 
-    // uint64_t completed = 0;
-    // device_completed.FromDevice(&completed);
     device_out.FromDevice(out.data());
 
     // SR
@@ -143,12 +131,9 @@ TEST(MXFP4, FP32ToFP4SRFailing)
     std::vector<float> out(2, -1.0f);
 
     DeviceMem device_out(2 * sizeof(float));
-    // DeviceMem device_completed(sizeof(uint64_t));
 
     run_test_mx_fp32_to_fp4_sr_failing<<<1, 1>>>(static_cast<float*>(device_out.GetDeviceBuffer()));
 
-    // uint64_t completed = 0;
-    // device_completed.FromDevice(&completed);
     device_out.FromDevice(out.data());
 
     // SR