@@ -562,7 +562,7 @@ void testHle2(float *const Result, __half2 Input1, __half2 Input2) {
562
562
563
563
void testHle2Cases (const vector<pair<half2_pair, h2i_pair>> &TestCases) {
564
564
float *Result;
565
- cudaMallocManaged (&Result, sizeof (*Result));
565
+ cudaMallocManaged (&Result, 2 * sizeof (*Result));
566
566
for (const auto &TestCase : TestCases) {
567
567
testHle2 (Result, TestCase.first .first , TestCase.first .second );
568
568
checkResult (" __hle2" , {TestCase.first .first , TestCase.first .second },
@@ -584,7 +584,7 @@ void testHleu2(float *const Result, __half2 Input1, __half2 Input2) {
584
584
585
585
void testHleu2Cases (const vector<pair<half2_pair, h2i_pair>> &TestCases) {
586
586
float *Result;
587
- cudaMallocManaged (&Result, sizeof (*Result));
587
+ cudaMallocManaged (&Result, 2 * sizeof (*Result));
588
588
for (const auto &TestCase : TestCases) {
589
589
testHleu2 (Result, TestCase.first .first , TestCase.first .second );
590
590
checkResult (" __hleu2" , {TestCase.first .first , TestCase.first .second },
@@ -606,7 +606,7 @@ void testHlt2(float *const Result, __half2 Input1, __half2 Input2) {
606
606
607
607
void testHlt2Cases (const vector<pair<half2_pair, h2i_pair>> &TestCases) {
608
608
float *Result;
609
- cudaMallocManaged (&Result, sizeof (*Result));
609
+ cudaMallocManaged (&Result, 2 * sizeof (*Result));
610
610
for (const auto &TestCase : TestCases) {
611
611
testHlt2 (Result, TestCase.first .first , TestCase.first .second );
612
612
checkResult (" __hlt2" , {TestCase.first .first , TestCase.first .second },
@@ -628,7 +628,7 @@ void testHltu2(float *const Result, __half2 Input1, __half2 Input2) {
628
628
629
629
void testHltu2Cases (const vector<pair<half2_pair, h2i_pair>> &TestCases) {
630
630
float *Result;
631
- cudaMallocManaged (&Result, sizeof (*Result));
631
+ cudaMallocManaged (&Result, 2 * sizeof (*Result));
632
632
for (const auto &TestCase : TestCases) {
633
633
testHltu2 (Result, TestCase.first .first , TestCase.first .second );
634
634
checkResult (" __hltu2" , {TestCase.first .first , TestCase.first .second },
@@ -650,7 +650,7 @@ void testHne2(float *const Result, __half2 Input1, __half2 Input2) {
650
650
651
651
void testHne2Cases (const vector<pair<half2_pair, h2i_pair>> &TestCases) {
652
652
float *Result;
653
- cudaMallocManaged (&Result, sizeof (*Result));
653
+ cudaMallocManaged (&Result, 2 * sizeof (*Result));
654
654
for (const auto &TestCase : TestCases) {
655
655
testHne2 (Result, TestCase.first .first , TestCase.first .second );
656
656
checkResult (" __hne2" , {TestCase.first .first , TestCase.first .second },
@@ -672,7 +672,7 @@ void testHneu2(float *const Result, __half2 Input1, __half2 Input2) {
672
672
673
673
void testHneu2Cases (const vector<pair<half2_pair, h2i_pair>> &TestCases) {
674
674
float *Result;
675
- cudaMallocManaged (&Result, sizeof (*Result));
675
+ cudaMallocManaged (&Result, 2 * sizeof (*Result));
676
676
for (const auto &TestCase : TestCases) {
677
677
testHneu2 (Result, TestCase.first .first , TestCase.first .second );
678
678
checkResult (" __hneu2" , {TestCase.first .first , TestCase.first .second },
0 commit comments