diff --git a/benchmarks/benchGemv.cpp b/benchmarks/benchGemv.cpp
index cc03973a6..ad994d50a 100644
--- a/benchmarks/benchGemv.cpp
+++ b/benchmarks/benchGemv.cpp
@@ -117,23 +117,23 @@ static void BM_GemvAdj(benchmark::State& state) {
 
 static void GemvSizes(::benchmark::Benchmark* b) {
   // Square matrices: exercises balanced kernel behavior.
-  for (int size : {8, 16, 32, 64, 128, 256, 512, 1024, 4096}) {
+  for (int size : {8, 32, 128, 512, 1024}) {
     b->Args({size, size});
   }
   // Tall-thin (m >> n): in ColMajor kernel, the inner vectorized loop over rows
   // is long while the outer column loop is short. In RowMajor kernel (transpose),
   // there are many rows to process but short dot products.
-  for (int n : {1, 4, 16, 64}) {
-    for (int m : {256, 1024, 4096}) {
-      if (m != n) b->Args({m, n});
+  for (int n : {1, 16}) {
+    for (int m : {256, 1024}) {
+      b->Args({m, n});
     }
   }
   // Short-wide (m << n): in ColMajor kernel, the outer column loop is long but
   // the inner vectorized loop over rows is short. In RowMajor kernel (transpose),
   // there are few rows but long dot products.
-  for (int m : {1, 4, 16, 64}) {
-    for (int n : {256, 1024, 4096}) {
-      if (m != n) b->Args({m, n});
+  for (int m : {1, 16}) {
+    for (int n : {256, 1024}) {
+      b->Args({m, n});
     }
   }
 }
@@ -149,12 +149,9 @@ BENCHMARK(BM_GemvTrans<float>)->Apply(GemvSizes)->Name("GemvTrans_float");
 BENCHMARK(BM_GemvTrans<double>)->Apply(GemvSizes)->Name("GemvTrans_double");
 
 // Complex types: all four variants exercise distinct kernel code paths.
+// Only cfloat is benchmarked since cdouble exercises the same paths but slower.
 
 BENCHMARK(BM_Gemv<std::complex<float>>)->Apply(GemvSizes)->Name("Gemv_cfloat");
-BENCHMARK(BM_Gemv<std::complex<double>>)->Apply(GemvSizes)->Name("Gemv_cdouble");
 BENCHMARK(BM_GemvTrans<std::complex<float>>)->Apply(GemvSizes)->Name("GemvTrans_cfloat");
-BENCHMARK(BM_GemvTrans<std::complex<double>>)->Apply(GemvSizes)->Name("GemvTrans_cdouble");
 BENCHMARK(BM_GemvConj<std::complex<float>>)->Apply(GemvSizes)->Name("GemvConj_cfloat");
-BENCHMARK(BM_GemvConj<std::complex<double>>)->Apply(GemvSizes)->Name("GemvConj_cdouble");
 BENCHMARK(BM_GemvAdj<std::complex<float>>)->Apply(GemvSizes)->Name("GemvAdj_cfloat");
-BENCHMARK(BM_GemvAdj<std::complex<double>>)->Apply(GemvSizes)->Name("GemvAdj_cdouble");
diff --git a/benchmarks/bench_trsm.cpp b/benchmarks/bench_trsm.cpp
index 3da3a90c9..7b050b53d 100644
--- a/benchmarks/bench_trsm.cpp
+++ b/benchmarks/bench_trsm.cpp
@@ -64,36 +64,31 @@ static void BM_TRSM_Right(benchmark::State& state) {
 // ---------- Size configurations ----------
 
 static void TrsvSizes(::benchmark::Benchmark* b) {
-  for (int n : {32, 64, 128, 256, 512, 1024}) {
+  for (int n : {32, 128, 512}) {
     b->Args({n});
   }
 }
 
 static void TrsmSizes(::benchmark::Benchmark* b) {
-  for (int n : {32, 64, 128, 256, 512, 1024}) {
-    for (int nrhs : {1, 4, 16, 64, 256}) {
+  for (int n : {64, 256, 512}) {
+    for (int nrhs : {1, 16, 64}) {
       b->Args({n, nrhs});
     }
   }
 }
 
 // ---------- TRSV benchmarks ----------
+// Only Lower is benchmarked; Upper exercises the same kernel via transposed storage.
 
 BENCHMARK(BM_TRSV<float, Lower>)->Apply(TrsvSizes)->Name("TRSV_float_Lower");
-BENCHMARK(BM_TRSV<float, Upper>)->Apply(TrsvSizes)->Name("TRSV_float_Upper");
 BENCHMARK(BM_TRSV<double, Lower>)->Apply(TrsvSizes)->Name("TRSV_double_Lower");
-BENCHMARK(BM_TRSV<double, Upper>)->Apply(TrsvSizes)->Name("TRSV_double_Upper");
 
 // ---------- TRSM Left benchmarks ----------
 
 BENCHMARK(BM_TRSM_Left<float, Lower>)->Apply(TrsmSizes)->Name("TRSM_Left_float_Lower");
-BENCHMARK(BM_TRSM_Left<float, Upper>)->Apply(TrsmSizes)->Name("TRSM_Left_float_Upper");
 BENCHMARK(BM_TRSM_Left<double, Lower>)->Apply(TrsmSizes)->Name("TRSM_Left_double_Lower");
-BENCHMARK(BM_TRSM_Left<double, Upper>)->Apply(TrsmSizes)->Name("TRSM_Left_double_Upper");
 
 // ---------- TRSM Right benchmarks ----------
 
 BENCHMARK(BM_TRSM_Right<float, Lower>)->Apply(TrsmSizes)->Name("TRSM_Right_float_Lower");
-BENCHMARK(BM_TRSM_Right<float, Upper>)->Apply(TrsmSizes)->Name("TRSM_Right_float_Upper");
 BENCHMARK(BM_TRSM_Right<double, Lower>)->Apply(TrsmSizes)->Name("TRSM_Right_double_Lower");
-BENCHMARK(BM_TRSM_Right<double, Upper>)->Apply(TrsmSizes)->Name("TRSM_Right_double_Upper");