Merge branch 'develop' into issue1593-2

CMakeLists.txt

@@ -6,7 +6,7 @@ cmake_minimum_required(VERSION 2.8.5)
 project(OpenBLAS C ASM)
 set(OpenBLAS_MAJOR_VERSION 0)
 set(OpenBLAS_MINOR_VERSION 3)
-set(OpenBLAS_PATCH_VERSION 0.dev)
+set(OpenBLAS_PATCH_VERSION 1.dev)
 set(OpenBLAS_VERSION "${OpenBLAS_MAJOR_VERSION}.${OpenBLAS_MINOR_VERSION}.${OpenBLAS_PATCH_VERSION}")
 
 # Adhere to GNU filesystem layout conventions

Makefile

@@ -294,9 +294,10 @@ endif
 
 lapack-test :
 	(cd $(NETLIB_LAPACK_DIR)/TESTING && rm -f x* *.out)
-	$(MAKE) -j 1 -C $(NETLIB_LAPACK_DIR)/TESTING xeigtstc  xeigtstd  xeigtsts  xeigtstz  xlintstc  xlintstd  xlintstds  xlintstrfd  xlintstrfz  xlintsts  xlintstz  xlintstzc xlintstrfs xlintstrfc
+	$(MAKE) -j 1 -C $(NETLIB_LAPACK_DIR)/TESTING/EIG xeigtstc  xeigtstd  xeigtsts  xeigtstz 
+	$(MAKE) -j 1 -C $(NETLIB_LAPACK_DIR)/TESTING/LIN xlintstc  xlintstd  xlintstds  xlintstrfd  xlintstrfz  xlintsts  xlintstz  xlintstzc xlintstrfs xlintstrfc
 ifneq ($(CROSS), 1)
-	( cd $(NETLIB_LAPACK_DIR)/INSTALL; ./testlsame; ./testslamch; ./testdlamch; \
+	( cd $(NETLIB_LAPACK_DIR)/INSTALL; make all; ./testlsame; ./testslamch; ./testdlamch; \
         ./testsecond; ./testdsecnd; ./testieee; ./testversion )
 	(cd $(NETLIB_LAPACK_DIR); ./lapack_testing.py -r )
 endif
@@ -308,9 +309,9 @@ lapack-runtest:
 
 
 blas-test:
-	(cd $(NETLIB_LAPACK_DIR)/BLAS && rm -f x* *.out)
+	(cd $(NETLIB_LAPACK_DIR)/BLAS/TESTING && rm -f x* *.out)
 	$(MAKE) -j 1 -C $(NETLIB_LAPACK_DIR) blas_testing
-	(cd $(NETLIB_LAPACK_DIR)/BLAS && cat *.out)
+	(cd $(NETLIB_LAPACK_DIR)/BLAS/TESTING && cat *.out)
 
 
 dummy :

Makefile.rule

@@ -3,7 +3,7 @@
 #
 
 # This library's version
-VERSION = 0.3.0.dev
+VERSION = 0.3.1.dev
 
 # If you set the suffix, the library name will be libopenblas_$(LIBNAMESUFFIX).a
 # and libopenblas_$(LIBNAMESUFFIX).so. Meanwhile, the soname in shared library

c_check

@@ -201,6 +201,21 @@ $architecture = zarch  if ($data =~ /ARCH_ZARCH/);
 $binformat    = bin32;
 $binformat    = bin64  if ($data =~ /BINARY_64/);
 
+$no_avx512= 0;
+if (($architecture eq "x86") || ($architecture eq "x86_64")) {
+    $code = '"vaddps %zmm1, %zmm0, %zmm0"'; 
+    print $tmpf "void main(void){ __asm__ volatile($code); }\n";
+    $args = " -o $tmpf.o -x c $tmpf";
+    my @cmd = ("$compiler_name $args");
+    system(@cmd) == 0;
+    if ($? != 0) {
+	$no_avx512 = 1;
+    } else {
+	$no_avx512 = 0;
+    }
+    unlink("tmpf.o");
+}
+
 $data = `$compiler_name -S ctest1.c && grep globl ctest1.s | head -n 1 && rm -f ctest1.s`;
 
 $data =~ /globl\s([_\.]*)(.*)/;
@@ -288,6 +303,7 @@ print MAKEFILE "CROSS=1\n" if $cross != 0;
 print MAKEFILE "CEXTRALIB=$linker_L $linker_l $linker_a\n";
 print MAKEFILE "HAVE_MSA=1\n" if $have_msa eq 1;
 print MAKEFILE "MSA_FLAGS=$msa_flags\n" if $have_msa eq 1;
+print MAKEFILE "NO_AVX512=1\n" if $no_avx512 eq 1;
 
 $os           =~ tr/[a-z]/[A-Z]/;
 $architecture =~ tr/[a-z]/[A-Z]/;

cmake/system_check.cmake

@@ -66,3 +66,12 @@ else()
   set(BINARY32 1)
 endif()
 
+if (X86_64 OR X86)
+  file(WRITE ${PROJECT_BINARY_DIR}/avx512.tmp "void main(void){ __asm__ volatile(\"vaddps %zmm1, %zmm0, %zmm0\"); }")
+execute_process(COMMAND ${CMAKE_C_COMPILER} -v -o ${PROJECT_BINARY_DIR}/avx512.o -x c ${PROJECT_BINARY_DIR}/avx512.tmp RESULT_VARIABLE NO_AVX512)
+if (NO_AVX512 EQUAL 1)
+set (CCOMMON_OPT "${CCOMMON_OPT} -DNO_AVX512")
+endif()
+  file(REMOVE "avx512.tmp" "avx512.o")
+endif()
+

common.h

@@ -642,6 +642,7 @@ void gotoblas_profile_init(void);
 void gotoblas_profile_quit(void);
 
 #ifdef USE_OPENMP
+
 #ifndef C_MSVC
 int omp_in_parallel(void);
 int omp_get_num_procs(void);
@@ -663,7 +664,6 @@ __declspec(dllimport) int __cdecl omp_get_num_procs(void);
 #define _Atomic volatile
 #endif
 
-
 #else
 #ifdef __ELF__
 int omp_in_parallel  (void) __attribute__ ((weak));

driver/level3/Makefile

@@ -362,7 +362,7 @@ cgemm_ct.$(SUFFIX) : gemm.c level3.c  ../../param.h
 	$(CC) $(CFLAGS) $(BLOCKS) -c -UDOUBLE -DCOMPLEX -DCT $< -o $(@F)
 
 cgemm_cr.$(SUFFIX) : gemm.c level3.c  ../../param.h
-	$(CC) $(CFLAGS) $(BLOCKS) -c -UDOUBLE -DCOMPLEX -DCR $< -o $(@F)
+	$(CC) $(CFLAGS) $(BLOCKS) -c -UDOUBLE -DCOMPLEX -DCR=CR $< -o $(@F)
 
 cgemm_cc.$(SUFFIX) : gemm.c level3.c  ../../param.h
 	$(CC) $(CFLAGS) $(BLOCKS) -c -UDOUBLE -DCOMPLEX -DCC $< -o $(@F)
@@ -410,7 +410,7 @@ zgemm_ct.$(SUFFIX) : gemm.c level3.c ../../param.h
 	$(CC) $(CFLAGS) $(BLOCKS) -c -DDOUBLE -DCOMPLEX -DCT $< -o $(@F)
 
 zgemm_cr.$(SUFFIX) : gemm.c level3.c ../../param.h
-	$(CC) $(CFLAGS) $(BLOCKS) -c -DDOUBLE -DCOMPLEX -DCR $< -o $(@F)
+	$(CC) $(CFLAGS) $(BLOCKS) -c -DDOUBLE -DCOMPLEX -DCR=CR $< -o $(@F)
 
 zgemm_cc.$(SUFFIX) : gemm.c level3.c ../../param.h
 	$(CC) $(CFLAGS) $(BLOCKS) -c -DDOUBLE -DCOMPLEX -DCC $< -o $(@F)
@@ -458,7 +458,7 @@ xgemm_ct.$(SUFFIX) : gemm.c level3.c ../../param.h
 	$(CC) $(CFLAGS) $(BLOCKS) -c -DXDOUBLE -DCOMPLEX -DCT $< -o $(@F)
 
 xgemm_cr.$(SUFFIX) : gemm.c level3.c ../../param.h
-	$(CC) $(CFLAGS) $(BLOCKS) -c -DXDOUBLE -DCOMPLEX -DCR $< -o $(@F)
+	$(CC) $(CFLAGS) $(BLOCKS) -c -DXDOUBLE -DCOMPLEX -DCR=CR $< -o $(@F)
 
 xgemm_cc.$(SUFFIX) : gemm.c level3.c ../../param.h
 	$(CC) $(CFLAGS) $(BLOCKS) -c -DXDOUBLE -DCOMPLEX -DCC $< -o $(@F)
@@ -558,7 +558,7 @@ cgemm_thread_ct.$(SUFFIX) : gemm.c level3_thread.c ../../param.h
 	$(CC) $(CFLAGS) $(BLOCKS) -c -DTHREADED_LEVEL3 -UDOUBLE -DCOMPLEX -DCT $< -o $(@F)
 
 cgemm_thread_cr.$(SUFFIX) : gemm.c level3_thread.c ../../param.h
-	$(CC) $(CFLAGS) $(BLOCKS) -c -DTHREADED_LEVEL3 -UDOUBLE -DCOMPLEX -DCR $< -o $(@F)
+	$(CC) $(CFLAGS) $(BLOCKS) -c -DTHREADED_LEVEL3 -UDOUBLE -DCOMPLEX -DCR=CR $< -o $(@F)
 
 cgemm_thread_cc.$(SUFFIX) : gemm.c level3_thread.c ../../param.h
 	$(CC) $(CFLAGS) $(BLOCKS) -c -DTHREADED_LEVEL3 -UDOUBLE -DCOMPLEX -DCC $< -o $(@F)
@@ -606,7 +606,7 @@ zgemm_thread_ct.$(SUFFIX) : gemm.c level3_thread.c ../../param.h
 	$(CC) $(CFLAGS) $(BLOCKS) -c -DTHREADED_LEVEL3 -DDOUBLE -DCOMPLEX -DCT $< -o $(@F)
 
 zgemm_thread_cr.$(SUFFIX) : gemm.c level3_thread.c ../../param.h
-	$(CC) $(CFLAGS) $(BLOCKS) -c -DTHREADED_LEVEL3 -DDOUBLE -DCOMPLEX -DCR $< -o $(@F)
+	$(CC) $(CFLAGS) $(BLOCKS) -c -DTHREADED_LEVEL3 -DDOUBLE -DCOMPLEX -DCR=CR $< -o $(@F)
 
 zgemm_thread_cc.$(SUFFIX) : gemm.c level3_thread.c ../../param.h
 	$(CC) $(CFLAGS) $(BLOCKS) -c -DTHREADED_LEVEL3 -DDOUBLE -DCOMPLEX -DCC $< -o $(@F)
@@ -654,7 +654,7 @@ xgemm_thread_ct.$(SUFFIX) : gemm.c level3_thread.c ../../param.h
 	$(CC) $(CFLAGS) $(BLOCKS) -c -DTHREADED_LEVEL3 -DXDOUBLE -DCOMPLEX -DCT $< -o $(@F)
 
 xgemm_thread_cr.$(SUFFIX) : gemm.c level3_thread.c ../../param.h
-	$(CC) $(CFLAGS) $(BLOCKS) -c -DTHREADED_LEVEL3 -DXDOUBLE -DCOMPLEX -DCR $< -o $(@F)
+	$(CC) $(CFLAGS) $(BLOCKS) -c -DTHREADED_LEVEL3 -DXDOUBLE -DCOMPLEX -DCR=CR $< -o $(@F)
 
 xgemm_thread_cc.$(SUFFIX) : gemm.c level3_thread.c ../../param.h
 	$(CC) $(CFLAGS) $(BLOCKS) -c -DTHREADED_LEVEL3 -DXDOUBLE -DCOMPLEX -DCC $< -o $(@F)
@@ -1821,7 +1821,7 @@ cgemm3m_ct.$(SUFFIX) : gemm3m.c gemm3m_level3.c
 	$(CC) $(CFLAGS) $(BLOCKS) -c -UDOUBLE -DCOMPLEX -DCT $< -o $(@F)
 
 cgemm3m_cr.$(SUFFIX) : gemm3m.c gemm3m_level3.c
-	$(CC) $(CFLAGS) $(BLOCKS) -c -UDOUBLE -DCOMPLEX -DCR $< -o $(@F)
+	$(CC) $(CFLAGS) $(BLOCKS) -c -UDOUBLE -DCOMPLEX -DCR=CR $< -o $(@F)
 
 cgemm3m_cc.$(SUFFIX) : gemm3m.c gemm3m_level3.c
 	$(CC) $(CFLAGS) $(BLOCKS) -c -UDOUBLE -DCOMPLEX -DCC $< -o $(@F)
@@ -1869,7 +1869,7 @@ zgemm3m_ct.$(SUFFIX) : gemm3m.c gemm3m_level3.c
 	$(CC) $(CFLAGS) $(BLOCKS) -c -DDOUBLE -DCOMPLEX -DCT $< -o $(@F)
 
 zgemm3m_cr.$(SUFFIX) : gemm3m.c gemm3m_level3.c
-	$(CC) $(CFLAGS) $(BLOCKS) -c -DDOUBLE -DCOMPLEX -DCR $< -o $(@F)
+	$(CC) $(CFLAGS) $(BLOCKS) -c -DDOUBLE -DCOMPLEX -DCR=CR $< -o $(@F)
 
 zgemm3m_cc.$(SUFFIX) : gemm3m.c gemm3m_level3.c
 	$(CC) $(CFLAGS) $(BLOCKS) -c -DDOUBLE -DCOMPLEX -DCC $< -o $(@F)
@@ -1917,7 +1917,7 @@ xgemm3m_ct.$(SUFFIX) : gemm3m.c gemm3m_level3.c ../../param.h
 	$(CC) $(CFLAGS) $(BLOCKS) -c -DXDOUBLE -DCOMPLEX -DCT $< -o $(@F)
 
 xgemm3m_cr.$(SUFFIX) : gemm3m.c gemm3m_level3.c ../../param.h
-	$(CC) $(CFLAGS) $(BLOCKS) -c -DXDOUBLE -DCOMPLEX -DCR $< -o $(@F)
+	$(CC) $(CFLAGS) $(BLOCKS) -c -DXDOUBLE -DCOMPLEX -DCR=CR $< -o $(@F)
 
 xgemm3m_cc.$(SUFFIX) : gemm3m.c gemm3m_level3.c ../../param.h
 	$(CC) $(CFLAGS) $(BLOCKS) -c -DXDOUBLE -DCOMPLEX -DCC $< -o $(@F)
@@ -1974,7 +1974,7 @@ cgemm3m_thread_ct.$(SUFFIX) : gemm3m.c level3_gemm3m_thread.c ../../param.h
 	$(CC) $(CFLAGS) $(BLOCKS) -c -DTHREADED_LEVEL3 -UDOUBLE -DCOMPLEX -DCT $< -o $(@F)
 
 cgemm3m_thread_cr.$(SUFFIX) : gemm3m.c level3_gemm3m_thread.c ../../param.h
-	$(CC) $(CFLAGS) $(BLOCKS) -c -DTHREADED_LEVEL3 -UDOUBLE -DCOMPLEX -DCR $< -o $(@F)
+	$(CC) $(CFLAGS) $(BLOCKS) -c -DTHREADED_LEVEL3 -UDOUBLE -DCOMPLEX -DCR=CR $< -o $(@F)
 
 cgemm3m_thread_cc.$(SUFFIX) : gemm3m.c level3_gemm3m_thread.c ../../param.h
 	$(CC) $(CFLAGS) $(BLOCKS) -c -DTHREADED_LEVEL3 -UDOUBLE -DCOMPLEX -DCC $< -o $(@F)
@@ -2022,7 +2022,7 @@ zgemm3m_thread_ct.$(SUFFIX) : gemm3m.c level3_gemm3m_thread.c ../../param.h
 	$(CC) $(CFLAGS) $(BLOCKS) -c -DTHREADED_LEVEL3 -DDOUBLE -DCOMPLEX -DCT $< -o $(@F)
 
 zgemm3m_thread_cr.$(SUFFIX) : gemm3m.c level3_gemm3m_thread.c ../../param.h
-	$(CC) $(CFLAGS) $(BLOCKS) -c -DTHREADED_LEVEL3 -DDOUBLE -DCOMPLEX -DCR $< -o $(@F)
+	$(CC) $(CFLAGS) $(BLOCKS) -c -DTHREADED_LEVEL3 -DDOUBLE -DCOMPLEX -DCR=CR $< -o $(@F)
 
 zgemm3m_thread_cc.$(SUFFIX) : gemm3m.c level3_gemm3m_thread.c ../../param.h
 	$(CC) $(CFLAGS) $(BLOCKS) -c -DTHREADED_LEVEL3 -DDOUBLE -DCOMPLEX -DCC $< -o $(@F)
@@ -2070,7 +2070,7 @@ xgemm3m_thread_ct.$(SUFFIX) : gemm3m.c level3_gemm3m_thread.c ../../param.h
 	$(CC) $(CFLAGS) $(BLOCKS) -c -DTHREADED_LEVEL3 -DXDOUBLE -DCOMPLEX -DCT $< -o $(@F)
 
 xgemm3m_thread_cr.$(SUFFIX) : gemm3m.c level3_gemm3m_thread.c ../../param.h
-	$(CC) $(CFLAGS) $(BLOCKS) -c -DTHREADED_LEVEL3 -DXDOUBLE -DCOMPLEX -DCR $< -o $(@F)
+	$(CC) $(CFLAGS) $(BLOCKS) -c -DTHREADED_LEVEL3 -DXDOUBLE -DCOMPLEX -DCR=CR $< -o $(@F)
 
 xgemm3m_thread_cc.$(SUFFIX) : gemm3m.c level3_gemm3m_thread.c ../../param.h
 	$(CC) $(CFLAGS) $(BLOCKS) -c -DTHREADED_LEVEL3 -DXDOUBLE -DCOMPLEX -DCC $< -o $(@F)
@@ -2731,7 +2731,7 @@ cgemm_ct.$(PSUFFIX) : gemm.c level3.c  ../../param.h
 	$(CC) $(PFLAGS) $(BLOCKS) -c -UDOUBLE -DCOMPLEX -DCT $< -o $(@F)
 
 cgemm_cr.$(PSUFFIX) : gemm.c level3.c  ../../param.h
-	$(CC) $(PFLAGS) $(BLOCKS) -c -UDOUBLE -DCOMPLEX -DCR $< -o $(@F)
+	$(CC) $(PFLAGS) $(BLOCKS) -c -UDOUBLE -DCOMPLEX -DCR=CR $< -o $(@F)
 
 cgemm_cc.$(PSUFFIX) : gemm.c level3.c  ../../param.h
 	$(CC) $(PFLAGS) $(BLOCKS) -c -UDOUBLE -DCOMPLEX -DCC $< -o $(@F)
@@ -2779,7 +2779,7 @@ zgemm_ct.$(PSUFFIX) : gemm.c level3.c ../../param.h
 	$(CC) $(PFLAGS) $(BLOCKS) -c -DDOUBLE -DCOMPLEX -DCT $< -o $(@F)
 
 zgemm_cr.$(PSUFFIX) : gemm.c level3.c ../../param.h
-	$(CC) $(PFLAGS) $(BLOCKS) -c -DDOUBLE -DCOMPLEX -DCR $< -o $(@F)
+	$(CC) $(PFLAGS) $(BLOCKS) -c -DDOUBLE -DCOMPLEX -DCR=CR $< -o $(@F)
 
 zgemm_cc.$(PSUFFIX) : gemm.c level3.c ../../param.h
 	$(CC) $(PFLAGS) $(BLOCKS) -c -DDOUBLE -DCOMPLEX -DCC $< -o $(@F)
@@ -2827,7 +2827,7 @@ xgemm_ct.$(PSUFFIX) : gemm.c level3.c ../../param.h
 	$(CC) $(PFLAGS) $(BLOCKS) -c -DXDOUBLE -DCOMPLEX -DCT $< -o $(@F)
 
 xgemm_cr.$(PSUFFIX) : gemm.c level3.c ../../param.h
-	$(CC) $(PFLAGS) $(BLOCKS) -c -DXDOUBLE -DCOMPLEX -DCR $< -o $(@F)
+	$(CC) $(PFLAGS) $(BLOCKS) -c -DXDOUBLE -DCOMPLEX -DCR=CR $< -o $(@F)
 
 xgemm_cc.$(PSUFFIX) : gemm.c level3.c ../../param.h
 	$(CC) $(PFLAGS) $(BLOCKS) -c -DXDOUBLE -DCOMPLEX -DCC $< -o $(@F)
@@ -2927,7 +2927,7 @@ cgemm_thread_ct.$(PSUFFIX) : gemm.c level3_thread.c ../../param.h
 	$(CC) $(PFLAGS) $(BLOCKS) -c -DTHREADED_LEVEL3 -UDOUBLE -DCOMPLEX -DCT $< -o $(@F)
 
 cgemm_thread_cr.$(PSUFFIX) : gemm.c level3_thread.c ../../param.h
-	$(CC) $(PFLAGS) $(BLOCKS) -c -DTHREADED_LEVEL3 -UDOUBLE -DCOMPLEX -DCR $< -o $(@F)
+	$(CC) $(PFLAGS) $(BLOCKS) -c -DTHREADED_LEVEL3 -UDOUBLE -DCOMPLEX -DCR=CR $< -o $(@F)
 
 cgemm_thread_cc.$(PSUFFIX) : gemm.c level3_thread.c ../../param.h
 	$(CC) $(PFLAGS) $(BLOCKS) -c -DTHREADED_LEVEL3 -UDOUBLE -DCOMPLEX -DCC $< -o $(@F)
@@ -2975,7 +2975,7 @@ zgemm_thread_ct.$(PSUFFIX) : gemm.c level3_thread.c ../../param.h
 	$(CC) $(PFLAGS) $(BLOCKS) -c -DTHREADED_LEVEL3 -DDOUBLE -DCOMPLEX -DCT $< -o $(@F)
 
 zgemm_thread_cr.$(PSUFFIX) : gemm.c level3_thread.c ../../param.h
-	$(CC) $(PFLAGS) $(BLOCKS) -c -DTHREADED_LEVEL3 -DDOUBLE -DCOMPLEX -DCR $< -o $(@F)
+	$(CC) $(PFLAGS) $(BLOCKS) -c -DTHREADED_LEVEL3 -DDOUBLE -DCOMPLEX -DCR=CR $< -o $(@F)
 
 zgemm_thread_cc.$(PSUFFIX) : gemm.c level3_thread.c ../../param.h
 	$(CC) $(PFLAGS) $(BLOCKS) -c -DTHREADED_LEVEL3 -DDOUBLE -DCOMPLEX -DCC $< -o $(@F)
@@ -3023,7 +3023,7 @@ xgemm_thread_ct.$(PSUFFIX) : gemm.c level3_thread.c ../../param.h
 	$(CC) $(PFLAGS) $(BLOCKS) -c -DTHREADED_LEVEL3 -DXDOUBLE -DCOMPLEX -DCT $< -o $(@F)
 
 xgemm_thread_cr.$(PSUFFIX) : gemm.c level3_thread.c ../../param.h
-	$(CC) $(PFLAGS) $(BLOCKS) -c -DTHREADED_LEVEL3 -DXDOUBLE -DCOMPLEX -DCR $< -o $(@F)
+	$(CC) $(PFLAGS) $(BLOCKS) -c -DTHREADED_LEVEL3 -DXDOUBLE -DCOMPLEX -DCR=CR $< -o $(@F)
 
 xgemm_thread_cc.$(PSUFFIX) : gemm.c level3_thread.c ../../param.h
 	$(CC) $(PFLAGS) $(BLOCKS) -c -DTHREADED_LEVEL3 -DXDOUBLE -DCOMPLEX -DCC $< -o $(@F)
@@ -4190,7 +4190,7 @@ cgemm3m_ct.$(PSUFFIX) : gemm3m.c gemm3m_level3.c
 	$(CC) $(PFLAGS) $(BLOCKS) -c -UDOUBLE -DCOMPLEX -DCT $< -o $(@F)
 
 cgemm3m_cr.$(PSUFFIX) : gemm3m.c gemm3m_level3.c
-	$(CC) $(PFLAGS) $(BLOCKS) -c -UDOUBLE -DCOMPLEX -DCR $< -o $(@F)
+	$(CC) $(PFLAGS) $(BLOCKS) -c -UDOUBLE -DCOMPLEX -DCR=CR $< -o $(@F)
 
 cgemm3m_cc.$(PSUFFIX) : gemm3m.c gemm3m_level3.c
 	$(CC) $(PFLAGS) $(BLOCKS) -c -UDOUBLE -DCOMPLEX -DCC $< -o $(@F)
@@ -4238,7 +4238,7 @@ zgemm3m_ct.$(PSUFFIX) : gemm3m.c gemm3m_level3.c
 	$(CC) $(PFLAGS) $(BLOCKS) -c -DDOUBLE -DCOMPLEX -DCT $< -o $(@F)
 
 zgemm3m_cr.$(PSUFFIX) : gemm3m.c gemm3m_level3.c
-	$(CC) $(PFLAGS) $(BLOCKS) -c -DDOUBLE -DCOMPLEX -DCR $< -o $(@F)
+	$(CC) $(PFLAGS) $(BLOCKS) -c -DDOUBLE -DCOMPLEX -DCR=CR $< -o $(@F)
 
 zgemm3m_cc.$(PSUFFIX) : gemm3m.c gemm3m_level3.c
 	$(CC) $(PFLAGS) $(BLOCKS) -c -DDOUBLE -DCOMPLEX -DCC $< -o $(@F)
@@ -4286,7 +4286,7 @@ xgemm3m_ct.$(PSUFFIX) : gemm3m.c gemm3m_level3.c ../../param.h
 	$(CC) $(PFLAGS) $(BLOCKS) -c -DXDOUBLE -DCOMPLEX -DCT $< -o $(@F)
 
 xgemm3m_cr.$(PSUFFIX) : gemm3m.c gemm3m_level3.c ../../param.h
-	$(CC) $(PFLAGS) $(BLOCKS) -c -DXDOUBLE -DCOMPLEX -DCR $< -o $(@F)
+	$(CC) $(PFLAGS) $(BLOCKS) -c -DXDOUBLE -DCOMPLEX -DCR=CR $< -o $(@F)
 
 xgemm3m_cc.$(PSUFFIX) : gemm3m.c gemm3m_level3.c ../../param.h
 	$(CC) $(PFLAGS) $(BLOCKS) -c -DXDOUBLE -DCOMPLEX -DCC $< -o $(@F)
@@ -4343,7 +4343,7 @@ cgemm3m_thread_ct.$(PSUFFIX) : gemm3m.c level3_gemm3m_thread.c ../../param.h
 	$(CC) $(PFLAGS) $(BLOCKS) -c -DTHREADED_LEVEL3 -UDOUBLE -DCOMPLEX -DCT $< -o $(@F)
 
 cgemm3m_thread_cr.$(PSUFFIX) : gemm3m.c level3_gemm3m_thread.c ../../param.h
-	$(CC) $(PFLAGS) $(BLOCKS) -c -DTHREADED_LEVEL3 -UDOUBLE -DCOMPLEX -DCR $< -o $(@F)
+	$(CC) $(PFLAGS) $(BLOCKS) -c -DTHREADED_LEVEL3 -UDOUBLE -DCOMPLEX -DCR=CR $< -o $(@F)
 
 cgemm3m_thread_cc.$(PSUFFIX) : gemm3m.c level3_gemm3m_thread.c ../../param.h
 	$(CC) $(PFLAGS) $(BLOCKS) -c -DTHREADED_LEVEL3 -UDOUBLE -DCOMPLEX -DCC $< -o $(@F)
@@ -4391,7 +4391,7 @@ zgemm3m_thread_ct.$(PSUFFIX) : gemm3m.c level3_gemm3m_thread.c ../../param.h
 	$(CC) $(PFLAGS) $(BLOCKS) -c -DTHREADED_LEVEL3 -DDOUBLE -DCOMPLEX -DCT $< -o $(@F)
 
 zgemm3m_thread_cr.$(PSUFFIX) : gemm3m.c level3_gemm3m_thread.c ../../param.h
-	$(CC) $(PFLAGS) $(BLOCKS) -c -DTHREADED_LEVEL3 -DDOUBLE -DCOMPLEX -DCR $< -o $(@F)
+	$(CC) $(PFLAGS) $(BLOCKS) -c -DTHREADED_LEVEL3 -DDOUBLE -DCOMPLEX -DCR=CR $< -o $(@F)
 
 zgemm3m_thread_cc.$(PSUFFIX) : gemm3m.c level3_gemm3m_thread.c ../../param.h
 	$(CC) $(PFLAGS) $(BLOCKS) -c -DTHREADED_LEVEL3 -DDOUBLE -DCOMPLEX -DCC $< -o $(@F)
@@ -4439,7 +4439,7 @@ xgemm3m_thread_ct.$(PSUFFIX) : gemm3m.c level3_gemm3m_thread.c ../../param.h
 	$(CC) $(PFLAGS) $(BLOCKS) -c -DTHREADED_LEVEL3 -DXDOUBLE -DCOMPLEX -DCT $< -o $(@F)
 
 xgemm3m_thread_cr.$(PSUFFIX) : gemm3m.c level3_gemm3m_thread.c ../../param.h
-	$(CC) $(PFLAGS) $(BLOCKS) -c -DTHREADED_LEVEL3 -DXDOUBLE -DCOMPLEX -DCR $< -o $(@F)
+	$(CC) $(PFLAGS) $(BLOCKS) -c -DTHREADED_LEVEL3 -DXDOUBLE -DCOMPLEX -DCR=CR $< -o $(@F)
 
 xgemm3m_thread_cc.$(PSUFFIX) : gemm3m.c level3_gemm3m_thread.c ../../param.h
 	$(CC) $(PFLAGS) $(BLOCKS) -c -DTHREADED_LEVEL3 -DXDOUBLE -DCOMPLEX -DCC $< -o $(@F)

driver/others/memory.c

@@ -180,7 +180,7 @@ int get_num_procs(void) {
 cpu_set_t *cpusetp;
 size_t size;
 int ret;
-// int i,n;
+int i,n;
 
   if (!nums) nums = sysconf(_SC_NPROCESSORS_CONF);
 #if !defined(OS_LINUX)

exports/Makefile

@@ -128,6 +128,8 @@ so : ../$(LIBSONAME)
 
 ifeq ($(OSNAME), Android)
 INTERNALNAME = $(LIBPREFIX).so
+FEXTRALIB += -lm
+EXTRALIB += -lm
 else
 INTERNALNAME = $(LIBPREFIX).so.$(MAJOR_VERSION)
 endif

kernel/power/KERNEL.POWER8

@@ -133,7 +133,7 @@ ZNRM2KERNEL  = ../arm/znrm2.c
 #
 SROTKERNEL   = srot.c
 DROTKERNEL   = drot.c
-#CROTKERNEL   = ../arm/zrot.c
+CROTKERNEL   = zrot.c
 ZROTKERNEL   = zrot.c
 #
 SSCALKERNEL  = sscal.c

kernel/x86/KERNEL.NEHALEM

@@ -1,3 +1 @@
 include $(KERNELDIR)/KERNEL.PENRYN
-SSWAPKERNEL  = ../arm/swap.c
-DSWAPKERNEL  = ../arm/swap.c

kernel/x86/swap.S

@@ -138,6 +138,14 @@
 /* INCX != 1 or INCY != 1 */
 
 .L14:
+	cmpl	$0, %ebx
+	jne	.L141
+	cmpl	$0, %ecx
+	jne	.L141
+/* INCX == 0 and INCY == 0 */	
+	jmp	.L27
+
+.L141:	
 	movl	%edx, %eax
 	sarl	$2,   %eax
 	jle	.L28

lapack-netlib/LAPACKE/src/lapacke_chetrf_aa_work.c

@@ -41,7 +41,7 @@ lapack_int LAPACKE_chetrf_aa_work( int matrix_layout, char uplo, lapack_int n,
     lapack_int info = 0;
     if( matrix_layout == LAPACK_COL_MAJOR ) {
         /* Call LAPACK function and adjust info */
-        LAPACK_chetrf( &uplo, &n, a, &lda, ipiv, work, &lwork, &info );
+        LAPACK_chetrf_aa( &uplo, &n, a, &lda, ipiv, work, &lwork, &info );
         if( info < 0 ) {
             info = info - 1;
         }
@@ -56,7 +56,7 @@ lapack_int LAPACKE_chetrf_aa_work( int matrix_layout, char uplo, lapack_int n,
         }
         /* Query optimal working array(s) size if requested */
         if( lwork == -1 ) {
-            LAPACK_chetrf( &uplo, &n, a, &lda_t, ipiv, work, &lwork, &info );
+            LAPACK_chetrf_aa( &uplo, &n, a, &lda_t, ipiv, work, &lwork, &info );
             return (info < 0) ? (info - 1) : info;
         }
         /* Allocate memory for temporary array(s) */
@@ -69,7 +69,7 @@ lapack_int LAPACKE_chetrf_aa_work( int matrix_layout, char uplo, lapack_int n,
         /* Transpose input matrices */
         LAPACKE_che_trans( matrix_layout, uplo, n, a, lda, a_t, lda_t );
         /* Call LAPACK function and adjust info */
-        LAPACK_chetrf( &uplo, &n, a_t, &lda_t, ipiv, work, &lwork, &info );
+        LAPACK_chetrf_aa( &uplo, &n, a_t, &lda_t, ipiv, work, &lwork, &info );
         if( info < 0 ) {
             info = info - 1;
         }

lapack-netlib/LAPACKE/src/lapacke_csytrf_aa_work.c

@@ -41,7 +41,7 @@ lapack_int LAPACKE_csytrf_aa_work( int matrix_layout, char uplo, lapack_int n,
     lapack_int info = 0;
     if( matrix_layout == LAPACK_COL_MAJOR ) {
         /* Call LAPACK function and adjust info */
-        LAPACK_csytrf( &uplo, &n, a, &lda, ipiv, work, &lwork, &info );
+        LAPACK_csytrf_aa( &uplo, &n, a, &lda, ipiv, work, &lwork, &info );
         if( info < 0 ) {
             info = info - 1;
         }
@@ -56,7 +56,7 @@ lapack_int LAPACKE_csytrf_aa_work( int matrix_layout, char uplo, lapack_int n,
         }
         /* Query optimal working array(s) size if requested */
         if( lwork == -1 ) {
-            LAPACK_csytrf( &uplo, &n, a, &lda_t, ipiv, work, &lwork, &info );
+            LAPACK_csytrf_aa( &uplo, &n, a, &lda_t, ipiv, work, &lwork, &info );
             return (info < 0) ? (info - 1) : info;
         }
         /* Allocate memory for temporary array(s) */
@@ -69,7 +69,7 @@ lapack_int LAPACKE_csytrf_aa_work( int matrix_layout, char uplo, lapack_int n,
         /* Transpose input matrices */
         LAPACKE_csy_trans( matrix_layout, uplo, n, a, lda, a_t, lda_t );
         /* Call LAPACK function and adjust info */
-        LAPACK_csytrf( &uplo, &n, a_t, &lda_t, ipiv, work, &lwork, &info );
+        LAPACK_csytrf_aa( &uplo, &n, a_t, &lda_t, ipiv, work, &lwork, &info );
         if( info < 0 ) {
             info = info - 1;
         }

lapack-netlib/LAPACKE/src/lapacke_dsytrf_aa_work.c

@@ -40,7 +40,7 @@ lapack_int LAPACKE_dsytrf_aa_work( int matrix_layout, char uplo, lapack_int n,
     lapack_int info = 0;
     if( matrix_layout == LAPACK_COL_MAJOR ) {
         /* Call LAPACK function and adjust info */
-        LAPACK_dsytrf( &uplo, &n, a, &lda, ipiv, work, &lwork, &info );
+        LAPACK_dsytrf_aa( &uplo, &n, a, &lda, ipiv, work, &lwork, &info );
         if( info < 0 ) {
             info = info - 1;
         }
@@ -55,7 +55,7 @@ lapack_int LAPACKE_dsytrf_aa_work( int matrix_layout, char uplo, lapack_int n,
         }
         /* Query optimal working array(s) size if requested */
         if( lwork == -1 ) {
-            LAPACK_dsytrf( &uplo, &n, a, &lda_t, ipiv, work, &lwork, &info );
+            LAPACK_dsytrf_aa( &uplo, &n, a, &lda_t, ipiv, work, &lwork, &info );
             return (info < 0) ? (info - 1) : info;
         }
         /* Allocate memory for temporary array(s) */
@@ -67,7 +67,7 @@ lapack_int LAPACKE_dsytrf_aa_work( int matrix_layout, char uplo, lapack_int n,
         /* Transpose input matrices */
         LAPACKE_dsy_trans( matrix_layout, uplo, n, a, lda, a_t, lda_t );
         /* Call LAPACK function and adjust info */
-        LAPACK_dsytrf( &uplo, &n, a_t, &lda_t, ipiv, work, &lwork, &info );
+        LAPACK_dsytrf_aa( &uplo, &n, a_t, &lda_t, ipiv, work, &lwork, &info );
         if( info < 0 ) {
             info = info - 1;
         }

lapack-netlib/LAPACKE/src/lapacke_ssytrf_aa_work.c

@@ -40,7 +40,7 @@ lapack_int LAPACKE_ssytrf_aa_work( int matrix_layout, char uplo, lapack_int n,
     lapack_int info = 0;
     if( matrix_layout == LAPACK_COL_MAJOR ) {
         /* Call LAPACK function and adjust info */
-        LAPACK_ssytrf( &uplo, &n, a, &lda, ipiv, work, &lwork, &info );
+        LAPACK_ssytrf_aa( &uplo, &n, a, &lda, ipiv, work, &lwork, &info );
         if( info < 0 ) {
             info = info - 1;
         }
@@ -55,7 +55,7 @@ lapack_int LAPACKE_ssytrf_aa_work( int matrix_layout, char uplo, lapack_int n,
         }
         /* Query optimal working array(s) size if requested */
         if( lwork == -1 ) {
-            LAPACK_ssytrf( &uplo, &n, a, &lda_t, ipiv, work, &lwork, &info );
+            LAPACK_ssytrf_aa( &uplo, &n, a, &lda_t, ipiv, work, &lwork, &info );
             return (info < 0) ? (info - 1) : info;
         }
         /* Allocate memory for temporary array(s) */
@@ -67,7 +67,7 @@ lapack_int LAPACKE_ssytrf_aa_work( int matrix_layout, char uplo, lapack_int n,
         /* Transpose input matrices */
         LAPACKE_ssy_trans( matrix_layout, uplo, n, a, lda, a_t, lda_t );
         /* Call LAPACK function and adjust info */
-        LAPACK_ssytrf( &uplo, &n, a_t, &lda_t, ipiv, work, &lwork, &info );
+        LAPACK_ssytrf_aa( &uplo, &n, a_t, &lda_t, ipiv, work, &lwork, &info );
         if( info < 0 ) {
             info = info - 1;
         }

lapack-netlib/LAPACKE/src/lapacke_zhetrf_aa_work.c

@@ -41,7 +41,7 @@ lapack_int LAPACKE_zhetrf_aa_work( int matrix_layout, char uplo, lapack_int n,
     lapack_int info = 0;
     if( matrix_layout == LAPACK_COL_MAJOR ) {
         /* Call LAPACK function and adjust info */
-        LAPACK_zhetrf( &uplo, &n, a, &lda, ipiv, work, &lwork, &info );
+        LAPACK_zhetrf_aa( &uplo, &n, a, &lda, ipiv, work, &lwork, &info );
         if( info < 0 ) {
             info = info - 1;
         }
@@ -56,7 +56,7 @@ lapack_int LAPACKE_zhetrf_aa_work( int matrix_layout, char uplo, lapack_int n,
         }
         /* Query optimal working array(s) size if requested */
         if( lwork == -1 ) {
-            LAPACK_zhetrf( &uplo, &n, a, &lda_t, ipiv, work, &lwork, &info );
+            LAPACK_zhetrf_aa( &uplo, &n, a, &lda_t, ipiv, work, &lwork, &info );
             return (info < 0) ? (info - 1) : info;
         }
         /* Allocate memory for temporary array(s) */
@@ -69,7 +69,7 @@ lapack_int LAPACKE_zhetrf_aa_work( int matrix_layout, char uplo, lapack_int n,
         /* Transpose input matrices */
         LAPACKE_zhe_trans( matrix_layout, uplo, n, a, lda, a_t, lda_t );
         /* Call LAPACK function and adjust info */
-        LAPACK_zhetrf( &uplo, &n, a_t, &lda_t, ipiv, work, &lwork, &info );
+        LAPACK_zhetrf_aa( &uplo, &n, a_t, &lda_t, ipiv, work, &lwork, &info );
         if( info < 0 ) {
             info = info - 1;
         }

lapack-netlib/LAPACKE/src/lapacke_zsytrf_aa_work.c

@@ -41,7 +41,7 @@ lapack_int LAPACKE_zsytrf_aa_work( int matrix_layout, char uplo, lapack_int n,
     lapack_int info = 0;
     if( matrix_layout == LAPACK_COL_MAJOR ) {
         /* Call LAPACK function and adjust info */
-        LAPACK_zsytrf( &uplo, &n, a, &lda, ipiv, work, &lwork, &info );
+        LAPACK_zsytrf_aa( &uplo, &n, a, &lda, ipiv, work, &lwork, &info );
         if( info < 0 ) {
             info = info - 1;
         }
@@ -56,7 +56,7 @@ lapack_int LAPACKE_zsytrf_aa_work( int matrix_layout, char uplo, lapack_int n,
         }
         /* Query optimal working array(s) size if requested */
         if( lwork == -1 ) {
-            LAPACK_zsytrf( &uplo, &n, a, &lda_t, ipiv, work, &lwork, &info );
+            LAPACK_zsytrf_aa( &uplo, &n, a, &lda_t, ipiv, work, &lwork, &info );
             return (info < 0) ? (info - 1) : info;
         }
         /* Allocate memory for temporary array(s) */
@@ -69,7 +69,7 @@ lapack_int LAPACKE_zsytrf_aa_work( int matrix_layout, char uplo, lapack_int n,
         /* Transpose input matrices */
         LAPACKE_zsy_trans( matrix_layout, uplo, n, a, lda, a_t, lda_t );
         /* Call LAPACK function and adjust info */
-        LAPACK_zsytrf( &uplo, &n, a_t, &lda_t, ipiv, work, &lwork, &info );
+        LAPACK_zsytrf_aa( &uplo, &n, a_t, &lda_t, ipiv, work, &lwork, &info );
         if( info < 0 ) {
             info = info - 1;
         }

lapack-netlib/SRC/cgejsv.f

@@ -701,7 +701,7 @@
           LWSVDJ  = MAX( 2 * N, 1 )         
           LWSVDJV = MAX( 2 * N, 1 )
 *         .. minimal REAL workspace length for CGEQP3, CPOCON, CGESVJ
-          LRWQP3  = N 
+          LRWQP3  = 2 * N 
           LRWCON  = N 
           LRWSVDJ = N 
           IF ( LQUERY ) THEN 
@@ -939,7 +939,7 @@
              END IF 
           END IF
           MINWRK = MAX( 2, MINWRK )
-          OPTWRK = MAX( 2, OPTWRK )
+          OPTWRK = MAX( OPTWRK, MINWRK )
           IF ( LWORK  .LT. MINWRK  .AND. (.NOT.LQUERY) ) INFO = - 17
           IF ( LRWORK .LT. MINRWRK .AND. (.NOT.LQUERY) ) INFO = - 19   
       END IF

lapack-netlib/SRC/chesv_aa.f

@@ -209,6 +209,8 @@
          INFO = -5
       ELSE IF( LDB.LT.MAX( 1, N ) ) THEN
          INFO = -8
+      ELSE IF( LWORK.LT.MAX( 2*N, 3*N-2 ) .AND. .NOT.LQUERY ) THEN
+         INFO = -10
       END IF
 *
       IF( INFO.EQ.0 ) THEN
@@ -219,9 +221,6 @@
          LWKOPT_HETRS = INT( WORK(1) )
          LWKOPT = MAX( LWKOPT_HETRF, LWKOPT_HETRS )
          WORK( 1 ) = LWKOPT
-         IF( LWORK.LT.LWKOPT .AND. .NOT.LQUERY ) THEN
-            INFO = -10
-         END IF
       END IF
 *
       IF( INFO.NE.0 ) THEN

lapack-netlib/SRC/chesv_aa_2stage.f

@@ -105,6 +105,7 @@
 *>
 *> \param[in] LTB
 *> \verbatim
+*>          LTB is INTEGER
 *>          The size of the array TB. LTB >= 4*N, internally
 *>          used to select NB such that LTB >= (3*NB+1)*N.
 *>
@@ -124,7 +125,7 @@
 *>
 *> \param[out] IPIV2
 *> \verbatim
-*>          IPIV is INTEGER array, dimension (N)
+*>          IPIV2 is INTEGER array, dimension (N)
 *>          On exit, it contains the details of the interchanges, i.e.,
 *>          the row and column k of T were interchanged with the
 *>          row and column IPIV(k).
@@ -150,6 +151,7 @@
 *>
 *> \param[in] LWORK
 *> \verbatim
+*>          LWORK is INTEGER
 *>          The size of WORK. LWORK >= N, internally used to select NB
 *>          such that LWORK >= N*NB.
 *>
@@ -233,19 +235,18 @@
          INFO = -3
       ELSE IF( LDA.LT.MAX( 1, N ) ) THEN
          INFO = -5
+      ELSE IF( LTB.LT.( 4*N ) .AND. .NOT.TQUERY ) THEN
+         INFO = -7
       ELSE IF( LDB.LT.MAX( 1, N ) ) THEN
          INFO = -11
+      ELSE IF( LWORK.LT.N .AND. .NOT.WQUERY ) THEN
+         INFO = -13
       END IF
 *
       IF( INFO.EQ.0 ) THEN
          CALL CHETRF_AA_2STAGE( UPLO, N, A, LDA, TB, -1, IPIV,
      $                          IPIV2, WORK, -1, INFO )
          LWKOPT = INT( WORK(1) )
-         IF( LTB.LT.INT( TB(1) ) .AND. .NOT.TQUERY ) THEN
-            INFO = -7
-         ELSE IF( LWORK.LT.LWKOPT .AND. .NOT.WQUERY ) THEN
-            INFO = -13
-         END IF
       END IF
 *
       IF( INFO.NE.0 ) THEN
@@ -270,6 +271,8 @@
       END IF
 *
       WORK( 1 ) = LWKOPT
+*
+      RETURN
 *
 *     End of CHESV_AA_2STAGE
 *

lapack-netlib/SRC/chetrf_aa_2stage.f

@@ -93,6 +93,7 @@
 *>
 *> \param[in] LTB
 *> \verbatim
+*>          LTB is INTEGER
 *>          The size of the array TB. LTB >= 4*N, internally
 *>          used to select NB such that LTB >= (3*NB+1)*N.
 *>
@@ -112,7 +113,7 @@
 *>
 *> \param[out] IPIV2
 *> \verbatim
-*>          IPIV is INTEGER array, dimension (N)
+*>          IPIV2 is INTEGER array, dimension (N)
 *>          On exit, it contains the details of the interchanges, i.e.,
 *>          the row and column k of T were interchanged with the
 *>          row and column IPIV(k).
@@ -125,6 +126,7 @@
 *>
 *> \param[in] LWORK
 *> \verbatim
+*>          LWORK is INTEGER
 *>          The size of WORK. LWORK >= N, internally used to select NB
 *>          such that LWORK >= N*NB.
 *>
@@ -658,6 +660,8 @@ c     $                     (J+1)*NB+1, (J+1)*NB+KB, IPIV, 1 )
 *
 *     Factor the band matrix
       CALL CGBTRF( N, N, NB, NB, TB, LDTB, IPIV2, INFO )
+*
+      RETURN
 *
 *     End of CHETRF_AA_2STAGE
 *

lapack-netlib/SRC/chetrs_aa_2stage.f

@@ -87,6 +87,7 @@
 *>
 *> \param[in] LTB
 *> \verbatim
+*>          LTB is INTEGER
 *>          The size of the array TB. LTB >= 4*N.
 *> \endverbatim
 *>

lapack-netlib/SRC/cla_syamv.f

@@ -241,7 +241,7 @@
          INFO = 10
       END IF
       IF( INFO.NE.0 )THEN
-         CALL XERBLA( 'SSYMV ', INFO )
+         CALL XERBLA( 'CLA_SYAMV', INFO )
          RETURN
       END IF
 *

lapack-netlib/SRC/claqr1.f

@@ -142,6 +142,13 @@
       CABS1( CDUM ) = ABS( REAL( CDUM ) ) + ABS( AIMAG( CDUM ) )
 *     ..
 *     .. Executable Statements ..
+*
+*     Quick return if possible
+*
+      IF( N.NE.2 .AND. N.NE.3 ) THEN
+         RETURN
+      END IF
+*
       IF( N.EQ.2 ) THEN
          S = CABS1( H( 1, 1 )-S2 ) + CABS1( H( 2, 1 ) )
          IF( S.EQ.RZERO ) THEN

lapack-netlib/SRC/csysv_aa.f

@@ -221,9 +221,6 @@
          LWKOPT_SYTRS = INT( WORK(1) )
          LWKOPT = MAX( LWKOPT_SYTRF, LWKOPT_SYTRS )
          WORK( 1 ) = LWKOPT
-         IF( LWORK.LT.LWKOPT .AND. .NOT.LQUERY ) THEN
-            INFO = -10
-         END IF
       END IF
 *
       IF( INFO.NE.0 ) THEN

lapack-netlib/SRC/csysv_aa_2stage.f

@@ -105,6 +105,7 @@
 *>
 *> \param[in] LTB
 *> \verbatim
+*>          LTB is INTEGER
 *>          The size of the array TB. LTB >= 4*N, internally
 *>          used to select NB such that LTB >= (3*NB+1)*N.
 *>
@@ -124,7 +125,7 @@
 *>
 *> \param[out] IPIV2
 *> \verbatim
-*>          IPIV is INTEGER array, dimension (N)
+*>          IPIV2 is INTEGER array, dimension (N)
 *>          On exit, it contains the details of the interchanges, i.e.,
 *>          the row and column k of T were interchanged with the
 *>          row and column IPIV(k).
@@ -150,6 +151,7 @@
 *>
 *> \param[in] LWORK
 *> \verbatim
+*>          LWORK is INTEGER
 *>          The size of WORK. LWORK >= N, internally used to select NB
 *>          such that LWORK >= N*NB.
 *>
@@ -233,19 +235,18 @@
          INFO = -3
       ELSE IF( LDA.LT.MAX( 1, N ) ) THEN
          INFO = -5
+      ELSE IF( LTB.LT.( 4*N ) .AND. .NOT.TQUERY ) THEN
+         INFO = -7
       ELSE IF( LDB.LT.MAX( 1, N ) ) THEN
          INFO = -11
+      ELSE IF( LWORK.LT.N .AND. .NOT.WQUERY ) THEN
+         INFO = -13
       END IF
 *
       IF( INFO.EQ.0 ) THEN
          CALL CSYTRF_AA_2STAGE( UPLO, N, A, LDA, TB, -1, IPIV,
      $                          IPIV2, WORK, -1, INFO )
          LWKOPT = INT( WORK(1) )
-         IF( LTB.LT.INT( TB(1) ) .AND. .NOT.TQUERY ) THEN
-            INFO = -7
-         ELSE IF( LWORK.LT.LWKOPT .AND. .NOT.WQUERY ) THEN
-            INFO = -13
-         END IF
       END IF
 *
       IF( INFO.NE.0 ) THEN
@@ -270,6 +271,8 @@
       END IF
 *
       WORK( 1 ) = LWKOPT
+*
+      RETURN
 *
 *     End of CSYSV_AA_2STAGE
 *

lapack-netlib/SRC/csytrf_aa_2stage.f

@@ -93,6 +93,7 @@
 *>
 *> \param[in] LTB
 *> \verbatim
+*>          LTB is INTEGER
 *>          The size of the array TB. LTB >= 4*N, internally
 *>          used to select NB such that LTB >= (3*NB+1)*N.
 *>
@@ -112,7 +113,7 @@
 *>
 *> \param[out] IPIV2
 *> \verbatim
-*>          IPIV is INTEGER array, dimension (N)
+*>          IPIV2 is INTEGER array, dimension (N)
 *>          On exit, it contains the details of the interchanges, i.e.,
 *>          the row and column k of T were interchanged with the
 *>          row and column IPIV(k).
@@ -125,6 +126,7 @@
 *>
 *> \param[in] LWORK
 *> \verbatim
+*>          LWORK is INTEGER
 *>          The size of WORK. LWORK >= N, internally used to select NB
 *>          such that LWORK >= N*NB.
 *>
@@ -662,6 +664,8 @@ c     $                     (J+1)*NB+1, (J+1)*NB+KB, IPIV, 1 )
 *
 *     Factor the band matrix
       CALL CGBTRF( N, N, NB, NB, TB, LDTB, IPIV2, INFO )
+*
+      RETURN
 *
 *     End of CSYTRF_AA_2STAGE
 *

lapack-netlib/SRC/csytri2.f

@@ -96,11 +96,11 @@
 *>          LWORK is INTEGER
 *>          The dimension of the array WORK.
 *>          WORK is size >= (N+NB+1)*(NB+3)
-*>          If LDWORK = -1, then a workspace query is assumed; the routine
+*>          If LWORK = -1, then a workspace query is assumed; the routine
 *>           calculates:
 *>              - the optimal size of the WORK array, returns
 *>          this value as the first entry of the WORK array,
-*>              - and no error message related to LDWORK is issued by XERBLA.
+*>              - and no error message related to LWORK is issued by XERBLA.
 *> \endverbatim
 *>
 *> \param[out] INFO
@@ -163,7 +163,7 @@
       UPPER = LSAME( UPLO, 'U' )
       LQUERY = ( LWORK.EQ.-1 )
 *     Get blocksize
-      NBMAX = ILAENV( 1, 'CSYTRF', UPLO, N, -1, -1, -1 )
+      NBMAX = ILAENV( 1, 'CSYTRI2', UPLO, N, -1, -1, -1 )
       IF ( NBMAX .GE. N ) THEN
          MINSIZE = N
       ELSE

lapack-netlib/SRC/csytrs_aa_2stage.f

@@ -85,6 +85,7 @@
 *>
 *> \param[in] LTB
 *> \verbatim
+*>          LTB is INTEGER
 *>          The size of the array TB. LTB >= 4*N.
 *> \endverbatim
 *>

lapack-netlib/SRC/ctrevc3.f

@@ -27,8 +27,8 @@
 *       ..
 *       .. Array Arguments ..
 *       LOGICAL            SELECT( * )
-*       REAL   RWORK( * )
-*       COMPLEX         T( LDT, * ), VL( LDVL, * ), VR( LDVR, * ),
+*       REAL               RWORK( * )
+*       COMPLEX            T( LDT, * ), VL( LDVL, * ), VR( LDVR, * ),
 *      $                   WORK( * )
 *       ..
 *
@@ -258,17 +258,17 @@
 *     ..
 *     .. Array Arguments ..
       LOGICAL            SELECT( * )
-      REAL   RWORK( * )
-      COMPLEX         T( LDT, * ), VL( LDVL, * ), VR( LDVR, * ),
+      REAL               RWORK( * )
+      COMPLEX            T( LDT, * ), VL( LDVL, * ), VR( LDVR, * ),
      $                   WORK( * )
 *     ..
 *
 *  =====================================================================
 *
 *     .. Parameters ..
-      REAL   ZERO, ONE
+      REAL               ZERO, ONE
       PARAMETER          ( ZERO = 0.0E+0, ONE = 1.0E+0 )
-      COMPLEX         CZERO, CONE
+      COMPLEX            CZERO, CONE
       PARAMETER          ( CZERO = ( 0.0E+0, 0.0E+0 ),
      $                     CONE  = ( 1.0E+0, 0.0E+0 ) )
       INTEGER            NBMIN, NBMAX
@@ -277,13 +277,13 @@
 *     .. Local Scalars ..
       LOGICAL            ALLV, BOTHV, LEFTV, LQUERY, OVER, RIGHTV, SOMEV
       INTEGER            I, II, IS, J, K, KI, IV, MAXWRK, NB
-      REAL   OVFL, REMAX, SCALE, SMIN, SMLNUM, ULP, UNFL
-      COMPLEX         CDUM
+      REAL               OVFL, REMAX, SCALE, SMIN, SMLNUM, ULP, UNFL
+      COMPLEX            CDUM
 *     ..
 *     .. External Functions ..
       LOGICAL            LSAME
       INTEGER            ILAENV, ICAMAX
-      REAL   SLAMCH, SCASUM
+      REAL               SLAMCH, SCASUM
       EXTERNAL           LSAME, ILAENV, ICAMAX, SLAMCH, SCASUM
 *     ..
 *     .. External Subroutines ..

lapack-netlib/SRC/dgelqt.f

@@ -158,7 +158,7 @@
       INTEGER    I, IB, IINFO, K
 *     ..
 *     .. External Subroutines ..
-      EXTERNAL   DGEQRT2, DGELQT3, DGEQRT3, DLARFB, XERBLA
+      EXTERNAL   DGELQT3, DLARFB, XERBLA
 *     ..
 *     .. Executable Statements ..
 *

lapack-netlib/SRC/dla_syamv.f

@@ -230,7 +230,7 @@
          INFO = 10
       END IF
       IF( INFO.NE.0 )THEN
-         CALL XERBLA( 'DSYMV ', INFO )
+         CALL XERBLA( 'DLA_SYAMV', INFO )
          RETURN
       END IF
 *

lapack-netlib/SRC/dlaqr1.f

@@ -147,6 +147,13 @@
       INTRINSIC          ABS
 *     ..
 *     .. Executable Statements ..
+*
+*     Quick return if possible
+*
+      IF( N.NE.2 .AND. N.NE.3 ) THEN
+         RETURN
+      END IF
+*
       IF( N.EQ.2 ) THEN
          S = ABS( H( 1, 1 )-SR2 ) + ABS( SI2 ) + ABS( H( 2, 1 ) )
          IF( S.EQ.ZERO ) THEN

lapack-netlib/SRC/dsysv_aa.f

@@ -221,9 +221,6 @@
          LWKOPT_SYTRS = INT( WORK(1) )
          LWKOPT = MAX( LWKOPT_SYTRF, LWKOPT_SYTRS )
          WORK( 1 ) = LWKOPT
-         IF( LWORK.LT.LWKOPT .AND. .NOT.LQUERY ) THEN
-            INFO = -10
-         END IF
       END IF
 *
       IF( INFO.NE.0 ) THEN

lapack-netlib/SRC/dsysv_aa_2stage.f

@@ -107,6 +107,7 @@
 *>
 *> \param[in] LTB
 *> \verbatim
+*>          LTB is INTEGER
 *>          The size of the array TB. LTB >= 4*N, internally
 *>          used to select NB such that LTB >= (3*NB+1)*N.
 *>
@@ -126,7 +127,7 @@
 *>
 *> \param[out] IPIV2
 *> \verbatim
-*>          IPIV is INTEGER array, dimension (N)
+*>          IPIV2 is INTEGER array, dimension (N)
 *>          On exit, it contains the details of the interchanges, i.e.,
 *>          the row and column k of T were interchanged with the
 *>          row and column IPIV(k).
@@ -152,6 +153,7 @@
 *>
 *> \param[in] LWORK
 *> \verbatim
+*>          LWORK is INTEGER
 *>          The size of WORK. LWORK >= N, internally used to select NB
 *>          such that LWORK >= N*NB.
 *>
@@ -235,19 +237,18 @@
          INFO = -3
       ELSE IF( LDA.LT.MAX( 1, N ) ) THEN
          INFO = -5
+      ELSE IF( LTB.LT.( 4*N ) .AND. .NOT.TQUERY ) THEN
+         INFO = -7
       ELSE IF( LDB.LT.MAX( 1, N ) ) THEN
          INFO = -11
+      ELSE IF( LWORK.LT.N .AND. .NOT.WQUERY ) THEN
+         INFO = -13
       END IF
 *
       IF( INFO.EQ.0 ) THEN
          CALL DSYTRF_AA_2STAGE( UPLO, N, A, LDA, TB, -1, IPIV,
      $                          IPIV2, WORK, -1, INFO )
          LWKOPT = INT( WORK(1) )
-         IF( LTB.LT.INT( TB(1) ) .AND. .NOT.TQUERY ) THEN
-            INFO = -7
-         ELSE IF( LWORK.LT.LWKOPT .AND. .NOT.WQUERY ) THEN
-            INFO = -13
-         END IF
       END IF
 *
       IF( INFO.NE.0 ) THEN

lapack-netlib/SRC/dsytrf_aa_2stage.f

@@ -93,6 +93,7 @@
 *>
 *> \param[in] LTB
 *> \verbatim
+*>          LTB is INTEGER
 *>          The size of the array TB. LTB >= 4*N, internally
 *>          used to select NB such that LTB >= (3*NB+1)*N.
 *>
@@ -109,6 +110,7 @@
 *>
 *> \param[in] LWORK
 *> \verbatim
+*>          LWORK is INTEGER
 *>          The size of WORK. LWORK >= N, internally used to select NB
 *>          such that LWORK >= N*NB.
 *>
@@ -128,10 +130,10 @@
 *>
 *> \param[out] IPIV2
 *> \verbatim
-*>          IPIV is INTEGER array, dimension (N)
+*>          IPIV2 is INTEGER array, dimension (N)
 *>          On exit, it contains the details of the interchanges, i.e.,
 *>          the row and column k of T were interchanged with the
-*>          row and column IPIV(k).
+*>          row and column IPIV2(k).
 *> \endverbatim
 *>
 *> \param[out] INFO
@@ -641,6 +643,8 @@ c     $                     (J+1)*NB+1, (J+1)*NB+KB, IPIV, 1 )
 *
 *     Factor the band matrix
       CALL DGBTRF( N, N, NB, NB, TB, LDTB, IPIV2, INFO )
+*
+      RETURN
 *
 *     End of DSYTRF_AA_2STAGE
 *

lapack-netlib/SRC/dsytri2.f

@@ -96,11 +96,11 @@
 *>          LWORK is INTEGER
 *>          The dimension of the array WORK.
 *>          WORK is size >= (N+NB+1)*(NB+3)
-*>          If LDWORK = -1, then a workspace query is assumed; the routine
+*>          If LWORK = -1, then a workspace query is assumed; the routine
 *>           calculates:
 *>              - the optimal size of the WORK array, returns
 *>          this value as the first entry of the WORK array,
-*>              - and no error message related to LDWORK is issued by XERBLA.
+*>              - and no error message related to LWORK is issued by XERBLA.
 *> \endverbatim
 *>
 *> \param[out] INFO
@@ -163,7 +163,7 @@
       UPPER = LSAME( UPLO, 'U' )
       LQUERY = ( LWORK.EQ.-1 )
 *     Get blocksize
-      NBMAX = ILAENV( 1, 'DSYTRF', UPLO, N, -1, -1, -1 )
+      NBMAX = ILAENV( 1, 'DSYTRI2', UPLO, N, -1, -1, -1 )
       IF ( NBMAX .GE. N ) THEN
          MINSIZE = N
       ELSE

lapack-netlib/SRC/dsytrs_aa_2stage.f

@@ -85,6 +85,7 @@
 *>
 *> \param[in] LTB
 *> \verbatim
+*>          LTB is INTEGER
 *>          The size of the array TB. LTB >= 4*N.
 *> \endverbatim
 *>

lapack-netlib/SRC/dtrevc3.f

@@ -45,9 +45,9 @@
 *> The right eigenvector x and the left eigenvector y of T corresponding
 *> to an eigenvalue w are defined by:
 *>
-*>    T*x = w*x,     (y**H)*T = w*(y**H)
+*>    T*x = w*x,     (y**T)*T = w*(y**T)
 *>
-*> where y**H denotes the conjugate transpose of y.
+*> where y**T denotes the transpose of the vector y.
 *> The eigenvalues are not input to this routine, but are read directly
 *> from the diagonal blocks of T.
 *>

lapack-netlib/SRC/iparmq.f

@@ -104,13 +104,13 @@
 *>
 *> \param[in] NAME
 *> \verbatim
-*>          NAME is character string
+*>          NAME is CHARACTER string
 *>               Name of the calling subroutine
 *> \endverbatim
 *>
 *> \param[in] OPTS
 *> \verbatim
-*>          OPTS is character string
+*>          OPTS is CHARACTER string
 *>               This is a concatenation of the string arguments to
 *>               TTQRE.
 *> \endverbatim

lapack-netlib/SRC/sla_syamv.f

@@ -230,7 +230,7 @@
          INFO = 10
       END IF
       IF( INFO.NE.0 )THEN
-         CALL XERBLA( 'SSYMV ', INFO )
+         CALL XERBLA( 'SLA_SYAMV', INFO )
          RETURN
       END IF
 *

lapack-netlib/SRC/slaqr1.f

@@ -147,6 +147,13 @@
       INTRINSIC          ABS
 *     ..
 *     .. Executable Statements ..
+*
+*     Quick return if possible
+*
+      IF( N.NE.2 .AND. N.NE.3 ) THEN
+         RETURN
+      END IF
+*
       IF( N.EQ.2 ) THEN
          S = ABS( H( 1, 1 )-SR2 ) + ABS( SI2 ) + ABS( H( 2, 1 ) )
          IF( S.EQ.ZERO ) THEN

lapack-netlib/SRC/ssysv_aa.f

@@ -220,9 +220,6 @@
          LWKOPT_SYTRS = INT( WORK(1) )
          LWKOPT = MAX( LWKOPT_SYTRF, LWKOPT_SYTRS )
          WORK( 1 ) = LWKOPT
-         IF( LWORK.LT.LWKOPT .AND. .NOT.LQUERY ) THEN
-            INFO = -10
-         END IF
       END IF
 *
       IF( INFO.NE.0 ) THEN

lapack-netlib/SRC/ssysv_aa_2stage.f

@@ -106,6 +106,7 @@
 *>
 *> \param[in] LTB
 *> \verbatim
+*>          LTB is INTEGER
 *>          The size of the array TB. LTB >= 4*N, internally
 *>          used to select NB such that LTB >= (3*NB+1)*N.
 *>
@@ -125,7 +126,7 @@
 *>
 *> \param[out] IPIV2
 *> \verbatim
-*>          IPIV is INTEGER array, dimension (N)
+*>          IPIV2 is INTEGER array, dimension (N)
 *>          On exit, it contains the details of the interchanges, i.e.,
 *>          the row and column k of T were interchanged with the
 *>          row and column IPIV(k).
@@ -151,6 +152,7 @@
 *>
 *> \param[in] LWORK
 *> \verbatim
+*>          LWORK is INTEGER
 *>          The size of WORK. LWORK >= N, internally used to select NB
 *>          such that LWORK >= N*NB.
 *>
@@ -234,19 +236,18 @@
          INFO = -3
       ELSE IF( LDA.LT.MAX( 1, N ) ) THEN
          INFO = -5
+      ELSE IF( LTB.LT.( 4*N ) .AND. .NOT.TQUERY ) THEN
+         INFO = -7
       ELSE IF( LDB.LT.MAX( 1, N ) ) THEN
          INFO = -11
+      ELSE IF( LWORK.LT.N .AND. .NOT.WQUERY ) THEN
+         INFO = -13
       END IF
 *
       IF( INFO.EQ.0 ) THEN
          CALL SSYTRF_AA_2STAGE( UPLO, N, A, LDA, TB, -1, IPIV,
      $                          IPIV2, WORK, -1, INFO )
          LWKOPT = INT( WORK(1) )
-         IF( LTB.LT.INT( TB(1) ) .AND. .NOT.TQUERY ) THEN
-            INFO = -7
-         ELSE IF( LWORK.LT.LWKOPT .AND. .NOT.WQUERY ) THEN
-            INFO = -13
-         END IF
       END IF
 *
       IF( INFO.NE.0 ) THEN

lapack-netlib/SRC/ssytrf_aa_2stage.f

@@ -93,6 +93,7 @@
 *>
 *> \param[in] LTB
 *> \verbatim
+*>          LTB is INTEGER
 *>          The size of the array TB. LTB >= 4*N, internally
 *>          used to select NB such that LTB >= (3*NB+1)*N.
 *>
@@ -112,7 +113,7 @@
 *>
 *> \param[out] IPIV2
 *> \verbatim
-*>          IPIV is INTEGER array, dimension (N)
+*>          IPIV2 is INTEGER array, dimension (N)
 *>          On exit, it contains the details of the interchanges, i.e.,
 *>          the row and column k of T were interchanged with the
 *>          row and column IPIV(k).
@@ -125,6 +126,7 @@
 *>
 *> \param[in] LWORK
 *> \verbatim
+*>          LWORK is INTEGER
 *>          The size of WORK. LWORK >= N, internally used to select NB
 *>          such that LWORK >= N*NB.
 *>
@@ -641,6 +643,8 @@ c     $                     (J+1)*NB+1, (J+1)*NB+KB, IPIV, 1 )
 *
 *     Factor the band matrix
       CALL SGBTRF( N, N, NB, NB, TB, LDTB, IPIV2, INFO )
+*
+      RETURN
 *
 *     End of SSYTRF_AA_2STAGE
 *

lapack-netlib/SRC/ssytri2.f

@@ -96,11 +96,11 @@
 *>          LWORK is INTEGER
 *>          The dimension of the array WORK.
 *>          WORK is size >= (N+NB+1)*(NB+3)
-*>          If LDWORK = -1, then a workspace query is assumed; the routine
+*>          If LWORK = -1, then a workspace query is assumed; the routine
 *>           calculates:
 *>              - the optimal size of the WORK array, returns
 *>          this value as the first entry of the WORK array,
-*>              - and no error message related to LDWORK is issued by XERBLA.
+*>              - and no error message related to LWORK is issued by XERBLA.
 *> \endverbatim
 *>
 *> \param[out] INFO

lapack-netlib/SRC/ssytrs_aa_2stage.f

@@ -85,6 +85,7 @@
 *>
 *> \param[in] LTB
 *> \verbatim
+*>          LTB is INTEGER
 *>          The size of the array TB. LTB >= 4*N.
 *> \endverbatim
 *>

lapack-netlib/SRC/strevc3.f

@@ -27,7 +27,7 @@
 *       ..
 *       .. Array Arguments ..
 *       LOGICAL            SELECT( * )
-*       REAL   T( LDT, * ), VL( LDVL, * ), VR( LDVR, * ),
+*       REAL               T( LDT, * ), VL( LDVL, * ), VR( LDVR, * ),
 *      $                   WORK( * )
 *       ..
 *
@@ -45,9 +45,9 @@
 *> The right eigenvector x and the left eigenvector y of T corresponding
 *> to an eigenvalue w are defined by:
 *>
-*>    T*x = w*x,     (y**H)*T = w*(y**H)
+*>    T*x = w*x,     (y**T)*T = w*(y**T)
 *>
-*> where y**H denotes the conjugate transpose of y.
+*> where y**T denotes the transpose of the vector y.
 *> The eigenvalues are not input to this routine, but are read directly
 *> from the diagonal blocks of T.
 *>
@@ -251,14 +251,14 @@
 *     ..
 *     .. Array Arguments ..
       LOGICAL            SELECT( * )
-      REAL   T( LDT, * ), VL( LDVL, * ), VR( LDVR, * ),
+      REAL               T( LDT, * ), VL( LDVL, * ), VR( LDVR, * ),
      $                   WORK( * )
 *     ..
 *
 *  =====================================================================
 *
 *     .. Parameters ..
-      REAL   ZERO, ONE
+      REAL               ZERO, ONE
       PARAMETER          ( ZERO = 0.0E+0, ONE = 1.0E+0 )
       INTEGER            NBMIN, NBMAX
       PARAMETER          ( NBMIN = 8, NBMAX = 128 )
@@ -268,7 +268,7 @@
      $                   RIGHTV, SOMEV
       INTEGER            I, IERR, II, IP, IS, J, J1, J2, JNXT, K, KI,
      $                   IV, MAXWRK, NB, KI2
-      REAL   BETA, BIGNUM, EMAX, OVFL, REC, REMAX, SCALE,
+      REAL               BETA, BIGNUM, EMAX, OVFL, REC, REMAX, SCALE,
      $                   SMIN, SMLNUM, ULP, UNFL, VCRIT, VMAX, WI, WR,
      $                   XNORM
 *     ..

lapack-netlib/SRC/zgejsv.f

@@ -704,7 +704,7 @@
           LWSVDJ  = MAX( 2 * N, 1 )         
           LWSVDJV = MAX( 2 * N, 1 )
 *         .. minimal REAL workspace length for ZGEQP3, ZPOCON, ZGESVJ
-          LRWQP3  = N 
+          LRWQP3  = 2 * N 
           LRWCON  = N 
           LRWSVDJ = N 
           IF ( LQUERY ) THEN 
@@ -942,7 +942,7 @@
              END IF 
           END IF
           MINWRK = MAX( 2, MINWRK )
-          OPTWRK = MAX( 2, OPTWRK )
+          OPTWRK = MAX( MINWRK, OPTWRK )
           IF ( LWORK  .LT. MINWRK  .AND. (.NOT.LQUERY) ) INFO = - 17
           IF ( LRWORK .LT. MINRWRK .AND. (.NOT.LQUERY) ) INFO = - 19   
       END IF

lapack-netlib/SRC/zhesv_aa.f

@@ -209,6 +209,8 @@
          INFO = -5
       ELSE IF( LDB.LT.MAX( 1, N ) ) THEN
          INFO = -8
+      ELSE IF( LWORK.LT.MAX(2*N, 3*N-2) .AND. .NOT.LQUERY ) THEN
+         INFO = -10
       END IF
 *
       IF( INFO.EQ.0 ) THEN
@@ -219,9 +221,6 @@
          LWKOPT_HETRS = INT( WORK(1) )
          LWKOPT = MAX( LWKOPT_HETRF, LWKOPT_HETRS )
          WORK( 1 ) = LWKOPT
-         IF( LWORK.LT.LWKOPT .AND. .NOT.LQUERY ) THEN
-             INFO = -10
-         END IF
       END IF
 *
       IF( INFO.NE.0 ) THEN

lapack-netlib/SRC/zhesv_aa_2stage.f

@@ -106,6 +106,7 @@
 *>
 *> \param[in] LTB
 *> \verbatim
+*>          LTB is INTEGER
 *>          The size of the array TB. LTB >= 4*N, internally
 *>          used to select NB such that LTB >= (3*NB+1)*N.
 *>
@@ -125,7 +126,7 @@
 *>
 *> \param[out] IPIV2
 *> \verbatim
-*>          IPIV is INTEGER array, dimension (N)
+*>          IPIV2 is INTEGER array, dimension (N)
 *>          On exit, it contains the details of the interchanges, i.e.,
 *>          the row and column k of T were interchanged with the
 *>          row and column IPIV(k).
@@ -151,6 +152,7 @@
 *>
 *> \param[in] LWORK
 *> \verbatim
+*>          LWORK is INTEGER
 *>          The size of WORK. LWORK >= N, internally used to select NB
 *>          such that LWORK >= N*NB.
 *>
@@ -240,19 +242,18 @@
          INFO = -3
       ELSE IF( LDA.LT.MAX( 1, N ) ) THEN
          INFO = -5
+      ELSE IF( LTB.LT.( 4*N ) .AND. .NOT.TQUERY ) THEN
+         INFO = -7
       ELSE IF( LDB.LT.MAX( 1, N ) ) THEN
          INFO = -11
+      ELSE IF( LWORK.LT.N .AND. .NOT.WQUERY ) THEN
+         INFO = -13
       END IF
 *
       IF( INFO.EQ.0 ) THEN
          CALL ZHETRF_AA_2STAGE( UPLO, N, A, LDA, TB, -1, IPIV,
      $                          IPIV2, WORK, -1, INFO )
          LWKOPT = INT( WORK(1) )
-         IF( LTB.LT.INT( TB(1) ) .AND. .NOT.TQUERY ) THEN
-            INFO = -7
-         ELSE IF( LWORK.LT.LWKOPT .AND. .NOT.WQUERY ) THEN
-            INFO = -13
-         END IF
       END IF
 *
       IF( INFO.NE.0 ) THEN

lapack-netlib/SRC/zhetrf_aa_2stage.f

@@ -93,6 +93,7 @@
 *>
 *> \param[in] LTB
 *> \verbatim
+*>          LTB is INTEGER
 *>          The size of the array TB. LTB >= 4*N, internally
 *>          used to select NB such that LTB >= (3*NB+1)*N.
 *>
@@ -112,7 +113,7 @@
 *>
 *> \param[out] IPIV2
 *> \verbatim
-*>          IPIV is INTEGER array, dimension (N)
+*>          IPIV2 is INTEGER array, dimension (N)
 *>          On exit, it contains the details of the interchanges, i.e.,
 *>          the row and column k of T were interchanged with the
 *>          row and column IPIV(k).
@@ -125,6 +126,7 @@
 *>
 *> \param[in] LWORK
 *> \verbatim
+*>          LWORK is INTEGER
 *>          The size of WORK. LWORK >= N, internally used to select NB
 *>          such that LWORK >= N*NB.
 *>
@@ -657,6 +659,8 @@ c     $                     (J+1)*NB+1, (J+1)*NB+KB, IPIV, 1 )
 *
 *     Factor the band matrix
       CALL ZGBTRF( N, N, NB, NB, TB, LDTB, IPIV2, INFO )
+*
+      RETURN
 *
 *     End of ZHETRF_AA_2STAGE
 *

lapack-netlib/SRC/zhetrs_aa_2stage.f

@@ -69,7 +69,7 @@
 *>
 *> \param[in] A
 *> \verbatim
-*>          A is COMPLEX*16array, dimension (LDA,N)
+*>          A is COMPLEX*16 array, dimension (LDA,N)
 *>          Details of factors computed by ZHETRF_AA_2STAGE.
 *> \endverbatim
 *>
@@ -81,12 +81,13 @@
 *>
 *> \param[out] TB
 *> \verbatim
-*>          TB is COMPLEX*16array, dimension (LTB)
+*>          TB is COMPLEX*16 array, dimension (LTB)
 *>          Details of factors computed by ZHETRF_AA_2STAGE.
 *> \endverbatim
 *>
 *> \param[in] LTB
 *> \verbatim
+*>          LTB is INTEGER
 *>          The size of the array TB. LTB >= 4*N.
 *> \endverbatim
 *>
@@ -106,7 +107,7 @@
 *>
 *> \param[in,out] B
 *> \verbatim
-*>          B is COMPLEX*16array, dimension (LDB,NRHS)
+*>          B is COMPLEX*16 array, dimension (LDB,NRHS)
 *>          On entry, the right hand side matrix B.
 *>          On exit, the solution matrix X.
 *> \endverbatim

lapack-netlib/SRC/zla_syamv.f

@@ -241,7 +241,7 @@
          INFO = 10
       END IF
       IF( INFO.NE.0 )THEN
-         CALL XERBLA( 'DSYMV ', INFO )
+         CALL XERBLA( 'ZLA_SYAMV', INFO )
          RETURN
       END IF
 *

lapack-netlib/SRC/zlaqr1.f

@@ -142,6 +142,13 @@
       CABS1( CDUM ) = ABS( DBLE( CDUM ) ) + ABS( DIMAG( CDUM ) )
 *     ..
 *     .. Executable Statements ..
+*
+*     Quick return if possible
+*
+      IF( N.NE.2 .AND. N.NE.3 ) THEN
+         RETURN
+      END IF
+*
       IF( N.EQ.2 ) THEN
          S = CABS1( H( 1, 1 )-S2 ) + CABS1( H( 2, 1 ) )
          IF( S.EQ.RZERO ) THEN

lapack-netlib/SRC/zsysv_aa.f

@@ -221,9 +221,6 @@
          LWKOPT_SYTRS = INT( WORK(1) )
          LWKOPT = MAX( LWKOPT_SYTRF, LWKOPT_SYTRS )
          WORK( 1 ) = LWKOPT
-         IF( LWORK.LT.LWKOPT .AND. .NOT.LQUERY ) THEN
-            INFO = -10
-         END IF
       END IF
 *
       IF( INFO.NE.0 ) THEN

lapack-netlib/SRC/zsysv_aa_2stage.f

@@ -105,6 +105,7 @@
 *>
 *> \param[in] LTB
 *> \verbatim
+*>          LTB is INTEGER
 *>          The size of the array TB. LTB >= 4*N, internally
 *>          used to select NB such that LTB >= (3*NB+1)*N.
 *>
@@ -124,7 +125,7 @@
 *>
 *> \param[out] IPIV2
 *> \verbatim
-*>          IPIV is INTEGER array, dimension (N)
+*>          IPIV2 is INTEGER array, dimension (N)
 *>          On exit, it contains the details of the interchanges, i.e.,
 *>          the row and column k of T were interchanged with the
 *>          row and column IPIV(k).
@@ -150,6 +151,7 @@
 *>
 *> \param[in] LWORK
 *> \verbatim
+*>          LWORK is INTEGER
 *>          The size of WORK. LWORK >= N, internally used to select NB
 *>          such that LWORK >= N*NB.
 *>
@@ -233,19 +235,18 @@
          INFO = -3
       ELSE IF( LDA.LT.MAX( 1, N ) ) THEN
          INFO = -5
+      ELSE IF( LTB.LT.( 4*N ) .AND. .NOT.TQUERY ) THEN
+         INFO = -7
       ELSE IF( LDB.LT.MAX( 1, N ) ) THEN
          INFO = -11
+      ELSE IF( LWORK.LT.N .AND. .NOT.WQUERY ) THEN
+         INFO = -13
       END IF
 *
       IF( INFO.EQ.0 ) THEN
          CALL ZSYTRF_AA_2STAGE( UPLO, N, A, LDA, TB, -1, IPIV,
      $                          IPIV2, WORK, -1, INFO )
          LWKOPT = INT( WORK(1) )
-         IF( LTB.LT.INT( TB(1) ) .AND. .NOT.TQUERY ) THEN
-            INFO = -7
-         ELSE IF( LWORK.LT.LWKOPT .AND. .NOT.WQUERY ) THEN
-            INFO = -13
-         END IF
       END IF
 *
       IF( INFO.NE.0 ) THEN

lapack-netlib/SRC/zsytrf_aa_2stage.f

@@ -93,6 +93,7 @@
 *>
 *> \param[in] LTB
 *> \verbatim
+*>          LTB is INTEGER
 *>          The size of the array TB. LTB >= 4*N, internally
 *>          used to select NB such that LTB >= (3*NB+1)*N.
 *>
@@ -112,7 +113,7 @@
 *>
 *> \param[out] IPIV2
 *> \verbatim
-*>          IPIV is INTEGER array, dimension (N)
+*>          IPIV2 is INTEGER array, dimension (N)
 *>          On exit, it contains the details of the interchanges, i.e.,
 *>          the row and column k of T were interchanged with the
 *>          row and column IPIV(k).
@@ -125,6 +126,7 @@
 *>
 *> \param[in] LWORK
 *> \verbatim
+*>          LWORK is INTEGER
 *>          The size of WORK. LWORK >= N, internally used to select NB
 *>          such that LWORK >= N*NB.
 *>
@@ -662,6 +664,8 @@ c     $                     (J+1)*NB+1, (J+1)*NB+KB, IPIV, 1 )
 *
 *     Factor the band matrix
       CALL ZGBTRF( N, N, NB, NB, TB, LDTB, IPIV2, INFO )
+*
+      RETURN
 *
 *     End of ZSYTRF_AA_2STAGE
 *

lapack-netlib/SRC/zsytri2.f

@@ -163,7 +163,7 @@
       UPPER = LSAME( UPLO, 'U' )
       LQUERY = ( LWORK.EQ.-1 )
 *     Get blocksize
-      NBMAX = ILAENV( 1, 'ZSYTRF', UPLO, N, -1, -1, -1 )
+      NBMAX = ILAENV( 1, 'ZSYTRI2', UPLO, N, -1, -1, -1 )
       IF ( NBMAX .GE. N ) THEN
          MINSIZE = N
       ELSE

lapack-netlib/SRC/zsytrs_aa_2stage.f

@@ -85,6 +85,7 @@
 *>
 *> \param[in] LTB
 *> \verbatim
+*>          LTB is INTEGER
 *>          The size of the array TB. LTB >= 4*N.
 *> \endverbatim
 *>

lapack-netlib/TESTING/LIN/dchksy_aa_2stage.f

@@ -218,7 +218,7 @@
 *     ..
 *     .. External Subroutines ..
       EXTERNAL           ALAERH, ALAHD, ALASUM, DERRSY, DLACPY, DLARHS,
-     $                   DLATB4, DLATMS, DPOT02, DSYTRF_AA_2STAGE
+     $                   DLATB4, DLATMS, DPOT02, DSYTRF_AA_2STAGE,
      $                   DSYTRS_AA_2STAGE, XLAENV
 *     ..
 *     .. Intrinsic Functions ..

lapack-netlib/TESTING/LIN/ddrvsy_aa_2stage.f

@@ -204,7 +204,7 @@
 *     .. External Subroutines ..
       EXTERNAL           ALADHD, ALAERH, ALASVM, XLAENV, DERRVX,
      $                   DGET04, DLACPY, DLARHS, DLATB4, DLATMS,
-     $                   DSYSV_AA_2STAGE, CHET01_AA, DPOT02,
+     $                   DSYSV_AA_2STAGE, DPOT02,
      $                   DSYTRF_AA_2STAGE
 *     ..
 *     .. Scalars in Common ..

lapack-netlib/TESTING/LIN/sdrvsy_aa_2stage.f

@@ -203,7 +203,7 @@
 *     ..
 *     .. External Subroutines ..
       EXTERNAL           ALADHD, ALAERH, ALASVM, XLAENV, SERRVX,
-     $                   CGET04, SLACPY, SLARHS, SLATB4, SLATMS,
+     $                   SLACPY, SLARHS, SLATB4, SLATMS,
      $                   SSYSV_AA_2STAGE, SSYT01_AA, SPOT02,
      $                   SSYTRF_AA_2STAGE
 *     ..

lapack-netlib/TESTING/LIN/zchksy_aa_2stage.f

@@ -217,8 +217,8 @@
       DOUBLE PRECISION   RESULT( NTESTS )
 *     ..
 *     .. External Subroutines ..
-      EXTERNAL           ALAERH, ALAHD, ALASUM, CERRSY, ZLACPY, ZLARHS,
-     $                   CLATB4, ZLATMS, ZSYT02, ZSYT01, 
+      EXTERNAL           ALAERH, ALAHD, ALASUM, ZERRSY, ZLACPY, ZLARHS,
+     $                   ZLATB4, ZLATMS, ZSYT02, ZSYT01, 
      $                   ZSYTRF_AA_2STAGE, ZSYTRS_AA_2STAGE,
      $                   XLAENV
 *     ..

lapack-netlib/lapack_testing.py

@@ -257,16 +257,16 @@ for dtype in range_prec:
         else:
             if dtest==16:
                 # LIN TESTS
-                cmdbase="xlintst"+letter+" < "+dtests[0][dtest]+".in > "+dtests[2][dtest]+".out"
+                cmdbase="LIN/xlintst"+letter+" < "+dtests[0][dtest]+".in > "+dtests[2][dtest]+".out"
             elif dtest==17:
                 # PROTO LIN TESTS
-                cmdbase="xlintst"+letter+dtypes[0][dtype-1]+" < "+dtests[0][dtest]+".in > "+dtests[2][dtest]+".out"
+                cmdbase="LIN/xlintst"+letter+dtypes[0][dtype-1]+" < "+dtests[0][dtest]+".in > "+dtests[2][dtest]+".out"
             elif dtest==18:
                 # PROTO LIN TESTS
-                cmdbase="xlintstrf"+letter+" < "+dtests[0][dtest]+".in > "+dtests[2][dtest]+".out"
+                cmdbase="LIN/xlintstrf"+letter+" < "+dtests[0][dtest]+".in > "+dtests[2][dtest]+".out"
             else:
                 # EIG TESTS
-                cmdbase="xeigtst"+letter+" < "+dtests[0][dtest]+".in > "+dtests[2][dtest]+".out"
+                cmdbase="EIG/xeigtst"+letter+" < "+dtests[0][dtest]+".in > "+dtests[2][dtest]+".out"
         if (not just_errors and not short_summary):
             print("Testing "+name+" "+dtests[1][dtest]+"-"+cmdbase, end=' ')
         # Run the process: either to read the file or run the LAPACK testing