Refs #330. Fixed the compatible issue with clang on Mac OSX.

kernel/x86_64/dgemm_kernel_16x2_haswell.S

@@ -192,8 +192,8 @@
 	VFMADD231PD_  	%ymm13,%ymm1,%ymm0
 	VFMADD231PD_  	%ymm14,%ymm2,%ymm0
 	VFMADD231PD_  	%ymm15,%ymm3,%ymm0
-	addq	$3*SIZE , BO	
-	addq	$16*SIZE, AO
+	addq	$ 3*SIZE , BO	
+	addq	$ 16*SIZE, AO
 .endm
 
 
@@ -212,8 +212,8 @@
 	VFMADD231PD_  	%ymm8,%ymm2,%ymm0
 	VFMADD231PD_  	%ymm9,%ymm3,%ymm0
 	prefetcht0	B_PR1(BO)
-	addq	$3*SIZE , BO	
-	addq	$8*SIZE, AO
+	addq	$ 3*SIZE , BO	
+	addq	$ 8*SIZE, AO
 .endm
 
 .macro KERNEL4x3_SUBN
@@ -224,8 +224,8 @@
 	VFMADD231PD_  	%ymm5,%ymm2,%ymm0
 	vbroadcastsd	-10 * SIZE(BO), %ymm3
 	VFMADD231PD_  	%ymm6,%ymm3,%ymm0
-	addq	$3*SIZE , BO	
-	addq	$4*SIZE, AO
+	addq	$ 3*SIZE , BO	
+	addq	$ 4*SIZE, AO
 .endm
 
 .macro KERNEL2x3_SUBN
@@ -240,8 +240,8 @@
 	VFMADD231SD_  	%xmm8,%xmm1,%xmm0
 	VFMADD231SD_  	%xmm10,%xmm2,%xmm0
 	VFMADD231SD_  	%xmm12,%xmm3,%xmm0
-	addq	$3*SIZE , BO	
-	addq	$2*SIZE, AO
+	addq	$ 3*SIZE , BO	
+	addq	$ 2*SIZE, AO
 .endm
 
 .macro KERNEL1x3_SUBN
@@ -252,8 +252,8 @@
 	VFMADD231SD_  	%xmm5,%xmm2,%xmm0
 	vmovsd	-10 * SIZE(BO), %xmm3
 	VFMADD231SD_  	%xmm6,%xmm3,%xmm0
-	addq	$3*SIZE , BO	
-	addq	$1*SIZE, AO
+	addq	$ 3*SIZE , BO	
+	addq	$ 1*SIZE, AO
 .endm
 
 
@@ -1602,16 +1602,16 @@
 	vmovsd	  1 * SIZE(BO, BI, SIZE), %xmm1
 	vmovsd 	-29 * SIZE(AO, %rax, SIZE), %xmm0
 	VFMADD231SD_  	%xmm4,%xmm1,%xmm0
-	addq	$4, BI				 
-	addq	$4, %rax 			 
+	addq	$ 4, BI				 
+	addq	$ 4, %rax 			 
 .endm
 
 .macro KERNEL1x1_SUB
 	vmovsd	 -2 * SIZE(BO, BI, SIZE), %xmm1
 	vmovsd 	-32 * SIZE(AO, %rax, SIZE), %xmm0
 	VFMADD231SD_  	%xmm4,%xmm1,%xmm0
-	addq	$1, BI				 
-	addq	$1 , %rax 			 
+	addq	$ 1, BI				 
+	addq	$ 1 , %rax 			 
 .endm
 
 .macro SAVE1x1
@@ -1749,9 +1749,9 @@
 	vmovsd	%xmm5, 8*SIZE(BO)
 	vmovups	%xmm6, 9*SIZE(BO)
 	vmovsd	%xmm7,11*SIZE(BO)
-	addq	$8*SIZE,BO1
-	addq	$8*SIZE,BO2
-	addq	$12*SIZE,BO
+	addq	$ 8*SIZE,BO1
+	addq	$ 8*SIZE,BO2
+	addq	$ 12*SIZE,BO
 
 	vmovups	0 * SIZE(BO1), %xmm0
 	vmovups	2 * SIZE(BO1), %xmm2
@@ -1769,9 +1769,9 @@
 	vmovsd	%xmm5, 8*SIZE(BO)
 	vmovups	%xmm6, 9*SIZE(BO)
 	vmovsd	%xmm7,11*SIZE(BO)
-	addq	$8*SIZE,BO1
-	addq	$8*SIZE,BO2
-	addq	$12*SIZE,BO
+	addq	$ 8*SIZE,BO1
+	addq	$ 8*SIZE,BO2
+	addq	$ 12*SIZE,BO
 
 	decq	%rax
 	jnz	.L6_01a_1
@@ -1792,9 +1792,9 @@
 	vmovsd  0 * SIZE(BO2), %xmm2
 	vmovups	%xmm0, 0*SIZE(BO)
 	vmovsd	%xmm2, 2*SIZE(BO)
-	addq	$2*SIZE,BO1
-	addq	$2*SIZE,BO2
-	addq	$3*SIZE,BO
+	addq	$ 2*SIZE,BO1
+	addq	$ 2*SIZE,BO2
+	addq	$ 3*SIZE,BO
 	decq	%rax
 	jnz	.L6_02b
 

kernel/x86_64/dgemm_kernel_4x4_haswell.S

@@ -107,22 +107,22 @@ USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
 #if defined(OS_WINDOWS)
 #if   L_BUFFER_SIZE > 16384
 #define STACK_TOUCH \
-        movl    $0,  4096 * 4(%rsp);\
-        movl    $0,  4096 * 3(%rsp);\
-        movl    $0,  4096 * 2(%rsp);\
-        movl    $0,  4096 * 1(%rsp);
+        movl    $ 0,  4096 * 4(%rsp);\
+        movl    $ 0,  4096 * 3(%rsp);\
+        movl    $ 0,  4096 * 2(%rsp);\
+        movl    $ 0,  4096 * 1(%rsp);
 #elif L_BUFFER_SIZE > 12288
 #define STACK_TOUCH \
-        movl    $0,  4096 * 3(%rsp);\
-        movl    $0,  4096 * 2(%rsp);\
-        movl    $0,  4096 * 1(%rsp);
+        movl    $ 0,  4096 * 3(%rsp);\
+        movl    $ 0,  4096 * 2(%rsp);\
+        movl    $ 0,  4096 * 1(%rsp);
 #elif L_BUFFER_SIZE > 8192
 #define STACK_TOUCH \
-        movl    $0,  4096 * 2(%rsp);\
-        movl    $0,  4096 * 1(%rsp);
+        movl    $ 0,  4096 * 2(%rsp);\
+        movl    $ 0,  4096 * 1(%rsp);
 #elif L_BUFFER_SIZE > 4096
 #define STACK_TOUCH \
-        movl    $0,  4096 * 1(%rsp);
+        movl    $ 0,  4096 * 1(%rsp);
 #else
 #define STACK_TOUCH
 #endif
@@ -168,17 +168,17 @@ USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
 	vmulpd  	%ymm0 ,%ymm2  , %ymm8
 	vmulpd  	%ymm0 ,%ymm3  , %ymm12
 	prefetcht0	B_PR1+256(BO)
-	vpermpd		$0xb1, %ymm0  , %ymm0
+	vpermpd		$ 0xb1, %ymm0  , %ymm0
 	vmulpd  	%ymm0 ,%ymm1  , %ymm5
 	vmulpd  	%ymm0 ,%ymm2  , %ymm9
 	vmulpd  	%ymm0 ,%ymm3  , %ymm13
-	vpermpd		$0x1b, %ymm0  , %ymm0
+	vpermpd		$ 0x1b, %ymm0  , %ymm0
 	vmulpd  	%ymm0 ,%ymm1  , %ymm6
 	vmulpd  	%ymm0 ,%ymm2  , %ymm10
 
-	addq		$12*SIZE, BO
+	addq		$ 12*SIZE, BO
 	vmulpd  	%ymm0 ,%ymm3  , %ymm14
-	vpermpd		$0xb1, %ymm0  , %ymm0
+	vpermpd		$ 0xb1, %ymm0  , %ymm0
 	vmulpd  	%ymm0 ,%ymm1  , %ymm7
 	vmovups		-12 * SIZE(BO), %ymm1
 	vmulpd  	%ymm0 ,%ymm2  , %ymm11
@@ -197,16 +197,16 @@ USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
 	vfmadd231pd  	%ymm0 ,%ymm2  , %ymm8
 	prefetcht0	B_PR1+128(BO)
 	vfmadd231pd  	%ymm0 ,%ymm3  , %ymm12
-	vpermpd		$0xb1, %ymm0  , %ymm0
+	vpermpd		$ 0xb1, %ymm0  , %ymm0
 	vfmadd231pd  	%ymm0 ,%ymm1  , %ymm5
 	vfmadd231pd  	%ymm0 ,%ymm2  , %ymm9
 	vfmadd231pd  	%ymm0 ,%ymm3  , %ymm13
-	vpermpd		$0x1b, %ymm0  , %ymm0
+	vpermpd		$ 0x1b, %ymm0  , %ymm0
 	vfmadd231pd  	%ymm0 ,%ymm1  , %ymm6
 	vfmadd231pd  	%ymm0 ,%ymm2  , %ymm10
 
 	vfmadd231pd  	%ymm0 ,%ymm3  , %ymm14
-	vpermpd		$0xb1, %ymm0  , %ymm0
+	vpermpd		$ 0xb1, %ymm0  , %ymm0
 	vfmadd231pd  	%ymm0 ,%ymm1  , %ymm7
 	vmovups		-12 * SIZE(BO), %ymm1
 	vfmadd231pd  	%ymm0 ,%ymm2  , %ymm11
@@ -221,24 +221,24 @@ USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
 	vfmadd231pd  	%ymm0 ,%ymm1  , %ymm4
 	vfmadd231pd  	%ymm0 ,%ymm2  , %ymm8
 	vfmadd231pd  	%ymm0 ,%ymm3  , %ymm12
-	vpermpd		$0xb1, %ymm0  , %ymm0
+	vpermpd		$ 0xb1, %ymm0  , %ymm0
 	vfmadd231pd  	%ymm0 ,%ymm1  , %ymm5
 	vfmadd231pd  	%ymm0 ,%ymm2  , %ymm9
 	vfmadd231pd  	%ymm0 ,%ymm3  , %ymm13
-	vpermpd		$0x1b, %ymm0  , %ymm0
+	vpermpd		$ 0x1b, %ymm0  , %ymm0
 	vfmadd231pd  	%ymm0 ,%ymm1  , %ymm6
 	vfmadd231pd  	%ymm0 ,%ymm2  , %ymm10
 
-	addq		$8*SIZE, AO
+	addq		$ 8*SIZE, AO
 	vfmadd231pd  	%ymm0 ,%ymm3  , %ymm14
-	vpermpd		$0xb1, %ymm0  , %ymm0
+	vpermpd		$ 0xb1, %ymm0  , %ymm0
 	vfmadd231pd  	%ymm0 ,%ymm1  , %ymm7
 	vmovups		  0 * SIZE(BO), %ymm1
 	vfmadd231pd  	%ymm0 ,%ymm2  , %ymm11
 	vmovups		  4 * SIZE(BO), %ymm2
 	vfmadd231pd  	%ymm0 ,%ymm3  , %ymm15
 	vmovups		  8 * SIZE(BO), %ymm3
-	addq		$24*SIZE, BO
+	addq		$ 24*SIZE, BO
 .endm
 
 
@@ -247,21 +247,21 @@ USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
 	vfmadd231pd  	%ymm0 ,%ymm1  , %ymm4
 	vfmadd231pd  	%ymm0 ,%ymm2  , %ymm8
 	vfmadd231pd  	%ymm0 ,%ymm3  , %ymm12
-	vpermpd		$0xb1, %ymm0  , %ymm0
+	vpermpd		$ 0xb1, %ymm0  , %ymm0
 	vfmadd231pd  	%ymm0 ,%ymm1  , %ymm5
 	vfmadd231pd  	%ymm0 ,%ymm2  , %ymm9
 	vfmadd231pd  	%ymm0 ,%ymm3  , %ymm13
-	vpermpd		$0x1b, %ymm0  , %ymm0
+	vpermpd		$ 0x1b, %ymm0  , %ymm0
 	vfmadd231pd  	%ymm0 ,%ymm1  , %ymm6
 	vfmadd231pd  	%ymm0 ,%ymm2  , %ymm10
 
-	addq		$8*SIZE, AO
+	addq		$ 8*SIZE, AO
 	vfmadd231pd  	%ymm0 ,%ymm3  , %ymm14
-	vpermpd		$0xb1, %ymm0  , %ymm0
+	vpermpd		$ 0xb1, %ymm0  , %ymm0
 	vfmadd231pd  	%ymm0 ,%ymm1  , %ymm7
 	vfmadd231pd  	%ymm0 ,%ymm2  , %ymm11
 	vfmadd231pd  	%ymm0 ,%ymm3  , %ymm15
-	addq		$12*SIZE, BO
+	addq		$ 12*SIZE, BO
 .endm
 
 .macro KERNEL4x12_SUB
@@ -272,17 +272,17 @@ USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
 	vfmadd231pd  	%ymm0 ,%ymm2  , %ymm8
 	vmovups		 -4 * SIZE(BO), %ymm3
 	vfmadd231pd  	%ymm0 ,%ymm3  , %ymm12
-	vpermpd		$0xb1, %ymm0  , %ymm0
+	vpermpd		$ 0xb1, %ymm0  , %ymm0
 	vfmadd231pd  	%ymm0 ,%ymm1  , %ymm5
 	vfmadd231pd  	%ymm0 ,%ymm2  , %ymm9
-	addq		$12*SIZE, BO
+	addq		$ 12*SIZE, BO
 	vfmadd231pd  	%ymm0 ,%ymm3  , %ymm13
-	vpermpd		$0x1b, %ymm0  , %ymm0
+	vpermpd		$ 0x1b, %ymm0  , %ymm0
 	vfmadd231pd  	%ymm0 ,%ymm1  , %ymm6
 	vfmadd231pd  	%ymm0 ,%ymm2  , %ymm10
-	addq		$4*SIZE, AO
+	addq		$ 4*SIZE, AO
 	vfmadd231pd  	%ymm0 ,%ymm3  , %ymm14
-	vpermpd		$0xb1, %ymm0  , %ymm0
+	vpermpd		$ 0xb1, %ymm0  , %ymm0
 	vfmadd231pd  	%ymm0 ,%ymm1  , %ymm7
 	vfmadd231pd  	%ymm0 ,%ymm2  , %ymm11
 	vfmadd231pd  	%ymm0 ,%ymm3  , %ymm15
@@ -309,23 +309,23 @@ USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
 	vmulpd	%ymm0 , %ymm14, %ymm14
 	vmulpd	%ymm0 , %ymm15, %ymm15
 
-	vpermpd $0xb1 , %ymm5, %ymm5
-	vpermpd $0xb1 , %ymm7, %ymm7
+	vpermpd $ 0xb1 , %ymm5, %ymm5
+	vpermpd $ 0xb1 , %ymm7, %ymm7
 
-	vblendpd $0x0a, %ymm5, %ymm4, %ymm0
-	vblendpd $0x05, %ymm5, %ymm4, %ymm1
-	vblendpd $0x0a, %ymm7, %ymm6, %ymm2
-	vblendpd $0x05, %ymm7, %ymm6, %ymm3
+	vblendpd $ 0x0a, %ymm5, %ymm4, %ymm0
+	vblendpd $ 0x05, %ymm5, %ymm4, %ymm1
+	vblendpd $ 0x0a, %ymm7, %ymm6, %ymm2
+	vblendpd $ 0x05, %ymm7, %ymm6, %ymm3
 
-	vpermpd $0x1b , %ymm2, %ymm2
-	vpermpd $0x1b , %ymm3, %ymm3
-	vpermpd $0xb1 , %ymm2, %ymm2
-	vpermpd $0xb1 , %ymm3, %ymm3
+	vpermpd $ 0x1b , %ymm2, %ymm2
+	vpermpd $ 0x1b , %ymm3, %ymm3
+	vpermpd $ 0xb1 , %ymm2, %ymm2
+	vpermpd $ 0xb1 , %ymm3, %ymm3
 
-	vblendpd $0x03, %ymm0, %ymm2 , %ymm4
-	vblendpd $0x03, %ymm1, %ymm3 , %ymm5
-	vblendpd $0x03, %ymm2, %ymm0 , %ymm6
-	vblendpd $0x03, %ymm3, %ymm1 , %ymm7
+	vblendpd $ 0x03, %ymm0, %ymm2 , %ymm4
+	vblendpd $ 0x03, %ymm1, %ymm3 , %ymm5
+	vblendpd $ 0x03, %ymm2, %ymm0 , %ymm6
+	vblendpd $ 0x03, %ymm3, %ymm1 , %ymm7
 
         leaq    (CO1, LDC, 2), %rax     
 	
@@ -349,23 +349,23 @@ USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
 	prefetcht0	32(%rax)
 	prefetcht0	32(%rax,LDC)
 
-	vpermpd $0xb1 , %ymm9 , %ymm9
-	vpermpd $0xb1 , %ymm11, %ymm11
+	vpermpd $ 0xb1 , %ymm9 , %ymm9
+	vpermpd $ 0xb1 , %ymm11, %ymm11
 
-	vblendpd $0x0a, %ymm9 , %ymm8 , %ymm0
-	vblendpd $0x05, %ymm9 , %ymm8 , %ymm1
-	vblendpd $0x0a, %ymm11, %ymm10, %ymm2
-	vblendpd $0x05, %ymm11, %ymm10, %ymm3
+	vblendpd $ 0x0a, %ymm9 , %ymm8 , %ymm0
+	vblendpd $ 0x05, %ymm9 , %ymm8 , %ymm1
+	vblendpd $ 0x0a, %ymm11, %ymm10, %ymm2
+	vblendpd $ 0x05, %ymm11, %ymm10, %ymm3
 
-	vpermpd $0x1b , %ymm2, %ymm2
-	vpermpd $0x1b , %ymm3, %ymm3
-	vpermpd $0xb1 , %ymm2, %ymm2
-	vpermpd $0xb1 , %ymm3, %ymm3
+	vpermpd $ 0x1b , %ymm2, %ymm2
+	vpermpd $ 0x1b , %ymm3, %ymm3
+	vpermpd $ 0xb1 , %ymm2, %ymm2
+	vpermpd $ 0xb1 , %ymm3, %ymm3
 
-	vblendpd $0x03, %ymm0, %ymm2 , %ymm4
-	vblendpd $0x03, %ymm1, %ymm3 , %ymm5
-	vblendpd $0x03, %ymm2, %ymm0 , %ymm6
-	vblendpd $0x03, %ymm3, %ymm1 , %ymm7
+	vblendpd $ 0x03, %ymm0, %ymm2 , %ymm4
+	vblendpd $ 0x03, %ymm1, %ymm3 , %ymm5
+	vblendpd $ 0x03, %ymm2, %ymm0 , %ymm6
+	vblendpd $ 0x03, %ymm3, %ymm1 , %ymm7
 
 
 	leaq	(%rax, LDC, 2), %rax
@@ -390,23 +390,23 @@ USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
 	prefetcht0	32(%rbp)
 	prefetcht0	32(%rbp,LDC)
 
-	vpermpd $0xb1 , %ymm13, %ymm13
-	vpermpd $0xb1 , %ymm15, %ymm15
+	vpermpd $ 0xb1 , %ymm13, %ymm13
+	vpermpd $ 0xb1 , %ymm15, %ymm15
 
-	vblendpd $0x0a, %ymm13, %ymm12, %ymm0
-	vblendpd $0x05, %ymm13, %ymm12, %ymm1
-	vblendpd $0x0a, %ymm15, %ymm14, %ymm2
-	vblendpd $0x05, %ymm15, %ymm14, %ymm3
+	vblendpd $ 0x0a, %ymm13, %ymm12, %ymm0
+	vblendpd $ 0x05, %ymm13, %ymm12, %ymm1
+	vblendpd $ 0x0a, %ymm15, %ymm14, %ymm2
+	vblendpd $ 0x05, %ymm15, %ymm14, %ymm3
 
-	vpermpd $0x1b , %ymm2, %ymm2
-	vpermpd $0x1b , %ymm3, %ymm3
-	vpermpd $0xb1 , %ymm2, %ymm2
-	vpermpd $0xb1 , %ymm3, %ymm3
+	vpermpd $ 0x1b , %ymm2, %ymm2
+	vpermpd $ 0x1b , %ymm3, %ymm3
+	vpermpd $ 0xb1 , %ymm2, %ymm2
+	vpermpd $ 0xb1 , %ymm3, %ymm3
 
-	vblendpd $0x03, %ymm0, %ymm2 , %ymm4
-	vblendpd $0x03, %ymm1, %ymm3 , %ymm5
-	vblendpd $0x03, %ymm2, %ymm0 , %ymm6
-	vblendpd $0x03, %ymm3, %ymm1 , %ymm7
+	vblendpd $ 0x03, %ymm0, %ymm2 , %ymm4
+	vblendpd $ 0x03, %ymm1, %ymm3 , %ymm5
+	vblendpd $ 0x03, %ymm2, %ymm0 , %ymm6
+	vblendpd $ 0x03, %ymm3, %ymm1 , %ymm7
 
 
 	leaq	(%rax, LDC, 4), %rax
@@ -431,7 +431,7 @@ USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
 	prefetcht0	32(%rbp)
 	prefetcht0	32(%rbp,LDC)
 
-	addq	$4*SIZE, CO1
+	addq	$ 4*SIZE, CO1
 .endm
 
 /******************************************************************************************/
@@ -477,9 +477,9 @@ USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
 	vfmadd231pd  	%xmm0 ,%xmm3  , %xmm12
 	vmovddup	 -1 * SIZE(BO), %xmm3
 	vfmadd231pd  	%xmm0 ,%xmm1  , %xmm13
-	addq		$12*SIZE, BO
+	addq		$ 12*SIZE, BO
 	vfmadd231pd  	%xmm0 ,%xmm2  , %xmm14
-	addq		$2*SIZE, AO
+	addq		$ 2*SIZE, AO
 	vfmadd231pd  	%xmm0 ,%xmm3  , %xmm15
 
 .endm
@@ -557,7 +557,7 @@ USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
 	vmovups	%xmm6 ,  	(%rbp)
 	vmovups	%xmm7 ,  	(%rbp, LDC)
 
-	addq	$2*SIZE, CO1
+	addq	$ 2*SIZE, CO1
 .endm
 
 
@@ -604,9 +604,9 @@ USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
 	vfmadd231sd  	%xmm0 ,%xmm3  , %xmm12
 	vmovsd	 -1 * SIZE(BO), %xmm3
 	vfmadd231sd  	%xmm0 ,%xmm1  , %xmm13
-	addq		$12*SIZE, BO
+	addq		$ 12*SIZE, BO
 	vfmadd231sd  	%xmm0 ,%xmm2  , %xmm14
-	addq		$1*SIZE, AO
+	addq		$ 1*SIZE, AO
 	vfmadd231sd  	%xmm0 ,%xmm3  , %xmm15
 
 .endm
@@ -684,7 +684,7 @@ USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
 	vmovsd	%xmm6 ,  	(%rbp)
 	vmovsd	%xmm7 ,  	(%rbp, LDC)
 
-	addq	$1*SIZE, CO1
+	addq	$ 1*SIZE, CO1
 .endm
 
 
@@ -707,13 +707,13 @@ USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
 	vmovups		-12 * SIZE(BO), %ymm1
 	vmovups 	-16 * SIZE(AO), %ymm0
 	vmulpd  	%ymm0 ,%ymm1  , %ymm4
-	vpermpd		$0xb1, %ymm0  , %ymm0
+	vpermpd		$ 0xb1, %ymm0  , %ymm0
 	vmulpd  	%ymm0 ,%ymm1  , %ymm5
-	vpermpd		$0x1b, %ymm0  , %ymm0
+	vpermpd		$ 0x1b, %ymm0  , %ymm0
 	vmulpd  	%ymm0 ,%ymm1  , %ymm6
 
-	addq		$4*SIZE, BO
-	vpermpd		$0xb1, %ymm0  , %ymm0
+	addq		$ 4*SIZE, BO
+	vpermpd		$ 0xb1, %ymm0  , %ymm0
 	vmulpd  	%ymm0 ,%ymm1  , %ymm7
 	vmovups		-12 * SIZE(BO), %ymm1
 
@@ -723,12 +723,12 @@ USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
 	prefetcht0	A_PR1(AO)
 	vmovups 	-16 * SIZE(AO), %ymm0
 	vfmadd231pd  	%ymm0 ,%ymm1  , %ymm4
-	vpermpd		$0xb1, %ymm0  , %ymm0
+	vpermpd		$ 0xb1, %ymm0  , %ymm0
 	vfmadd231pd  	%ymm0 ,%ymm1  , %ymm5
-	vpermpd		$0x1b, %ymm0  , %ymm0
+	vpermpd		$ 0x1b, %ymm0  , %ymm0
 	vfmadd231pd  	%ymm0 ,%ymm1  , %ymm6
 
-	vpermpd		$0xb1, %ymm0  , %ymm0
+	vpermpd		$ 0xb1, %ymm0  , %ymm0
 	vfmadd231pd  	%ymm0 ,%ymm1  , %ymm7
 	vmovups		-12 * SIZE(BO), %ymm1
 
@@ -737,44 +737,44 @@ USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
 .macro KERNEL4x4_M2
 	vmovups 	-12 * SIZE(AO), %ymm0
 	vfmadd231pd  	%ymm0 ,%ymm1  , %ymm4
-	vpermpd		$0xb1, %ymm0  , %ymm0
+	vpermpd		$ 0xb1, %ymm0  , %ymm0
 	vfmadd231pd  	%ymm0 ,%ymm1  , %ymm5
-	vpermpd		$0x1b, %ymm0  , %ymm0
+	vpermpd		$ 0x1b, %ymm0  , %ymm0
 	vfmadd231pd  	%ymm0 ,%ymm1  , %ymm6
 
-	addq		$8*SIZE, AO
-	vpermpd		$0xb1, %ymm0  , %ymm0
+	addq		$ 8*SIZE, AO
+	vpermpd		$ 0xb1, %ymm0  , %ymm0
 	vfmadd231pd  	%ymm0 ,%ymm1  , %ymm7
 	vmovups		 -8 * SIZE(BO), %ymm1
-	addq		$8*SIZE, BO
+	addq		$ 8*SIZE, BO
 .endm
 
 
 .macro KERNEL4x4_E
 	vmovups 	-12 * SIZE(AO), %ymm0
 	vfmadd231pd  	%ymm0 ,%ymm1  , %ymm4
-	vpermpd		$0xb1, %ymm0  , %ymm0
+	vpermpd		$ 0xb1, %ymm0  , %ymm0
 	vfmadd231pd  	%ymm0 ,%ymm1  , %ymm5
-	vpermpd		$0x1b, %ymm0  , %ymm0
+	vpermpd		$ 0x1b, %ymm0  , %ymm0
 	vfmadd231pd  	%ymm0 ,%ymm1  , %ymm6
 
-	addq		$8*SIZE, AO
-	vpermpd		$0xb1, %ymm0  , %ymm0
+	addq		$ 8*SIZE, AO
+	vpermpd		$ 0xb1, %ymm0  , %ymm0
 	vfmadd231pd  	%ymm0 ,%ymm1  , %ymm7
-	addq		$4*SIZE, BO
+	addq		$ 4*SIZE, BO
 .endm
 
 .macro KERNEL4x4_SUB
 	vmovups		-12 * SIZE(BO), %ymm1
 	vmovups 	-16 * SIZE(AO), %ymm0
 	vfmadd231pd  	%ymm0 ,%ymm1  , %ymm4
-	vpermpd		$0xb1, %ymm0  , %ymm0
+	vpermpd		$ 0xb1, %ymm0  , %ymm0
 	vfmadd231pd  	%ymm0 ,%ymm1  , %ymm5
-	addq		$4*SIZE, BO
-	vpermpd		$0x1b, %ymm0  , %ymm0
+	addq		$ 4*SIZE, BO
+	vpermpd		$ 0x1b, %ymm0  , %ymm0
 	vfmadd231pd  	%ymm0 ,%ymm1  , %ymm6
-	addq		$4*SIZE, AO
-	vpermpd		$0xb1, %ymm0  , %ymm0
+	addq		$ 4*SIZE, AO
+	vpermpd		$ 0xb1, %ymm0  , %ymm0
 	vfmadd231pd  	%ymm0 ,%ymm1  , %ymm7
 
 .endm
@@ -788,23 +788,23 @@ USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
 	vmulpd	%ymm0 , %ymm5 , %ymm5
 	vmulpd	%ymm0 , %ymm6 , %ymm6
 
-	vpermpd $0xb1 , %ymm5, %ymm5
-	vpermpd $0xb1 , %ymm7, %ymm7
+	vpermpd $ 0xb1 , %ymm5, %ymm5
+	vpermpd $ 0xb1 , %ymm7, %ymm7
 
-	vblendpd $0x0a, %ymm5, %ymm4, %ymm0
-	vblendpd $0x05, %ymm5, %ymm4, %ymm1
-	vblendpd $0x0a, %ymm7, %ymm6, %ymm2
-	vblendpd $0x05, %ymm7, %ymm6, %ymm3
+	vblendpd $ 0x0a, %ymm5, %ymm4, %ymm0
+	vblendpd $ 0x05, %ymm5, %ymm4, %ymm1
+	vblendpd $ 0x0a, %ymm7, %ymm6, %ymm2
+	vblendpd $ 0x05, %ymm7, %ymm6, %ymm3
 
-	vpermpd $0x1b , %ymm2, %ymm2
-	vpermpd $0x1b , %ymm3, %ymm3
-	vpermpd $0xb1 , %ymm2, %ymm2
-	vpermpd $0xb1 , %ymm3, %ymm3
+	vpermpd $ 0x1b , %ymm2, %ymm2
+	vpermpd $ 0x1b , %ymm3, %ymm3
+	vpermpd $ 0xb1 , %ymm2, %ymm2
+	vpermpd $ 0xb1 , %ymm3, %ymm3
 
-	vblendpd $0x03, %ymm0, %ymm2 , %ymm4
-	vblendpd $0x03, %ymm1, %ymm3 , %ymm5
-	vblendpd $0x03, %ymm2, %ymm0 , %ymm6
-	vblendpd $0x03, %ymm3, %ymm1 , %ymm7
+	vblendpd $ 0x03, %ymm0, %ymm2 , %ymm4
+	vblendpd $ 0x03, %ymm1, %ymm3 , %ymm5
+	vblendpd $ 0x03, %ymm2, %ymm0 , %ymm6
+	vblendpd $ 0x03, %ymm3, %ymm1 , %ymm7
 
         leaq    (CO1, LDC, 2), %rax     
 	
@@ -823,7 +823,7 @@ USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
 	vmovups	%ymm6 ,  	(%rax)
 	vmovups	%ymm7 ,  	(%rax, LDC)
 
-	addq	$4*SIZE, CO1
+	addq	$ 4*SIZE, CO1
 .endm
 
 /******************************************************************************************/
@@ -848,9 +848,9 @@ USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
 	vfmadd231pd  	%xmm0 ,%xmm2  , %xmm5
 	vmovddup	 -9 * SIZE(BO), %xmm8
 	vfmadd231pd  	%xmm0 ,%xmm3  , %xmm6
-	addq		$4*SIZE, BO
+	addq		$ 4*SIZE, BO
 	vfmadd231pd  	%xmm0 ,%xmm8  , %xmm7
-	addq		$2*SIZE, AO
+	addq		$ 2*SIZE, AO
 
 .endm
 
@@ -880,7 +880,7 @@ USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
 	vmovups	%xmm6 ,  	(%rax)
 	vmovups	%xmm7 ,  	(%rax, LDC)
 
-	addq	$2*SIZE, CO1
+	addq	$ 2*SIZE, CO1
 .endm
 
 /******************************************************************************************/
@@ -905,9 +905,9 @@ USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
 	vfmadd231sd  	%xmm0 ,%xmm2  , %xmm5
 	vmovsd	 -9 * SIZE(BO), %xmm8
 	vfmadd231sd  	%xmm0 ,%xmm3  , %xmm6
-	addq		$4*SIZE, BO
+	addq		$ 4*SIZE, BO
 	vfmadd231sd  	%xmm0 ,%xmm8  , %xmm7
-	addq		$1*SIZE, AO
+	addq		$ 1*SIZE, AO
 
 .endm
 
@@ -937,7 +937,7 @@ USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
 	vmovsd	%xmm6 ,  	(%rax)
 	vmovsd	%xmm7 ,  	(%rax, LDC)
 
-	addq	$1*SIZE, CO1
+	addq	$ 1*SIZE, CO1
 .endm
 
 
@@ -963,8 +963,8 @@ USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
 	vfmadd231pd  	%xmm1 ,%xmm2  , %xmm5
 	vfmadd231pd  	%xmm0 ,%xmm3  , %xmm6
 	vfmadd231pd  	%xmm1 ,%xmm3  , %xmm7
-	addq		$2*SIZE, BO
-	addq		$4*SIZE, AO
+	addq		$ 2*SIZE, BO
+	addq		$ 4*SIZE, AO
 
 .endm
 
@@ -993,7 +993,7 @@ USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
 	vmovups	%xmm6 ,  	(CO1, LDC)
 	vmovups	%xmm7 , 2 * SIZE(CO1, LDC)
 
-	addq	$4*SIZE, CO1
+	addq	$ 4*SIZE, CO1
 .endm
 
 
@@ -1014,8 +1014,8 @@ USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
 	vmovddup	-11 * SIZE(BO), %xmm3
 	vfmadd231pd  	%xmm0 ,%xmm2  , %xmm4
 	vfmadd231pd  	%xmm0 ,%xmm3  , %xmm6
-	addq		$2*SIZE, BO
-	addq		$2*SIZE, AO
+	addq		$ 2*SIZE, BO
+	addq		$ 2*SIZE, AO
 
 .endm
 
@@ -1038,7 +1038,7 @@ USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
 	vmovups	%xmm4 ,  	(CO1)
 	vmovups	%xmm6 ,  	(CO1, LDC)
 
-	addq	$2*SIZE, CO1
+	addq	$ 2*SIZE, CO1
 .endm
 
 /******************************************************************************************/
@@ -1058,8 +1058,8 @@ USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
 	vmovsd	-11 * SIZE(BO), %xmm2
 	vfmadd231sd  	%xmm0 ,%xmm1  , %xmm4
 	vfmadd231sd  	%xmm0 ,%xmm2  , %xmm5
-	addq		$2*SIZE, BO
-	addq		$1*SIZE, AO
+	addq		$ 2*SIZE, BO
+	addq		$ 1*SIZE, AO
 
 .endm
 
@@ -1082,7 +1082,7 @@ USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
 	vmovsd	%xmm4 ,  	(CO1)
 	vmovsd	%xmm5 ,  	(CO1, LDC)
 
-	addq	$1*SIZE, CO1
+	addq	$ 1*SIZE, CO1
 .endm
 
 
@@ -1103,8 +1103,8 @@ USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
 	vmovups 	-14 * SIZE(AO), %xmm1
 	vfmadd231pd  	%xmm0 ,%xmm2  , %xmm4
 	vfmadd231pd  	%xmm1 ,%xmm2  , %xmm5
-	addq		$1*SIZE, BO
-	addq		$4*SIZE, AO
+	addq		$ 1*SIZE, BO
+	addq		$ 4*SIZE, AO
 
 .endm
 
@@ -1127,7 +1127,7 @@ USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
 	vmovups	%xmm4 ,  	(CO1)
 	vmovups	%xmm5 , 2 * SIZE(CO1)
 
-	addq	$4*SIZE, CO1
+	addq	$ 4*SIZE, CO1
 .endm
 
 
@@ -1145,8 +1145,8 @@ USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
 	vmovddup	-12 * SIZE(BO), %xmm2
 	vmovups 	-16 * SIZE(AO), %xmm0
 	vfmadd231pd  	%xmm0 ,%xmm2  , %xmm4
-	addq		$1*SIZE, BO
-	addq		$2*SIZE, AO
+	addq		$ 1*SIZE, BO
+	addq		$ 2*SIZE, AO
 
 .endm
 
@@ -1166,7 +1166,7 @@ USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
 
 	vmovups	%xmm4 ,  	(CO1)
 
-	addq	$2*SIZE, CO1
+	addq	$ 2*SIZE, CO1
 .endm
 
 
@@ -1184,8 +1184,8 @@ USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
 	vmovsd	-12 * SIZE(BO), %xmm1
 	vmovsd 	-16 * SIZE(AO), %xmm0
 	vfmadd231sd  	%xmm0 ,%xmm1  , %xmm4
-	addq		$1*SIZE, BO
-	addq		$1*SIZE, AO
+	addq		$ 1*SIZE, BO
+	addq		$ 1*SIZE, AO
 
 .endm
 
@@ -1205,7 +1205,7 @@ USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
 
 	vmovsd	%xmm4 ,  	(CO1)
 
-	addq	$1*SIZE, CO1
+	addq	$ 1*SIZE, CO1
 .endm
 
 
@@ -1262,13 +1262,13 @@ USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
 
         STACK_TOUCH
 
-	cmpq	$0, OLD_M
+	cmpq	$ 0, OLD_M
 	je	.L999
 
-	cmpq	$0, OLD_N
+	cmpq	$ 0, OLD_N
 	je	.L999
 
-	cmpq	$0, OLD_K
+	cmpq	$ 0, OLD_K
 	je	.L999
 
 	movq	OLD_M, M
@@ -1288,7 +1288,7 @@ USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
 
 
 	movq	Ndiv12,  J
-	cmpq	$0, J
+	cmpq	$ 0, J
 	je	.L4_0
 	ALIGN_4
 
@@ -1330,10 +1330,10 @@ USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
 	vmovups	%ymm6, 16 * SIZE(BO)
 	vmovups	%ymm7, 20 * SIZE(BO)
 
-	addq	$8 * SIZE ,BO1
-	addq	$8 * SIZE ,BO2
-	addq	$8 * SIZE ,BO3
-	addq    $24 *SIZE ,BO
+	addq	$ 8 * SIZE ,BO1
+	addq	$ 8 * SIZE ,BO2
+	addq	$ 8 * SIZE ,BO3
+	addq    $ 24 *SIZE ,BO
 
 	decq	%rax
 	jnz	.L12_01a_1
@@ -1356,10 +1356,10 @@ USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
 	vmovups	%ymm1, 0 * SIZE(BO)
 	vmovups	%ymm2, 4 * SIZE(BO)
 	vmovups	%ymm3, 8 * SIZE(BO)
-	addq	$4*SIZE,BO1
-	addq	$4*SIZE,BO2
-	addq	$4*SIZE,BO3
-	addq	$12*SIZE,BO
+	addq	$ 4*SIZE,BO1
+	addq	$ 4*SIZE,BO2
+	addq	$ 4*SIZE,BO3
+	addq	$ 12*SIZE,BO
 	decq	%rax
 	jnz	.L12_02b
 
@@ -1407,8 +1407,7 @@ USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
 	subq $2, %rax
 	je	.L12_12a
 
-	.align 32
-
+	ALIGN_5
 .L12_12:
 
 	KERNEL4x12_M1
@@ -1621,7 +1620,7 @@ USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
 
 .L4_0:
 
-	cmpq	$0, Nmod12		// N % 12 == 0
+	cmpq	$ 0, Nmod12		// N % 12 == 0
 	je	.L999
 
 	movq	Nmod12, J		
@@ -1666,7 +1665,7 @@ USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
 	subq $2, %rax
 	je	.L4_12a
 
-	.align 32
+	ALIGN_5
 
 .L4_12:
 
@@ -1912,7 +1911,7 @@ USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
 
 	je	.L2_16
 
-	.align 32
+	ALIGN_5
 
 .L2_12:
 
@@ -2108,7 +2107,7 @@ USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
 	sarq	$3, %rax			//  K / 8
 	je	.L1_16
 
-	.align 32
+	ALIGN_5
 
 .L1_12:
 
@@ -2362,13 +2361,13 @@ USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
 
         STACK_TOUCH
 
-	cmpq	$0, OLD_M
+	cmpq	$ 0, OLD_M
 	je	.L999
 
-	cmpq	$0, OLD_N
+	cmpq	$ 0, OLD_N
 	je	.L999
 
-	cmpq	$0, OLD_K
+	cmpq	$ 0, OLD_K
 	je	.L999
 
 	movq	OLD_M, M
@@ -2397,7 +2396,7 @@ USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
 
 
 	movq	Ndiv12,  J
-	cmpq	$0, J
+	cmpq	$ 0, J
 	je	.L2_0
 	ALIGN_4
 
@@ -2471,7 +2470,7 @@ USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
 	subq $2, %rax
 	je	.L4_12a
 
-	.align 32
+	ALIGN_5
 
 .L4_12:
 
@@ -2848,7 +2847,7 @@ USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
 
 	je	.L2_16
 
-	.align 32
+	ALIGN_5
 
 .L2_12:
 
@@ -3176,7 +3175,7 @@ USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
 	sarq	$3, %rax			//  K / 8
 	je	.L1_16
 
-	.align 32
+	ALIGN_5
 
 .L1_12:
 

kernel/x86_64/dgemm_kernel_8x2_piledriver.S

@@ -196,7 +196,7 @@ USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
 	vmovups 	-12 * SIZE(AO), %xmm0
 	vmulpd  	%xmm1,%xmm0,%xmm10
 	vmulpd  	%xmm2,%xmm0,%xmm11
-	addq		$3*SIZE, BO
+	addq		$ 3 * SIZE, BO
 	vmulpd  	%xmm3,%xmm0,%xmm12
 	vmovups 	-10 * SIZE(AO), %xmm0
 	vmulpd  	%xmm1,%xmm0,%xmm13
@@ -294,7 +294,7 @@ USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
 	vmovups 	 14 * SIZE(AO), %xmm0
 	VFMADD231PD_(  	%xmm1,%xmm0,%xmm13 )
 	vmovddup	 -3 * SIZE(BO), %xmm1
-	addq		$32 * SIZE, AO
+	addq		$ 32 * SIZE, AO
 	VFMADD231PD_(  	%xmm2,%xmm0,%xmm14 )
 	vmovddup	 -2 * SIZE(BO), %xmm2
 	VFMADD231PD_(  	%xmm3,%xmm0,%xmm15 )
@@ -392,8 +392,8 @@ USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
 	vmovddup	 10 * SIZE(BO), %xmm2
 	VFMADD231PD_(  	%xmm3,%xmm0,%xmm15 )
 	vmovddup	 11 * SIZE(BO), %xmm3
-	addq		$32 * SIZE, AO
-	addq		$24 * SIZE, BO
+	addq		$ 32 * SIZE, AO
+	addq		$ 24 * SIZE, BO
 .endm
 
 
@@ -414,9 +414,9 @@ USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
 	VFMADD231PD_(  	%xmm3,%xmm0,%xmm12 )
 	vmovups 	 14 * SIZE(AO), %xmm0
 	VFMADD231PD_(  	%xmm1,%xmm0,%xmm13 )
-	addq		$32*SIZE, AO
+	addq		$ 32 * SIZE, AO
 	VFMADD231PD_(  	%xmm2,%xmm0,%xmm14 )
-	addq		$21*SIZE, BO
+	addq		$ 21 * SIZE, BO
 	VFMADD231PD_(  	%xmm3,%xmm0,%xmm15 )
 .endm
 
@@ -438,9 +438,9 @@ USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
 	VFMADD231PD_(  	%xmm3,%xmm0,%xmm12 )
 	vmovups 	-10 * SIZE(AO), %xmm0
 	VFMADD231PD_(  	%xmm1,%xmm0,%xmm13 )
-	addq		$3*SIZE, BO
+	addq		$ 3 * SIZE, BO
 	VFMADD231PD_(  	%xmm2,%xmm0,%xmm14 )
-	addq		$8*SIZE, AO
+	addq		$ 8 * SIZE, AO
 	VFMADD231PD_(  	%xmm3,%xmm0,%xmm15 )
 .endm
 
@@ -483,7 +483,7 @@ USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
 	prefetcht0	 C_PR1(CO1,LDC)
 	prefetcht0	 C_PR1(CO1,LDC,2)
 
-	addq	$8 * SIZE, CO1		# coffset += 8
+	addq	$ 8 * SIZE, CO1		# coffset += 8
 .endm
 
 
@@ -1165,9 +1165,9 @@ USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
 	vmovsd	%xmm5, 8*SIZE(BO)
 	vmovups	%xmm6, 9*SIZE(BO)
 	vmovsd	%xmm7,11*SIZE(BO)
-	addq	$8*SIZE,BO1
-	addq	$8*SIZE,BO2
-	addq	$12*SIZE,BO
+	addq	$ 8*SIZE,BO1
+	addq	$ 8*SIZE,BO2
+	addq	$ 12*SIZE,BO
 	decq	%rax
 	jnz	.L6_02
 
@@ -1184,9 +1184,9 @@ USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
 	vmovsd  (BO2), %xmm1
 	vmovups	%xmm0,       (BO)
 	vmovsd	%xmm1, 2*SIZE(BO)
-	addq	$2*SIZE,BO1
-	addq	$2*SIZE,BO2
-	addq	$3*SIZE,BO
+	addq	$ 2*SIZE,BO1
+	addq	$ 2*SIZE,BO2
+	addq	$ 3*SIZE,BO
 	decq	%rax
 	jnz	.L6_02b
 
@@ -1223,9 +1223,9 @@ USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
 	vmovups	%xmm4, 7*SIZE(BO)
 	vmovsd	%xmm7, 9*SIZE(BO)
 	vmovups	%xmm6,10*SIZE(BO)
-	addq	$8*SIZE,BO1
-	addq	$8*SIZE,BO2
-	addq	$12*SIZE,BO
+	addq	$ 8*SIZE,BO1
+	addq	$ 8*SIZE,BO2
+	addq	$ 12*SIZE,BO
 	decq	%rax
 	jnz	.L6_03
 
@@ -1243,9 +1243,9 @@ USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
 	vmovups  	(BO2), %xmm1
 	vmovsd	%xmm0,       (BO)
 	vmovups %xmm1, 1*SIZE(BO)
-	addq	$2*SIZE,BO1
-	addq	$2*SIZE,BO2
-	addq	$3*SIZE,BO
+	addq	$ 2*SIZE,BO1
+	addq	$ 2*SIZE,BO2
+	addq	$ 3*SIZE,BO
 	decq	%rax
 	jnz	.L6_03b
 

kernel/x86_64/sgemm_kernel_16x4_haswell.S

@@ -166,8 +166,8 @@ USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
 	VFMADD231PS_(  	%ymm9,%ymm2,%ymm1  )
 	VFMADD231PS_(  	%ymm10,%ymm3,%ymm0 )
 	VFMADD231PS_(  	%ymm11,%ymm3,%ymm1 )
-	addq	$4 , BI	
-	addq	$16, %rax 
+	addq	$ 4 , BI	
+	addq	$ 16, %rax 
 .endm
 
 .macro SAVE16x4
@@ -233,8 +233,8 @@ USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
 	vbroadcastss	 -1 * SIZE(BO, BI, SIZE), %ymm3
 	VFMADD231PS_(  	%ymm8,%ymm2,%ymm0  )
 	VFMADD231PS_(  	%ymm10,%ymm3,%ymm0 )
-	addq	$4 , BI	
-	addq	$8 , %rax 
+	addq	$ 4 , BI	
+	addq	$ 8 , %rax 
 .endm
 
 .macro SAVE8x4
@@ -277,8 +277,8 @@ USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
 	vbroadcastss	 -1 * SIZE(BO, BI, SIZE), %xmm3
 	VFMADD231PS_(  	%xmm8,%xmm2,%xmm0  )
 	VFMADD231PS_(  	%xmm10,%xmm3,%xmm0 )
-	addq	$4 , BI	
-	addq	$4 , %rax 
+	addq	$ 4 , BI	
+	addq	$ 4 , %rax 
 .endm
 
 .macro SAVE4x4
@@ -325,8 +325,8 @@ USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
 	VFMADD231SS_(  	%xmm9,%xmm2,%xmm1  )
 	VFMADD231SS_(  	%xmm10,%xmm3,%xmm0 )
 	VFMADD231SS_(  	%xmm11,%xmm3,%xmm1 )
-	addq	$4 , BI	
-	addq	$2, %rax 
+	addq	$ 4 , BI	
+	addq	$ 2, %rax 
 .endm
 
 .macro SAVE2x4
@@ -386,8 +386,8 @@ USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
 	vmovss	 -1 * SIZE(BO, BI, SIZE), %xmm3
 	VFMADD231SS_(  	%xmm8,%xmm2,%xmm0  )
 	VFMADD231SS_(  	%xmm10,%xmm3,%xmm0 )
-	addq	$4 , BI	
-	addq	$1, %rax 
+	addq	$ 4 , BI	
+	addq	$ 1, %rax 
 .endm
 
 .macro SAVE1x4
@@ -432,8 +432,8 @@ USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
 	VFMADD231PS_(  	%ymm5,%ymm2,%ymm1 )
 	VFMADD231PS_(  	%ymm6,%ymm3,%ymm0 )
 	VFMADD231PS_(  	%ymm7,%ymm3,%ymm1 )
-	addq	$2 , BI	
-	addq	$16, %rax 
+	addq	$ 2 , BI	
+	addq	$ 16, %rax 
 .endm
 
 .macro SAVE16x2
@@ -474,8 +474,8 @@ USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
 	vbroadcastss	 -3 * SIZE(BO, BI, SIZE), %ymm3
 	VFMADD231PS_(  	%ymm4,%ymm2,%ymm0 )
 	VFMADD231PS_(  	%ymm6,%ymm3,%ymm0 )
-	addq	$2 , BI	
-	addq	$8 , %rax 
+	addq	$ 2 , BI	
+	addq	$ 8 , %rax 
 .endm
 
 .macro SAVE8x2
@@ -507,8 +507,8 @@ USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
 	vbroadcastss	 -3 * SIZE(BO, BI, SIZE), %xmm3
 	VFMADD231PS_(  	%xmm4,%xmm2,%xmm0 )
 	VFMADD231PS_(  	%xmm6,%xmm3,%xmm0 )
-	addq	$2 , BI	
-	addq	$4 , %rax 
+	addq	$ 2 , BI	
+	addq	$ 4 , %rax 
 .endm
 
 .macro SAVE4x2
@@ -542,8 +542,8 @@ USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
 	VFMADD231SS_(  	%xmm5,%xmm2,%xmm1 )
 	VFMADD231SS_(  	%xmm6,%xmm3,%xmm0 )
 	VFMADD231SS_(  	%xmm7,%xmm3,%xmm1 )
-	addq	$2 , BI	
-	addq	$2, %rax 
+	addq	$ 2 , BI	
+	addq	$ 2, %rax 
 .endm
 
 .macro SAVE2x2
@@ -583,8 +583,8 @@ USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
 	vmovss	 -3 * SIZE(BO, BI, SIZE), %xmm3
 	VFMADD231SS_(  	%xmm4,%xmm2,%xmm0 )
 	VFMADD231SS_(  	%xmm6,%xmm3,%xmm0 )
-	addq	$2 , BI	
-	addq	$1, %rax 
+	addq	$ 2 , BI	
+	addq	$ 1, %rax 
 .endm
 
 .macro SAVE1x2
@@ -619,8 +619,8 @@ USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
 	vbroadcastss	 -4 * SIZE(BO, BI, SIZE), %ymm2
 	VFMADD231PS_(  	%ymm4,%ymm2,%ymm0 )
 	VFMADD231PS_(  	%ymm5,%ymm2,%ymm1 )
-	addq	$1 , BI	
-	addq	$16, %rax 
+	addq	$ 1 , BI	
+	addq	$ 16, %rax 
 .endm
 
 .macro SAVE16x1
@@ -649,8 +649,8 @@ USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
 	vmovups 	-16 * SIZE(AO, %rax, SIZE), %ymm0
 	vbroadcastss	 -4 * SIZE(BO, BI, SIZE), %ymm2
 	VFMADD231PS_(  	%ymm4,%ymm2,%ymm0 )
-	addq	$1 , BI	
-	addq	$8 , %rax 
+	addq	$ 1 , BI	
+	addq	$ 8 , %rax 
 .endm
 
 .macro SAVE8x1
@@ -677,8 +677,8 @@ USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
 	vmovups 	-16 * SIZE(AO, %rax, SIZE), %xmm0
 	vbroadcastss	 -4 * SIZE(BO, BI, SIZE), %xmm2
 	VFMADD231PS_(  	%xmm4,%xmm2,%xmm0 )
-	addq	$1 , BI	
-	addq	$4 , %rax 
+	addq	$ 1 , BI	
+	addq	$ 4 , %rax 
 .endm
 
 .macro SAVE4x1
@@ -706,8 +706,8 @@ USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
 	vmovss	 -4 * SIZE(BO, BI, SIZE), %xmm2
 	VFMADD231SS_(  	%xmm4,%xmm2,%xmm0 )
 	VFMADD231SS_(  	%xmm5,%xmm2,%xmm1 )
-	addq	$1 , BI	
-	addq	$2, %rax 
+	addq	$ 1 , BI	
+	addq	$ 2 , %rax 
 .endm
 
 .macro SAVE2x1
@@ -736,8 +736,8 @@ USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
 	vmovss 	-16 * SIZE(AO, %rax, SIZE), %xmm0
 	vmovss	 -4 * SIZE(BO, BI, SIZE), %xmm2
 	VFMADD231SS_(  	%xmm4,%xmm2,%xmm0 )
-	addq	$1 , BI	
-	addq	$1, %rax 
+	addq	$ 1 , BI	
+	addq	$ 1 , %rax 
 .endm
 
 .macro SAVE1x1
@@ -882,8 +882,8 @@ USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
 	vmovups	%xmm2, 8*SIZE(BO)
 	vmovups	%xmm3,12*SIZE(BO)
 
-	addq	$16*SIZE,BO1
-	addq	$16*SIZE,BO
+	addq	$ 16*SIZE,BO1
+	addq	$ 16*SIZE,BO
 	decq	%rax
 	jnz	.L4_01a
 
@@ -899,8 +899,8 @@ USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
 
 	vmovups	(BO1), %xmm0
 	vmovups	%xmm0, (BO)
-	addq	$4*SIZE,BO1
-	addq	$4*SIZE,BO
+	addq	$ 4*SIZE,BO1
+	addq	$ 4*SIZE,BO
 	decq	%rax
 	jnz	.L4_02c
 
@@ -919,7 +919,7 @@ USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
 #endif
 	
 	movq	A, AO		 	// aoffset = a
-	addq	$16 * SIZE, AO
+	addq	$ 16 * SIZE, AO
 
 	movq	M,  I
 	sarq	$4, I			// i = (m >> 4)

kernel/x86_64/zgemm_kernel_4x2_haswell.S

@@ -109,22 +109,22 @@ USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
 #if defined(OS_WINDOWS)
 #if   L_BUFFER_SIZE > 16384
 #define STACK_TOUCH \
-        movl    $0,  4096 * 4(%rsp);\
-        movl    $0,  4096 * 3(%rsp);\
-        movl    $0,  4096 * 2(%rsp);\
-        movl    $0,  4096 * 1(%rsp);
+        movl    $ 0,  4096 * 4(%rsp);\
+        movl    $ 0,  4096 * 3(%rsp);\
+        movl    $ 0,  4096 * 2(%rsp);\
+        movl    $ 0,  4096 * 1(%rsp);
 #elif L_BUFFER_SIZE > 12288
 #define STACK_TOUCH \
-        movl    $0,  4096 * 3(%rsp);\
-        movl    $0,  4096 * 2(%rsp);\
-        movl    $0,  4096 * 1(%rsp);
+        movl    $ 0,  4096 * 3(%rsp);\
+        movl    $ 0,  4096 * 2(%rsp);\
+        movl    $ 0,  4096 * 1(%rsp);
 #elif L_BUFFER_SIZE > 8192
 #define STACK_TOUCH \
-        movl    $0,  4096 * 2(%rsp);\
-        movl    $0,  4096 * 1(%rsp);
+        movl    $ 0,  4096 * 2(%rsp);\
+        movl    $ 0,  4096 * 1(%rsp);
 #elif L_BUFFER_SIZE > 4096
 #define STACK_TOUCH \
-        movl    $0,  4096 * 1(%rsp);
+        movl    $ 0,  4096 * 1(%rsp);
 #else
 #define STACK_TOUCH
 #endif
@@ -212,8 +212,8 @@ USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
         VFMADDPD_I(        %ymm11,%ymm7,%ymm0 )
         VFMADDPD_I(        %ymm15,%ymm7,%ymm1 )
 
-        addq    $4, BI                           
-        addq    $8, %rax                         
+        addq    $ 4, BI                           
+        addq    $ 8, %rax                         
 .endm
 
 .macro SAVE4x2
@@ -222,10 +222,10 @@ USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
 	vbroadcastsd	ALPHA_I, %ymm1
 
 	// swap high and low 8 bytes
-        vshufpd $0x05, %ymm9 , %ymm9, %ymm9
-        vshufpd $0x05, %ymm11, %ymm11, %ymm11
-        vshufpd $0x05, %ymm13, %ymm13, %ymm13
-        vshufpd $0x05, %ymm15, %ymm15, %ymm15
+        vshufpd $ 0x05, %ymm9 , %ymm9, %ymm9
+        vshufpd $ 0x05, %ymm11, %ymm11, %ymm11
+        vshufpd $ 0x05, %ymm13, %ymm13, %ymm13
+        vshufpd $ 0x05, %ymm15, %ymm15, %ymm15
 
 #if defined(NN) || defined(NT) || defined(TN) || defined(TT) || \
     defined(NR) || defined(NC) || defined(TR) || defined(TC)
@@ -235,10 +235,10 @@ USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
         vaddsubpd %ymm13,%ymm12, %ymm12
         vaddsubpd %ymm15,%ymm14, %ymm14
 
-        vshufpd $0x05, %ymm8 , %ymm8, %ymm9
-        vshufpd $0x05, %ymm10, %ymm10, %ymm11
-        vshufpd $0x05, %ymm12, %ymm12, %ymm13
-        vshufpd $0x05, %ymm14, %ymm14, %ymm15
+        vshufpd $ 0x05, %ymm8 , %ymm8, %ymm9
+        vshufpd $ 0x05, %ymm10, %ymm10, %ymm11
+        vshufpd $ 0x05, %ymm12, %ymm12, %ymm13
+        vshufpd $ 0x05, %ymm14, %ymm14, %ymm15
 
 #else
         vaddsubpd %ymm8,  %ymm9 ,%ymm9
@@ -252,10 +252,10 @@ USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
         vmovapd   %ymm15, %ymm14
 
 	// swap high and low 8 bytes
-        vshufpd $0x05, %ymm9 , %ymm9, %ymm9
-        vshufpd $0x05, %ymm11, %ymm11, %ymm11
-        vshufpd $0x05, %ymm13, %ymm13, %ymm13
-        vshufpd $0x05, %ymm15, %ymm15, %ymm15
+        vshufpd $ 0x05, %ymm9 , %ymm9, %ymm9
+        vshufpd $ 0x05, %ymm11, %ymm11, %ymm11
+        vshufpd $ 0x05, %ymm13, %ymm13, %ymm13
+        vshufpd $ 0x05, %ymm15, %ymm15, %ymm15
 
 #endif
 
@@ -316,8 +316,8 @@ USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
         vmovddup         -5 * SIZE(BO, BI, SIZE), %xmm7
         VFMADDPD_I(        %xmm11,%xmm7,%xmm0 )
         VFMADDPD_I(        %xmm15,%xmm7,%xmm1 )
-        addq    $4, BI                           
-        addq    $4, %rax                         
+        addq    $ 4, BI                           
+        addq    $ 4, %rax                         
 .endm
 
 .macro SAVE2x2
@@ -326,10 +326,10 @@ USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
 	vmovddup	ALPHA_I, %xmm1
 
 	// swap high and low 64 bytes
-        vshufpd $0x01, %xmm9 , %xmm9, %xmm9
-        vshufpd $0x01, %xmm11, %xmm11, %xmm11
-        vshufpd $0x01, %xmm13, %xmm13, %xmm13
-        vshufpd $0x01, %xmm15, %xmm15, %xmm15
+        vshufpd $ 0x01, %xmm9 , %xmm9, %xmm9
+        vshufpd $ 0x01, %xmm11, %xmm11, %xmm11
+        vshufpd $ 0x01, %xmm13, %xmm13, %xmm13
+        vshufpd $ 0x01, %xmm15, %xmm15, %xmm15
 
 #if defined(NN) || defined(NT) || defined(TN) || defined(TT) || \
     defined(NR) || defined(NC) || defined(TR) || defined(TC)
@@ -339,10 +339,10 @@ USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
         vaddsubpd %xmm13,%xmm12, %xmm12
         vaddsubpd %xmm15,%xmm14, %xmm14
 
-        vshufpd $0x01, %xmm8 , %xmm8, %xmm9
-        vshufpd $0x01, %xmm10, %xmm10, %xmm11
-        vshufpd $0x01, %xmm12, %xmm12, %xmm13
-        vshufpd $0x01, %xmm14, %xmm14, %xmm15
+        vshufpd $ 0x01, %xmm8 , %xmm8, %xmm9
+        vshufpd $ 0x01, %xmm10, %xmm10, %xmm11
+        vshufpd $ 0x01, %xmm12, %xmm12, %xmm13
+        vshufpd $ 0x01, %xmm14, %xmm14, %xmm15
 
 #else
         vaddsubpd %xmm8,  %xmm9 ,%xmm9
@@ -356,10 +356,10 @@ USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
         vmovapd   %xmm15, %xmm14
 
 	// swap high and low 64 bytes
-        vshufpd $0x01, %xmm9 , %xmm9, %xmm9
-        vshufpd $0x01, %xmm11, %xmm11, %xmm11
-        vshufpd $0x01, %xmm13, %xmm13, %xmm13
-        vshufpd $0x01, %xmm15, %xmm15, %xmm15
+        vshufpd $ 0x01, %xmm9 , %xmm9, %xmm9
+        vshufpd $ 0x01, %xmm11, %xmm11, %xmm11
+        vshufpd $ 0x01, %xmm13, %xmm13, %xmm13
+        vshufpd $ 0x01, %xmm15, %xmm15, %xmm15
 
 #endif
 
@@ -415,8 +415,8 @@ USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
         vmovddup         -5 * SIZE(BO, BI, SIZE), %xmm7
         VFMADDPD_R(        %xmm10,%xmm6,%xmm0 )
         VFMADDPD_I(        %xmm11,%xmm7,%xmm0 )
-        addq    $4, BI                           
-        addq    $2, %rax                         
+        addq    $ 4, BI                           
+        addq    $ 2, %rax                         
 .endm
 
 .macro SAVE1x2
@@ -425,8 +425,8 @@ USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
 	vmovddup	ALPHA_I, %xmm1
 
 	// swap high and low 64 bytes
-        vshufpd $0x01, %xmm9 , %xmm9, %xmm9
-        vshufpd $0x01, %xmm11, %xmm11, %xmm11
+        vshufpd $ 0x01, %xmm9 , %xmm9, %xmm9
+        vshufpd $ 0x01, %xmm11, %xmm11, %xmm11
 
 #if defined(NN) || defined(NT) || defined(TN) || defined(TT) || \
     defined(NR) || defined(NC) || defined(TR) || defined(TC)
@@ -434,8 +434,8 @@ USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
         vaddsubpd %xmm9, %xmm8 , %xmm8
         vaddsubpd %xmm11,%xmm10, %xmm10
 
-        vshufpd $0x01, %xmm8 , %xmm8, %xmm9
-        vshufpd $0x01, %xmm10, %xmm10, %xmm11
+        vshufpd $ 0x01, %xmm8 , %xmm8, %xmm9
+        vshufpd $ 0x01, %xmm10, %xmm10, %xmm11
 
 #else
         vaddsubpd %xmm8, %xmm9, %xmm9
@@ -445,8 +445,8 @@ USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
         vmovapd   %xmm11, %xmm10
 
 	// swap high and low 64 bytes
-        vshufpd $0x01, %xmm9 , %xmm9, %xmm9
-        vshufpd $0x01, %xmm11, %xmm11, %xmm11
+        vshufpd $ 0x01, %xmm9 , %xmm9, %xmm9
+        vshufpd $ 0x01, %xmm11, %xmm11, %xmm11
 
 #endif
 
@@ -486,8 +486,8 @@ USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
         VFMADDPD_I(        %ymm9 ,%ymm5,%ymm0 )
         VFMADDPD_I(        %ymm13,%ymm5,%ymm1 )
 
-        addq    $2, BI                           
-        addq    $8, %rax                         
+        addq    $ 2, BI                           
+        addq    $ 8, %rax                         
 .endm
 
 .macro SAVE4x1
@@ -496,8 +496,8 @@ USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
 	vbroadcastsd	ALPHA_I, %ymm1
 
 	// swap high and low 8 bytes
-        vshufpd $0x05, %ymm9 , %ymm9, %ymm9
-        vshufpd $0x05, %ymm13, %ymm13, %ymm13
+        vshufpd $ 0x05, %ymm9 , %ymm9, %ymm9
+        vshufpd $ 0x05, %ymm13, %ymm13, %ymm13
 
 #if defined(NN) || defined(NT) || defined(TN) || defined(TT) || \
     defined(NR) || defined(NC) || defined(TR) || defined(TC)
@@ -505,8 +505,8 @@ USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
         vaddsubpd %ymm9, %ymm8  , %ymm8
         vaddsubpd %ymm13,%ymm12 , %ymm12
 
-        vshufpd $0x05, %ymm8 , %ymm8, %ymm9
-        vshufpd $0x05, %ymm12, %ymm12, %ymm13
+        vshufpd $ 0x05, %ymm8 , %ymm8, %ymm9
+        vshufpd $ 0x05, %ymm12, %ymm12, %ymm13
 
 #else
         vaddsubpd %ymm8, %ymm9 , %ymm9
@@ -516,8 +516,8 @@ USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
         vmovapd   %ymm13, %ymm12
 
 	// swap high and low 8 bytes
-        vshufpd $0x05, %ymm9 , %ymm9, %ymm9
-        vshufpd $0x05, %ymm13, %ymm13, %ymm13
+        vshufpd $ 0x05, %ymm9 , %ymm9, %ymm9
+        vshufpd $ 0x05, %ymm13, %ymm13, %ymm13
 
 #endif
 
@@ -559,8 +559,8 @@ USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
         vmovddup         -3 * SIZE(BO, BI, SIZE), %xmm5
         VFMADDPD_I(        %xmm9,%xmm5,%xmm0  )
         VFMADDPD_I(        %xmm13,%xmm5,%xmm1 )
-        addq    $2, BI                           
-        addq    $4, %rax                         
+        addq    $ 2, BI                           
+        addq    $ 4, %rax                         
 .endm
 
 .macro SAVE2x1
@@ -569,8 +569,8 @@ USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
 	vmovddup	ALPHA_I, %xmm1
 
 	// swap high and low 64 bytes
-        vshufpd $0x01, %xmm9 , %xmm9, %xmm9
-        vshufpd $0x01, %xmm13, %xmm13, %xmm13
+        vshufpd $ 0x01, %xmm9 , %xmm9, %xmm9
+        vshufpd $ 0x01, %xmm13, %xmm13, %xmm13
 
 #if defined(NN) || defined(NT) || defined(TN) || defined(TT) || \
     defined(NR) || defined(NC) || defined(TR) || defined(TC)
@@ -578,8 +578,8 @@ USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
         vaddsubpd %xmm9, %xmm8  , %xmm8
         vaddsubpd %xmm13,%xmm12 , %xmm12
 
-        vshufpd $0x01, %xmm8 , %xmm8, %xmm9
-        vshufpd $0x01, %xmm12, %xmm12, %xmm13
+        vshufpd $ 0x01, %xmm8 , %xmm8, %xmm9
+        vshufpd $ 0x01, %xmm12, %xmm12, %xmm13
 
 #else
         vaddsubpd %xmm8, %xmm9 , %xmm9
@@ -589,8 +589,8 @@ USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
         vmovapd   %xmm13, %xmm12
 
 	// swap high and low 64 bytes
-        vshufpd $0x01, %xmm9 , %xmm9, %xmm9
-        vshufpd $0x01, %xmm13, %xmm13, %xmm13
+        vshufpd $ 0x01, %xmm9 , %xmm9, %xmm9
+        vshufpd $ 0x01, %xmm13, %xmm13, %xmm13
 
 #endif
 
@@ -626,8 +626,8 @@ USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
         VFMADDPD_R(        %xmm8,%xmm4,%xmm0 )
         vmovddup         -3 * SIZE(BO, BI, SIZE), %xmm5
         VFMADDPD_I(        %xmm9,%xmm5,%xmm0 )
-        addq    $2, BI                           
-        addq    $2, %rax                         
+        addq    $ 2, BI                           
+        addq    $ 2, %rax                         
 .endm
 
 .macro SAVE1x1
@@ -636,14 +636,14 @@ USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
 	vmovddup	ALPHA_I, %xmm1
 
 	// swap high and low 64 bytes
-        vshufpd $0x01, %xmm9 , %xmm9, %xmm9
+        vshufpd $ 0x01, %xmm9 , %xmm9, %xmm9
 
 #if defined(NN) || defined(NT) || defined(TN) || defined(TT) || \
     defined(NR) || defined(NC) || defined(TR) || defined(TC)
 
         vaddsubpd %xmm9, %xmm8,  %xmm8
 
-        vshufpd $0x01, %xmm8 , %xmm8, %xmm9
+        vshufpd $ 0x01, %xmm8 , %xmm8, %xmm9
 
 #else
         vaddsubpd %xmm8, %xmm9,  %xmm9
@@ -651,7 +651,7 @@ USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
         vmovapd   %xmm9,  %xmm8
 
 	// swap high and low 64 bytes
-        vshufpd $0x01, %xmm9 , %xmm9, %xmm9
+        vshufpd $ 0x01, %xmm9 , %xmm9, %xmm9
 
 #endif
 
@@ -682,7 +682,7 @@ USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
 	PROLOGUE
 	PROFCODE
 	
-	subq	$STACKSIZE, %rsp
+	subq	$ STACKSIZE, %rsp
 	movq	%rbx,   (%rsp)
 	movq	%rbp,  8(%rsp)
 	movq	%r12, 16(%rsp)
@@ -727,18 +727,18 @@ USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
 #endif
 
 	movq    %rsp, SP      # save old stack
-        subq    $128 + L_BUFFER_SIZE, %rsp
-        andq    $-4096, %rsp    # align stack
+        subq    $ 128 + L_BUFFER_SIZE, %rsp
+        andq    $ -4096, %rsp    # align stack
 
         STACK_TOUCH
 
-	cmpq	$0, OLD_M
+	cmpq	$ 0, OLD_M
 	je	.L999
 
-	cmpq	$0, OLD_N
+	cmpq	$ 0, OLD_N
 	je	.L999
 
-	cmpq	$0, OLD_K
+	cmpq	$ 0, OLD_K
 	je	.L999
 
 	movq	OLD_M, M
@@ -748,11 +748,11 @@ USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
 	vmovsd	 %xmm0, ALPHA_R
 	vmovsd	 %xmm1, ALPHA_I
 
-	salq	$ZBASE_SHIFT, LDC
+	salq	$ ZBASE_SHIFT, LDC
 
 	movq    N, %rax
         xorq    %rdx, %rdx
-        movq    $2,  %rdi
+        movq    $ 2,  %rdi
         divq    %rdi                    //    N / 2
         movq    %rax, Ndiv6             //    N / 2
         movq    %rdx, Nmod6             //    N % 2
@@ -770,7 +770,7 @@ USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
 .L2_00_0:
 
 	movq	Ndiv6,  J
-	cmpq	$0, J
+	cmpq	$ 0, J
 	je	.L1_2_0
 	ALIGN_4
 
@@ -789,8 +789,8 @@ USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
 	vmovups	2 * SIZE(BO1), %xmm1
 	vmovups	%xmm0,       (BO)
 	vmovups	%xmm1, 2 * SIZE(BO)
-	addq	$4*SIZE,BO1
-	addq	$4*SIZE,BO
+	addq	$ 4*SIZE,BO1
+	addq	$ 4*SIZE,BO
 	decq	%rax
 	jnz	.L2_00_02b
 
@@ -809,10 +809,10 @@ USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
 #endif
 	
 	movq	A, AO		 	// aoffset = a
-	addq	$8 * SIZE, AO
+	addq	$ 8 * SIZE, AO
 
 	movq	M,  I
-	sarq	$2, I			// i = (m >> 2)
+	sarq	$ 2, I			// i = (m >> 2)
 	je	.L2_2_10
 
 	ALIGN_4
@@ -825,15 +825,15 @@ USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
         (defined(TRMMKERNEL) &&  defined(LEFT) &&  defined(TRANSA)) || \
         (defined(TRMMKERNEL) && !defined(LEFT) && !defined(TRANSA))
 	leaq	BUFFER1, BO		// first buffer to BO
-	addq	$8 * SIZE, BO
+	addq	$ 8 * SIZE, BO
 #else
         movq    KK, %rax
 	leaq	BUFFER1, BO			// first buffer to BO
-	addq	$8 * SIZE, BO
+	addq	$ 8 * SIZE, BO
 	movq    %rax, BI                        //  Index for BO
         leaq    (,BI,4), BI                     //  BI = BI * 4 ; number of values
         leaq    (BO, BI, SIZE), BO
-	salq	$3, %rax			// rax = rax * 8 ; number of values
+	salq	$ 3, %rax			// rax = rax * 8 ; number of values
         leaq    (AO, %rax, SIZE), AO
 #endif
 
@@ -848,20 +848,20 @@ USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
 #else
         movq    KK, %rax
 #ifdef LEFT
-        addq    $4, %rax        // number of values in AO
+        addq    $ 4, %rax        // number of values in AO
 #else
-        addq    $2, %rax        // number of values in BO
+        addq    $ 2, %rax        // number of values in BO
 #endif
         movq    %rax, KKK
 #endif
 
 
-	andq	$-8, %rax			//  K = K - ( K % 8 )
+	andq	$ -8, %rax			//  K = K - ( K % 8 )
 	je	.L2_4_16
 	movq    %rax, BI                        //  Index for BO
         leaq    ( ,BI,4), BI                    //  BI = BI * 4 ; number of values
 
-	salq	$3, %rax			// rax = rax * 8 ; number of values
+	salq	$ 3, %rax			// rax = rax * 8 ; number of values
 	leaq	(AO, %rax, SIZE), AO
 	leaq	(BO, BI, SIZE), BO
 	negq	BI
@@ -928,13 +928,13 @@ USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
         movq    KKK, %rax
 #endif
 
-	andq	$7, %rax		# if (k & 1)
+	andq	$ 7, %rax		# if (k & 1)
 	je .L2_4_19
 
 	movq    %rax, BI                        //  Index for BO
         leaq    ( ,BI,4), BI                    //  BI = BI * 4 ; number of values
 
-	salq	$3, %rax			// rax = rax * 8 ; number of values
+	salq	$ 3, %rax			// rax = rax * 8 ; number of values
 	leaq	(AO, %rax, SIZE), AO
 	leaq	(BO, BI, SIZE), BO
 	negq	BI
@@ -960,16 +960,16 @@ USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
 	movq    %rax, BI                        //  Index for BO
         leaq    ( ,BI,4), BI                    //  BI = BI * 4 ; number of values
         leaq    (BO, BI, SIZE), BO
-	salq	$3, %rax			// rax = rax * 8 ; number of values
+	salq	$ 3, %rax			// rax = rax * 8 ; number of values
         leaq    (AO, %rax, SIZE), AO
 #endif
 
 
 #if defined(TRMMKERNEL) && defined(LEFT)
-        addq    $4, KK
+        addq    $ 4, KK
 #endif
 
-	addq	$8 * SIZE, CO1		# coffset += 8
+	addq	$ 8 * SIZE, CO1		# coffset += 8
 	decq	I			# i --
 	jg	.L2_4_11
 	ALIGN_4	
@@ -982,7 +982,7 @@ USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
 
 /******************************************************************************************************************/
 .L2_2_10:
-	testq	$2, M		
+	testq	$ 2, M		
 	jz	.L2_2_40		// to next 2 lines of N
 
 .L2_2_11:
@@ -991,15 +991,15 @@ USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
         (defined(TRMMKERNEL) &&  defined(LEFT) &&  defined(TRANSA)) || \
         (defined(TRMMKERNEL) && !defined(LEFT) && !defined(TRANSA))
 	leaq	BUFFER1, BO		// first buffer to BO
-	addq	$8 * SIZE, BO
+	addq	$ 8 * SIZE, BO
 #else
         movq    KK, %rax
 	leaq	BUFFER1, BO			// first buffer to BO
-	addq	$8 * SIZE, BO
+	addq	$ 8 * SIZE, BO
 	movq    %rax, BI                        //  Index for BO
         leaq    (,BI,4), BI                     //  BI = BI * 4 ; number of values
         leaq    (BO, BI, SIZE), BO
-	salq	$2, %rax			// rax = rax * 4 ; number of values
+	salq	$ 2, %rax			// rax = rax * 4 ; number of values
         leaq    (AO, %rax, SIZE), AO
 #endif
 
@@ -1014,20 +1014,20 @@ USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
 #else
         movq    KK, %rax
 #ifdef LEFT
-        addq    $2, %rax        // number of values in AO
+        addq    $ 2, %rax        // number of values in AO
 #else
-        addq    $2, %rax        // number of values in BO
+        addq    $ 2, %rax        // number of values in BO
 #endif
         movq    %rax, KKK
 #endif
 
 
-	andq	$-8, %rax			//  K = K - ( K % 8 )
+	andq	$ -8, %rax			//  K = K - ( K % 8 )
 	je	.L2_2_16
 	movq    %rax, BI                        //  Index for BO
         leaq    ( ,BI,4), BI                    //  BI = BI * 4 ; number of values
 
-	salq	$2, %rax			// rax = rax * 4 ; number of values
+	salq	$ 2, %rax			// rax = rax * 4 ; number of values
 	leaq	(AO, %rax, SIZE), AO
 	leaq	(BO, BI, SIZE), BO
 	negq	BI
@@ -1086,13 +1086,13 @@ USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
         movq    KKK, %rax
 #endif
 
-	andq	$7, %rax		# if (k & 1)
+	andq	$ 7, %rax		# if (k & 1)
 	je .L2_2_19
 
 	movq    %rax, BI                        //  Index for BO
         leaq    ( ,BI,4), BI                    //  BI = BI * 4 ; number of values
 
-	salq	$2, %rax			// rax = rax * 4 ; number of values
+	salq	$ 2, %rax			// rax = rax * 4 ; number of values
 	leaq	(AO, %rax, SIZE), AO
 	leaq	(BO, BI, SIZE), BO
 	negq	BI
@@ -1118,16 +1118,16 @@ USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
 	movq    %rax, BI                        //  Index for BO
         leaq    ( ,BI,4), BI                    //  BI = BI * 4 ; number of values
         leaq    (BO, BI, SIZE), BO
-	salq	$2, %rax			// rax = rax * 4 ; number of values
+	salq	$ 2, %rax			// rax = rax * 4 ; number of values
         leaq    (AO, %rax, SIZE), AO
 #endif
 
 
 #if defined(TRMMKERNEL) && defined(LEFT)
-        addq    $2, KK
+        addq    $ 2, KK
 #endif
 
-	addq	$4 * SIZE, CO1		# coffset += 4
+	addq	$ 4 * SIZE, CO1		# coffset += 4
 	ALIGN_4	
 
 
@@ -1135,7 +1135,7 @@ USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
 * Rest of M 
 ***************************************************************************/
 .L2_2_40:
-	testq	$1, M		
+	testq	$ 1, M		
 	jz	.L2_2_60		// to next 2 lines of N
 
 	ALIGN_4
@@ -1146,15 +1146,15 @@ USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
         (defined(TRMMKERNEL) &&  defined(LEFT) &&  defined(TRANSA)) || \
         (defined(TRMMKERNEL) && !defined(LEFT) && !defined(TRANSA))
 	leaq	BUFFER1, BO		// first buffer to BO
-	addq	$8 * SIZE, BO
+	addq	$ 8 * SIZE, BO
 #else
         movq    KK, %rax
 	leaq	BUFFER1, BO			// first buffer to BO
-	addq	$8 * SIZE, BO
+	addq	$ 8 * SIZE, BO
 	movq    %rax, BI                        //  Index for BO
         leaq    (,BI,4), BI                     //  BI = BI * 4 ; number of values
         leaq    (BO, BI, SIZE), BO
-	salq	$1, %rax			// rax = rax * 2 ; number of values
+	salq	$ 1, %rax			// rax = rax * 2 ; number of values
         leaq    (AO, %rax, SIZE), AO
 #endif
 
@@ -1169,20 +1169,20 @@ USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
 #else
         movq    KK, %rax
 #ifdef LEFT
-        addq    $1, %rax        // number of values in AO
+        addq    $ 1, %rax        // number of values in AO
 #else
-        addq    $2, %rax        // number of values in BO
+        addq    $ 2, %rax        // number of values in BO
 #endif
         movq    %rax, KKK
 #endif
 
 
-	andq	$-8, %rax			//  K = K - ( K % 8 )
+	andq	$ -8, %rax			//  K = K - ( K % 8 )
 	je	.L2_2_46
 	movq    %rax, BI                        //  Index for BO
         leaq    ( ,BI,4), BI                    //  BI = BI * 4 ; number of values
 
-	salq	$1, %rax			// rax = rax * 2 ; number of values
+	salq	$ 1, %rax			// rax = rax * 2 ; number of values
 	leaq	(AO, %rax, SIZE), AO
 	leaq	(BO, BI, SIZE), BO
 	negq	BI
@@ -1237,13 +1237,13 @@ USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
         movq    KKK, %rax
 #endif
 
-	andq	$7, %rax		# if (k & 1)
+	andq	$ 7, %rax		# if (k & 1)
 	je .L2_2_49
 
 	movq    %rax, BI                        //  Index for BO
         leaq    ( ,BI,4), BI                    //  BI = BI * 4 ; number of values
 
-	salq	$1, %rax			// rax = rax * 2 ; number of values
+	salq	$ 1, %rax			// rax = rax * 2 ; number of values
 	leaq	(AO, %rax, SIZE), AO
 	leaq	(BO, BI, SIZE), BO
 	negq	BI
@@ -1269,16 +1269,16 @@ USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
 	movq    %rax, BI                        //  Index for BO
         leaq    ( ,BI,4), BI                    //  BI = BI * 4 ; number of values
         leaq    (BO, BI, SIZE), BO
-	salq	$1, %rax			// rax = rax * 2 ; number of values
+	salq	$ 1, %rax			// rax = rax * 2 ; number of values
         leaq    (AO, %rax, SIZE), AO
 #endif
 
 
 #if defined(TRMMKERNEL) && defined(LEFT)
-        addq    $1, KK
+        addq    $ 1, KK
 #endif
 
-	addq	$2 * SIZE, CO1		# coffset += 2
+	addq	$ 2 * SIZE, CO1		# coffset += 2
 	decq	I			# i --
 	jg	.L2_2_41
 	ALIGN_4	
@@ -1288,7 +1288,7 @@ USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
 	
 .L2_2_60:
 #if defined(TRMMKERNEL) && !defined(LEFT)
-        addq    $2, KK
+        addq    $ 2, KK
 #endif
 
 	decq	J			// j --
@@ -1303,7 +1303,7 @@ USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
 *************************************************************************************************/
 
 	movq	Nmod6, J		
-	andq	$1, J			// j % 2
+	andq	$ 1, J			// j % 2
 	je	.L999
 	ALIGN_4
 
@@ -1318,8 +1318,8 @@ USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
 
 	vmovups		(BO1), %xmm0
 	vmovups	%xmm0,       (BO)
-	addq	$2*SIZE,BO1
-	addq	$2*SIZE,BO
+	addq	$ 2*SIZE,BO1
+	addq	$ 2*SIZE,BO
 	decq	%rax
 	jnz	.L1_00_02b
 
@@ -1337,10 +1337,10 @@ USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
 #endif
 	
 	movq	A, AO		 	// aoffset = a
-	addq	$8 * SIZE, AO
+	addq	$ 8 * SIZE, AO
 
 	movq	M,  I
-	sarq	$2, I			// i = (m >> 2)
+	sarq	$ 2, I			// i = (m >> 2)
 	je	.L1_2_10
 
 	ALIGN_4
@@ -1354,15 +1354,15 @@ USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
         (defined(TRMMKERNEL) &&  defined(LEFT) &&  defined(TRANSA)) || \
         (defined(TRMMKERNEL) && !defined(LEFT) && !defined(TRANSA))
 	leaq	BUFFER1, BO		// first buffer to BO
-	addq	$4 * SIZE, BO
+	addq	$ 4 * SIZE, BO
 #else
         movq    KK, %rax
 	leaq	BUFFER1, BO			// first buffer to BO
-	addq	$4 * SIZE, BO
+	addq	$ 4 * SIZE, BO
 	movq    %rax, BI                        //  Index for BO
         leaq    (,BI,2), BI                     //  BI = BI * 2 ; number of values
         leaq    (BO, BI, SIZE), BO
-	salq	$3, %rax			// rax = rax * 8 ; number of values
+	salq	$ 3, %rax			// rax = rax * 8 ; number of values
         leaq    (AO, %rax, SIZE), AO
 #endif
 
@@ -1377,20 +1377,20 @@ USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
 #else
         movq    KK, %rax
 #ifdef LEFT
-        addq    $4, %rax        // number of values in AO
+        addq    $ 4, %rax        // number of values in AO
 #else
-        addq    $1, %rax        // number of values in BO
+        addq    $ 1, %rax        // number of values in BO
 #endif
         movq    %rax, KKK
 #endif
 
 
-	andq	$-8, %rax			//  K = K - ( K % 8 )
+	andq	$ -8, %rax			//  K = K - ( K % 8 )
 	je	.L1_4_16
 	movq    %rax, BI                        //  Index for BO
         leaq    ( ,BI,2), BI                    //  BI = BI * 2 ; number of values
 
-	salq	$3, %rax			// rax = rax * 8 ; number of values
+	salq	$ 3, %rax			// rax = rax * 8 ; number of values
 	leaq	(AO, %rax, SIZE), AO
 	leaq	(BO, BI, SIZE), BO
 	negq	BI
@@ -1433,13 +1433,13 @@ USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
         movq    KKK, %rax
 #endif
 
-	andq	$7, %rax		# if (k & 1)
+	andq	$ 7, %rax		# if (k & 1)
 	je .L1_4_19
 
 	movq    %rax, BI                        //  Index for BO
         leaq    ( ,BI,2), BI                    //  BI = BI * 2 ; number of values
 
-	salq	$3, %rax			// rax = rax * 8 ; number of values
+	salq	$ 3, %rax			// rax = rax * 8 ; number of values
 	leaq	(AO, %rax, SIZE), AO
 	leaq	(BO, BI, SIZE), BO
 	negq	BI
@@ -1466,16 +1466,16 @@ USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
 	movq    %rax, BI                        //  Index for BO
         leaq    ( ,BI,2), BI                    //  BI = BI * 2 ; number of values
         leaq    (BO, BI, SIZE), BO
-	salq	$3, %rax			// rax = rax * 8 ; number of values
+	salq	$ 3, %rax			// rax = rax * 8 ; number of values
         leaq    (AO, %rax, SIZE), AO
 #endif
 
 
 #if defined(TRMMKERNEL) && defined(LEFT)
-        addq    $4, KK
+        addq    $ 4, KK
 #endif
 
-	addq	$8 * SIZE, CO1		# coffset += 8
+	addq	$ 8 * SIZE, CO1		# coffset += 8
 	decq	I			# i --
 	jg	.L1_4_11
 	ALIGN_4	
@@ -1485,7 +1485,7 @@ USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
 
 /*******************************************************************************************************/
 .L1_2_10:
-	testq	$2, M		
+	testq	$ 2, M		
 	jz	.L1_2_40
 
 
@@ -1495,15 +1495,15 @@ USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
         (defined(TRMMKERNEL) &&  defined(LEFT) &&  defined(TRANSA)) || \
         (defined(TRMMKERNEL) && !defined(LEFT) && !defined(TRANSA))
 	leaq	BUFFER1, BO		// first buffer to BO
-	addq	$4 * SIZE, BO
+	addq	$ 4 * SIZE, BO
 #else
         movq    KK, %rax
 	leaq	BUFFER1, BO			// first buffer to BO
-	addq	$4 * SIZE, BO
+	addq	$ 4 * SIZE, BO
 	movq    %rax, BI                        //  Index for BO
         leaq    (,BI,2), BI                     //  BI = BI * 2 ; number of values
         leaq    (BO, BI, SIZE), BO
-	salq	$2, %rax			// rax = rax * 4 ; number of values
+	salq	$ 2, %rax			// rax = rax * 4 ; number of values
         leaq    (AO, %rax, SIZE), AO
 #endif
 
@@ -1518,20 +1518,20 @@ USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
 #else
         movq    KK, %rax
 #ifdef LEFT
-        addq    $2, %rax        // number of values in AO
+        addq    $ 2, %rax        // number of values in AO
 #else
-        addq    $1, %rax        // number of values in BO
+        addq    $ 1, %rax        // number of values in BO
 #endif
         movq    %rax, KKK
 #endif
 
 
-	andq	$-8, %rax			//  K = K - ( K % 8 )
+	andq	$ -8, %rax			//  K = K - ( K % 8 )
 	je	.L1_2_16
 	movq    %rax, BI                        //  Index for BO
         leaq    ( ,BI,2), BI                    //  BI = BI * 2 ; number of values
 
-	salq	$2, %rax			// rax = rax * 4 ; number of values
+	salq	$ 2, %rax			// rax = rax * 4 ; number of values
 	leaq	(AO, %rax, SIZE), AO
 	leaq	(BO, BI, SIZE), BO
 	negq	BI
@@ -1583,13 +1583,13 @@ USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
         movq    KKK, %rax
 #endif
 
-	andq	$7, %rax		# if (k & 1)
+	andq	$ 7, %rax		# if (k & 1)
 	je .L1_2_19
 
 	movq    %rax, BI                        //  Index for BO
         leaq    ( ,BI,2), BI                    //  BI = BI * 2 ; number of values
 
-	salq	$2, %rax			// rax = rax * 4 ; number of values
+	salq	$ 2, %rax			// rax = rax * 4 ; number of values
 	leaq	(AO, %rax, SIZE), AO
 	leaq	(BO, BI, SIZE), BO
 	negq	BI
@@ -1615,16 +1615,16 @@ USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
 	movq    %rax, BI                        //  Index for BO
         leaq    ( ,BI,2), BI                    //  BI = BI * 2 ; number of values
         leaq    (BO, BI, SIZE), BO
-	salq	$2, %rax			// rax = rax * 4 ; number of values
+	salq	$ 2, %rax			// rax = rax * 4 ; number of values
         leaq    (AO, %rax, SIZE), AO
 #endif
 
 
 #if defined(TRMMKERNEL) && defined(LEFT)
-        addq    $2, KK
+        addq    $ 2, KK
 #endif
 
-	addq	$4 * SIZE, CO1		# coffset += 4
+	addq	$ 4 * SIZE, CO1		# coffset += 4
 
 	ALIGN_4	
 
@@ -1633,7 +1633,7 @@ USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
 * Rest of M 
 ***************************************************************************/
 .L1_2_40:
-	testq	$1, M		
+	testq	$ 1, M		
 	jz	.L999
 
 	ALIGN_4
@@ -1644,15 +1644,15 @@ USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
         (defined(TRMMKERNEL) &&  defined(LEFT) &&  defined(TRANSA)) || \
         (defined(TRMMKERNEL) && !defined(LEFT) && !defined(TRANSA))
 	leaq	BUFFER1, BO		// first buffer to BO
-	addq	$4 * SIZE, BO
+	addq	$ 4 * SIZE, BO
 #else
         movq    KK, %rax
 	leaq	BUFFER1, BO			// first buffer to BO
-	addq	$4 * SIZE, BO
+	addq	$ 4 * SIZE, BO
 	movq    %rax, BI                        //  Index for BO
         leaq    (,BI,2), BI                     //  BI = BI * 2 ; number of values
         leaq    (BO, BI, SIZE), BO
-	salq	$1, %rax			// rax = rax * 2 ; number of values
+	salq	$ 1, %rax			// rax = rax * 2 ; number of values
         leaq    (AO, %rax, SIZE), AO
 #endif
 
@@ -1667,20 +1667,20 @@ USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
 #else
         movq    KK, %rax
 #ifdef LEFT
-        addq    $1, %rax        // number of values in AO
+        addq    $ 1, %rax        // number of values in AO
 #else
-        addq    $1, %rax        // number of values in BO
+        addq    $ 1, %rax        // number of values in BO
 #endif
         movq    %rax, KKK
 #endif
 
 
-	andq	$-8, %rax			//  K = K - ( K % 8 )
+	andq	$ -8, %rax			//  K = K - ( K % 8 )
 	je	.L1_2_46
 	movq    %rax, BI                        //  Index for BO
         leaq    ( ,BI,2), BI                    //  BI = BI * 2 ; number of values
 
-	salq	$1, %rax			// rax = rax * 2 ; number of values
+	salq	$ 1, %rax			// rax = rax * 2 ; number of values
 	leaq	(AO, %rax, SIZE), AO
 	leaq	(BO, BI, SIZE), BO
 	negq	BI
@@ -1731,13 +1731,13 @@ USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
         movq    KKK, %rax
 #endif
 
-	andq	$7, %rax		# if (k & 1)
+	andq	$ 7, %rax		# if (k & 1)
 	je .L1_2_49
 
 	movq    %rax, BI                        //  Index for BO
         leaq    ( ,BI,2), BI                    //  BI = BI * 2 ; number of values
 
-	salq	$1, %rax			// rax = rax * 2 ; number of values
+	salq	$ 1, %rax			// rax = rax * 2 ; number of values
 	leaq	(AO, %rax, SIZE), AO
 	leaq	(BO, BI, SIZE), BO
 	negq	BI
@@ -1763,16 +1763,16 @@ USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
 	movq    %rax, BI                        //  Index for BO
         leaq    ( ,BI,2), BI                    //  BI = BI * 2 ; number of values
         leaq    (BO, BI, SIZE), BO
-	salq	$1, %rax			// rax = rax * 2 ; number of values
+	salq	$ 1, %rax			// rax = rax * 2 ; number of values
         leaq    (AO, %rax, SIZE), AO
 #endif
 
 
 #if defined(TRMMKERNEL) && defined(LEFT)
-        addq    $1, KK
+        addq    $ 1, KK
 #endif
 
-	addq	$2 * SIZE, CO1		# coffset += 2
+	addq	$ 2 * SIZE, CO1		# coffset += 2
 	decq	I			# i --
 	jg	.L1_2_41
 	ALIGN_4	
@@ -1806,7 +1806,7 @@ USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
 	movups	208(%rsp), %xmm15
 #endif
 
-	addq	$STACKSIZE, %rsp
+	addq	$ STACKSIZE, %rsp
 	ret
 
 	EPILOGUE