[use movq; movhpd instead of movdqa to allow unaligned stores
peter@cordes.ca**20080319053541] {
hunk ./rshift.asm 214
-C size 10000	5.244	c/l (no ALIGN: 5.257)
-C size 10000000	14.053	c/l. (no ALIGN: 14.529)
+C size 10000	5.244	c/l (no ALIGN: 5.257. movhpd: 5.248)
+C size 10000000	14.053	c/l. (no ALIGN: 14.529. movhpd: 14.005)
hunk ./rshift.asm 220
-C size 3	5.376	c/l  (no ALIGN: 5.312)
-C size 4	4.040	c/l  (no ALIGN: 3.980, and one less icache line touched?)
+C size 3	5.376	c/l  (no ALIGN: 5.312. unaligned movhpd stores: 5.355)
+C size 4	4.040	c/l  (no ALIGN: 3.980, and one less icache line touched?) (movhpd: 4.180)
hunk ./rshift.asm 225
-C size 496:	2.052	c/l. (no ALIGN: 2.052)
-C size 10000	2.320	c/l  (no ALIGN: 2.320)
-C size 10000000 11.178  c/l. (no ALIGN: 11.195)  (2.4GHz, g965, dual channel DDR800)
+C size 496:	2.052	c/l. (no ALIGN: 2.052) (movhpd: 2.132)
+C size 10000	2.320	c/l  (no ALIGN: 2.320) (movhpd: 2.656)
+C size 10000000 11.178  c/l. (no ALIGN: 11.195. movhpd: 11.520))  (2.4GHz, g965, dual channel DDR800)
hunk ./rshift.asm 239
-C size 10000	2.847	c/l  (no ALIGN: 2.968)
-C size 10000000	14.460	c/l. (no ALIGN: 14.403)
+C size 10000	2.847	c/l  (no ALIGN: 2.968. movhpd: 3.566)
+C size 10000000	14.460	c/l. (no ALIGN: 14.403. movhpd: 14.294)
hunk ./rshift.asm 275
-	movdqa	%xmm6,	(%rdi,%rdx,8) C store the result.
+	movq	%xmm6,	(%rdi,%rdx,8) C store the result.
+	movhpd	%xmm6,	8(%rdi,%rdx,8) C store the result.
+
hunk ./rshift.asm 279
-C	movq	16(%rsi,%rdx,8),	%xmm6	C %6 = limb3  (about to become limb1)
hunk ./rshift.asm 294
-	movdqa	%xmm6, (%rdi)	C store the result.
+C	movdqa	%xmm6, (%rdi)	C store the result.
+	movq	%xmm6,	(%rdi)	C store the result.
+	movhpd	%xmm6,	8(%rdi)	C store the result.
}