aarch64: ACLE I8MM multiply-accumulate intrinsics

aarch64: ACLE I8MM multiply-accumulate intrinsics

This patch adds intrinsics for 8-bit integer matrix multiply-accumulate
operations including vmmlaq_s32, vmmlaq_u32, and vusmmlaq_s32.

gcc/ChangeLog:

2020-02-07  Dennis Zhang  <dennis.zhang@arm.com>

	* config/aarch64/aarch64-simd-builtins.def (simd_smmla): New entry.
	(simd_ummla, simd_usmmla): Likewise.
	* config/aarch64/aarch64-simd.md (aarch64_simd_<sur>mmlav16qi): New.
	* config/aarch64/arm_neon.h (vmmlaq_s32, vmmlaq_u32): New.
	(vusmmlaq_s32): New.

gcc/testsuite/ChangeLog:

2020-02-07  Dennis Zhang  <dennis.zhang@arm.com>

	* gcc.target/aarch64/simd/vmmla.c: New test.