x86: Add high bit-depth film grain AVX-512 (Ice Lake) asm

fgy_32x32xn_16bpc_avx2:                 1217.0
fgy_32x32xn_16bpc_avx512icl:            1175.0

fguv_32x32xn_16bpc_420_csfl0_avx2:       714.9
fguv_32x32xn_16bpc_420_csfl0_avx512icl:  498.1
fguv_32x32xn_16bpc_420_csfl1_avx2:       420.9
fguv_32x32xn_16bpc_420_csfl1_avx512icl:  413.0

fguv_32x32xn_16bpc_422_csfl0_avx2:      1400.8
fguv_32x32xn_16bpc_422_csfl0_avx512icl:  967.5
fguv_32x32xn_16bpc_422_csfl1_avx2:       821.8
fguv_32x32xn_16bpc_422_csfl1_avx512icl:  818.0

fguv_32x32xn_16bpc_444_csfl0_avx2:      2691.0
fguv_32x32xn_16bpc_444_csfl0_avx512icl: 1530.8
fguv_32x32xn_16bpc_444_csfl1_avx2:      1378.8
fguv_32x32xn_16bpc_444_csfl1_avx512icl: 1193.8

Helps the most in the csfl0 chroma functions due to those being the most computationally expensive relative to the number of memory loads.

The other ones are bottlenecked by the number of memory loads that can be executed per cycle, at least on RKL which has two load ports. Some newer µarchs has three load ports so it might help more on such systems. The number of retired instructions is much lower compared to AVX2 in any case though, which should free up more resources for the other SMT-thread in multi-threaded scenarios.

added performance x86 labels

requested review from @psilokos

mentioned in issue #316

added 8 commits

• 18d80e98...28a9c46e - 6 commits from branch videolan:master
• ca6fe6a6 - x86: Add high bit-depth film grain AVX-512 (Ice Lake) asm
• 0b0daff8 - x86: Reduce code size in 8-bit film grain AVX-512 asm

Compare with previous version

I really like the code sharing between the different overlap variants, very nice.

approved this merge request

resolved all threads

added 2 commits

• 949b8902 - x86: Add high bit-depth film grain AVX-512 (Ice Lake) asm
• b1a5189c - x86: Reduce code size in 8-bit film grain AVX-512 asm

Compare with previous version

merged

changed milestone to %1.0.0