Using only 16 registers of the 32 registers

Hello

Any updates that fixed the problem of using only 16 registers of the 32 available for KNL and Skylake?

I tried qopt-zmm-usage=low/high and so on and it seems that nothing works... Here is  an assembly example:

  40122e:       62 71 7c 48 28 fd       vmovaps %zmm5,%zmm15
  401234:       62 d1 85 48 5c ec       vsubpd %zmm12,%zmm15,%zmm5
  40123a:       62 51 ed 48 58 fa       vaddpd %zmm10,%zmm2,%zmm15
  401240:       62 c1 7c 48 28 ca       vmovaps %zmm10,%zmm17
  401246:       62 31 7c 48 28 e0       vmovaps %zmm16,%zmm12
  40124c:       62 61 7c 48 28 e9       vmovaps %zmm1,%zmm29
  401252:       62 d1 9d 48 58 c8       vaddpd %zmm8,%zmm12,%zmm1
  401258:       62 61 7c 48 28 e4       vmovaps %zmm4,%zmm28
  40125e:       62 b1 7c 48 28 e1       vmovaps %zmm17,%zmm4
  401264:       62 61 7c 48 28 d9       vmovaps %zmm1,%zmm27
  40126a:       62 f1 dd 48 5c ca       vsubpd %zmm2,%zmm4,%zmm1

An alternative is to use only the 16 registers... is there a flag that actually works and restricts the number of registers to 16?

Thanks