Mercurial > jdk8-mips64-public > hotspot / file comparison
comparison: src/cpu/x86/vm/assembler_x86.hpp
src/cpu/x86/vm/assembler_x86.hpp
- changeset 4103
- 137868b7aa6f
- parent 4052
- 75f33eecc1b3
- child 4159
- 8e47bac5643a
equal
deleted
inserted
replaced
| 4102:8d3cc6612bd1 | 4103:137868b7aa6f |
|---|---|
| 1741 | 1741 |
| 1742 // Copy low 128bit into high 128bit of YMM registers. | 1742 // Copy low 128bit into high 128bit of YMM registers. |
| 1743 void vinsertf128h(XMMRegister dst, XMMRegister nds, XMMRegister src); | 1743 void vinsertf128h(XMMRegister dst, XMMRegister nds, XMMRegister src); |
| 1744 void vinserti128h(XMMRegister dst, XMMRegister nds, XMMRegister src); | 1744 void vinserti128h(XMMRegister dst, XMMRegister nds, XMMRegister src); |
| 1745 | 1745 |
| 1746 // Load/store high 128bit of YMM registers which does not destroy other half. | |
| 1747 void vinsertf128h(XMMRegister dst, Address src); | |
| 1748 void vinserti128h(XMMRegister dst, Address src); | |
| 1749 void vextractf128h(Address dst, XMMRegister src); | |
| 1750 void vextracti128h(Address dst, XMMRegister src); | |
| 1751 | |
| 1746 // AVX instruction which is used to clear upper 128 bits of YMM registers and | 1752 // AVX instruction which is used to clear upper 128 bits of YMM registers and |
| 1747 // to avoid transaction penalty between AVX and SSE states. There is no | 1753 // to avoid transaction penalty between AVX and SSE states. There is no |
| 1748 // penalty if legacy SSE instructions are encoded using VEX prefix because | 1754 // penalty if legacy SSE instructions are encoded using VEX prefix because |
| 1749 // they always clear upper 128 bits. It should be used before calling | 1755 // they always clear upper 128 bits. It should be used before calling |
| 1750 // runtime code and native libraries. | 1756 // runtime code and native libraries. |
