RFR: 8339910: RISC-V: crc32 intrinsic with carry-less multiplication

[Andrew Haley]

On Mon, 2 Dec 2024 10:00:11 GMT, Hamlin Li <mli at openjdk.org> wrote:

> Hi,
> Can you review this patch to implement CRC32 with `vclmul` (Zvbc)?
> Thanks!
> 
> 
> ## Test
> test/hotspot/jtreg/compiler/intrinsics/zip/TestCRC32.java, 
> test/jdk/java/util/zip/TestCRC32.java

src/hotspot/cpu/riscv/macroAssembler_riscv.cpp line 1761:

> 1759:   bind(L_16_bytes_loop);
> 1760:   {
> 1761:     #define CRC32_VCLMUL_FOLD_16_BYTES(vx, vt, vtmp1, vtmp2, vtmp3, vtmp4) \

It makes no sense for this to be a macro, because it's called infrequently. Making it a function would save space.

src/hotspot/cpu/riscv/macroAssembler_riscv.cpp line 1833:

> 1831:     vslideup_vi(vy, vtmp1, 1); \
> 1832:     vmv_x_s(tmp2, vtmp3); \
> 1833:     vmv_s_x(vy, tmp2);

Again, not a macro.

-------------

PR Review Comment: https://git.openjdk.org/jdk/pull/22475#discussion_r1865652614
PR Review Comment: https://git.openjdk.org/jdk/pull/22475#discussion_r1865653545