This is an automated email from the ASF dual-hosted git repository.
chenBright pushed a commit to branch master
in repository https://gitbox.apache.org/repos/asf/brpc.git
The following commit(s) were added to refs/heads/master by this push:
new 7ef1d00e CRC32C RISC-V: pick clmul or vclmul by size when both are
available (#3601)
7ef1d00e is described below
commit 7ef1d00efd9a35f1ba3a6017f1cf4f856b21905d
Author: Felix-Gong <[email protected]>
AuthorDate: Sun Oct 18 17:05:02 2026 +0800
CRC32C RISC-V: pick clmul or vclmul by size when both are available (#3601)
PR3332 sent every Zbc+Zvbc CPU to the vclmul path. After #3571
rewrote it, vclmul beats clmul from ~256 B up (1.5x-7.5x) but costs
more below that, where vector setup dominates. A fixed order to
either path is wrong for these CPUs.
Select by size in the dispatcher: clmul under 256 B, vclmul from
256 B on. CPUs with only Zbc or only Zvbc keep their current path,
and Extend() is untouched.
k3 (X100, VLEN=256, 20-round median):
256 B clmul 1534 vclmul 2155
1 MiB clmul 1555 vclmul 11702 (MB/s)
Signed-off-by: Xiaofei Gong <[email protected]>
Signed-off-by: YuanSheng <[email protected]>
---
src/butil/crc32c.cc | 28 +++++++++++++++++++++++++---
1 file changed, 25 insertions(+), 3 deletions(-)
diff --git a/src/butil/crc32c.cc b/src/butil/crc32c.cc
index ac81c7f6..47157093 100644
--- a/src/butil/crc32c.cc
+++ b/src/butil/crc32c.cc
@@ -814,6 +814,15 @@ static bool isSSE42() {
typedef uint32_t (*Function)(uint32_t, const char*, size_t);
+#if defined(__riscv) && (__riscv_xlen == 64) && defined(__riscv_zbc) &&
defined(__riscv_zvbc)
+// Both Zbc and Zvbc: scalar is cheaper for small inputs, the vector fold
+// wins from ~256 B on.
+static constexpr size_t kRiscvVclmulThresholdBytes = 256;
+static uint32_t RiscvExtend(uint32_t crc, const char* buf, size_t size) {
+ return (size >= kRiscvVclmulThresholdBytes ? rv_crc32c_vclmul :
rv_crc32c_clmul)(crc, buf, size);
+}
+#endif
+
static inline Function Choose_Extend() {
#ifdef __SSE4_2__
if (isSSE42()) {
@@ -821,12 +830,25 @@ static inline Function Choose_Extend() {
}
#endif
#if defined(__riscv) && (__riscv_xlen == 64) && (defined(__riscv_zbc) ||
defined(__riscv_zvbc))
-#if defined(__riscv_zvbc)
+#if defined(__riscv_zbc) && defined(__riscv_zvbc)
+ {
+ static const bool has_zbc = is_zbc();
+ static const bool has_zvbc = is_zvbc();
+ if (has_zbc && has_zvbc) {
+ return (Function)RiscvExtend;
+ }
+ if (has_zvbc) {
+ return (Function)rv_crc32c_vclmul;
+ }
+ if (has_zbc) {
+ return (Function)rv_crc32c_clmul;
+ }
+ }
+#elif defined(__riscv_zvbc)
if (is_zvbc()) {
return (Function)rv_crc32c_vclmul;
}
-#endif
-#if defined(__riscv_zbc)
+#elif defined(__riscv_zbc)
if (is_zbc()) {
return (Function)rv_crc32c_clmul;
}
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
