[dpdk-dev] [PATCH v2 3/3] app/testpmd: further improve MAC swap performance for x86

Qi Zhang qi.z.zhang at intel.com
Tue Dec 11 06:55:11 CET 2018


Do four packets macswap in same loop iterate to squeeze more
CPU cycles.

Signed-off-by: Qi Zhang <qi.z.zhang at intel.com>
---
 app/test-pmd/macswap_sse.h | 62 +++++++++++++++++++++++++++++++++++++---------
 1 file changed, 50 insertions(+), 12 deletions(-)

diff --git a/app/test-pmd/macswap_sse.h b/app/test-pmd/macswap_sse.h
index 79f4f9a7c..df2875ace 100644
--- a/app/test-pmd/macswap_sse.h
+++ b/app/test-pmd/macswap_sse.h
@@ -11,11 +11,12 @@ static inline void
 do_macswap(struct rte_mbuf *pkts[], uint16_t nb,
 		struct rte_port *txp)
 {
-	struct ether_hdr *eth_hdr;
-	struct rte_mbuf *mb;
+	struct ether_hdr *eth_hdr[4];
+	struct rte_mbuf *mb[4];
 	uint64_t ol_flags;
 	int i;
-	__m128i addr;
+	int r;
+	__m128i addr0, addr1, addr2, addr3;
 	__m128i shfl_msk = _mm_set_epi8(15, 14, 13, 12,
 					5, 4, 3, 2,
 					1, 0, 11, 10,
@@ -25,19 +26,56 @@ do_macswap(struct rte_mbuf *pkts[], uint16_t nb,
 	vlan_qinq_set(pkts, nb, ol_flags,
 			txp->tx_vlan_id, txp->tx_vlan_id_outer);
 
-	for (i = 0; i < nb; i++) {
-		if (likely(i < nb - 1))
-			rte_prefetch0(rte_pktmbuf_mtod(pkts[i+1], void *));
-		mb = pkts[i];
+	i = 0;
+	r = nb;
+
+	while (r >= 4) {
+		mb[0] = pkts[i++];
+		eth_hdr[0] = rte_pktmbuf_mtod(mb[0], struct ether_hdr *);
+		addr0 = _mm_loadu_si128((__m128i *)eth_hdr[0]);
+
+		mb[1] = pkts[i++];
+		eth_hdr[1] = rte_pktmbuf_mtod(mb[1], struct ether_hdr *);
+		addr1 = _mm_loadu_si128((__m128i *)eth_hdr[1]);
+
+
+		mb[2] = pkts[i++];
+		eth_hdr[2] = rte_pktmbuf_mtod(mb[2], struct ether_hdr *);
+		addr2 = _mm_loadu_si128((__m128i *)eth_hdr[2]);
+
+		mb[3] = pkts[i++];
+		eth_hdr[3] = rte_pktmbuf_mtod(mb[3], struct ether_hdr *);
+		addr3 = _mm_loadu_si128((__m128i *)eth_hdr[3]);
 
-		eth_hdr = rte_pktmbuf_mtod(mb, struct ether_hdr *);
+		addr0 = _mm_shuffle_epi8(addr0, shfl_msk);
+		addr1 = _mm_shuffle_epi8(addr1, shfl_msk);
+		addr2 = _mm_shuffle_epi8(addr2, shfl_msk);
+		addr3 = _mm_shuffle_epi8(addr3, shfl_msk);
+
+		_mm_storeu_si128((__m128i *)eth_hdr[0], addr0);
+		_mm_storeu_si128((__m128i *)eth_hdr[1], addr1);
+		_mm_storeu_si128((__m128i *)eth_hdr[2], addr2);
+		_mm_storeu_si128((__m128i *)eth_hdr[3], addr3);
+
+		mbuf_field_set(mb[0], ol_flags);
+		mbuf_field_set(mb[1], ol_flags);
+		mbuf_field_set(mb[2], ol_flags);
+		mbuf_field_set(mb[3], ol_flags);
+		r -= 4;
+	}
+
+	for ( ; i < nb; i++) {
+		if (i < nb - 1)
+			rte_prefetch0(rte_pktmbuf_mtod(pkts[i+1], void *));
+		mb[0] = pkts[i];
+		eth_hdr[0] = rte_pktmbuf_mtod(mb[0], struct ether_hdr *);
 
 		/* Swap dest and src mac addresses. */
-		addr = _mm_loadu_si128((__m128i *)eth_hdr);
-		addr = _mm_shuffle_epi8(addr, shfl_msk);
-		_mm_storeu_si128((__m128i *)eth_hdr, addr);
+		addr0 = _mm_loadu_si128((__m128i *)eth_hdr);
+		addr0 = _mm_shuffle_epi8(addr0, shfl_msk);
+		_mm_storeu_si128((__m128i *)eth_hdr[0], addr0);
 
-		mbuf_field_set(mb, ol_flags);
+		mbuf_field_set(mb[0], ol_flags);
 	}
 }
 
-- 
2.13.6



More information about the dev mailing list