From mboxrd@z Thu Jan 1 00:00:00 1970
Return-Path: <xiaoyun.li@intel.com>
Received: from mga01.intel.com (mga01.intel.com [192.55.52.88])
by dpdk.org (Postfix) with ESMTP id 8818237B1
for <dev@dpdk.org>; Thu, 5 Oct 2017 14:35:03 +0200 (CEST)
Received: from orsmga004.jf.intel.com ([10.7.209.38])
by fmsmga101.fm.intel.com with ESMTP/TLS/DHE-RSA-AES256-GCM-SHA384;
05 Oct 2017 05:35:03 -0700
X-ExtLoop1: 1
X-IronPort-AV: E=Sophos;i="5.42,481,1500966000"; d="scan'208";a="135502582"
Received: from dpdk-lixiaoyun.sh.intel.com ([10.67.111.93])
by orsmga004.jf.intel.com with ESMTP; 05 Oct 2017 05:35:01 -0700
From: Xiaoyun Li <xiaoyun.li@intel.com>
To: konstantin.ananyev@intel.com,
bruce.richardson@intel.com
Cc: wenzhuo.lu@intel.com, helin.zhang@intel.com, dev@dpdk.org,
Xiaoyun Li <xiaoyun.li@intel.com>
Date: Thu, 5 Oct 2017 20:33:13 +0800
Message-Id: <1507206794-79941-3-git-send-email-xiaoyun.li@intel.com>
X-Mailer: git-send-email 2.7.4
In-Reply-To: <1507206794-79941-1-git-send-email-xiaoyun.li@intel.com>
References: <1507157911-8702-1-git-send-email-xiaoyun.li@intel.com>
<1507206794-79941-1-git-send-email-xiaoyun.li@intel.com>
Subject: [dpdk-dev] [PATCH v7 2/3] app/test: run-time dispatch over memcpy
perf test
X-BeenThere: dev@dpdk.org
X-Mailman-Version: 2.1.15
Precedence: list
List-Id: DPDK patches and discussions <dev.dpdk.org>
List-Unsubscribe: <http://dpdk.org/ml/options/dev>,
<mailto:dev-request@dpdk.org?subject=unsubscribe>
List-Archive: <http://dpdk.org/ml/archives/dev/>
List-Post: <mailto:dev@dpdk.org>
List-Help: <mailto:dev-request@dpdk.org?subject=help>
List-Subscribe: <http://dpdk.org/ml/listinfo/dev>,
<mailto:dev-request@dpdk.org?subject=subscribe>
X-List-Received-Date: Thu, 05 Oct 2017 12:35:04 -0000
This patch modifies assignment of alignment unit from build-time
to run-time based on CPU flags that machine supports.
Signed-off-by: Xiaoyun Li <xiaoyun.li@intel.com>
---
test/test/test_memcpy_perf.c | 40 +++++++++++++++++++++++++++-------------
1 file changed, 27 insertions(+), 13 deletions(-)
diff --git a/test/test/test_memcpy_perf.c b/test/test/test_memcpy_perf.c
index ff3aaaa..33def3b 100644
--- a/test/test/test_memcpy_perf.c
+++ b/test/test/test_memcpy_perf.c
@@ -79,13 +79,7 @@ static size_t buf_sizes[TEST_VALUE_RANGE];
#define TEST_BATCH_SIZE 100
/* Data is aligned on this many bytes (power of 2) */
-#ifdef RTE_MACHINE_CPUFLAG_AVX512F
-#define ALIGNMENT_UNIT 64
-#elif defined RTE_MACHINE_CPUFLAG_AVX2
-#define ALIGNMENT_UNIT 32
-#else /* RTE_MACHINE_CPUFLAG */
-#define ALIGNMENT_UNIT 16
-#endif /* RTE_MACHINE_CPUFLAG */
+static uint8_t alignment_unit = 16;
/*
* Pointers used in performance tests. The two large buffers are for uncached
@@ -100,20 +94,39 @@ static int
init_buffers(void)
{
unsigned i;
+#ifdef CC_SUPPORT_AVX512
+ if (rte_cpu_get_flag_enabled(RTE_CPUFLAG_AVX512F))
+ alignment_unit = 64;
+ else
+#endif
+#ifdef CC_SUPPORT_AVX2
+ if (rte_cpu_get_flag_enabled(RTE_CPUFLAG_AVX2))
+ alignment_unit = 32;
+ else
+#endif
+ alignment_unit = 16;
- large_buf_read = rte_malloc("memcpy", LARGE_BUFFER_SIZE + ALIGNMENT_UNIT, ALIGNMENT_UNIT);
+ large_buf_read = rte_malloc("memcpy",
+ LARGE_BUFFER_SIZE + alignment_unit,
+ alignment_unit);
if (large_buf_read == NULL)
goto error_large_buf_read;
- large_buf_write = rte_malloc("memcpy", LARGE_BUFFER_SIZE + ALIGNMENT_UNIT, ALIGNMENT_UNIT);
+ large_buf_write = rte_malloc("memcpy",
+ LARGE_BUFFER_SIZE + alignment_unit,
+ alignment_unit);
if (large_buf_write == NULL)
goto error_large_buf_write;
- small_buf_read = rte_malloc("memcpy", SMALL_BUFFER_SIZE + ALIGNMENT_UNIT, ALIGNMENT_UNIT);
+ small_buf_read = rte_malloc("memcpy",
+ SMALL_BUFFER_SIZE + alignment_unit,
+ alignment_unit);
if (small_buf_read == NULL)
goto error_small_buf_read;
- small_buf_write = rte_malloc("memcpy", SMALL_BUFFER_SIZE + ALIGNMENT_UNIT, ALIGNMENT_UNIT);
+ small_buf_write = rte_malloc("memcpy",
+ SMALL_BUFFER_SIZE + alignment_unit,
+ alignment_unit);
if (small_buf_write == NULL)
goto error_small_buf_write;
@@ -153,7 +166,7 @@ static inline size_t
get_rand_offset(size_t uoffset)
{
return ((rte_rand() % (LARGE_BUFFER_SIZE - SMALL_BUFFER_SIZE)) &
- ~(ALIGNMENT_UNIT - 1)) + uoffset;
+ ~(alignment_unit - 1)) + uoffset;
}
/* Fill in source and destination addresses. */
@@ -321,7 +334,8 @@ perf_test(void)
"(bytes) (ticks) (ticks) (ticks) (ticks)\n"
"------- -------------- -------------- -------------- --------------");
- printf("\n========================== %2dB aligned ============================", ALIGNMENT_UNIT);
+ printf("\n========================= %2dB aligned ============================",
+ alignment_unit);
/* Do aligned tests where size is a variable */
perf_test_variable_aligned();
printf("\n------- -------------- -------------- -------------- --------------");
--
2.7.4