SRCS-$(CONFIG_RTE_APP_TEST) += test_memory.c
SRCS-$(CONFIG_RTE_APP_TEST) += test_memzone.c
SRCS-$(CONFIG_RTE_APP_TEST) += test_ring.c
+SRCS-$(CONFIG_RTE_APP_TEST) += test_ring_perf.c
SRCS-$(CONFIG_RTE_APP_TEST) += test_rwlock.c
SRCS-$(CONFIG_RTE_APP_TEST) += test_timer.c
SRCS-$(CONFIG_RTE_APP_TEST) += test_mempool.c
"Func" : default_autotest,
"Report" : None,
},
+ {
+ "Name" : "Ring autotest",
+ "Command" : "ring_autotest",
+ "Func" : default_autotest,
+ "Report" : None,
+ },
]
},
{
"Tests" :
[
{
- "Name" : "Ring autotest",
- "Command" : "ring_autotest",
- "Func" : ring_autotest,
+ "Name" : "Ring performance autotest",
+ "Command" : "ring_perf_autotest",
+ "Func" : default_autotest,
"Report" : None,
},
]
def ring_autotest(child, test_name):
child.sendline(test_name)
index = child.expect(["Test OK", "Test Failed",
- pexpect.TIMEOUT], timeout = 1500)
+ pexpect.TIMEOUT], timeout = 15)
if index == 1:
return -1, "Fail"
elif index == 2:
ret |= test_cycles();
if (all || !strcmp(res->autotest, "ring_autotest"))
ret |= test_ring();
+ if (all || !strcmp(res->autotest, "ring_perf_autotest"))
+ ret |= test_ring_perf();
if (all || !strcmp(res->autotest, "timer_autotest"))
ret |= test_timer();
if (all || !strcmp(res->autotest, "mempool_autotest"))
"version_autotest#eal_fs_autotest#"
"cmdline_autotest#func_reentrancy_autotest#"
"mempool_perf_autotest#hash_perf_autotest#"
+ "memcpy_perf_autotest#ring_perf_autotest#"
"red_autotest#meter_autotest#sched_autotest#"
"memcpy_perf_autotest#kni_autotest#"
"pm_autotest#acl_autotest#power_autotest#"
int test_logs(void);
int test_memzone(void);
int test_ring(void);
+int test_ring_perf(void);
int test_mempool(void);
int test_mempool_perf(void);
int test_mbuf(void);
*
* #. Performance tests.
*
- * This test is done on the following configurations:
- *
- * - One core enqueuing, one core dequeuing
- * - One core enqueuing, other cores dequeuing
- * - One core dequeuing, other cores enqueuing
- * - Half of the cores enqueuing, the other half dequeuing
- *
- * When only one core enqueues/dequeues, the test is done with the
- * SP/SC functions in addition to the MP/MC functions.
- *
- * The test is done with different bulk size.
- *
- * On each core, the test enqueues or dequeues objects during
- * TIME_S seconds. The number of successes and failures are stored on
- * each core, then summed and displayed.
- *
- * The test checks that the number of enqueues is equal to the
- * number of dequeues.
+ * Tests done in test_ring_perf.c
*/
#define RING_SIZE 4096
static struct rte_ring *r;
-struct test_stats {
- unsigned enq_success ;
- unsigned enq_quota;
- unsigned enq_fail;
-
- unsigned deq_success;
- unsigned deq_fail;
-} __rte_cache_aligned;
-
-static struct test_stats test_stats[RTE_MAX_LCORE];
-
-static int
-ring_enqueue_test(int (que_func)(struct rte_ring*, void * const *, unsigned),
- void* arg, unsigned bulk_or_burst)
-{
- unsigned success = 0;
- unsigned quota = 0;
- unsigned fail = 0;
- unsigned i;
- unsigned long dummy_obj;
- void *obj_table[MAX_BULK];
- int ret;
- unsigned lcore_id = rte_lcore_id();
- unsigned count = *((unsigned*)arg);
- uint64_t start_cycles, end_cycles;
- uint64_t time_diff = 0, hz = rte_get_hpet_hz();
-
- /* init dummy object table */
- for (i = 0; i< MAX_BULK; i++) {
- dummy_obj = lcore_id + 0x1000 + i;
- obj_table[i] = (void *)dummy_obj;
- }
-
- /* wait synchro for slaves */
- if (lcore_id != rte_get_master_lcore())
- while (rte_atomic32_read(&synchro) == 0);
-
- start_cycles = rte_get_hpet_cycles();
-
- /* enqueue as many object as possible */
- while (time_diff/hz < TIME_S) {
- for (i = 0; likely(i < N); i++) {
- ret = que_func(r, obj_table, count);
- /*
- * bulk_or_burst
- * 1: for bulk operation
- * 0: for burst operation
- */
- if (bulk_or_burst) {
- /* The *count* objects enqueued, unless fail */
- if (ret == 0)
- success += count;
- else if (ret == -EDQUOT)
- quota += count;
- else
- fail++;
- } else {
- /* The actual objects enqueued */
- if (ret != 0)
- success += (ret & RTE_RING_SZ_MASK);
- else
- fail++;
- }
- }
- end_cycles = rte_get_hpet_cycles();
- time_diff = end_cycles - start_cycles;
- }
-
- /* write statistics in a shared structure */
- test_stats[lcore_id].enq_success = success;
- test_stats[lcore_id].enq_quota = quota;
- test_stats[lcore_id].enq_fail = fail;
-
- return 0;
-}
-
-static int
-ring_dequeue_test(int (que_func)(struct rte_ring*, void **, unsigned),
- void* arg, unsigned bulk_or_burst)
-{
- unsigned success = 0;
- unsigned fail = 0;
- unsigned i;
- void *obj_table[MAX_BULK];
- int ret;
- unsigned lcore_id = rte_lcore_id();
- unsigned count = *((unsigned*)arg);
- uint64_t start_cycles, end_cycles;
- uint64_t time_diff = 0, hz = rte_get_hpet_hz();
-
- /* wait synchro for slaves */
- if (lcore_id != rte_get_master_lcore())
- while (rte_atomic32_read(&synchro) == 0);
-
- start_cycles = rte_get_hpet_cycles();
-
- /* dequeue as many object as possible */
- while (time_diff/hz < TIME_S) {
- for (i = 0; likely(i < N); i++) {
- ret = que_func(r, obj_table, count);
- /*
- * bulk_or_burst
- * 1: for bulk operation
- * 0: for burst operation
- */
- if (bulk_or_burst) {
- if (ret == 0)
- success += count;
- else
- fail++;
- } else {
- if (ret != 0)
- success += ret;
- else
- fail++;
- }
- }
- end_cycles = rte_get_hpet_cycles();
- time_diff = end_cycles - start_cycles;
- }
-
- /* write statistics in a shared structure */
- test_stats[lcore_id].deq_success = success;
- test_stats[lcore_id].deq_fail = fail;
-
- return 0;
-}
-
-static int
-test_ring_per_core_sp_enqueue(void *arg)
-{
- return ring_enqueue_test(&rte_ring_sp_enqueue_bulk, arg, 1);
-}
-
-static int
-test_ring_per_core_mp_enqueue(void *arg)
-{
- return ring_enqueue_test(&rte_ring_mp_enqueue_bulk, arg, 1);
-}
-
-static int
-test_ring_per_core_mc_dequeue(void *arg)
-{
- return ring_dequeue_test(&rte_ring_mc_dequeue_bulk, arg, 1);
-}
-
-static int
-test_ring_per_core_sc_dequeue(void *arg)
-{
- return ring_dequeue_test(&rte_ring_sc_dequeue_bulk, arg, 1);
-}
-
-static int
-test_ring_per_core_sp_enqueue_burst(void *arg)
-{
- return ring_enqueue_test(&rte_ring_sp_enqueue_burst, arg, 0);
-}
-
-static int
-test_ring_per_core_mp_enqueue_burst(void *arg)
-{
- return ring_enqueue_test(&rte_ring_mp_enqueue_burst, arg, 0);
-}
-
-static int
-test_ring_per_core_mc_dequeue_burst(void *arg)
-{
- return ring_dequeue_test(&rte_ring_mc_dequeue_burst, arg, 0);
-}
-
-static int
-test_ring_per_core_sc_dequeue_burst(void *arg)
-{
- return ring_dequeue_test(&rte_ring_sc_dequeue_burst, arg, 0);
-}
-
#define TEST_RING_VERIFY(exp) \
if (!(exp)) { \
printf("error at %s:%d\tcondition " #exp " failed\n", \
#define TEST_RING_FULL_EMTPY_ITER 8
-
-static int
-launch_cores(unsigned enq_core_count, unsigned deq_core_count,
- unsigned n_enq_bulk, unsigned n_deq_bulk,
- int sp, int sc, int bulk_not_burst)
-{
- void *obj;
- unsigned lcore_id;
- unsigned rate, deq_remain = 0;
- unsigned enq_total, deq_total;
- struct test_stats sum;
- int (*enq_f)(void *);
- int (*deq_f)(void *);
- unsigned cores = enq_core_count + deq_core_count;
- int ret;
-
- rte_atomic32_set(&synchro, 0);
-
- printf("ring_autotest e/d_core=%u,%u e/d_bulk=%u,%u ",
- enq_core_count, deq_core_count, n_enq_bulk, n_deq_bulk);
- printf("sp=%d sc=%d ", sp, sc);
-
- if (bulk_not_burst) {
- /* set enqueue function to be used */
- if (sp)
- enq_f = test_ring_per_core_sp_enqueue;
- else
- enq_f = test_ring_per_core_mp_enqueue;
-
- /* set dequeue function to be used */
- if (sc)
- deq_f = test_ring_per_core_sc_dequeue;
- else
- deq_f = test_ring_per_core_mc_dequeue;
-
- } else {
- /* set enqueue function to be used */
- if (sp)
- enq_f = test_ring_per_core_sp_enqueue_burst;
- else
- enq_f = test_ring_per_core_mp_enqueue_burst;
-
- /* set dequeue function to be used */
- if (sc)
- deq_f = test_ring_per_core_sc_dequeue_burst;
- else
- deq_f = test_ring_per_core_mc_dequeue_burst;
- }
-
- RTE_LCORE_FOREACH_SLAVE(lcore_id) {
- if (enq_core_count != 0) {
- enq_core_count--;
- rte_eal_remote_launch(enq_f, &n_enq_bulk, lcore_id);
- }
- if (deq_core_count != 1) {
- deq_core_count--;
- rte_eal_remote_launch(deq_f, &n_deq_bulk, lcore_id);
- }
- }
-
- memset(test_stats, 0, sizeof(test_stats));
-
- /* start synchro and launch test on master */
- rte_atomic32_set(&synchro, 1);
- ret = deq_f(&n_deq_bulk);
-
- /* wait all cores */
- RTE_LCORE_FOREACH_SLAVE(lcore_id) {
- if (cores == 1)
- break;
- cores--;
- if (rte_eal_wait_lcore(lcore_id) < 0)
- ret = -1;
- }
-
- memset(&sum, 0, sizeof(sum));
- for (lcore_id = 0; lcore_id < RTE_MAX_LCORE; lcore_id++) {
- sum.enq_success += test_stats[lcore_id].enq_success;
- sum.enq_quota += test_stats[lcore_id].enq_quota;
- sum.enq_fail += test_stats[lcore_id].enq_fail;
- sum.deq_success += test_stats[lcore_id].deq_success;
- sum.deq_fail += test_stats[lcore_id].deq_fail;
- }
-
- /* empty the ring */
- while (rte_ring_sc_dequeue(r, &obj) == 0)
- deq_remain += 1;
-
- if (ret < 0) {
- printf("per-lcore test returned -1\n");
- return -1;
- }
-
- enq_total = sum.enq_success + sum.enq_quota;
- deq_total = sum.deq_success + deq_remain;
-
- rate = deq_total/TIME_S;
-
- printf("rate_persec=%u\n", rate);
-
- if (enq_total != deq_total) {
- printf("invalid enq/deq_success counter: %u %u\n",
- enq_total, deq_total);
- return -1;
- }
-
- return 0;
-}
-
-static int
-do_one_ring_test2(unsigned enq_core_count, unsigned deq_core_count,
- unsigned n_enq_bulk, unsigned n_deq_bulk, unsigned bulk_or_burst)
-{
- int sp, sc;
- int do_sp, do_sc;
- int ret;
-
- do_sp = (enq_core_count == 1) ? 1 : 0;
- do_sc = (deq_core_count == 1) ? 1 : 0;
-
- for (sp = 0; sp <= do_sp; sp ++) {
- for (sc = 0; sc <= do_sc; sc ++) {
- ret = launch_cores(enq_core_count, deq_core_count,
- n_enq_bulk, n_deq_bulk, sp, sc, bulk_or_burst);
- if (ret < 0)
- return -1;
- }
- }
- return 0;
-}
-
-static int
-do_one_ring_test(unsigned enq_core_count, unsigned deq_core_count,
- unsigned bulk_or_burst)
-{
- unsigned bulk_enqueue_tab[] = { 1, 2, 4, 32, 0 };
- unsigned bulk_dequeue_tab[] = { 1, 2, 4, 32, 0 };
- unsigned *bulk_enqueue_ptr;
- unsigned *bulk_dequeue_ptr;
- int ret;
-
- for (bulk_enqueue_ptr = bulk_enqueue_tab;
- *bulk_enqueue_ptr;
- bulk_enqueue_ptr++) {
-
- for (bulk_dequeue_ptr = bulk_dequeue_tab;
- *bulk_dequeue_ptr;
- bulk_dequeue_ptr++) {
-
- ret = do_one_ring_test2(enq_core_count, deq_core_count,
- *bulk_enqueue_ptr,
- *bulk_dequeue_ptr,
- bulk_or_burst);
- if (ret < 0)
- return -1;
- }
- }
- return 0;
-}
-
static int
check_live_watermark_change(__attribute__((unused)) void *dummy)
{
int
test_ring(void)
{
- unsigned enq_core_count, deq_core_count;
-
/* some more basic operations */
if (test_ring_basic_ex() < 0)
return -1;
else
printf ( "Test detected NULL ring lookup \n");
- printf("start performance tests \n");
-
- /* one lcore for enqueue, one for dequeue */
- enq_core_count = 1;
- deq_core_count = 1;
- if (do_one_ring_test(enq_core_count, deq_core_count, 1) < 0)
- return -1;
-
- /* max cores for enqueue, one for dequeue */
- enq_core_count = rte_lcore_count() - 1;
- deq_core_count = 1;
- if (do_one_ring_test(enq_core_count, deq_core_count, 1) < 0)
- return -1;
-
- /* max cores for dequeue, one for enqueue */
- enq_core_count = 1;
- deq_core_count = rte_lcore_count() - 1;
- if (do_one_ring_test(enq_core_count, deq_core_count, 1) < 0)
- return -1;
-
- /* half for enqueue and half for dequeue */
- enq_core_count = rte_lcore_count() / 2;
- deq_core_count = rte_lcore_count() / 2;
- if (do_one_ring_test(enq_core_count, deq_core_count, 1) < 0)
- return -1;
-
- printf("start performance tests - burst operations \n");
-
- /* one lcore for enqueue, one for dequeue */
- enq_core_count = 1;
- deq_core_count = 1;
- if (do_one_ring_test(enq_core_count, deq_core_count, 0) < 0)
- return -1;
-
- /* max cores for enqueue, one for dequeue */
- enq_core_count = rte_lcore_count() - 1;
- deq_core_count = 1;
- if (do_one_ring_test(enq_core_count, deq_core_count, 0) < 0)
- return -1;
-
- /* max cores for dequeue, one for enqueue */
- enq_core_count = 1;
- deq_core_count = rte_lcore_count() - 1;
- if (do_one_ring_test(enq_core_count, deq_core_count, 0) < 0)
- return -1;
-
- /* half for enqueue and half for dequeue */
- enq_core_count = rte_lcore_count() / 2;
- deq_core_count = rte_lcore_count() / 2;
- if (do_one_ring_test(enq_core_count, deq_core_count, 0) < 0)
- return -1;
-
/* test of creating ring with wrong size */
if (test_ring_creation_with_wrong_size() < 0)
return -1;
--- /dev/null
+/*-
+ * BSD LICENSE
+ *
+ * Copyright(c) 2010-2013 Intel Corporation. All rights reserved.
+ * All rights reserved.
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions
+ * are met:
+ *
+ * * Redistributions of source code must retain the above copyright
+ * notice, this list of conditions and the following disclaimer.
+ * * Redistributions in binary form must reproduce the above copyright
+ * notice, this list of conditions and the following disclaimer in
+ * the documentation and/or other materials provided with the
+ * distribution.
+ * * Neither the name of Intel Corporation nor the names of its
+ * contributors may be used to endorse or promote products derived
+ * from this software without specific prior written permission.
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
+ * "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
+ * LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
+ * A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
+ * OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
+ * SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
+ * LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
+ * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
+ * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
+ * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
+ * OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
+ *
+ */
+
+
+#include <stdio.h>
+#include <inttypes.h>
+#include <rte_ring.h>
+#include <rte_cycles.h>
+#include <rte_launch.h>
+
+#include <cmdline_parse.h>
+
+#include "test.h"
+
+/*
+ * Ring
+ * ====
+ *
+ * Measures performance of various operations using rdtsc
+ * * Empty ring dequeue
+ * * Enqueue/dequeue of bursts in 1 threads
+ * * Enqueue/dequeue of bursts in 2 threads
+ */
+
+#define RING_NAME "RING_PERF"
+#define RING_SIZE 4096
+#define MAX_BURST 32
+
+/*
+ * the sizes to enqueue and dequeue in testing
+ * (marked volatile so they won't be seen as compile-time constants)
+ */
+static const volatile unsigned bulk_sizes[] = { 8, 32 };
+
+/* The ring structure used for tests */
+static struct rte_ring *r;
+
+struct lcore_pair {
+ unsigned c1, c2;
+};
+
+static volatile unsigned lcore_count = 0;
+
+/**** Functions to analyse our core mask to get cores for different tests ***/
+
+static int
+get_two_hyperthreads(struct lcore_pair *lcp)
+{
+ unsigned id1, id2;
+ unsigned c1, c2, s1, s2;
+ RTE_LCORE_FOREACH(id1) {
+ /* inner loop just re-reads all id's. We could skip the first few
+ * elements, but since number of cores is small there is little point
+ */
+ RTE_LCORE_FOREACH(id2) {
+ if (id1 == id2)
+ continue;
+ c1 = lcore_config[id1].core_id;
+ c2 = lcore_config[id2].core_id;
+ s1 = lcore_config[id1].socket_id;
+ s2 = lcore_config[id2].socket_id;
+ if ((c1 == c2) && (s1 == s2)){
+ lcp->c1 = id1;
+ lcp->c2 = id2;
+ return 0;
+ }
+ }
+ }
+ return 1;
+}
+
+static int
+get_two_cores(struct lcore_pair *lcp)
+{
+ unsigned id1, id2;
+ unsigned c1, c2, s1, s2;
+ RTE_LCORE_FOREACH(id1) {
+ RTE_LCORE_FOREACH(id2) {
+ if (id1 == id2)
+ continue;
+ c1 = lcore_config[id1].core_id;
+ c2 = lcore_config[id2].core_id;
+ s1 = lcore_config[id1].socket_id;
+ s2 = lcore_config[id2].socket_id;
+ if ((c1 != c2) && (s1 == s2)){
+ lcp->c1 = id1;
+ lcp->c2 = id2;
+ return 0;
+ }
+ }
+ }
+ return 1;
+}
+
+static int
+get_two_sockets(struct lcore_pair *lcp)
+{
+ unsigned id1, id2;
+ unsigned s1, s2;
+ RTE_LCORE_FOREACH(id1) {
+ RTE_LCORE_FOREACH(id2) {
+ if (id1 == id2)
+ continue;
+ s1 = lcore_config[id1].socket_id;
+ s2 = lcore_config[id2].socket_id;
+ if (s1 != s2){
+ lcp->c1 = id1;
+ lcp->c2 = id2;
+ return 0;
+ }
+ }
+ }
+ return 1;
+}
+
+/* Get cycle counts for dequeuing from an empty ring. Should be 2 or 3 cycles */
+static void
+test_empty_dequeue(void)
+{
+ const unsigned iter_shift = 26;
+ const unsigned iterations = 1<<iter_shift;
+ unsigned i = 0;
+ void *burst[MAX_BURST];
+
+ const uint64_t sc_start = rte_rdtsc();
+ for (i = 0; i < iterations; i++)
+ rte_ring_sc_dequeue_bulk(r, burst, bulk_sizes[0]);
+ const uint64_t sc_end = rte_rdtsc();
+
+ const uint64_t mc_start = rte_rdtsc();
+ for (i = 0; i < iterations; i++)
+ rte_ring_mc_dequeue_bulk(r, burst, bulk_sizes[0]);
+ const uint64_t mc_end = rte_rdtsc();
+
+ printf("SC empty dequeue: %.2F\n",
+ (double)(sc_end-sc_start) / iterations);
+ printf("MC empty dequeue: %.2F\n",
+ (double)(mc_end-mc_start) / iterations);
+}
+
+/*
+ * for the separate enqueue and dequeue threads they take in one param
+ * and return two. Input = burst size, output = cycle average for sp/sc & mp/mc
+ */
+struct thread_params {
+ unsigned size; /* input value, the burst size */
+ double spsc, mpmc; /* output value, the single or multi timings */
+};
+
+/*
+ * Function that uses rdtsc to measure timing for ring enqueue. Needs pair
+ * thread running dequeue_bulk function
+ */
+static int
+enqueue_bulk(void *p)
+{
+ const unsigned iter_shift = 23;
+ const unsigned iterations = 1<<iter_shift;
+ struct thread_params *params = p;
+ const unsigned size = params->size;
+ unsigned i;
+ void *burst[MAX_BURST] = {0};
+
+ if ( __sync_add_and_fetch(&lcore_count, 1) != 2 )
+ while(lcore_count != 2)
+ rte_pause();
+
+ const uint64_t sp_start = rte_rdtsc();
+ for (i = 0; i < iterations; i++)
+ while (rte_ring_sp_enqueue_bulk(r, burst, size) != 0)
+ rte_pause();
+ const uint64_t sp_end = rte_rdtsc();
+
+ const uint64_t mp_start = rte_rdtsc();
+ for (i = 0; i < iterations; i++)
+ while (rte_ring_mp_enqueue_bulk(r, burst, size) != 0)
+ rte_pause();
+ const uint64_t mp_end = rte_rdtsc();
+
+ params->spsc = ((double)(sp_end - sp_start))/(iterations*size);
+ params->mpmc = ((double)(mp_end - mp_start))/(iterations*size);
+ return 0;
+}
+
+/*
+ * Function that uses rdtsc to measure timing for ring dequeue. Needs pair
+ * thread running enqueue_bulk function
+ */
+static int
+dequeue_bulk(void *p)
+{
+ const unsigned iter_shift = 23;
+ const unsigned iterations = 1<<iter_shift;
+ struct thread_params *params = p;
+ const unsigned size = params->size;
+ unsigned i;
+ void *burst[MAX_BURST] = {0};
+
+ if ( __sync_add_and_fetch(&lcore_count, 1) != 2 )
+ while(lcore_count != 2)
+ rte_pause();
+
+ const uint64_t sc_start = rte_rdtsc();
+ for (i = 0; i < iterations; i++)
+ while (rte_ring_sc_dequeue_bulk(r, burst, size) != 0)
+ rte_pause();
+ const uint64_t sc_end = rte_rdtsc();
+
+ const uint64_t mc_start = rte_rdtsc();
+ for (i = 0; i < iterations; i++)
+ while (rte_ring_mc_dequeue_bulk(r, burst, size) != 0)
+ rte_pause();
+ const uint64_t mc_end = rte_rdtsc();
+
+ params->spsc = ((double)(sc_end - sc_start))/(iterations*size);
+ params->mpmc = ((double)(mc_end - mc_start))/(iterations*size);
+ return 0;
+}
+
+/*
+ * Function that calls the enqueue and dequeue bulk functions on pairs of cores.
+ * used to measure ring perf between hyperthreads, cores and sockets.
+ */
+static void
+run_on_core_pair(struct lcore_pair *cores,
+ lcore_function_t f1, lcore_function_t f2)
+{
+ struct thread_params param1 = {.size = 0}, param2 = {.size = 0};
+ unsigned i;
+ for (i = 0; i < sizeof(bulk_sizes)/sizeof(bulk_sizes[0]); i++) {
+ lcore_count = 0;
+ param1.size = param2.size = bulk_sizes[i];
+ if (cores->c1 == rte_get_master_lcore()) {
+ rte_eal_remote_launch(f2, ¶m2, cores->c2);
+ f1(¶m1);
+ rte_eal_wait_lcore(cores->c2);
+ } else {
+ rte_eal_remote_launch(f1, ¶m1, cores->c1);
+ rte_eal_remote_launch(f2, ¶m2, cores->c2);
+ rte_eal_wait_lcore(cores->c1);
+ rte_eal_wait_lcore(cores->c2);
+ }
+ printf("SP/SC bulk enq/dequeue (size: %u): %.2F\n", bulk_sizes[i],
+ param1.spsc + param2.spsc);
+ printf("MP/MC bulk enq/dequeue (size: %u): %.2F\n", bulk_sizes[i],
+ param1.mpmc + param2.mpmc);
+ }
+}
+
+/*
+ * Test function that determines how long an enqueue + dequeue of a single item
+ * takes on a single lcore. Result is for comparison with the bulk enq+deq.
+ */
+static void
+test_single_enqueue_dequeue(void)
+{
+ const unsigned iter_shift = 24;
+ const unsigned iterations = 1<<iter_shift;
+ unsigned i = 0;
+ void *burst = NULL;
+
+ const uint64_t sc_start = rte_rdtsc();
+ for (i = 0; i < iterations; i++) {
+ rte_ring_sp_enqueue(r, burst);
+ rte_ring_sc_dequeue(r, &burst);
+ }
+ const uint64_t sc_end = rte_rdtsc();
+
+ const uint64_t mc_start = rte_rdtsc();
+ for (i = 0; i < iterations; i++) {
+ rte_ring_mp_enqueue(r, burst);
+ rte_ring_mc_dequeue(r, &burst);
+ }
+ const uint64_t mc_end = rte_rdtsc();
+
+ printf("SP/SC single enq/dequeue: %"PRIu64"\n",
+ (sc_end-sc_start) >> iter_shift);
+ printf("MP/MC single enq/dequeue: %"PRIu64"\n",
+ (mc_end-mc_start) >> iter_shift);
+}
+
+/*
+ * Test that does both enqueue and dequeue on a core using the burst() API calls
+ * instead of the bulk() calls used in other tests. Results should be the same
+ * as for the bulk function called on a single lcore.
+ */
+static void
+test_burst_enqueue_dequeue(void)
+{
+ const unsigned iter_shift = 23;
+ const unsigned iterations = 1<<iter_shift;
+ unsigned sz, i = 0;
+ void *burst[MAX_BURST] = {0};
+
+ for (sz = 0; sz < sizeof(bulk_sizes)/sizeof(bulk_sizes[0]); sz++) {
+ const uint64_t sc_start = rte_rdtsc();
+ for (i = 0; i < iterations; i++) {
+ rte_ring_sp_enqueue_burst(r, burst, bulk_sizes[sz]);
+ rte_ring_sc_dequeue_burst(r, burst, bulk_sizes[sz]);
+ }
+ const uint64_t sc_end = rte_rdtsc();
+
+ const uint64_t mc_start = rte_rdtsc();
+ for (i = 0; i < iterations; i++) {
+ rte_ring_mp_enqueue_burst(r, burst, bulk_sizes[sz]);
+ rte_ring_mc_dequeue_burst(r, burst, bulk_sizes[sz]);
+ }
+ const uint64_t mc_end = rte_rdtsc();
+
+ uint64_t mc_avg = ((mc_end-mc_start) >> iter_shift) / bulk_sizes[sz];
+ uint64_t sc_avg = ((sc_end-sc_start) >> iter_shift) / bulk_sizes[sz];
+
+ printf("SP/SC burst enq/dequeue (size: %u): %"PRIu64"\n", bulk_sizes[sz],
+ sc_avg);
+ printf("MP/MC burst enq/dequeue (size: %u): %"PRIu64"\n", bulk_sizes[sz],
+ mc_avg);
+ }
+}
+
+/* Times enqueue and dequeue on a single lcore */
+static void
+test_bulk_enqueue_dequeue(void)
+{
+ const unsigned iter_shift = 23;
+ const unsigned iterations = 1<<iter_shift;
+ unsigned sz, i = 0;
+ void *burst[MAX_BURST] = {0};
+
+ for (sz = 0; sz < sizeof(bulk_sizes)/sizeof(bulk_sizes[0]); sz++) {
+ const uint64_t sc_start = rte_rdtsc();
+ for (i = 0; i < iterations; i++) {
+ rte_ring_sp_enqueue_bulk(r, burst, bulk_sizes[sz]);
+ rte_ring_sc_dequeue_bulk(r, burst, bulk_sizes[sz]);
+ }
+ const uint64_t sc_end = rte_rdtsc();
+
+ const uint64_t mc_start = rte_rdtsc();
+ for (i = 0; i < iterations; i++) {
+ rte_ring_mp_enqueue_bulk(r, burst, bulk_sizes[sz]);
+ rte_ring_mc_dequeue_bulk(r, burst, bulk_sizes[sz]);
+ }
+ const uint64_t mc_end = rte_rdtsc();
+
+ double sc_avg = ((double)(sc_end-sc_start) /
+ (iterations * bulk_sizes[sz]));
+ double mc_avg = ((double)(mc_end-mc_start) /
+ (iterations * bulk_sizes[sz]));
+
+ printf("SP/SC bulk enq/dequeue (size: %u): %.2F\n", bulk_sizes[sz],
+ sc_avg);
+ printf("MP/MC bulk enq/dequeue (size: %u): %.2F\n", bulk_sizes[sz],
+ mc_avg);
+ }
+}
+
+int
+test_ring_perf(void)
+{
+ struct lcore_pair cores;
+ r = rte_ring_create(RING_NAME, RING_SIZE, rte_socket_id(), 0);
+ if (r == NULL && (r = rte_ring_lookup(RING_NAME)) == NULL)
+ return -1;
+
+ printf("### Testing single element and burst enq/deq ###\n");
+ test_single_enqueue_dequeue();
+ test_burst_enqueue_dequeue();
+
+ printf("\n### Testing empty dequeue ###\n");
+ test_empty_dequeue();
+
+ printf("\n### Testing using a single lcore ###\n");
+ test_bulk_enqueue_dequeue();
+
+ if (get_two_hyperthreads(&cores) == 0) {
+ printf("\n### Testing using two hyperthreads ###\n");
+ run_on_core_pair(&cores, enqueue_bulk, dequeue_bulk);
+ }
+ if (get_two_cores(&cores) == 0) {
+ printf("\n### Testing using two physical cores ###\n");
+ run_on_core_pair(&cores, enqueue_bulk, dequeue_bulk);
+ }
+ if (get_two_sockets(&cores) == 0) {
+ printf("\n### Testing using two NUMA nodes ###\n");
+ run_on_core_pair(&cores, enqueue_bulk, dequeue_bulk);
+ }
+ return 0;
+}
PHONY: test fast_test
coverage: BLACKLIST=-Mempool_perf,Memcpy_perf,Hash_perf
-fast_test: BLACKLIST=-Ring,Mempool_perf,Memcpy_perf,Hash_perf,Lpm6
-ring_test: WHITELIST=Ring
+fast_test: BLACKLIST=-Ring_perf,Mempool_perf,Memcpy_perf,Hash_perf,Lpm6
+ring_test: WHITELIST=Ring,Ring_perf
mempool_test: WHITELIST=Mempool,Mempool_perf
-perf_test:WHITELIST=Mempool_perf,Memcpy_perf,Hash_perf,Ring
+perf_test:WHITELIST=Mempool_perf,Memcpy_perf,Hash_perf,Ring_perf
test fast_test ring_test mempool_test perf_test:
@mkdir -p $(AUTOTEST_DIR) ; \
cd $(AUTOTEST_DIR) ; \
git.droids-corp.org / dpdk.git / commitdiff