subject:"\[dpdk\-dev\] \[memnic PATCH v2 6\/7\] pmd\: add branch hint in recv\/xmit"

[dpdk-dev] [memnic PATCH v2 6/7] pmd: add branch hint in recv/xmit

2014-10-02 Thread Hiroshi Shimamoto

> Subject: Re: [dpdk-dev] [memnic PATCH v2 6/7] pmd: add branch hint in 
> recv/xmit
> 
> On Wed, Oct 01, 2014 at 11:33:23PM +, Hiroshi Shimamoto wrote:
> > > Subject: Re: [dpdk-dev] [memnic PATCH v2 6/7] pmd: add branch hint in 
> > > recv/xmit
> > >
> > > On Wed, Oct 01, 2014 at 09:12:44AM +, Hiroshi Shimamoto wrote:
> > > > > Subject: Re: [dpdk-dev] [memnic PATCH v2 6/7] pmd: add branch hint in 
> > > > > recv/xmit
> > > > >
> > > > > On Tue, Sep 30, 2014 at 11:52:00PM +0000, Hiroshi Shimamoto wrote:
> > > > > > Hi,
> > > > > >
> > > > > > > Subject: Re: [dpdk-dev] [memnic PATCH v2 6/7] pmd: add branch 
> > > > > > > hint in recv/xmit
> > > > > > >
> > > > > > > On Tue, Sep 30, 2014 at 11:14:40AM +, Hiroshi Shimamoto wrote:
> > > > > > > > From: Hiroshi Shimamoto 
> > > > > > > >
> > > > > > > > To reduce instruction cache miss, add branch condition hints 
> > > > > > > > into
> > > > > > > > recv/xmit functions. This improves a bit performance.
> > > > > > > >
> > > > > > > > We can see performance improvements with memnic-tester.
> > > > > > > > Using Xeon E5-2697 v2 @ 2.70GHz, 4 vCPU.
> > > > > > > >  size |  before  |  after
> > > > > > > >64 | 5.54Mpps | 5.55Mpps
> > > > > > > >   128 | 5.46Mpps | 5.44Mpps
> > > > > > > >   256 | 5.21Mpps | 5.22Mpps
> > > > > > > >   512 | 4.50Mpps | 4.52Mpps
> > > > > > > >  1024 | 3.71Mpps | 3.73Mpps
> > > > > > > >  1280 | 3.21Mpps | 3.22Mpps
> > > > > > > >  1518 | 2.92Mpps | 2.93Mpps
> > > > > > > >
> > > > > > > > Signed-off-by: Hiroshi Shimamoto 
> > > > > > > > Reviewed-by: Hayato Momma 
> > > > > > > > ---
> > > > > > > >  pmd/pmd_memnic.c | 18 +-
> > > > > > > >  1 file changed, 9 insertions(+), 9 deletions(-)
> > > > > > > >
> > > > > > > > diff --git a/pmd/pmd_memnic.c b/pmd/pmd_memnic.c
> > > > > > > > index 7fc3093..875d3ea 100644
> > > > > > > > --- a/pmd/pmd_memnic.c
> > > > > > > > +++ b/pmd/pmd_memnic.c
> > > > > > > > @@ -289,26 +289,26 @@ static uint16_t memnic_recv_pkts(void 
> > > > > > > > *rx_queue,
> > > > > > > > int idx, next;
> > > > > > > > struct rte_eth_stats *st = 
> > > > > > > > &adapter->stats[rte_lcore_id()];
> > > > > > > >
> > > > > > > > -   if (!adapter->nic->hdr.valid)
> > > > > > > > +   if (unlikely(!adapter->nic->hdr.valid))
> > > > > > > > return 0;
> > > > > > > >
> > > > > > > > pkts = bytes = errs = 0;
> > > > > > > > idx = adapter->up_idx;
> > > > > > > > for (nr = 0; nr < nb_pkts; nr++) {
> > > > > > > > p = &data->packets[idx];
> > > > > > > > -   if (p->status != MEMNIC_PKT_ST_FILLED)
> > > > > > > > +   if (unlikely(p->status != MEMNIC_PKT_ST_FILLED))
> > > > > > > > break;
> > > > > > > > /* prefetch the next area */
> > > > > > > > next = idx;
> > > > > > > > -   if (++next >= MEMNIC_NR_PACKET)
> > > > > > > > +   if (unlikely(++next >= MEMNIC_NR_PACKET))
> > > > > > > > next = 0;
> > > > > > > > rte_prefetch0(&data->packets[next]);
> > > > > > > > -   if (p->len > framesz) {
> > > > > > > > +   if (unlikely(p->len > framesz)) {
> > > > > > > > errs++;
> > > > > > > > goto drop;
> > > > > > > >

[dpdk-dev] [memnic PATCH v2 6/7] pmd: add branch hint in recv/xmit

2014-10-01 Thread Hiroshi Shimamoto

> Subject: Re: [dpdk-dev] [memnic PATCH v2 6/7] pmd: add branch hint in 
> recv/xmit
> 
> On Wed, Oct 01, 2014 at 09:12:44AM +, Hiroshi Shimamoto wrote:
> > > Subject: Re: [dpdk-dev] [memnic PATCH v2 6/7] pmd: add branch hint in 
> > > recv/xmit
> > >
> > > On Tue, Sep 30, 2014 at 11:52:00PM +, Hiroshi Shimamoto wrote:
> > > > Hi,
> > > >
> > > > > Subject: Re: [dpdk-dev] [memnic PATCH v2 6/7] pmd: add branch hint in 
> > > > > recv/xmit
> > > > >
> > > > > On Tue, Sep 30, 2014 at 11:14:40AM +, Hiroshi Shimamoto wrote:
> > > > > > From: Hiroshi Shimamoto 
> > > > > >
> > > > > > To reduce instruction cache miss, add branch condition hints into
> > > > > > recv/xmit functions. This improves a bit performance.
> > > > > >
> > > > > > We can see performance improvements with memnic-tester.
> > > > > > Using Xeon E5-2697 v2 @ 2.70GHz, 4 vCPU.
> > > > > >  size |  before  |  after
> > > > > >64 | 5.54Mpps | 5.55Mpps
> > > > > >   128 | 5.46Mpps | 5.44Mpps
> > > > > >   256 | 5.21Mpps | 5.22Mpps
> > > > > >   512 | 4.50Mpps | 4.52Mpps
> > > > > >  1024 | 3.71Mpps | 3.73Mpps
> > > > > >  1280 | 3.21Mpps | 3.22Mpps
> > > > > >  1518 | 2.92Mpps | 2.93Mpps
> > > > > >
> > > > > > Signed-off-by: Hiroshi Shimamoto 
> > > > > > Reviewed-by: Hayato Momma 
> > > > > > ---
> > > > > >  pmd/pmd_memnic.c | 18 +-
> > > > > >  1 file changed, 9 insertions(+), 9 deletions(-)
> > > > > >
> > > > > > diff --git a/pmd/pmd_memnic.c b/pmd/pmd_memnic.c
> > > > > > index 7fc3093..875d3ea 100644
> > > > > > --- a/pmd/pmd_memnic.c
> > > > > > +++ b/pmd/pmd_memnic.c
> > > > > > @@ -289,26 +289,26 @@ static uint16_t memnic_recv_pkts(void 
> > > > > > *rx_queue,
> > > > > > int idx, next;
> > > > > > struct rte_eth_stats *st = &adapter->stats[rte_lcore_id()];
> > > > > >
> > > > > > -   if (!adapter->nic->hdr.valid)
> > > > > > +   if (unlikely(!adapter->nic->hdr.valid))
> > > > > > return 0;
> > > > > >
> > > > > > pkts = bytes = errs = 0;
> > > > > > idx = adapter->up_idx;
> > > > > > for (nr = 0; nr < nb_pkts; nr++) {
> > > > > > p = &data->packets[idx];
> > > > > > -   if (p->status != MEMNIC_PKT_ST_FILLED)
> > > > > > +   if (unlikely(p->status != MEMNIC_PKT_ST_FILLED))
> > > > > > break;
> > > > > > /* prefetch the next area */
> > > > > > next = idx;
> > > > > > -   if (++next >= MEMNIC_NR_PACKET)
> > > > > > +   if (unlikely(++next >= MEMNIC_NR_PACKET))
> > > > > > next = 0;
> > > > > > rte_prefetch0(&data->packets[next]);
> > > > > > -   if (p->len > framesz) {
> > > > > > +   if (unlikely(p->len > framesz)) {
> > > > > > errs++;
> > > > > > goto drop;
> > > > > > }
> > > > > > mb = rte_pktmbuf_alloc(adapter->mp);
> > > > > > -   if (!mb)
> > > > > > +   if (unlikely(!mb))
> > > > > > break;
> > > > > >
> > > > > > rte_memcpy(rte_pktmbuf_mtod(mb, void *), p->data, 
> > > > > > p->len);
> > > > > > @@ -350,7 +350,7 @@ static uint16_t memnic_xmit_pkts(void *tx_queue,
> > > > > > uint64_t pkts, bytes, errs;
> > > > > > uint32_t framesz = adapter->framesz;
> > > > > >
> > > > > > -   if (!adapter->nic->hdr.valid)
> > > > > > +   if (unlikely(!adapter->nic->hdr.valid))
> > > > > > return 0;
> > > > > >
> > > > > > pkts = bytes = errs = 0;
> > > > > > @

[dpdk-dev] [memnic PATCH v2 6/7] pmd: add branch hint in recv/xmit

2014-10-01 Thread Neil Horman

On Thu, Oct 02, 2014 at 02:07:09AM +, Hiroshi Shimamoto wrote:
> > Subject: Re: [dpdk-dev] [memnic PATCH v2 6/7] pmd: add branch hint in 
> > recv/xmit
> > 
> > On Wed, Oct 01, 2014 at 11:33:23PM +, Hiroshi Shimamoto wrote:
> > > > Subject: Re: [dpdk-dev] [memnic PATCH v2 6/7] pmd: add branch hint in 
> > > > recv/xmit
> > > >
> > > > On Wed, Oct 01, 2014 at 09:12:44AM +0000, Hiroshi Shimamoto wrote:
> > > > > > Subject: Re: [dpdk-dev] [memnic PATCH v2 6/7] pmd: add branch hint 
> > > > > > in recv/xmit
> > > > > >
> > > > > > On Tue, Sep 30, 2014 at 11:52:00PM +, Hiroshi Shimamoto wrote:
> > > > > > > Hi,
> > > > > > >
> > > > > > > > Subject: Re: [dpdk-dev] [memnic PATCH v2 6/7] pmd: add branch 
> > > > > > > > hint in recv/xmit
> > > > > > > >
> > > > > > > > On Tue, Sep 30, 2014 at 11:14:40AM +, Hiroshi Shimamoto 
> > > > > > > > wrote:
> > > > > > > > > From: Hiroshi Shimamoto 
> > > > > > > > >
> > > > > > > > > To reduce instruction cache miss, add branch condition hints 
> > > > > > > > > into
> > > > > > > > > recv/xmit functions. This improves a bit performance.
> > > > > > > > >
> > > > > > > > > We can see performance improvements with memnic-tester.
> > > > > > > > > Using Xeon E5-2697 v2 @ 2.70GHz, 4 vCPU.
> > > > > > > > >  size |  before  |  after
> > > > > > > > >64 | 5.54Mpps | 5.55Mpps
> > > > > > > > >   128 | 5.46Mpps | 5.44Mpps
> > > > > > > > >   256 | 5.21Mpps | 5.22Mpps
> > > > > > > > >   512 | 4.50Mpps | 4.52Mpps
> > > > > > > > >  1024 | 3.71Mpps | 3.73Mpps
> > > > > > > > >  1280 | 3.21Mpps | 3.22Mpps
> > > > > > > > >  1518 | 2.92Mpps | 2.93Mpps
> > > > > > > > >
> > > > > > > > > Signed-off-by: Hiroshi Shimamoto  > > > > > > > > ct.jp.nec.com>
> > > > > > > > > Reviewed-by: Hayato Momma 
> > > > > > > > > ---
> > > > > > > > >  pmd/pmd_memnic.c | 18 +-
> > > > > > > > >  1 file changed, 9 insertions(+), 9 deletions(-)
> > > > > > > > >
> > > > > > > > > diff --git a/pmd/pmd_memnic.c b/pmd/pmd_memnic.c
> > > > > > > > > index 7fc3093..875d3ea 100644
> > > > > > > > > --- a/pmd/pmd_memnic.c
> > > > > > > > > +++ b/pmd/pmd_memnic.c
> > > > > > > > > @@ -289,26 +289,26 @@ static uint16_t memnic_recv_pkts(void 
> > > > > > > > > *rx_queue,
> > > > > > > > >   int idx, next;
> > > > > > > > >   struct rte_eth_stats *st = 
> > > > > > > > > &adapter->stats[rte_lcore_id()];
> > > > > > > > >
> > > > > > > > > - if (!adapter->nic->hdr.valid)
> > > > > > > > > + if (unlikely(!adapter->nic->hdr.valid))
> > > > > > > > >   return 0;
> > > > > > > > >
> > > > > > > > >   pkts = bytes = errs = 0;
> > > > > > > > >   idx = adapter->up_idx;
> > > > > > > > >   for (nr = 0; nr < nb_pkts; nr++) {
> > > > > > > > >   p = &data->packets[idx];
> > > > > > > > > - if (p->status != MEMNIC_PKT_ST_FILLED)
> > > > > > > > > + if (unlikely(p->status != MEMNIC_PKT_ST_FILLED))
> > > > > > > > >   break;
> > > > > > > > >   /* prefetch the next area */
> > > > > > > > >   next = idx;
> > > > > > > > > - if (++next >= MEMNIC_NR_PACKET)
> > > > > > > > > + if (unlikely(++next >= MEMNIC_NR_PACKET))
> > > > > > > > >   ne

[dpdk-dev] [memnic PATCH v2 6/7] pmd: add branch hint in recv/xmit

2014-10-01 Thread Neil Horman

On Wed, Oct 01, 2014 at 11:33:23PM +, Hiroshi Shimamoto wrote:
> > Subject: Re: [dpdk-dev] [memnic PATCH v2 6/7] pmd: add branch hint in 
> > recv/xmit
> > 
> > On Wed, Oct 01, 2014 at 09:12:44AM +, Hiroshi Shimamoto wrote:
> > > > Subject: Re: [dpdk-dev] [memnic PATCH v2 6/7] pmd: add branch hint in 
> > > > recv/xmit
> > > >
> > > > On Tue, Sep 30, 2014 at 11:52:00PM +, Hiroshi Shimamoto wrote:
> > > > > Hi,
> > > > >
> > > > > > Subject: Re: [dpdk-dev] [memnic PATCH v2 6/7] pmd: add branch hint 
> > > > > > in recv/xmit
> > > > > >
> > > > > > On Tue, Sep 30, 2014 at 11:14:40AM +, Hiroshi Shimamoto wrote:
> > > > > > > From: Hiroshi Shimamoto 
> > > > > > >
> > > > > > > To reduce instruction cache miss, add branch condition hints into
> > > > > > > recv/xmit functions. This improves a bit performance.
> > > > > > >
> > > > > > > We can see performance improvements with memnic-tester.
> > > > > > > Using Xeon E5-2697 v2 @ 2.70GHz, 4 vCPU.
> > > > > > >  size |  before  |  after
> > > > > > >64 | 5.54Mpps | 5.55Mpps
> > > > > > >   128 | 5.46Mpps | 5.44Mpps
> > > > > > >   256 | 5.21Mpps | 5.22Mpps
> > > > > > >   512 | 4.50Mpps | 4.52Mpps
> > > > > > >  1024 | 3.71Mpps | 3.73Mpps
> > > > > > >  1280 | 3.21Mpps | 3.22Mpps
> > > > > > >  1518 | 2.92Mpps | 2.93Mpps
> > > > > > >
> > > > > > > Signed-off-by: Hiroshi Shimamoto 
> > > > > > > Reviewed-by: Hayato Momma 
> > > > > > > ---
> > > > > > >  pmd/pmd_memnic.c | 18 +-
> > > > > > >  1 file changed, 9 insertions(+), 9 deletions(-)
> > > > > > >
> > > > > > > diff --git a/pmd/pmd_memnic.c b/pmd/pmd_memnic.c
> > > > > > > index 7fc3093..875d3ea 100644
> > > > > > > --- a/pmd/pmd_memnic.c
> > > > > > > +++ b/pmd/pmd_memnic.c
> > > > > > > @@ -289,26 +289,26 @@ static uint16_t memnic_recv_pkts(void 
> > > > > > > *rx_queue,
> > > > > > >   int idx, next;
> > > > > > >   struct rte_eth_stats *st = &adapter->stats[rte_lcore_id()];
> > > > > > >
> > > > > > > - if (!adapter->nic->hdr.valid)
> > > > > > > + if (unlikely(!adapter->nic->hdr.valid))
> > > > > > >   return 0;
> > > > > > >
> > > > > > >   pkts = bytes = errs = 0;
> > > > > > >   idx = adapter->up_idx;
> > > > > > >   for (nr = 0; nr < nb_pkts; nr++) {
> > > > > > >   p = &data->packets[idx];
> > > > > > > - if (p->status != MEMNIC_PKT_ST_FILLED)
> > > > > > > + if (unlikely(p->status != MEMNIC_PKT_ST_FILLED))
> > > > > > >   break;
> > > > > > >   /* prefetch the next area */
> > > > > > >   next = idx;
> > > > > > > - if (++next >= MEMNIC_NR_PACKET)
> > > > > > > + if (unlikely(++next >= MEMNIC_NR_PACKET))
> > > > > > >   next = 0;
> > > > > > >   rte_prefetch0(&data->packets[next]);
> > > > > > > - if (p->len > framesz) {
> > > > > > > + if (unlikely(p->len > framesz)) {
> > > > > > >   errs++;
> > > > > > >   goto drop;
> > > > > > >   }
> > > > > > >   mb = rte_pktmbuf_alloc(adapter->mp);
> > > > > > > - if (!mb)
> > > > > > > + if (unlikely(!mb))
> > > > > > >   break;
> > > > > > >
> > > > > > >   rte_memcpy(rte_pktmbuf_mtod(mb, void *), p->data, 
> > > > > > > p->len);
> > > > > > > @@ -350,7 +350,7 @@ static uint16_t memnic_xmit_pkts(void 
> > > > > > > *tx_queue,
> > > > > > >   uint64_t pkts, bytes, errs

[dpdk-dev] [memnic PATCH v2 6/7] pmd: add branch hint in recv/xmit

2014-10-01 Thread Bruce Richardson

On Tue, Sep 30, 2014 at 11:52:00PM +, Hiroshi Shimamoto wrote:
> Hi,
> 
> > Subject: Re: [dpdk-dev] [memnic PATCH v2 6/7] pmd: add branch hint in 
> > recv/xmit
> > 
> > On Tue, Sep 30, 2014 at 11:14:40AM +, Hiroshi Shimamoto wrote:
> > > From: Hiroshi Shimamoto 
> > >
> > > To reduce instruction cache miss, add branch condition hints into
> > > recv/xmit functions. This improves a bit performance.
> > >
> > > We can see performance improvements with memnic-tester.
> > > Using Xeon E5-2697 v2 @ 2.70GHz, 4 vCPU.
> > >  size |  before  |  after
> > >64 | 5.54Mpps | 5.55Mpps
> > >   128 | 5.46Mpps | 5.44Mpps
> > >   256 | 5.21Mpps | 5.22Mpps
> > >   512 | 4.50Mpps | 4.52Mpps
> > >  1024 | 3.71Mpps | 3.73Mpps
> > >  1280 | 3.21Mpps | 3.22Mpps
> > >  1518 | 2.92Mpps | 2.93Mpps
> > >
> > > Signed-off-by: Hiroshi Shimamoto 
> > > Reviewed-by: Hayato Momma 
> > > ---
> > >  pmd/pmd_memnic.c | 18 +-
> > >  1 file changed, 9 insertions(+), 9 deletions(-)
> > >
> > > diff --git a/pmd/pmd_memnic.c b/pmd/pmd_memnic.c
> > > index 7fc3093..875d3ea 100644
> > > --- a/pmd/pmd_memnic.c
> > > +++ b/pmd/pmd_memnic.c
> > > @@ -289,26 +289,26 @@ static uint16_t memnic_recv_pkts(void *rx_queue,
> > >   int idx, next;
> > >   struct rte_eth_stats *st = &adapter->stats[rte_lcore_id()];
> > >
> > > - if (!adapter->nic->hdr.valid)
> > > + if (unlikely(!adapter->nic->hdr.valid))
> > >   return 0;
> > >
> > >   pkts = bytes = errs = 0;
> > >   idx = adapter->up_idx;
> > >   for (nr = 0; nr < nb_pkts; nr++) {
> > >   p = &data->packets[idx];
> > > - if (p->status != MEMNIC_PKT_ST_FILLED)
> > > + if (unlikely(p->status != MEMNIC_PKT_ST_FILLED))
> > >   break;
> > >   /* prefetch the next area */
> > >   next = idx;
> > > - if (++next >= MEMNIC_NR_PACKET)
> > > + if (unlikely(++next >= MEMNIC_NR_PACKET))
> > >   next = 0;
> > >   rte_prefetch0(&data->packets[next]);
> > > - if (p->len > framesz) {
> > > + if (unlikely(p->len > framesz)) {
> > >   errs++;
> > >   goto drop;
> > >   }
> > >   mb = rte_pktmbuf_alloc(adapter->mp);
> > > - if (!mb)
> > > + if (unlikely(!mb))
> > >   break;
> > >
> > >   rte_memcpy(rte_pktmbuf_mtod(mb, void *), p->data, p->len);
> > > @@ -350,7 +350,7 @@ static uint16_t memnic_xmit_pkts(void *tx_queue,
> > >   uint64_t pkts, bytes, errs;
> > >   uint32_t framesz = adapter->framesz;
> > >
> > > - if (!adapter->nic->hdr.valid)
> > > + if (unlikely(!adapter->nic->hdr.valid))
> > >   return 0;
> > >
> > >   pkts = bytes = errs = 0;
> > > @@ -360,7 +360,7 @@ static uint16_t memnic_xmit_pkts(void *tx_queue,
> > >   struct rte_mbuf *sg;
> > >   void *ptr;
> > >
> > > - if (pkt_len > framesz) {
> > > + if (unlikely(pkt_len > framesz)) {
> > >   errs++;
> > >   break;
> > >   }
> > > @@ -379,7 +379,7 @@ retry:
> > >   goto retry;
> > >   }
> > >
> > > - if (idx != ACCESS_ONCE(adapter->down_idx)) {
> > > + if (unlikely(idx != ACCESS_ONCE(adapter->down_idx))) {
> > Why are you using ACCESS_ONCE here?  Or for that matter, anywhere else in 
> > this
> > PMD?  The whole idea of the ACCESS_ONCE macro is to assign a value to a 
> > variable
> > once and prevent it from getting reloaded from memory at a later time, this 
> > is
> > exactly contrary to that, both in the sense that you're explicitly 
> > reloading the
> > same variable multiple times, and that you're using it as part of a 
> > comparison
> > operation, rather than an asignment operation
> 
> ACCESS_ONCE prevents compiler optimization and ensures load from memory.
> There could be multiple threads which read/write that index.
> We should compare the value previous and the current value in memory.
> In that reason, I use ACCESS_ONCE macro to get value in the memory.

Should you not just make the variable volatile? That's the normal way to 
guarantee reads from memory and prevent the compiler caching things in 
registers.

/Bruce

> 
> thanks,
> Hiroshi
> 
> > 
> > Neil
>

[dpdk-dev] [memnic PATCH v2 6/7] pmd: add branch hint in recv/xmit

2014-10-01 Thread Hiroshi Shimamoto

> Subject: Re: [dpdk-dev] [memnic PATCH v2 6/7] pmd: add branch hint in 
> recv/xmit
> 
> On Tue, Sep 30, 2014 at 11:52:00PM +, Hiroshi Shimamoto wrote:
> > Hi,
> >
> > > Subject: Re: [dpdk-dev] [memnic PATCH v2 6/7] pmd: add branch hint in 
> > > recv/xmit
> > >
> > > On Tue, Sep 30, 2014 at 11:14:40AM +, Hiroshi Shimamoto wrote:
> > > > From: Hiroshi Shimamoto 
> > > >
> > > > To reduce instruction cache miss, add branch condition hints into
> > > > recv/xmit functions. This improves a bit performance.
> > > >
> > > > We can see performance improvements with memnic-tester.
> > > > Using Xeon E5-2697 v2 @ 2.70GHz, 4 vCPU.
> > > >  size |  before  |  after
> > > >64 | 5.54Mpps | 5.55Mpps
> > > >   128 | 5.46Mpps | 5.44Mpps
> > > >   256 | 5.21Mpps | 5.22Mpps
> > > >   512 | 4.50Mpps | 4.52Mpps
> > > >  1024 | 3.71Mpps | 3.73Mpps
> > > >  1280 | 3.21Mpps | 3.22Mpps
> > > >  1518 | 2.92Mpps | 2.93Mpps
> > > >
> > > > Signed-off-by: Hiroshi Shimamoto 
> > > > Reviewed-by: Hayato Momma 
> > > > ---
> > > >  pmd/pmd_memnic.c | 18 +-
> > > >  1 file changed, 9 insertions(+), 9 deletions(-)
> > > >
> > > > diff --git a/pmd/pmd_memnic.c b/pmd/pmd_memnic.c
> > > > index 7fc3093..875d3ea 100644
> > > > --- a/pmd/pmd_memnic.c
> > > > +++ b/pmd/pmd_memnic.c
> > > > @@ -289,26 +289,26 @@ static uint16_t memnic_recv_pkts(void *rx_queue,
> > > > int idx, next;
> > > > struct rte_eth_stats *st = &adapter->stats[rte_lcore_id()];
> > > >
> > > > -   if (!adapter->nic->hdr.valid)
> > > > +   if (unlikely(!adapter->nic->hdr.valid))
> > > > return 0;
> > > >
> > > > pkts = bytes = errs = 0;
> > > > idx = adapter->up_idx;
> > > > for (nr = 0; nr < nb_pkts; nr++) {
> > > > p = &data->packets[idx];
> > > > -   if (p->status != MEMNIC_PKT_ST_FILLED)
> > > > +   if (unlikely(p->status != MEMNIC_PKT_ST_FILLED))
> > > > break;
> > > > /* prefetch the next area */
> > > > next = idx;
> > > > -   if (++next >= MEMNIC_NR_PACKET)
> > > > +   if (unlikely(++next >= MEMNIC_NR_PACKET))
> > > > next = 0;
> > > > rte_prefetch0(&data->packets[next]);
> > > > -   if (p->len > framesz) {
> > > > +   if (unlikely(p->len > framesz)) {
> > > > errs++;
> > > > goto drop;
> > > > }
> > > > mb = rte_pktmbuf_alloc(adapter->mp);
> > > > -   if (!mb)
> > > > +   if (unlikely(!mb))
> > > > break;
> > > >
> > > > rte_memcpy(rte_pktmbuf_mtod(mb, void *), p->data, 
> > > > p->len);
> > > > @@ -350,7 +350,7 @@ static uint16_t memnic_xmit_pkts(void *tx_queue,
> > > > uint64_t pkts, bytes, errs;
> > > > uint32_t framesz = adapter->framesz;
> > > >
> > > > -   if (!adapter->nic->hdr.valid)
> > > > +   if (unlikely(!adapter->nic->hdr.valid))
> > > > return 0;
> > > >
> > > > pkts = bytes = errs = 0;
> > > > @@ -360,7 +360,7 @@ static uint16_t memnic_xmit_pkts(void *tx_queue,
> > > > struct rte_mbuf *sg;
> > > > void *ptr;
> > > >
> > > > -   if (pkt_len > framesz) {
> > > > +   if (unlikely(pkt_len > framesz)) {
> > > > errs++;
> > > > break;
> > > > }
> > > > @@ -379,7 +379,7 @@ retry:
> > > > goto retry;
> > > > }
> > > >
> > > > -   if (idx != ACCESS_ONCE(adapter->down_idx)) {
> > > > +   if (unlikely(idx != ACCESS_ONCE(adapter->down_idx))) {
> > > Why are you using AC

[dpdk-dev] [memnic PATCH v2 6/7] pmd: add branch hint in recv/xmit

2014-10-01 Thread Thomas Monjalon

2014-10-01 00:01, Hiroshi Shimamoto:
> > The patch is ok. For the commit message, is it better
> > "to reduce branch mispredication"?
> 
> yes, that seems more suitable to explain the situation.
> 
> Thomas, what do you think? Can you replace the message when you apply
> this patch?

Yes, I did it ;)

-- 
Thomas

[dpdk-dev] [memnic PATCH v2 6/7] pmd: add branch hint in recv/xmit

2014-10-01 Thread Neil Horman

On Wed, Oct 01, 2014 at 09:12:44AM +, Hiroshi Shimamoto wrote:
> > Subject: Re: [dpdk-dev] [memnic PATCH v2 6/7] pmd: add branch hint in 
> > recv/xmit
> > 
> > On Tue, Sep 30, 2014 at 11:52:00PM +, Hiroshi Shimamoto wrote:
> > > Hi,
> > >
> > > > Subject: Re: [dpdk-dev] [memnic PATCH v2 6/7] pmd: add branch hint in 
> > > > recv/xmit
> > > >
> > > > On Tue, Sep 30, 2014 at 11:14:40AM +, Hiroshi Shimamoto wrote:
> > > > > From: Hiroshi Shimamoto 
> > > > >
> > > > > To reduce instruction cache miss, add branch condition hints into
> > > > > recv/xmit functions. This improves a bit performance.
> > > > >
> > > > > We can see performance improvements with memnic-tester.
> > > > > Using Xeon E5-2697 v2 @ 2.70GHz, 4 vCPU.
> > > > >  size |  before  |  after
> > > > >64 | 5.54Mpps | 5.55Mpps
> > > > >   128 | 5.46Mpps | 5.44Mpps
> > > > >   256 | 5.21Mpps | 5.22Mpps
> > > > >   512 | 4.50Mpps | 4.52Mpps
> > > > >  1024 | 3.71Mpps | 3.73Mpps
> > > > >  1280 | 3.21Mpps | 3.22Mpps
> > > > >  1518 | 2.92Mpps | 2.93Mpps
> > > > >
> > > > > Signed-off-by: Hiroshi Shimamoto 
> > > > > Reviewed-by: Hayato Momma 
> > > > > ---
> > > > >  pmd/pmd_memnic.c | 18 +-
> > > > >  1 file changed, 9 insertions(+), 9 deletions(-)
> > > > >
> > > > > diff --git a/pmd/pmd_memnic.c b/pmd/pmd_memnic.c
> > > > > index 7fc3093..875d3ea 100644
> > > > > --- a/pmd/pmd_memnic.c
> > > > > +++ b/pmd/pmd_memnic.c
> > > > > @@ -289,26 +289,26 @@ static uint16_t memnic_recv_pkts(void *rx_queue,
> > > > >   int idx, next;
> > > > >   struct rte_eth_stats *st = &adapter->stats[rte_lcore_id()];
> > > > >
> > > > > - if (!adapter->nic->hdr.valid)
> > > > > + if (unlikely(!adapter->nic->hdr.valid))
> > > > >   return 0;
> > > > >
> > > > >   pkts = bytes = errs = 0;
> > > > >   idx = adapter->up_idx;
> > > > >   for (nr = 0; nr < nb_pkts; nr++) {
> > > > >   p = &data->packets[idx];
> > > > > - if (p->status != MEMNIC_PKT_ST_FILLED)
> > > > > + if (unlikely(p->status != MEMNIC_PKT_ST_FILLED))
> > > > >   break;
> > > > >   /* prefetch the next area */
> > > > >   next = idx;
> > > > > - if (++next >= MEMNIC_NR_PACKET)
> > > > > + if (unlikely(++next >= MEMNIC_NR_PACKET))
> > > > >   next = 0;
> > > > >   rte_prefetch0(&data->packets[next]);
> > > > > - if (p->len > framesz) {
> > > > > + if (unlikely(p->len > framesz)) {
> > > > >   errs++;
> > > > >   goto drop;
> > > > >   }
> > > > >   mb = rte_pktmbuf_alloc(adapter->mp);
> > > > > - if (!mb)
> > > > > + if (unlikely(!mb))
> > > > >   break;
> > > > >
> > > > >   rte_memcpy(rte_pktmbuf_mtod(mb, void *), p->data, 
> > > > > p->len);
> > > > > @@ -350,7 +350,7 @@ static uint16_t memnic_xmit_pkts(void *tx_queue,
> > > > >   uint64_t pkts, bytes, errs;
> > > > >   uint32_t framesz = adapter->framesz;
> > > > >
> > > > > - if (!adapter->nic->hdr.valid)
> > > > > + if (unlikely(!adapter->nic->hdr.valid))
> > > > >   return 0;
> > > > >
> > > > >   pkts = bytes = errs = 0;
> > > > > @@ -360,7 +360,7 @@ static uint16_t memnic_xmit_pkts(void *tx_queue,
> > > > >   struct rte_mbuf *sg;
> > > > >   void *ptr;
> > > > >
> > > > > - if (pkt_len > framesz) {
> > > > > + if (unlikely(pkt_len > framesz)) {
> > > > >   errs++;
> > > > >   break;
>

[dpdk-dev] [memnic PATCH v2 6/7] pmd: add branch hint in recv/xmit

2014-10-01 Thread Neil Horman

On Wed, Oct 01, 2014 at 09:44:45AM +0100, Bruce Richardson wrote:
> On Tue, Sep 30, 2014 at 11:52:00PM +, Hiroshi Shimamoto wrote:
> > Hi,
> > 
> > > Subject: Re: [dpdk-dev] [memnic PATCH v2 6/7] pmd: add branch hint in 
> > > recv/xmit
> > > 
> > > On Tue, Sep 30, 2014 at 11:14:40AM +, Hiroshi Shimamoto wrote:
> > > > From: Hiroshi Shimamoto 
> > > >
> > > > To reduce instruction cache miss, add branch condition hints into
> > > > recv/xmit functions. This improves a bit performance.
> > > >
> > > > We can see performance improvements with memnic-tester.
> > > > Using Xeon E5-2697 v2 @ 2.70GHz, 4 vCPU.
> > > >  size |  before  |  after
> > > >64 | 5.54Mpps | 5.55Mpps
> > > >   128 | 5.46Mpps | 5.44Mpps
> > > >   256 | 5.21Mpps | 5.22Mpps
> > > >   512 | 4.50Mpps | 4.52Mpps
> > > >  1024 | 3.71Mpps | 3.73Mpps
> > > >  1280 | 3.21Mpps | 3.22Mpps
> > > >  1518 | 2.92Mpps | 2.93Mpps
> > > >
> > > > Signed-off-by: Hiroshi Shimamoto 
> > > > Reviewed-by: Hayato Momma 
> > > > ---
> > > >  pmd/pmd_memnic.c | 18 +-
> > > >  1 file changed, 9 insertions(+), 9 deletions(-)
> > > >
> > > > diff --git a/pmd/pmd_memnic.c b/pmd/pmd_memnic.c
> > > > index 7fc3093..875d3ea 100644
> > > > --- a/pmd/pmd_memnic.c
> > > > +++ b/pmd/pmd_memnic.c
> > > > @@ -289,26 +289,26 @@ static uint16_t memnic_recv_pkts(void *rx_queue,
> > > > int idx, next;
> > > > struct rte_eth_stats *st = &adapter->stats[rte_lcore_id()];
> > > >
> > > > -   if (!adapter->nic->hdr.valid)
> > > > +   if (unlikely(!adapter->nic->hdr.valid))
> > > > return 0;
> > > >
> > > > pkts = bytes = errs = 0;
> > > > idx = adapter->up_idx;
> > > > for (nr = 0; nr < nb_pkts; nr++) {
> > > > p = &data->packets[idx];
> > > > -   if (p->status != MEMNIC_PKT_ST_FILLED)
> > > > +   if (unlikely(p->status != MEMNIC_PKT_ST_FILLED))
> > > > break;
> > > > /* prefetch the next area */
> > > > next = idx;
> > > > -   if (++next >= MEMNIC_NR_PACKET)
> > > > +   if (unlikely(++next >= MEMNIC_NR_PACKET))
> > > > next = 0;
> > > > rte_prefetch0(&data->packets[next]);
> > > > -   if (p->len > framesz) {
> > > > +   if (unlikely(p->len > framesz)) {
> > > > errs++;
> > > > goto drop;
> > > > }
> > > > mb = rte_pktmbuf_alloc(adapter->mp);
> > > > -   if (!mb)
> > > > +   if (unlikely(!mb))
> > > > break;
> > > >
> > > > rte_memcpy(rte_pktmbuf_mtod(mb, void *), p->data, 
> > > > p->len);
> > > > @@ -350,7 +350,7 @@ static uint16_t memnic_xmit_pkts(void *tx_queue,
> > > > uint64_t pkts, bytes, errs;
> > > > uint32_t framesz = adapter->framesz;
> > > >
> > > > -   if (!adapter->nic->hdr.valid)
> > > > +   if (unlikely(!adapter->nic->hdr.valid))
> > > > return 0;
> > > >
> > > > pkts = bytes = errs = 0;
> > > > @@ -360,7 +360,7 @@ static uint16_t memnic_xmit_pkts(void *tx_queue,
> > > > struct rte_mbuf *sg;
> > > > void *ptr;
> > > >
> > > > -   if (pkt_len > framesz) {
> > > > +   if (unlikely(pkt_len > framesz)) {
> > > > errs++;
> > > > break;
> > > > }
> > > > @@ -379,7 +379,7 @@ retry:
> > > > goto retry;
> > > > }
> > > >
> > > > -   if (idx != ACCESS_ONCE(adapter->down_idx)) {
> > > > +   if (unlikely(idx != ACCESS_ONCE(adapter->down_idx))) {
> > > Why are you using ACCESS_ONCE here?  Or for that matter,

[dpdk-dev] [memnic PATCH v2 6/7] pmd: add branch hint in recv/xmit

2014-10-01 Thread Hiroshi Shimamoto

Hi,

> Subject: RE: [memnic PATCH v2 6/7] pmd: add branch hint in recv/xmit
> 
> The patch is ok. For the commit message, is it better
> "to reduce branch mispredication"?

yes, that seems more suitable to explain the situation.

Thomas, what do you think? Can you replace the message when you apply
this patch?

thanks,
Hiroshi

> 
> > -Original Message-
> > From: dev [mailto:dev-bounces at dpdk.org] On Behalf Of Hiroshi Shimamoto
> > Sent: Tuesday, September 30, 2014 7:15 PM
> > To: dev at dpdk.org
> > Cc: Hayato Momma
> > Subject: [dpdk-dev] [memnic PATCH v2 6/7] pmd: add branch hint in recv/xmit
> >
> > From: Hiroshi Shimamoto 
> >
> > To reduce instruction cache miss, add branch condition hints into
> > recv/xmit functions. This improves a bit performance.
> >
> > We can see performance improvements with memnic-tester.
> > Using Xeon E5-2697 v2 @ 2.70GHz, 4 vCPU.
> >  size |  before  |  after
> >64 | 5.54Mpps | 5.55Mpps
> >   128 | 5.46Mpps | 5.44Mpps
> >   256 | 5.21Mpps | 5.22Mpps
> >   512 | 4.50Mpps | 4.52Mpps
> >  1024 | 3.71Mpps | 3.73Mpps
> >  1280 | 3.21Mpps | 3.22Mpps
> >  1518 | 2.92Mpps | 2.93Mpps
> >
> > Signed-off-by: Hiroshi Shimamoto 
> > Reviewed-by: Hayato Momma 
> > ---
> >  pmd/pmd_memnic.c | 18 +-
> >  1 file changed, 9 insertions(+), 9 deletions(-)
> >
> > diff --git a/pmd/pmd_memnic.c b/pmd/pmd_memnic.c
> > index 7fc3093..875d3ea 100644
> > --- a/pmd/pmd_memnic.c
> > +++ b/pmd/pmd_memnic.c
> > @@ -289,26 +289,26 @@ static uint16_t memnic_recv_pkts(void *rx_queue,
> > int idx, next;
> > struct rte_eth_stats *st = &adapter->stats[rte_lcore_id()];
> >
> > -   if (!adapter->nic->hdr.valid)
> > +   if (unlikely(!adapter->nic->hdr.valid))
> > return 0;
> >
> > pkts = bytes = errs = 0;
> > idx = adapter->up_idx;
> > for (nr = 0; nr < nb_pkts; nr++) {
> > p = &data->packets[idx];
> > -   if (p->status != MEMNIC_PKT_ST_FILLED)
> > +   if (unlikely(p->status != MEMNIC_PKT_ST_FILLED))
> > break;
> > /* prefetch the next area */
> > next = idx;
> > -   if (++next >= MEMNIC_NR_PACKET)
> > +   if (unlikely(++next >= MEMNIC_NR_PACKET))
> > next = 0;
> > rte_prefetch0(&data->packets[next]);
> > -   if (p->len > framesz) {
> > +   if (unlikely(p->len > framesz)) {
> > errs++;
> > goto drop;
> > }
> > mb = rte_pktmbuf_alloc(adapter->mp);
> > -   if (!mb)
> > +   if (unlikely(!mb))
> > break;
> >
> > rte_memcpy(rte_pktmbuf_mtod(mb, void *), p->data, p->len);
> > @@ -350,7 +350,7 @@ static uint16_t memnic_xmit_pkts(void *tx_queue,
> > uint64_t pkts, bytes, errs;
> > uint32_t framesz = adapter->framesz;
> >
> > -   if (!adapter->nic->hdr.valid)
> > +   if (unlikely(!adapter->nic->hdr.valid))
> > return 0;
> >
> > pkts = bytes = errs = 0;
> > @@ -360,7 +360,7 @@ static uint16_t memnic_xmit_pkts(void *tx_queue,
> > struct rte_mbuf *sg;
> > void *ptr;
> >
> > -   if (pkt_len > framesz) {
> > +   if (unlikely(pkt_len > framesz)) {
> > errs++;
> > break;
> > }
> > @@ -379,7 +379,7 @@ retry:
> > goto retry;
> > }
> >
> > -   if (idx != ACCESS_ONCE(adapter->down_idx)) {
> > +   if (unlikely(idx != ACCESS_ONCE(adapter->down_idx))) {
> > /*
> >  * host freed this and got false positive,
> >  * need to recover the status and retry.
> > @@ -388,7 +388,7 @@ retry:
> > goto retry;
> > }
> >
> > -   if (++idx >= MEMNIC_NR_PACKET)
> > +   if (unlikely(++idx >= MEMNIC_NR_PACKET))
> > idx = 0;
> > adapter->down_idx = idx;
> >
> > --
> > 1.8.3.1

[dpdk-dev] [memnic PATCH v2 6/7] pmd: add branch hint in recv/xmit

2014-09-30 Thread Hiroshi Shimamoto

Hi,

> Subject: Re: [dpdk-dev] [memnic PATCH v2 6/7] pmd: add branch hint in 
> recv/xmit
> 
> On Tue, Sep 30, 2014 at 11:14:40AM +, Hiroshi Shimamoto wrote:
> > From: Hiroshi Shimamoto 
> >
> > To reduce instruction cache miss, add branch condition hints into
> > recv/xmit functions. This improves a bit performance.
> >
> > We can see performance improvements with memnic-tester.
> > Using Xeon E5-2697 v2 @ 2.70GHz, 4 vCPU.
> >  size |  before  |  after
> >64 | 5.54Mpps | 5.55Mpps
> >   128 | 5.46Mpps | 5.44Mpps
> >   256 | 5.21Mpps | 5.22Mpps
> >   512 | 4.50Mpps | 4.52Mpps
> >  1024 | 3.71Mpps | 3.73Mpps
> >  1280 | 3.21Mpps | 3.22Mpps
> >  1518 | 2.92Mpps | 2.93Mpps
> >
> > Signed-off-by: Hiroshi Shimamoto 
> > Reviewed-by: Hayato Momma 
> > ---
> >  pmd/pmd_memnic.c | 18 +-
> >  1 file changed, 9 insertions(+), 9 deletions(-)
> >
> > diff --git a/pmd/pmd_memnic.c b/pmd/pmd_memnic.c
> > index 7fc3093..875d3ea 100644
> > --- a/pmd/pmd_memnic.c
> > +++ b/pmd/pmd_memnic.c
> > @@ -289,26 +289,26 @@ static uint16_t memnic_recv_pkts(void *rx_queue,
> > int idx, next;
> > struct rte_eth_stats *st = &adapter->stats[rte_lcore_id()];
> >
> > -   if (!adapter->nic->hdr.valid)
> > +   if (unlikely(!adapter->nic->hdr.valid))
> > return 0;
> >
> > pkts = bytes = errs = 0;
> > idx = adapter->up_idx;
> > for (nr = 0; nr < nb_pkts; nr++) {
> > p = &data->packets[idx];
> > -   if (p->status != MEMNIC_PKT_ST_FILLED)
> > +   if (unlikely(p->status != MEMNIC_PKT_ST_FILLED))
> > break;
> > /* prefetch the next area */
> > next = idx;
> > -   if (++next >= MEMNIC_NR_PACKET)
> > +   if (unlikely(++next >= MEMNIC_NR_PACKET))
> > next = 0;
> > rte_prefetch0(&data->packets[next]);
> > -   if (p->len > framesz) {
> > +   if (unlikely(p->len > framesz)) {
> > errs++;
> > goto drop;
> > }
> > mb = rte_pktmbuf_alloc(adapter->mp);
> > -   if (!mb)
> > +   if (unlikely(!mb))
> > break;
> >
> > rte_memcpy(rte_pktmbuf_mtod(mb, void *), p->data, p->len);
> > @@ -350,7 +350,7 @@ static uint16_t memnic_xmit_pkts(void *tx_queue,
> > uint64_t pkts, bytes, errs;
> > uint32_t framesz = adapter->framesz;
> >
> > -   if (!adapter->nic->hdr.valid)
> > +   if (unlikely(!adapter->nic->hdr.valid))
> > return 0;
> >
> > pkts = bytes = errs = 0;
> > @@ -360,7 +360,7 @@ static uint16_t memnic_xmit_pkts(void *tx_queue,
> > struct rte_mbuf *sg;
> > void *ptr;
> >
> > -   if (pkt_len > framesz) {
> > +   if (unlikely(pkt_len > framesz)) {
> > errs++;
> > break;
> > }
> > @@ -379,7 +379,7 @@ retry:
> > goto retry;
> > }
> >
> > -   if (idx != ACCESS_ONCE(adapter->down_idx)) {
> > +   if (unlikely(idx != ACCESS_ONCE(adapter->down_idx))) {
> Why are you using ACCESS_ONCE here?  Or for that matter, anywhere else in this
> PMD?  The whole idea of the ACCESS_ONCE macro is to assign a value to a 
> variable
> once and prevent it from getting reloaded from memory at a later time, this is
> exactly contrary to that, both in the sense that you're explicitly reloading 
> the
> same variable multiple times, and that you're using it as part of a comparison
> operation, rather than an asignment operation

ACCESS_ONCE prevents compiler optimization and ensures load from memory.
There could be multiple threads which read/write that index.
We should compare the value previous and the current value in memory.
In that reason, I use ACCESS_ONCE macro to get value in the memory.

thanks,
Hiroshi

> 
> Neil

[dpdk-dev] [memnic PATCH v2 6/7] pmd: add branch hint in recv/xmit

2014-09-30 Thread Thomas Monjalon

2014-09-30 14:38, Xie, Huawei:
> > -   if (++next >= MEMNIC_NR_PACKET)
> > +   if (unlikely(++next >= MEMNIC_NR_PACKET))
> 
> On IA,  compiler can use add, cmp and cmov to avoid branch.
> But If MEMNIC_NR_PACKET is always power of 2, 
>   it is better just next = (next + 1) & (MEMNIC_NR_PACKET - 1)

Power of 2 is not enforced for MEMNIC_NR_PACKET.

-- 
Thomas

[dpdk-dev] [memnic PATCH v2 6/7] pmd: add branch hint in recv/xmit

2014-09-30 Thread Xie, Huawei



> -Original Message-
> From: dev [mailto:dev-bounces at dpdk.org] On Behalf Of Hiroshi Shimamoto
> Sent: Tuesday, September 30, 2014 7:15 PM
> To: dev at dpdk.org
> Cc: Hayato Momma
> Subject: [dpdk-dev] [memnic PATCH v2 6/7] pmd: add branch hint in recv/xmit
> 
> From: Hiroshi Shimamoto 
> 
> To reduce instruction cache miss, add branch condition hints into
> recv/xmit functions. This improves a bit performance.
> 
> We can see performance improvements with memnic-tester.
> Using Xeon E5-2697 v2 @ 2.70GHz, 4 vCPU.
>  size |  before  |  after
>64 | 5.54Mpps | 5.55Mpps
>   128 | 5.46Mpps | 5.44Mpps
>   256 | 5.21Mpps | 5.22Mpps
>   512 | 4.50Mpps | 4.52Mpps
>  1024 | 3.71Mpps | 3.73Mpps
>  1280 | 3.21Mpps | 3.22Mpps
>  1518 | 2.92Mpps | 2.93Mpps
> 
> Signed-off-by: Hiroshi Shimamoto 
> Reviewed-by: Hayato Momma 
> ---
>  pmd/pmd_memnic.c | 18 +-
>  1 file changed, 9 insertions(+), 9 deletions(-)
> 
> diff --git a/pmd/pmd_memnic.c b/pmd/pmd_memnic.c
> index 7fc3093..875d3ea 100644
> --- a/pmd/pmd_memnic.c
> +++ b/pmd/pmd_memnic.c
> @@ -289,26 +289,26 @@ static uint16_t memnic_recv_pkts(void *rx_queue,
>   int idx, next;
>   struct rte_eth_stats *st = &adapter->stats[rte_lcore_id()];
> 
> - if (!adapter->nic->hdr.valid)
> + if (unlikely(!adapter->nic->hdr.valid))
>   return 0;
> 
>   pkts = bytes = errs = 0;
>   idx = adapter->up_idx;
>   for (nr = 0; nr < nb_pkts; nr++) {
>   p = &data->packets[idx];
> - if (p->status != MEMNIC_PKT_ST_FILLED)
> + if (unlikely(p->status != MEMNIC_PKT_ST_FILLED))
>   break;
>   /* prefetch the next area */
>   next = idx;
> - if (++next >= MEMNIC_NR_PACKET)
> + if (unlikely(++next >= MEMNIC_NR_PACKET))
On IA,  compiler can use add, cmp and cmov to avoid branch.
But If MEMNIC_NR_PACKET is always power of 2, 
it is better just next = (next + 1) & (MEMNIC_NR_PACKET - 1)

>   next = 0;
>   rte_prefetch0(&data->packets[next]);
> - if (p->len > framesz) {
> + if (unlikely(p->len > framesz)) {
>   errs++;
>   goto drop;
>   }
>   mb = rte_pktmbuf_alloc(adapter->mp);
> - if (!mb)
> + if (unlikely(!mb))
>   break;
> 
>   rte_memcpy(rte_pktmbuf_mtod(mb, void *), p->data, p->len);
> @@ -350,7 +350,7 @@ static uint16_t memnic_xmit_pkts(void *tx_queue,
>   uint64_t pkts, bytes, errs;
>   uint32_t framesz = adapter->framesz;
> 
> - if (!adapter->nic->hdr.valid)
> + if (unlikely(!adapter->nic->hdr.valid))
>   return 0;
> 
>   pkts = bytes = errs = 0;
> @@ -360,7 +360,7 @@ static uint16_t memnic_xmit_pkts(void *tx_queue,
>   struct rte_mbuf *sg;
>   void *ptr;
> 
> - if (pkt_len > framesz) {
> + if (unlikely(pkt_len > framesz)) {
>   errs++;
>   break;
>   }
> @@ -379,7 +379,7 @@ retry:
>   goto retry;
>   }
> 
> - if (idx != ACCESS_ONCE(adapter->down_idx)) {
> + if (unlikely(idx != ACCESS_ONCE(adapter->down_idx))) {
>   /*
>* host freed this and got false positive,
>* need to recover the status and retry.
> @@ -388,7 +388,7 @@ retry:
>   goto retry;
>   }
> 
> - if (++idx >= MEMNIC_NR_PACKET)
> + if (unlikely(++idx >= MEMNIC_NR_PACKET))
>   idx = 0;
>   adapter->down_idx = idx;
> 
> --
> 1.8.3.1

[dpdk-dev] [memnic PATCH v2 6/7] pmd: add branch hint in recv/xmit

2014-09-30 Thread Xie, Huawei

The patch is ok. For the commit message, is it better 
"to reduce branch mispredication"?

> -Original Message-
> From: dev [mailto:dev-bounces at dpdk.org] On Behalf Of Hiroshi Shimamoto
> Sent: Tuesday, September 30, 2014 7:15 PM
> To: dev at dpdk.org
> Cc: Hayato Momma
> Subject: [dpdk-dev] [memnic PATCH v2 6/7] pmd: add branch hint in recv/xmit
> 
> From: Hiroshi Shimamoto 
> 
> To reduce instruction cache miss, add branch condition hints into
> recv/xmit functions. This improves a bit performance.
> 
> We can see performance improvements with memnic-tester.
> Using Xeon E5-2697 v2 @ 2.70GHz, 4 vCPU.
>  size |  before  |  after
>64 | 5.54Mpps | 5.55Mpps
>   128 | 5.46Mpps | 5.44Mpps
>   256 | 5.21Mpps | 5.22Mpps
>   512 | 4.50Mpps | 4.52Mpps
>  1024 | 3.71Mpps | 3.73Mpps
>  1280 | 3.21Mpps | 3.22Mpps
>  1518 | 2.92Mpps | 2.93Mpps
> 
> Signed-off-by: Hiroshi Shimamoto 
> Reviewed-by: Hayato Momma 
> ---
>  pmd/pmd_memnic.c | 18 +-
>  1 file changed, 9 insertions(+), 9 deletions(-)
> 
> diff --git a/pmd/pmd_memnic.c b/pmd/pmd_memnic.c
> index 7fc3093..875d3ea 100644
> --- a/pmd/pmd_memnic.c
> +++ b/pmd/pmd_memnic.c
> @@ -289,26 +289,26 @@ static uint16_t memnic_recv_pkts(void *rx_queue,
>   int idx, next;
>   struct rte_eth_stats *st = &adapter->stats[rte_lcore_id()];
> 
> - if (!adapter->nic->hdr.valid)
> + if (unlikely(!adapter->nic->hdr.valid))
>   return 0;
> 
>   pkts = bytes = errs = 0;
>   idx = adapter->up_idx;
>   for (nr = 0; nr < nb_pkts; nr++) {
>   p = &data->packets[idx];
> - if (p->status != MEMNIC_PKT_ST_FILLED)
> + if (unlikely(p->status != MEMNIC_PKT_ST_FILLED))
>   break;
>   /* prefetch the next area */
>   next = idx;
> - if (++next >= MEMNIC_NR_PACKET)
> + if (unlikely(++next >= MEMNIC_NR_PACKET))
>   next = 0;
>   rte_prefetch0(&data->packets[next]);
> - if (p->len > framesz) {
> + if (unlikely(p->len > framesz)) {
>   errs++;
>   goto drop;
>   }
>   mb = rte_pktmbuf_alloc(adapter->mp);
> - if (!mb)
> + if (unlikely(!mb))
>   break;
> 
>   rte_memcpy(rte_pktmbuf_mtod(mb, void *), p->data, p->len);
> @@ -350,7 +350,7 @@ static uint16_t memnic_xmit_pkts(void *tx_queue,
>   uint64_t pkts, bytes, errs;
>   uint32_t framesz = adapter->framesz;
> 
> - if (!adapter->nic->hdr.valid)
> + if (unlikely(!adapter->nic->hdr.valid))
>   return 0;
> 
>   pkts = bytes = errs = 0;
> @@ -360,7 +360,7 @@ static uint16_t memnic_xmit_pkts(void *tx_queue,
>   struct rte_mbuf *sg;
>   void *ptr;
> 
> - if (pkt_len > framesz) {
> + if (unlikely(pkt_len > framesz)) {
>   errs++;
>   break;
>   }
> @@ -379,7 +379,7 @@ retry:
>   goto retry;
>   }
> 
> - if (idx != ACCESS_ONCE(adapter->down_idx)) {
> + if (unlikely(idx != ACCESS_ONCE(adapter->down_idx))) {
>   /*
>* host freed this and got false positive,
>* need to recover the status and retry.
> @@ -388,7 +388,7 @@ retry:
>   goto retry;
>   }
> 
> - if (++idx >= MEMNIC_NR_PACKET)
> + if (unlikely(++idx >= MEMNIC_NR_PACKET))
>   idx = 0;
>   adapter->down_idx = idx;
> 
> --
> 1.8.3.1

[dpdk-dev] [memnic PATCH v2 6/7] pmd: add branch hint in recv/xmit

2014-09-30 Thread Hiroshi Shimamoto

From: Hiroshi Shimamoto 

To reduce instruction cache miss, add branch condition hints into
recv/xmit functions. This improves a bit performance.

We can see performance improvements with memnic-tester.
Using Xeon E5-2697 v2 @ 2.70GHz, 4 vCPU.
 size |  before  |  after
   64 | 5.54Mpps | 5.55Mpps
  128 | 5.46Mpps | 5.44Mpps
  256 | 5.21Mpps | 5.22Mpps
  512 | 4.50Mpps | 4.52Mpps
 1024 | 3.71Mpps | 3.73Mpps
 1280 | 3.21Mpps | 3.22Mpps
 1518 | 2.92Mpps | 2.93Mpps

Signed-off-by: Hiroshi Shimamoto 
Reviewed-by: Hayato Momma 
---
 pmd/pmd_memnic.c | 18 +-
 1 file changed, 9 insertions(+), 9 deletions(-)

diff --git a/pmd/pmd_memnic.c b/pmd/pmd_memnic.c
index 7fc3093..875d3ea 100644
--- a/pmd/pmd_memnic.c
+++ b/pmd/pmd_memnic.c
@@ -289,26 +289,26 @@ static uint16_t memnic_recv_pkts(void *rx_queue,
int idx, next;
struct rte_eth_stats *st = &adapter->stats[rte_lcore_id()];

-   if (!adapter->nic->hdr.valid)
+   if (unlikely(!adapter->nic->hdr.valid))
return 0;

pkts = bytes = errs = 0;
idx = adapter->up_idx;
for (nr = 0; nr < nb_pkts; nr++) {
p = &data->packets[idx];
-   if (p->status != MEMNIC_PKT_ST_FILLED)
+   if (unlikely(p->status != MEMNIC_PKT_ST_FILLED))
break;
/* prefetch the next area */
next = idx;
-   if (++next >= MEMNIC_NR_PACKET)
+   if (unlikely(++next >= MEMNIC_NR_PACKET))
next = 0;
rte_prefetch0(&data->packets[next]);
-   if (p->len > framesz) {
+   if (unlikely(p->len > framesz)) {
errs++;
goto drop;
}
mb = rte_pktmbuf_alloc(adapter->mp);
-   if (!mb)
+   if (unlikely(!mb))
break;

rte_memcpy(rte_pktmbuf_mtod(mb, void *), p->data, p->len);
@@ -350,7 +350,7 @@ static uint16_t memnic_xmit_pkts(void *tx_queue,
uint64_t pkts, bytes, errs;
uint32_t framesz = adapter->framesz;

-   if (!adapter->nic->hdr.valid)
+   if (unlikely(!adapter->nic->hdr.valid))
return 0;

pkts = bytes = errs = 0;
@@ -360,7 +360,7 @@ static uint16_t memnic_xmit_pkts(void *tx_queue,
struct rte_mbuf *sg;
void *ptr;

-   if (pkt_len > framesz) {
+   if (unlikely(pkt_len > framesz)) {
errs++;
break;
}
@@ -379,7 +379,7 @@ retry:
goto retry;
}

-   if (idx != ACCESS_ONCE(adapter->down_idx)) {
+   if (unlikely(idx != ACCESS_ONCE(adapter->down_idx))) {
/*
 * host freed this and got false positive,
 * need to recover the status and retry.
@@ -388,7 +388,7 @@ retry:
goto retry;
}

-   if (++idx >= MEMNIC_NR_PACKET)
+   if (unlikely(++idx >= MEMNIC_NR_PACKET))
idx = 0;
adapter->down_idx = idx;

-- 
1.8.3.1

[dpdk-dev] [memnic PATCH v2 6/7] pmd: add branch hint in recv/xmit

2014-09-30 Thread Neil Horman

On Tue, Sep 30, 2014 at 11:14:40AM +, Hiroshi Shimamoto wrote:
> From: Hiroshi Shimamoto 
> 
> To reduce instruction cache miss, add branch condition hints into
> recv/xmit functions. This improves a bit performance.
> 
> We can see performance improvements with memnic-tester.
> Using Xeon E5-2697 v2 @ 2.70GHz, 4 vCPU.
>  size |  before  |  after
>64 | 5.54Mpps | 5.55Mpps
>   128 | 5.46Mpps | 5.44Mpps
>   256 | 5.21Mpps | 5.22Mpps
>   512 | 4.50Mpps | 4.52Mpps
>  1024 | 3.71Mpps | 3.73Mpps
>  1280 | 3.21Mpps | 3.22Mpps
>  1518 | 2.92Mpps | 2.93Mpps
> 
> Signed-off-by: Hiroshi Shimamoto 
> Reviewed-by: Hayato Momma 
> ---
>  pmd/pmd_memnic.c | 18 +-
>  1 file changed, 9 insertions(+), 9 deletions(-)
> 
> diff --git a/pmd/pmd_memnic.c b/pmd/pmd_memnic.c
> index 7fc3093..875d3ea 100644
> --- a/pmd/pmd_memnic.c
> +++ b/pmd/pmd_memnic.c
> @@ -289,26 +289,26 @@ static uint16_t memnic_recv_pkts(void *rx_queue,
>   int idx, next;
>   struct rte_eth_stats *st = &adapter->stats[rte_lcore_id()];
>  
> - if (!adapter->nic->hdr.valid)
> + if (unlikely(!adapter->nic->hdr.valid))
>   return 0;
>  
>   pkts = bytes = errs = 0;
>   idx = adapter->up_idx;
>   for (nr = 0; nr < nb_pkts; nr++) {
>   p = &data->packets[idx];
> - if (p->status != MEMNIC_PKT_ST_FILLED)
> + if (unlikely(p->status != MEMNIC_PKT_ST_FILLED))
>   break;
>   /* prefetch the next area */
>   next = idx;
> - if (++next >= MEMNIC_NR_PACKET)
> + if (unlikely(++next >= MEMNIC_NR_PACKET))
>   next = 0;
>   rte_prefetch0(&data->packets[next]);
> - if (p->len > framesz) {
> + if (unlikely(p->len > framesz)) {
>   errs++;
>   goto drop;
>   }
>   mb = rte_pktmbuf_alloc(adapter->mp);
> - if (!mb)
> + if (unlikely(!mb))
>   break;
>  
>   rte_memcpy(rte_pktmbuf_mtod(mb, void *), p->data, p->len);
> @@ -350,7 +350,7 @@ static uint16_t memnic_xmit_pkts(void *tx_queue,
>   uint64_t pkts, bytes, errs;
>   uint32_t framesz = adapter->framesz;
>  
> - if (!adapter->nic->hdr.valid)
> + if (unlikely(!adapter->nic->hdr.valid))
>   return 0;
>  
>   pkts = bytes = errs = 0;
> @@ -360,7 +360,7 @@ static uint16_t memnic_xmit_pkts(void *tx_queue,
>   struct rte_mbuf *sg;
>   void *ptr;
>  
> - if (pkt_len > framesz) {
> + if (unlikely(pkt_len > framesz)) {
>   errs++;
>   break;
>   }
> @@ -379,7 +379,7 @@ retry:
>   goto retry;
>   }
>  
> - if (idx != ACCESS_ONCE(adapter->down_idx)) {
> + if (unlikely(idx != ACCESS_ONCE(adapter->down_idx))) {
Why are you using ACCESS_ONCE here?  Or for that matter, anywhere else in this
PMD?  The whole idea of the ACCESS_ONCE macro is to assign a value to a variable
once and prevent it from getting reloaded from memory at a later time, this is
exactly contrary to that, both in the sense that you're explicitly reloading the
same variable multiple times, and that you're using it as part of a comparison
operation, rather than an asignment operation

Neil

[dpdk-dev] [memnic PATCH v2 6/7] pmd: add branch hint in recv/xmit

[dpdk-dev] [memnic PATCH v2 6/7] pmd: add branch hint in recv/xmit

[dpdk-dev] [memnic PATCH v2 6/7] pmd: add branch hint in recv/xmit

[dpdk-dev] [memnic PATCH v2 6/7] pmd: add branch hint in recv/xmit

[dpdk-dev] [memnic PATCH v2 6/7] pmd: add branch hint in recv/xmit

[dpdk-dev] [memnic PATCH v2 6/7] pmd: add branch hint in recv/xmit

[dpdk-dev] [memnic PATCH v2 6/7] pmd: add branch hint in recv/xmit

[dpdk-dev] [memnic PATCH v2 6/7] pmd: add branch hint in recv/xmit

[dpdk-dev] [memnic PATCH v2 6/7] pmd: add branch hint in recv/xmit

[dpdk-dev] [memnic PATCH v2 6/7] pmd: add branch hint in recv/xmit

[dpdk-dev] [memnic PATCH v2 6/7] pmd: add branch hint in recv/xmit

[dpdk-dev] [memnic PATCH v2 6/7] pmd: add branch hint in recv/xmit

[dpdk-dev] [memnic PATCH v2 6/7] pmd: add branch hint in recv/xmit

[dpdk-dev] [memnic PATCH v2 6/7] pmd: add branch hint in recv/xmit

[dpdk-dev] [memnic PATCH v2 6/7] pmd: add branch hint in recv/xmit

[dpdk-dev] [memnic PATCH v2 6/7] pmd: add branch hint in recv/xmit

16 matches

Site Navigation

Mail list logo

Footer information