Re: [PATCH v8 1/3] eal/x86: run-time dispatch over memcpy

From: "Li, Xiaoyun" <xiaoyun.li@intel.com>
To: "Ananyev, Konstantin" <konstantin.ananyev@intel.com>,
	"Richardson, Bruce" <bruce.richardson@intel.com>,
	Thomas Monjalon <thomas@monjalon.net>
Cc: "dev@dpdk.org" <dev@dpdk.org>,
	"Lu, Wenzhuo" <wenzhuo.lu@intel.com>,
	"Zhang, Helin" <helin.zhang@intel.com>,
	"ophirmu@mellanox.com" <ophirmu@mellanox.com>
Subject: Re: [PATCH v8 1/3] eal/x86: run-time dispatch over memcpy
Date: Wed, 25 Oct 2017 08:54:10 +0000	[thread overview]
Message-ID: <B9E724F4CB7543449049E7AE7669D82F48176B@SHSMSX101.ccr.corp.intel.com> (raw)
In-Reply-To: <2601191342CEEE43887BDE71AB9772585FAAEB8E@IRSMSX103.ger.corp.intel.com>


> -----Original Message-----
> From: Ananyev, Konstantin
> Sent: Wednesday, October 25, 2017 16:51
> To: Li, Xiaoyun <xiaoyun.li@intel.com>; Richardson, Bruce
> <bruce.richardson@intel.com>; Thomas Monjalon <thomas@monjalon.net>
> Cc: dev@dpdk.org; Lu, Wenzhuo <wenzhuo.lu@intel.com>; Zhang, Helin
> <helin.zhang@intel.com>; ophirmu@mellanox.com
> Subject: RE: [dpdk-dev] [PATCH v8 1/3] eal/x86: run-time dispatch over
> memcpy
> 
> 
> 
> > -----Original Message-----
> > From: Li, Xiaoyun
> > Sent: Wednesday, October 25, 2017 7:55 AM
> > To: Li, Xiaoyun <xiaoyun.li@intel.com>; Richardson, Bruce
> > <bruce.richardson@intel.com>; Thomas Monjalon <thomas@monjalon.net>
> > Cc: Ananyev, Konstantin <konstantin.ananyev@intel.com>; dev@dpdk.org;
> > Lu, Wenzhuo <wenzhuo.lu@intel.com>; Zhang, Helin
> > <helin.zhang@intel.com>; ophirmu@mellanox.com
> > Subject: RE: [dpdk-dev] [PATCH v8 1/3] eal/x86: run-time dispatch over
> > memcpy
> >
> > Hi
> >
> > > -----Original Message-----
> > > From: dev [mailto:dev-bounces@dpdk.org] On Behalf Of Li, Xiaoyun
> > > Sent: Friday, October 20, 2017 09:03
> > > To: Richardson, Bruce <bruce.richardson@intel.com>; Thomas Monjalon
> > > <thomas@monjalon.net>
> > > Cc: Ananyev, Konstantin <konstantin.ananyev@intel.com>;
> > > dev@dpdk.org; Lu, Wenzhuo <wenzhuo.lu@intel.com>; Zhang, Helin
> > > <helin.zhang@intel.com>; ophirmu@mellanox.com
> > > Subject: Re: [dpdk-dev] [PATCH v8 1/3] eal/x86: run-time dispatch
> > > over memcpy
> > >
> > >
> > >
> > > > -----Original Message-----
> > > > From: Richardson, Bruce
> > > > Sent: Thursday, October 19, 2017 17:30
> > > > To: Thomas Monjalon <thomas@monjalon.net>
> > > > Cc: Li, Xiaoyun <xiaoyun.li@intel.com>; Ananyev, Konstantin
> > > > <konstantin.ananyev@intel.com>; dev@dpdk.org; Lu, Wenzhuo
> > > > <wenzhuo.lu@intel.com>; Zhang, Helin <helin.zhang@intel.com>;
> > > > ophirmu@mellanox.com
> > > > Subject: Re: [dpdk-dev] [PATCH v8 1/3] eal/x86: run-time dispatch
> > > > over memcpy
> > > >
> > > > On Thu, Oct 19, 2017 at 11:00:33AM +0200, Thomas Monjalon wrote:
> > > > > 19/10/2017 10:50, Li, Xiaoyun:
> > > > > >
> > > > > > > -----Original Message-----
> > > > > > > From: Thomas Monjalon [mailto:thomas@monjalon.net]
> > > > > > > Sent: Thursday, October 19, 2017 16:34
> > > > > > > To: Li, Xiaoyun <xiaoyun.li@intel.com>
> > > > > > > Cc: Ananyev, Konstantin <konstantin.ananyev@intel.com>;
> > > > > > > Richardson, Bruce <bruce.richardson@intel.com>;
> > > > > > > dev@dpdk.org; Lu, Wenzhuo <wenzhuo.lu@intel.com>; Zhang,
> > > > > > > Helin <helin.zhang@intel.com>; ophirmu@mellanox.com
> > > > > > > Subject: Re: [dpdk-dev] [PATCH v8 1/3] eal/x86: run-time
> > > > > > > dispatch over memcpy
> > > > > > >
> > > > > > > 19/10/2017 09:51, Li, Xiaoyun:
> > > > > > > > From: Thomas Monjalon [mailto:thomas@monjalon.net]
> > > > > > > > > 19/10/2017 04:45, Li, Xiaoyun:
> > > > > > > > > > Hi
> > > > > > > > > > > > >
> > > > > > > > > > > > > The significant change of this patch is to call
> > > > > > > > > > > > > a function pointer for packet size > 128
> > > > (RTE_X86_MEMCPY_THRESH).
> > > > > > > > > > > > The perf drop is due to function call replacing inline.
> > > > > > > > > > > >
> > > > > > > > > > > > > Please could you provide some benchmark numbers?
> > > > > > > > > > > > I ran memcpy_perf_test which would show the time
> > > > > > > > > > > > cost of memcpy. I ran it on broadwell with sse and avx2.
> > > > > > > > > > > > But I just draw pictures and looked at the trend
> > > > > > > > > > > > not computed the exact percentage. Sorry about that.
> > > > > > > > > > > > The picture shows results of copy size of 2, 4, 6,
> > > > > > > > > > > > 8, 9, 12, 16, 32, 64, 128, 192, 256, 320, 384,
> > > > > > > > > > > > 448, 512, 768, 1024, 1518, 1522, 1536, 1600, 2048,
> > > > > > > > > > > > 2560, 3072, 3584, 4096, 4608, 5120, 5632, 6144,
> > > > > > > > > > > > 6656, 7168,
> > > > > > > > > > > 7680, 8192.
> > > > > > > > > > > > In my test, the size grows, the drop degrades.
> > > > > > > > > > > > (Using copy time indicates the
> > > > > > > > > > > > perf.) From the trend picture, when the size is
> > > > > > > > > > > > smaller than
> > > > > > > > > > > > 128 bytes, the perf drops a lot, almost 50%. And
> > > > > > > > > > > > above
> > > > > > > > > > > > 128 bytes, it approaches the original dpdk.
> > > > > > > > > > > > I computed it right now, it shows that when
> > > > > > > > > > > > greater than
> > > > > > > > > > > > 128 bytes and smaller than 1024 bytes, the perf
> > > > > > > > > > > > drops about
> > > > 15%.
> > > > > > > > > > > > When above
> > > > > > > > > > > > 1024 bytes, the perf drops about 4%.
> > > > > > > > > > > >
> > > > > > > > > > > > > From a test done at Mellanox, there might be a
> > > > > > > > > > > > > performance degradation of about 15% in testpmd
> > > > > > > > > > > > > txonly
> > > > with AVX2.
> > > > > > > > > > >
> > > > > > > > > >
> > > > > > > > > > I did tests on X710, XXV710, X540 and MT27710 but
> > > > > > > > > > didn't see
> > > > > > > > > performance degradation.
> > > > > > > > > >
> > > > > > > > > > I used command
> > > > > > > > > > "./x86_64-native-linuxapp-gcc/app/testpmd
> > > > > > > > > > -c 0xf -n
> > > > > > > > > > 4 -- -
> > > > > > > > > I" and set fwd txonly.
> > > > > > > > > > I tested it on v17.11-rc1, then revert my patch and
> > > > > > > > > > tested it
> > > again.
> > > > > > > > > > Show port stats all and see the throughput pps. But
> > > > > > > > > > the results are similar
> > > > > > > > > and no drop.
> > > > > > > > > >
> > > > > > > > > > Did I miss something?
> > > > > > > > >
> > > > > > > > > I do not understand. Yesterday you confirmed a 15% drop
> > > > > > > > > with buffers between
> > > > > > > > > 128 and 1024 bytes.
> > > > > > > > > But you do not see this drop in your txonly tests, right?
> > > > > > > > >
> > > > > > > > Yes. The drop is using test.
> > > > > > > > Using command "make test -j" and then " ./build/app/test -c f -n
> 4 "
> > > > > > > > Then run "memcpy_perf_autotest"
> > > > > > > > The results are the cycles that memory copy costs.
> > > > > > > > But I just use it to show the trend because I heard that
> > > > > > > > it's not
> > > > > > > recommended to use micro benchmarks like test_memcpy_perf
> > > > > > > for memcpy performance report as they aren't likely able to
> > > > > > > reflect performance of real world applications.
> > > > > > >
> > > > > > > Yes real applications can hide the memcpy cost.
> > > > > > > Sometimes, the cost appear for real :)
> > > > > > >
> > > > > > > > Details can be seen at
> > > > > > > > https://software.intel.com/en-us/articles/performance-opti
> > > > > > > > miza
> > > > > > > > ti
> > > > > > > > on-of-
> > > > > > > > memcpy-in-dpdk
> > > > > > > >
> > > > > > > > And I didn't see drop in testpmd txonly test. Maybe it's
> > > > > > > > because not a lot
> > > > > > > memcpy calls.
> > > > > > >
> > > > > > > It has been seen in a mlx4 use-case using more memcpy.
> > > > > > > I think 15% in micro-benchmark is too much.
> > > > > > > What can we do? Raise the threshold?
> > > > > > >
> > > > > > I think so. If there is big drop, can try raise the threshold. Maybe
> 1024?
> > > > but not sure.
> > > > > > But I didn't reproduce the 15% drop on mellanox and not sure
> > > > > > how to
> > > > verify it.
> > > > >
> > > > > I think we should focus on micro-benchmark and find a
> > > > > reasonnable threshold for a reasonnable drop tradeoff.
> > > > >
> > > > Sadly, it may not be that simple. What shows best performance for
> > > > micro- benchmarks may not show the same effect in a real application.
> > > >
> > > > /Bruce
> > >
> > > Then how to measure the performance?
> > >
> > > And I cannot reproduce 15% drop on mellanox.
> > > Could the person who tested 15% drop help to do test again with 1024
> > > threshold and see if there is any improvement?
> >
> > As Bruce said, best performance on micro-benchmark may not show the
> same effect in real applications.
> > And I cannot reproduce the 15% drop.
> > And I don't know if raising the threshold can improve the perf or not.
> > Could the person who tested 15% drop help to do test again with 1024
> threshold and see if there is any improvement?
> 
> As I already asked before - why not to make that threshold dynamic?
> Konstantin
> 
I want to confirm that raising threshold is useful. Then can make it dynamic and set it very large as default.

> >
> > Best Regards
> > Xiaoyun Li
> >
> >