<html xmlns:o="urn:schemas-microsoft-com:office:office" xmlns:w="urn:schemas-microsoft-com:office:word" xmlns:m="http://schemas.microsoft.com/office/2004/12/omml" xmlns="http://www.w3.org/TR/REC-html40"> <head> <meta http-equiv="Content-Type" content="text/html; charset=us-ascii"> <meta name="Generator" content="Microsoft Word 15 (filtered medium)"> <style></style> </head> <body lang="en-IT" link="#0563C1" vlink="#954F72" style="word-wrap:break-word"> <div class="WordSection1"> <div> > From: Jerin Jacob <jerinjacobk@gmail.com><o:p></o:p> > Date: Tuesday, 16 November 2021 at 20:09<o:p></o:p> > To: Ferruh Yigit <ferruh.yigit@intel.com><o:p></o:p> > Cc: Elena Agostini <eagostini@nvidia.com>, dev@dpdk.org <dev@dpdk.org>, Slava Ovsiienko <viacheslavo@nvidia.com><o:p></o:p> > Subject: Re: [PATCH v2 1/1] app/testpmd: add GPU memory option in iofwd engine<o:p></o:p> > External email: Use caution opening links or attachments<o:p></o:p> > <o:p></o:p> > <o:p></o:p> > On Tue, Nov 16, 2021 at 11:42 PM Ferruh Yigit <ferruh.yigit@intel.com> wrote:<o:p></o:p> > ><o:p></o:p> > > On 11/16/2021 6:06 PM, Elena Agostini wrote:<o:p></o:p> > > > > From: Ferruh Yigit <ferruh.yigit@intel.com><o:p></o:p> > > ><o:p></o:p> > > > > Date: Tuesday, 16 November 2021 at 19:00<o:p></o:p> > > ><o:p></o:p> > > > > To: Elena Agostini <eagostini@nvidia.com><o:p></o:p> > > ><o:p></o:p> > > > > Cc: dev@dpdk.org <dev@dpdk.org>, Slava Ovsiienko <viacheslavo@nvidia.com><o:p></o:p> > > ><o:p></o:p> > > > > Subject: Re: [PATCH v2 1/1] app/testpmd: add GPU memory option in iofwd engine<o:p></o:p> > > ><o:p></o:p> > > > > External email: Use caution opening links or attachments><o:p></o:p> > > ><o:p></o:p> > > > ><o:p></o:p> > > ><o:p></o:p> > > > > On 11/11/2021 9:41 PM, eagostini@nvidia.com wrote:<o:p></o:p> > > ><o:p></o:p> > > > > > --- a/app/test-pmd/meson.build<o:p></o:p> > > ><o:p></o:p> > > > > > +++ b/app/test-pmd/meson.build<o:p></o:p> > > ><o:p></o:p> > > > > > @@ -32,7 +32,7 @@ if dpdk_conf.has('RTE_HAS_JANSSON')<o:p></o:p> > > ><o:p></o:p> > > > > > ext_deps += jansson_dep<o:p></o:p> > > ><o:p></o:p> > > > > > endif<o:p></o:p> > > ><o:p></o:p> > > > > ><o:p></o:p> > > ><o:p></o:p> > > > > > -deps += ['ethdev', 'gro', 'gso', 'cmdline', 'metrics', 'bus_pci']<o:p></o:p> > > ><o:p></o:p> > > > > > +deps += ['ethdev', 'gro', 'gso', 'cmdline', 'metrics', 'bus_pci', 'gpudev']><o:p></o:p> > > ><o:p></o:p> > > > > I didn't review the set, but in a very high level do we want to add<o:p></o:p> > > ><o:p></o:p> > > > > 'gpudev' as dependency? Isn't this like adding 'rawdev' as dependency.<o:p></o:p> > > ><o:p></o:p> > > > gpudev is a library that can be built without a gpu driver as all the other libraries<o:p></o:p> > > ><o:p></o:p> > > > and itis actually used only in case of GPU memory mempool.<o:p></o:p> > > ><o:p></o:p> > > > Reasons for this patch are:<o:p></o:p> > > ><o:p></o:p> > > > - Have an upstreamed benchmark tool to measure network metrics using GPU memory<o:p></o:p> > > ><o:p></o:p> > > > - Test some DPDK features not really tested anywhere like the external memory mempool feature<o:p></o:p> > > ><o:p></o:p> > ><o:p></o:p> > > I can see the reason, that is obvious, yet again why we are not adding rawdev<o:p></o:p> > > testing to the testpmd? But adding gpudev.<o:p></o:p> > > It is easier to add it to the testpmd, and for some testing perspective it<o:p></o:p> > > makes sense, but still I am not quite sure about this new dependency, I would<o:p></o:p> > > like to get more feedback.<o:p></o:p> > <o:p></o:p> > I had the similar concern earlier. IMO, It is better to have a<o:p></o:p> > separate test application for gpudev like<o:p></o:p> > other device classes. For eventdev cases when it needs to work with<o:p></o:p> > ethdev for Rx adapter cases,<o:p></o:p> > We have enabled such code in app/test-eventdev to make testpmd focus on ethdev.<o:p></o:p> <o:p> </o:p> gpudev already has a test app in app/test-gpudev.<o:p></o:p> <o:p> </o:p> gpudev needs to be also test with network card and today another application<o:p></o:p> decidated to test gpudev over the network would be very similar to testpmd io.<o:p></o:p> <o:p> </o:p> At this stage, there is no point in reinventing the wheel<o:p></o:p> </div> </div> </body> </html>