<div dir="ltr">Yes, what is different here is that the MBUF size is communicated from the PMD to the hardware which *changes its behavior* of data motion to optimize throughput and latency as a function of that setting. And it does that per-queue. And can be done at runtime (that's the dynamic part). ... To the best our knowledge, other PMDs use this as a host-software setting only - and their DPDK naive DMA engines just use the same fixed settings (respecting PCIe, of course).<div><br></div><div>Hope that helps. If it is contentious in any way, we are fine with removing that line. We added it as users have remarked it is a unique capability they think we should point out.</div><div><br></div><div>-Shep</div><div><br></div></div><br><div class="gmail_quote"><div dir="ltr" class="gmail_attr">On Mon, Feb 13, 2023 at 12:23 PM Ferruh Yigit <<a href="mailto:ferruh.yigit@amd.com">ferruh.yigit@amd.com</a>> wrote:<br></div><blockquote class="gmail_quote" style="margin:0px 0px 0px 0.8ex;border-left:1px solid rgb(204,204,204);padding-left:1ex">On 2/13/2023 5:09 PM, Shepard Siegel wrote:<br>
> Hi Ferruh,<br>
> <br>
> Yes, there will probably be next versions in the future. If you don't<br>
> mind making the marker length adjustment, that would be great.<br>
> <br>
> Regarding MBUF (re)sizing - Arkville supports the ability to configure<br>
> or reconfigure the MBUF size used on a per-queue basis. This feature is<br>
> useful when the are conflicting motivations for using smaller/larger<br>
> MBUF sizes. For example, user can switch a queue to use a size best for<br>
> that queue's application workload.<br>
> <br>
<br>
Application can allocate multiple mempool with different sizes and set<br>
these to specific queues, this is same for all PMDs, is ark PMD doing<br>
something specific here? Or are you referring to something else?<br>
<br>
And what does 'dynamic' emphasis means here?<br>
<br>
<br>
> -Shep<br>
> <br>
> <br>
> On Mon, Feb 13, 2023 at 10:46 AM Ferruh Yigit <<a href="mailto:ferruh.yigit@amd.com" target="_blank">ferruh.yigit@amd.com</a><br>
> <mailto:<a href="mailto:ferruh.yigit@amd.com" target="_blank">ferruh.yigit@amd.com</a>>> wrote:<br>
> <br>
> On 2/13/2023 2:58 PM, Shepard Siegel wrote:<br>
> > Add detail for the existing Arkville configurations FX0 and FX1.<br>
> > Corrected minor errors of omission.<br>
> ><br>
> > Signed-off-by: Shepard Siegel <<a href="mailto:shepard.siegel@atomicrules.com" target="_blank">shepard.siegel@atomicrules.com</a><br>
> <mailto:<a href="mailto:shepard.siegel@atomicrules.com" target="_blank">shepard.siegel@atomicrules.com</a>>><br>
> > ---<br>
> > doc/guides/nics/ark.rst | 18 ++++++++++++++++++<br>
> > 1 file changed, 18 insertions(+)<br>
> ><br>
> > diff --git a/doc/guides/nics/ark.rst b/doc/guides/nics/ark.rst<br>
> > index ba00f14e80..edaa02dc96 100644<br>
> > --- a/doc/guides/nics/ark.rst<br>
> > +++ b/doc/guides/nics/ark.rst<br>
> > @@ -52,6 +52,10 @@ board. While specific capabilities such as<br>
> number of physical<br>
> > hardware queue-pairs are negotiated; the driver is designed to<br>
> > remain constant over a broad and extendable feature set.<br>
> > <br>
> > +* FPGA Vendors Supported: AMD/Xilinx and Intel<br>
> > +* Number of RX/TX Queue-Pairs: up to 128<br>
> > +* PCIe Endpoint Technology: Gen3, Gen4, Gen5<br>
> > +<br>
> > Intentionally, Arkville by itself DOES NOT provide common NIC<br>
> > capabilities such as offload or receive-side scaling (RSS).<br>
> > These capabilities would be viewed as a gate-level "tax" on<br>
> > @@ -303,6 +307,18 @@ ARK PMD supports the following Arkville RTL<br>
> PCIe instances including:<br>
> > * ``1d6c:101e`` - AR-ARKA-FX1 [Arkville 64B DPDK Data Mover for<br>
> Agilex R-Tile]<br>
> > * ``1d6c:101f`` - AR-TK242 [2x100GbE Packet Capture Device]<br>
> > <br>
> > +Arkville RTL Core Configurations<br>
> > +-------------------------------------<br>
> > +<br>
> <br>
> The title marker length (-) should be same as title length, can you<br>
> please fix if there will be next version, if not I can fix while<br>
> merging.<br>
> <br>
> <br>
> > +Arkville's RTL core may be configured by the user with different<br>
> > +datapath widths to balance throughput against FPGA logic area.<br>
> The ARK PMD<br>
> > +has introspection on the RTL core configuration and acts accordingly.<br>
> > +All Arkville configurations present identical RTL user-facing AXI<br>
> stream<br>
> > +interfaces for both AMD/Xilinx and Intel FPGAs.<br>
> > +<br>
> > +* ARK-FX0 - 256-bit 32B datapath (PCIe Gen3, Gen4)<br>
> > +* ARK-FX1 - 512-bit 64B datapath (PCIe Gen3, Gen4, Gen5)<br>
> > +<br>
> > DPDK and Arkville Firmware Versioning<br>
> > -------------------------------------<br>
> > <br>
> > @@ -334,6 +350,8 @@ Supported Features<br>
> > ------------------<br>
> > <br>
> > * Dynamic ARK PMD extensions<br>
> > +* Dynamic per-queue MBUF (re)sizing up to 32KB<br>
> <br>
> What is this feature? What does it mean to size/resize mbuf dynamically?<br>
> <br>
> > +* SR-IOV, VF-based queue-segregation<br>
> > * Multiple receive and transmit queues<br>
> > * Jumbo frames up to 9K<br>
> > * Hardware Statistics<br>
> <br>
<br>
</blockquote></div>