<?xml version="1.0" encoding="UTF-8"?><rss version="2.0"
	xmlns:content="http://purl.org/rss/1.0/modules/content/"
	xmlns:dc="http://purl.org/dc/elements/1.1/"
	xmlns:atom="http://www.w3.org/2005/Atom"
	xmlns:sy="http://purl.org/rss/1.0/modules/syndication/"
		>
<channel>
	<title>Comments on: Ultrascale GTH transceivers: Advanced doesn&#8217;t necessarily mean better</title>
	<atom:link href="http://billauer.se/blog/2020/09/xilinx-gth-gtx-gtp-eye-scan/feed/" rel="self" type="application/rss+xml" />
	<link>https://billauer.se/blog/2020/09/xilinx-gth-gtx-gtp-eye-scan/</link>
	<description>Anything I found worthy to write down.</description>
	<lastBuildDate>Thu, 26 Mar 2026 13:15:15 +0000</lastBuildDate>
	<sy:updatePeriod>hourly</sy:updatePeriod>
	<sy:updateFrequency>1</sy:updateFrequency>
	<generator>http://wordpress.org/?v=3.1.2</generator>
	<item>
		<title>By: Daniels Kurgs</title>
		<link>https://billauer.se/blog/2020/09/xilinx-gth-gtx-gtp-eye-scan/comment-page-1/#comment-1535</link>
		<dc:creator>Daniels Kurgs</dc:creator>
		<pubDate>Wed, 24 Feb 2021 16:36:24 +0000</pubDate>
		<guid isPermaLink="false">https://billauer.se/blog/?p=6136#comment-1535</guid>
		<description>We&#039;re just now also in the situation, where the LPM equalizer shows better performance than the DFE on Ultrascale&#039;s GTH.

Eli, or maybe Evgeni, do you perhaps have any idea if we could override 10 out of the 11 DFE taps to 0? The idea is to try DisplayPort&#039;s reference DFE, which has just one tap.

Of course, Xilinx offers no support in modifying the DFE in their official documentation.</description>
		<content:encoded><![CDATA[<p>We&#8217;re just now also in the situation, where the LPM equalizer shows better performance than the DFE on Ultrascale&#8217;s GTH.</p>
<p>Eli, or maybe Evgeni, do you perhaps have any idea if we could override 10 out of the 11 DFE taps to 0? The idea is to try DisplayPort&#8217;s reference DFE, which has just one tap.</p>
<p>Of course, Xilinx offers no support in modifying the DFE in their official documentation.</p>
]]></content:encoded>
	</item>
	<item>
		<title>By: Evgeni Stavinov</title>
		<link>https://billauer.se/blog/2020/09/xilinx-gth-gtx-gtp-eye-scan/comment-page-1/#comment-1521</link>
		<dc:creator>Evgeni Stavinov</dc:creator>
		<pubDate>Thu, 10 Sep 2020 04:54:25 +0000</pubDate>
		<guid isPermaLink="false">https://billauer.se/blog/?p=6136#comment-1521</guid>
		<description>I&#039;ve been spending significant amount of time dealing with Xilinx Ultrascale transceivers, and squeezing every bit of performance out of them. Specifically, for supporting latest USB 4, PCIe Gen5, and 200/400 GbE specs. Whatever settings transceiver IP wizard provides are not going to achieve the best possible performance, that&#039;s for sure. You&#039;d need to work with your FAE to get a few settings for RXDFE_GC_CFG2 and other undocumented registers that might work better for your specific channel and SSC clock. I even found experimentally that completely turning off automatic DFE adaptation and running custom algorithm (reading DFE and gain stats from Digital Monitor interface) did improve performance.</description>
		<content:encoded><![CDATA[<p>I&#8217;ve been spending significant amount of time dealing with Xilinx Ultrascale transceivers, and squeezing every bit of performance out of them. Specifically, for supporting latest USB 4, PCIe Gen5, and 200/400 GbE specs. Whatever settings transceiver IP wizard provides are not going to achieve the best possible performance, that&#8217;s for sure. You&#8217;d need to work with your FAE to get a few settings for RXDFE_GC_CFG2 and other undocumented registers that might work better for your specific channel and SSC clock. I even found experimentally that completely turning off automatic DFE adaptation and running custom algorithm (reading DFE and gain stats from Digital Monitor interface) did improve performance.</p>
]]></content:encoded>
	</item>
</channel>
</rss>
