Fwiw, that approvedmodems page is a load of crap. It says a lot without saying anything at all, ie: its written to look like the author knows what he or she is talking about, but the text is meaningless.
Not defending Intel here, or now MaxLinear, but, the Puma 7 is not a problematic Puma 6. While they might share the same architecture type, and who knows except for Intel, the Puma 7 modem does not have the same issue with internal task caused latency that the Puma 6 suffers or suffered from. For a typical consumer, if you ran a series of ping tests, testing ICMP and TCP/IP latency, and continuous DNS queries for UDP latency, you would be very hard pressed to tell the difference between the Intel Puma 7 and Broadcom BCM-3390 results.
You can force both modems into a higher latency situation by running a ping test and then starting a speedtest. The Puma 7 modem will end up with a much higher latency spike compared to the BCM-3390 modem. That's really the only way to see the difference, run a multi-protocol test to force a different outcome in performance. For anyone who has Linux type capability, run a Flent test to see how your modem behaves:
https://flent.org/index.html
Today, that multi-protocol performance is more important given that parents are working from home, kids are participating in remote classes and then of course there's the usual simultaneous gaming, streaming, remote backups, etc, etc.
The one real problem area is DOS performance. Intel has released updates for Puma 5, 6 and 7 modems for latency issues and DOS issues, but, no company, including Intel, the modem manufacturers or ISPs have come out with test plots which definitively prove that Intel has solved the DOS issue for all Puma chipsets. You can run your own DOS test by going to the site created by xymox1 and selecting Puma 5/6/7 test:
https://www.badmodems.com/
Intel has sold the Home Gateway Platform Division to MaxLinear, who makes the front end cable tuner for the Puma 6 and 7 modems, so, now this is all Maxlinear's problem. Where MaxLinear takes the Puma chipset from here will be interesting to see. My hope is that they completely redesign the chipset and end up with something that works without any of the previous problems. Time will tell.
So, given that you have a choice, the TC-4400 should be the modem of choice for the reasons stated above, not because its on the so called "approvedmodems list". Fwiw, here's the original list that I created almost four years ago, which the approvedmodems author appears to piggyback off of without any accreditation to that original source or to the badmodems site which was created by xymox1:
https://www.dslreports.com/forum/r31134484-
