An update!
Added 2018-09-01 00:18:45 +0000 UTCHere’s a vloggish update for you. I *Just* got the 2990wx so I haven’t had a long time with it yet. Hope you enjoy the update. Going live this w/e!
Comments
Phoronix released docker image with a test suite, that should help: <a href="https://www.phoronix.com/scan.php?page=article&item=docker-phoronix-pts&num=1" rel="nofollow noopener" target="_blank">https://www.phoronix.com/scan.php?page=article&item=docker-phoronix-pts&num=1</a>
2018-09-02 20:21:40 +0000 UTCOne would hope but who knows this is new territory. Good server chip that actually cares about speed, idk how quickly they would want to move and margins on gaming ram is probably better
Mutation666
2018-09-01 07:45:16 +0000 UTCI found this RH Bugzilla thread that you were probably referring to in the video: <a href="https://bugzilla.redhat.com/show_bug.cgi?id=1608242" rel="nofollow noopener" target="_blank">https://bugzilla.redhat.com/show_bug.cgi?id=1608242</a> The consensus there seems to be to downgrade the BIOS to one with an older AGESA. Unfortunately my ASUS BIOS refuses to downgrade - it sees all previous versions that I've tried as being invalid. Ugh! This comment gives a work-around for the boot delay via adding a timeout to the udev-settle service unit: <a href="https://bugzilla.redhat.com/show_bug.cgi?id=1608242" rel="nofollow noopener" target="_blank">https://bugzilla.redhat.com/show_bug.cgi?id=1608242</a>#c19 I tried it out and it appears to work, although the shutdown delay is still present. This comment links to a patch to add timeouts to the relevant PSP calls in order to work around the issue in the kernel vs. waiting for AGESA fixes to be rolled out: <a href="https://bugzilla.redhat.com/show_bug.cgi?id=1608242" rel="nofollow noopener" target="_blank">https://bugzilla.redhat.com/show_bug.cgi?id=1608242</a>#c22 <a href="https://marc.info/?l=linux-crypto-vger&m=153436754612783&w=2" rel="nofollow noopener" target="_blank">https://marc.info/?l=linux-crypto-vger&m=153436754612783&w=2</a>
2018-09-01 04:14:39 +0000 UTCNo one cared. I can diy it. Samsung is doing registered 2933 I gotta figure unreg isn't far behind
Level1 Techs
2018-09-01 03:22:09 +0000 UTCIf you install an older ker el like 4.15 on f28 it'll boot fine. Just compile the kernel with the new AMD sp_psp set to n and it will also work fine. That's the corresponding kernel bit. Asrock and I are close to a bios update that disables sev via toggleable option so the kernel continues
Level1 Techs
2018-09-01 03:21:23 +0000 UTCThat bit about SEV might explain why I've had so many problems with Fedora 28 on my shiny new Threadripper hardware! Through much hair pulling and trial and error I found that I can install Fedora 27 and other distros such as Ubuntu, but I can't directly install Fedora 28, and if I install 27 and then upgrade to 28 boot/shutdown are buggy. I'm not seeing anything about SEV in my ASUS PRIME X399-A's BIOS, however I did find the 'mem_encrypt' kernel boot parameter which controls the AMD Secure Memory Encryption module. I'll try setting it to 'off' to see if that fixes the issues.
2018-09-01 02:40:04 +0000 UTCExcellent Vlog and news updates. Take as long as you need, from my point of view, love the setup you're looking to test in the DevOps system. Not got any configs myself, but your/viewers Docker experiences would/will be appreciated, as well as a fairly flexible virtualisation setup for the 128GB Linux tests.
2018-09-01 01:11:29 +0000 UTCWendell have you made any contacts at any ram companies? I know you did that poll for fast ECC, but haven't heard much past that.
Mutation666
2018-09-01 00:53:24 +0000 UTC
