Patrick - Congratulations ... googling "Tyan S5512WGM2NR" turns up servethehome.com as the first half dozen entries, well ahead of Tyan.com! Maybe google's page ranking reflects the amount of documentation and support that the two sites provide?
Anyway, here's a snapshot of where I am. I'd welcome any ideas/comments and especially success stories.
STATUS: Mine is on its way back to Egghead for an RMA swap. (The Egghead CSR assured me they'll send it back to Tyan rather than selling it to the next servethehome forum user as an Open Box item
I'll be happy to shoot pictures of its replacement when it arrives next week. I'm not losing faith in the product yet, at least partly because I don't see the features I want on any competing boards.
SYMPTOMS: As I mentioned before, I did get my board to power up eventually but it was intermittent. By the end, it was alive long enough to install ESXi, but the screen froze while it was running. I had no video about 2/3 of the time (Video was dead on both the physical VGA connector and the IPMI Java console.) During the periods when the board was alive, the IPMI web interface would not show sensor data but the console would, so I was able to see that the temperatures seemed acceptable, though surprisingly high. For example, the CPU was typically showing 50 deg C with no load and an oversized heat sink (Cooler Master Hyper 212+), and the LSI controller chip was running around 62 deg C despite the CPU fan pulling enough air over it that you could feel it with an unmoistened finger.
ATTEMPTED FIXES:
(1) Replace power supply. Twice, actually. Done on general principles when the board seemed to be totally dead, with no LED or fan activity.
(2) Remove everything but the Tyan board, RAM, IPMI LAN cable, and 4-pin CPU fan. Subjectively this seemed to help a bit, i.e. my impression is that it booted more often with no video or USB attached to the back panel.
(3) Reflash BIOS. As I mentioned earlier, Tyan posted an updated BIOS on their website. I know that reflashing was high-risk given the instability, but fortunately the board stayed alive for the entire process. Oddly, when I rebooted the old BIOS was still showing, but after clearing CMOS the replacement appeared. That's impossible unless Tyan retains a backup copy of the BIOS so that must be the case.
(2) Contemplate and pray over jumper and headers in the absence of documentation
My CMOS jumpers were definitely on pins 1-2, unfortunately ... that was one of the few things that the tiny content-free manual did show clearly. I'd welcome ideas from Patrick or anyone else passing through that knows about these things. Especially, are there any jumpers or headers that MUST be connected? I was doing this initial testing in a spare consumer-grade ATX case. It has no headers for a server status panel, and the only connections I made were for a power-on LED, the power switch, and the reset button. Eventually I removed even those, and just controlled it using the IPMI. So I'm implicitly assuming that all of the other control signals are non-essential and safely pulled up or down to default values that should let the board work. (The manual only provides the names of all those other signals and no clue about their meaning.)
So now I get to wait a week for a replacement. If the replacement is similarly dead, are there any other Socket 1155 boards out there that support Vt-d (For some bizarre reason Asus doesn't, right?) and that provide 14 SATA ports with great throughput for a couple hundred $$ total? Or should I switch strategies, e.g. back off to last-gen Supermicro or wait for Bulldozer?