I finally got the above system working, but am not totally thrilled. I have been using ASUS m/bs for
decades, but I thought I’d give GIGABYTE a try. Here are a few things I ran into that may
save you some headaches:
-
The RIPJAWS DRAM (128GB - 4x32GB) had lots of failures in memtest86 when I enabled
XMP (The M/B does not mention the AMD flavor, but this seems to work [now]). I eventually
realized that I was not running the latest BIOS, so installed that (F20a), and it now runs
clean as your basic whistle at 3200MHz. -
I managed to mangle some pins in the F_USB30 MB header. Oddly enough, as I was
cursing the MB , I was “watching”, with my third eye, this:https://www.youtube.com/watch?v=Wnptqz_177E
About 43 minutes into it, Linus stumbles onto exactly the same problem! Different MB,
different manufacturer. Different perp. I would probably have made things worse if
I tried to unbend the MB pins, but luckily, this MB has two of those headers, so I used
that. Now, the problem here is that this 20pin connector apparently managed to
become a “standard” along the way. Said connector makes a very poor connection, and
does not have any retention mechanism, so it tends to get loose with vibration or
cable “management”, which is my case was clearly cable “manglement”. This connector
wants redesign, badly. Maybe I worked with real computers too long… -
The F_PANEL header is horrible, and this is entirely, IMO, bad design on the part of GIGABYTE.
When I got the thing built, everything front-panel-ish was dead, so I checked connections.
I redid the switchboard at the fat end of the “front panel extension cable”, finding nothing wrong.
I unplugged that cable, and ran continuity tests on all the pins, on the assumption that I
had crushed something vital in the cable. I plugged it back in, to no avail. Shorting the
Power pins on the switchboard did nothing. Not good. Time to call GIGABYTE? Can’t return
the MB, thanks to some virus thing… Since I had traced the pinout on the cable
( conveniently only documented on the switchboard end by our friends at GIGABYTE), I knew
which MB pins were power. So, I pulled the cable again, and got suspicious. Well, looky there:
the extension cable can plug into the MB either way! In my country (Hogtown), we stupidproof
these things by plugging one socket hole, or doing something similar. Now, there are
wee, tiny ridges on the MB end of that cable, but the MB socket is, at this point, nearly
invisible and inaccessible. So, does the cable seat in the socket with the ridges facing
MB right? Yes. Does it power up? No. Does the cable seat with the ridges facing
MB left? Yes, my dear. Does it power up? Yup!! [Slight pause while I curse whoever “designed”
that cable and socket.] This made everything all the F_PANEL stuff work fine. -
The MB cheaped out on USB headers, so I have no place to plug in the PSU monitor.
I will have to get an internal USB hub or splitter, but that can wait until things open
up again here. -
I built this in a Fractal Design R6 case that comes with three case fans, and a NOCTUA
NH-D15S cpu cooler. The whole things runs extremely quietly, even w/open case,
but I have not been able to poke it very hard yet. Something about missing APEX and SAC
compilers. Looking at Linus’s build, I think that the Phanteks 600S case would be
a lot pleasanter to work in, and they are both $200CAD here.
can give all those threads some useful work.