Question E5 1680 V2 problems (stock and OC)
Question E5 1680 V2 problems (stock and OC)
So I am having an unusual issue while trying to install a new CPU. I replaced a i7 3930K with a E5 1680 V2. While doing my benchmark suite I noticed some weird points of instability at stock that get exaggerated when an OC is attempted. To put it simply if I leave turbo frequencies enabled at stock settings the system will crash in a couple of unique situations. If I turn off my TV (ie my PC is connected to an Onkyo receiver then to a 4K TV) and turn it back on again it appears to crash the video card (currently not overclocked either) if and only if my CPU is under load while turboing higher then 3.1ghz. if there is no load or it in only at 3.1ghz or less the TV turns on fine and I still get a video feed. So as you can imagine any over clock above 3.1 ghz causes the same issue even though the OC is stable with multiple benchmarks/stress tests (save one, sandra). Now I can switch between inputs and all that and still have a video signal at any turbo frequency and CPU load. It is only when turning the TV off I crash completely.
Now for the other oddball. Sandra sisoft benchmark suite crashes the CPU regardless of settings. Stock or otherwise. At stock I can get it to "sometimes" pass but not all the time. Yet intel burn test on maximum for 5 hours is rock stable, same for prime 95 (8 hours), aida64(5 hours), furturemark, memtest86...to name a few whether I am at stock or up to 4.3ghz (haven't found a stable 4.4ghz setting yet).
So my point is what am I missing? Do I just have a bad CPU or is there a setting I am missing? listed below is my overclock settings...stock settings mentioned above are well stock. Both OC's listed have been done with and without ram OC if Ram is OC'd it is DDR3 2400 cl10 T1 with 1.65 volts at "stock" it is running DDR3 1333 CL 9 T2 1.5 volts. OC temps are ≤ 62C
sync all cores
core ratio 43
Vcore 1.21
VTT 1.15 (went as high as 1.18125)
VCCSA 1.050
CPU spread spectrum disabled
PCIe spread spectrum disabled
C states disabled
Speedstep disabled
LLC ultra high
CPU current capability 180%
VCCSA LLC High
VCCSA Current Capablilty 130%
Dram Current Capability 140%
Basically I followed this guide aiming for a 4.4 ghz...
X79 Overclocking Tutorial - i7-3930K? No problems? 1680 v2? Easy! - Vloggest
Here is the overclocking tutorial for X79, which oddly enough you guys were r...
vloggest.com
I also did my "own" semi stable OC and it looks more like this
sync all cores
core ratio 43
Vcore 1.21
VTT auto
VCCSA auto
CPU spread spectrum disabled
PCIe spread spectrum disabled
C states disabled
Speedstep disabled
LLC ultra high
CPU current capability 180%
VCCSA LLC High
VCCSA Current Capablilty 130%
Dram Current Capability 140%
Am I missing something or do I have a bum CPU??
So there are some new findings. More testing indicates that even at 3.0ghz (turbo off or both disabled) the TV can still cause crashes in the video card or PC. This happens less often at 3.0ghz. Additionally, using DDU revealed that the regular VGA driver doesn’t crash when turning off the TV, whether it’s on or off, though I noticed instability with other drivers for my GTX 1080. I’ve tried several NVIDIA drivers after running DDU between each one, and it seems multiple versions are causing issues. An Event ID 41 task in category 63 helped identify the problem in the event viewer. It could be a hardware issue, but so far it seems to be limited to the E5 1680 V2 and not my i7 3930K. I’m almost ready to reinstall Windows now. My i7 has remained stable up to 4.6ghz, which has kept me from taking that path.
So another update... I’ve further refined the situation. After checking my video settings, I discovered what was causing the crashes. It turned out HDR was enabled through Windows, which messed things up when I turned off and back on the TV. In Windows it’s disabled, but using 4:4:4 subsampling in the NVIDIA control panel—another method to turn on HDR—caused problems. Everything worked fine otherwise. So basically, "Windows HDR bad, Nvidia HDR good" to put it simply. I never expected this. Anyway, that leaves me with Sandra Sisoft, which I might have to accept but will keep trying updates for fixes. If anyone has solid HDR knowledge to help, I’d really appreciate it.
I've finally resolved my issues. Sandra Sisoft clearly dislikes Fastboot with my Xeon, though it works with the i7. I previously tested Fastboot both enabled and disabled in UEFI, but it remained active in Windows without my notice. After removing both UEFI and Windows versions, Sandra Sisoft ran smoothly. Based on this, I think I'm stable at 4.3ghz, assuming the earlier fixes were correct (excluding the bugs I resolved). I plan to run an Intel burn test with 20 maximum attempts, which should take just over five hours to confirm.
I achieved a consistent 4.3ghz performance, though it needed a boost to 1.3v initially. Temperatures stayed below 62-65°C (about 1.3V depending on the session) since it was warmer today. Sandra's software only required 1.3V on each other bench, and stability at 1.24V was maintained in most cases, though still slightly above previous levels. I plan to use a bit more voltage when temperatures rise to ensure everything remains stable.