nice box, how fast does your cpu run now and which cooler do you use?
Could it have anything to do with the cold boot?
I'm having a hard time overclocking this chip on Linux, I have to disable speedstep otherwise it seems the turbo never kicks in, and overclocking on Sandy Bridge-E is done through the Turbo! Since a /proc/cpuinfo | grep MHz returns 3299.992 no matter what I do with speedstep disabled I'm basically having to refer to Windows overclock temps to see what is happening even though it's the same overclock - http://i.imgur.com/5UAXz.png
On the Windows OC I had speedstep and all C States enabled, but no power management through Windows. This let the chip idle @ 1.2GHz through C State and on load all cores were at 100x42 for a 4.2GHz overclock.
On Linux, the same settings stopped the turbo kicking in at all, as my cpuinfo grep always returns 3299.992 I had to use temps to see what was happening. With speedstep enabled and without having any frequency scaling set up in Linux (cpufreq/cpupower) I was getting temps a good 15C lower which seemed to indicate a lack of turbo. By disabling speeedstep I saw my temps correlate to the Windows overclock. Whether I'm Idling @ 1.2GHz within Linux is anybody's guess, I've absolutely no way of knowing. but it would be nice :/
Any news on this bug?
I am planning to purchase the processor discussed here (Intel Core i7 3960X "Sandy Bridge" E) and would like to run Ubuntu 12.04, but I am really worried about the bug. Does anybody know whether this bug has been confirmed or resolved?
Thanks a lot!
Originally Posted by phoronix
I had some serious serious bios issues with the 3930k system I built. Back in february/march I finally got a bios which allowed the system to boot cleanly every single time. I'm guessing the problems are related to ram compatibility with the memory controller.
thanks for your reply. Do you think that the new bios resolved the issues, so that now it should be safe to buy such a system?
The reason that I am asking is that I found a thread on the Intel forum which discusses a similar issue ( http://communities.intel.com/message/162695#162695 ). There, the problem seems to have been solved when a new bios update came out.
Originally Posted by bnolsen