Created attachment 105545 [details]
dump from /sys/class/drm/card0/error

It's happening from time to time.
The system frezees for a while and then resume (the WM is misbehaving after this, but this might be another issue).
The messages I found the in the logs are:

Sep  1 11:08:16 plambri kernel: [11194.320270] [drm] stuck on render ring
Sep  1 11:08:16 plambri kernel: [11194.320749] [drm] GPU HANG: ecode 0:0xf5f7fffe, in X [1308], reason: Ring hung, action: reset
Sep  1 11:08:16 plambri kernel: [11194.320752] [drm] GPU hangs can indicate a bug anywhere in the entire gfx stack, including userspace.
Sep  1 11:08:16 plambri kernel: [11194.320754] [drm] Please file a _new_ bug report on bugs.freedesktop.org against DRI -> DRM/Intel
Sep  1 11:08:16 plambri kernel: [11194.320759] [drm] drm/i915 developers can then reassign to the right component if it's not a kernel issue.
Sep  1 11:08:16 plambri kernel: [11194.320760] [drm] The gpu crash dump is required to analyze gpu hangs, so please always attach it.
Sep  1 11:08:16 plambri kernel: [11194.320762] [drm] GPU crash dump saved to /sys/class/drm/card0/error
Sep  1 11:08:16 plambri kernel: [drm] stuck on render ring
Sep  1 11:08:16 plambri kernel: [drm] GPU HANG: ecode 0:0xf5f7fffe, in X [1308], reason: Ring hung, action: reset
Sep  1 11:08:16 plambri kernel: [drm] GPU hangs can indicate a bug anywhere in the entire gfx stack, including userspace.
Sep  1 11:08:16 plambri kernel: [drm] Please file a _new_ bug report on bugs.freedesktop.org against DRI -> DRM/Intel
Sep  1 11:08:16 plambri kernel: [drm] drm/i915 developers can then reassign to the right component if it's not a kernel issue.
Sep  1 11:08:16 plambri kernel: [drm] The gpu crash dump is required to analyze gpu hangs, so please always attach it.
Sep  1 11:08:16 plambri kernel: [drm] GPU crash dump saved to /sys/class/drm/card0/error
Sep  1 11:08:18 plambri kernel: [11196.320804] [drm] Enabling RC6 states: RC6 on, RC6p on, RC6pp off
Sep  1 11:08:18 plambri kernel: [drm] Enabling RC6 states: RC6 on, RC6p on, RC6pp off

I'm using fedora with these packages:

xorg-x11-server-Xorg-1.14.4-11.fc20.x86_64
xorg-x11-drv-intel-2.21.15-7.fc20.x86_64

kernel-3.15.10-201.fc20.x86_64

I'll attach the dump of the card0/error