Am 15.08.2017 um 11:53 schrieb He, Roger: > [SNIP] > Don't adjust the global parameter here. > > Better use the default handling to amdgpu_vm_adjust_size() like you had in the last version of the patch. > > For Vega10 fragment handling works a bit differently, the L1 works like > GFX8 but the L2 is currently fixed to 2M. > > So different fragment sizes still make sense for Vega10, you should just not change the hardware setup of the L2. > [Roger]: which register do you mean here? The VM_L2_CNTL* registers. They work differently on Vega10 compared to earlier generations. Christian. > > Thanks > Roger(Hongbo.He)
