Updated this morning, played Titan Quest 2 for a couple of hours. Normally, I’m right at the max VRAM allocated (8 GB), with temps in the mid to upper 70s (major encounters, usually).
This session: 500 MB less VRAM used, temps in the mid to upper 60s, absolutely no lag/micro-stuttering. This has been the best single-update performance boost I’ve ever experienced. Between this and the switch to the linux-xanmod kernel (thanks, @UncleSpellbinder !), I could not be happier.
Very possible, yes. I am getting a few suspect journal entries (warnings, mostly). But these haven’t had any negative effects (happens only when I’m about to run out of NVRAM):
Sep 14 16:17:43 kernel: NVRM: intermapRegisterDmaMapping: Failed to insert new mapping node for range 0xDF7D820000-0xDF7D8AFFFF!
Sep 14 16:17:43 kernel: NVRM: GPU0 nvAssertFailedNoLog: Assertion failed: NV_OK == status @ gpu_vaspace.c:4586
Sep 14 16:17:43 kernel: NVRM: GPU0 nvAssertFailedNoLog: Assertion failed: NV_OK == status @ gpu_vaspace.c:2080
Sep 14 16:17:43 kernel: NVRM: nvCheckFailedNoLog: Check failed: 0 == (pMapNode->gpuMask & gpuMask) @ gpu_vaspace.c:4503
Sep 14 16:17:43 kernel: NVRM: nvCheckOkFailedNoLog: Check failed: Invalid argument to call [NV_ERR_INVALID_ARGUMENT] (0x0000001F) returned from _gvaspaceMappingInsert(pGVAS, pGpu, pVASBlock, vaLo, vaHi, flags) @ gpu_vaspace.c:2036
Sep 14 16:17:43 kernel: NVRM: nvCheckFailedNoLog: Check failed: NV_OK == status @ virt_mem_allocator_gm107.c:2563
Sep 14 16:17:43 kernel: NVRM: dmaAllocMapping_GM107: can't update VA space for mapping @vaddr=0x1d4000000
Sep 14 16:26:47 kernel: [drm:nv_drm_gem_alloc_nvkms_memory_ioctl [nvidia_drm]] *ERROR* [nvidia-drm] [GPU ID 0x00000100] Failed to allocate NVKMS memory for GEM object
