Multi-tile Devices

Different vendors use the term “tile” a bit differently, but in the Intel world, a ‘tile’ is pretty close to what most people would think of as being a complete GPU. When multiple GPUs are placed behind a single PCI device, that’s what is referred to as a “multi-tile device.” In such cases, pretty much all hardware is replicated per-tile, although certain responsibilities like PCI communication, reporting of interrupts to the OS, etc. are handled solely by the “root tile.” A multi-tile platform takes care of tying the tiles together in a way such that interrupt notifications from remote tiles are forwarded to the root tile, the per-tile vram is combined into a single address space, etc.

In contrast, a “GT” (which officially stands for “Graphics Technology”) is the subset of a GPU/tile that is responsible for implementing graphics and/or media operations. The GT is where a lot of the driver implementation happens since it’s where the hardware engines, the execution units, and the GuC all reside.

Historically most Intel devices were single-tile devices that contained a single GT. PVC is an example of an Intel platform built on a multi-tile design (i.e., multiple GPUs behind a single PCI device); each PVC tile only has a single GT. In contrast, platforms like MTL that have separate chips for render and media IP are still only a single logical GPU, but the graphics and media IP blocks are each exposed as a separate GT within that single GPU. This is important from a software perspective because multi-GT platforms like MTL only replicate a subset of the GPU hardware and behave differently than multi-tile platforms like PVC where nearly everything is replicated.

Per-tile functionality (shared by all GTs within the tile):
  • Complete 4MB MMIO space (containing SGunit/SoC registers, GT registers, display registers, etc.)

  • Global GTT

  • VRAM (if discrete)

  • Interrupt flows

  • Migration context

  • kernel batchbuffer pool

  • Primary GT

  • Media GT (if media version >= 13)

Per-GT functionality:
  • GuC

  • Hardware engines

  • Programmable hardware units (subslices, EUs)

  • GSI subset of registers (multiple copies of these registers reside within the complete MMIO space provided by the tile, but at different offsets --- 0 for render, 0x380000 for media)

  • Multicast register steering

  • TLBs to cache page table translations

  • Reset capability

  • Low-level power management (e.g., C6)

  • Clock frequency

  • MOCS and PAT programming

Internal API

int xe_tile_alloc(struct xe_tile *tile)

Perform per-tile memory allocation

Parameters

struct xe_tile *tile

Tile to perform allocations for

Description

Allocates various per-tile data structures using DRM-managed allocations. Does not touch the hardware.

Returns -ENOMEM if allocations fail, otherwise 0.

int xe_tile_init_early(struct xe_tile *tile, struct xe_device *xe, u8 id)

Initialize the tile and primary GT

Parameters

struct xe_tile *tile

Tile to initialize

struct xe_device *xe

Parent Xe device

u8 id

Tile ID

Description

Initializes per-tile resources that don’t require any interactions with the hardware or any knowledge about the Graphics/Media IP version.

Return

0 on success, negative error code on error.

int xe_tile_init_noalloc(struct xe_tile *tile)

Init tile up to the point where allocations can happen.

Parameters

struct xe_tile *tile

The tile to initialize.

Description

This function prepares the tile to allow memory allocations to VRAM, but is not allowed to allocate memory itself. This state is useful for display readout, because the inherited display framebuffer will otherwise be overwritten as it is usually put at the start of VRAM.

Note that since this is tile initialization, it should not perform any GT-specific operations, and thus does not need to hold GT forcewake.

Return

0 on success, negative error code on error.