ywkang
32b29a1e5c
NOC topology is an implementation choice (mesh, ring, crossbar, etc.). ADR-0017 covers the current 2D mesh choice; ADRs at the system-level shouldn't bind to that specific implementation. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
69 lines
2.4 KiB
Markdown
69 lines
2.4 KiB
Markdown
# ADR-0003: Target System Hierarchy & Modeling Scope
|
||
|
||
## Status
|
||
|
||
Accepted
|
||
|
||
## Context
|
||
|
||
We need a system-level simulator to evaluate LLM kernel performance on our AI Accelerator platform.
|
||
The platform is organized as a compute tray containing multiple identical SIPs connected via PCIe or UAL
|
||
through switching fabrics, with a host CPU issuing commands/kernels.
|
||
|
||
## Decision
|
||
|
||
We model the system hierarchy explicitly:
|
||
|
||
### D1. Tray-level
|
||
|
||
- A compute tray contains:
|
||
- Host CPU (issues requests / coordinates runtime & data placement)
|
||
- Multiple identical SIPs (accelerators)
|
||
- Interconnect fabric between SIPs (PCIe and/or UAL via switches)
|
||
|
||
### D2. SIP-level
|
||
|
||
- A SIP is a multi-die package composed of:
|
||
- Multiple CUBEs (HBM die + compute PEs + UCIe)
|
||
- One or more IO chiplets (host/SIP interfaces)
|
||
- IO chiplets:
|
||
- provide interfaces: PCIe-EP, IO_CPU, optionally UAL-EP
|
||
- can be multiple per SIP
|
||
- placement constrained to SIP shoreline (top/bottom/left/right); each shoreline may host 1–2 IO chiplets
|
||
|
||
### D3. CUBE-level
|
||
|
||
- A CUBE contains:
|
||
- HBM + memory controller (HBM_CTRL)
|
||
- NOC (on-die fabric): carries all intra-cube traffic including HBM data,
|
||
inter-cube (UCIe), command (M_CPU↔PE_CPU), and shared SRAM access.
|
||
Must provide: full-BW PE↔local HBM path, PE↔SRAM connectivity,
|
||
PE↔UCIe connectivity, M_CPU↔PE command path.
|
||
NOC topology is an implementation choice (e.g., 2D mesh, ring, crossbar);
|
||
current implementation uses a 2D mesh with XY routing (see ADR-0017).
|
||
HBM_CTRL is attached to each PE's local NOC port (local HBM = minimal hop).
|
||
- Shared SRAM: cube-level shared memory accessible by all PEs via NOC
|
||
- management/control CPU (M_CPU) coordinating PE command distribution and completion aggregation
|
||
- multiple PEs
|
||
- up to 4 UCIe endpoints (N/E/W/S) for CUBE↔CUBE and CUBE↔IO connectivity
|
||
|
||
### D4. PE-level
|
||
|
||
- A PE can execute one kernel instance
|
||
- PE contains internal control + accelerators (modeled at PE view granularity):
|
||
- PE_CPU, command handler, PE_TCM, DMA/GEMM/MATH engines, internal queues
|
||
|
||
## Consequences
|
||
|
||
- The simulator supports abstraction by “views”:
|
||
- SIP view hides PE internals
|
||
- CUBE view treats each PE as a single block
|
||
- PE view expands PE internals
|
||
- Topology remains parameterized; sizes/counts/links come from configuration.
|
||
|
||
## Links
|
||
|
||
- SPEC R3/R5
|
||
- ADR-0005 (diagram views)
|
||
- ADR-0017 (cube NOC 2D mesh architecture)
|