GPU Design · All levels

Instruction Issue & Scoreboard

Shader Multiprocessor & Core Pipeline: Scoreboards track data hazards and memory readiness so schedulers issue only safe instructions; scoreboard pressure directly limits ILP extraction.

What this topic teaches

Instruction Issue & Scoreboard converts GPU architecture concepts into review-ready engineering decisions. Scoreboards track data hazards and memory readiness so schedulers issue only safe instructions; scoreboard pressure directly limits ILP extraction. The practical goal is to tie counters and traces to a specific mechanism, owner, and closure action.

The senior-engineer question

When issue slot utilization, dependency stall ratio, and scoreboard wait depth shifts, can you prove whether the root cause is SIMT control flow, SM scheduling, memory traffic, interconnect pressure, or graphics stage imbalance?

diagram

SIMT EXECUTION — Instruction Issue & Scoreboard

warp 0 lanes:  0 1 2 3 4 5 6 7 ... 31
active mask :  1 1 1 1 0 0 1 1 ...  1
instruction :  IF branch taken on active lanes

cycle 10: issue warp 0
cycle 11: issue warp 3
cycle 12: warp 0 reconverges

Focus: anchor discussion in lane masks and warp progress
Metric tracked: issue slot utilization, dependency stall ratio, and scoreboard wait depth

Picture the architecture

Begin with an architecture sketch before touching tuning knobs. These diagrams are for design reviews, interview whiteboards, and closure discussions.

Issue slots vs dependency stalls

diagram

WARP SCHEDULER VIEW — Instruction Issue & Scoreboard

cycle ->      0    1    2    3    4
eligible   [W1,W2,W5] [W2] [W2,W7] [W7] [W3,W7]
issued         W1      W2    W7      W7    W3
stall reason    -    dep wait  -   mem wait  -

Scheduler objective: keep issue slots non-empty.
Focus: map scoreboard wait reasons to lost issue opportunities

Scoreboard stall triage

diagram

ROOT-CAUSE TREE — Instruction Issue & Scoreboard

issue slot utilization, dependency stall ratio, and scoreboard wait depth regressed
        |
  reproducible on replay?
      /              \
    no                yes
    |                  |
env/test noise    counter triage
                   |
             compute-bound or memory-bound?
                /                  \
             compute            memory/interconnect
             issue stalls       cache/NoC/DRAM stalls

Stop at first failing mechanism, then patch.

SM and datapath context

diagram

SM BLOCK DIAGRAM — Instruction Issue & Scoreboard

        +---------------------------+
        | Warp Schedulers / Dispatch|
        +------------+--------------+
                     |
     +---------------+----------------+
     |  Register File / Operand Cross |
     +--------+---------------+-------+
              |               |
           [ALU/FPU]       [LD/ST]
              |               |
              +-------+-------+
                      |
                L1 / Shared Mem

Focus: front-end to execute dataflow

Memory hierarchy context

diagram

GPU MEMORY HIERARCHY — Instruction Issue & Scoreboard

                [ Registers ]
              latency:   1-2 cycles
                     |
                [ Shared/L1 ]
              latency:  20-40 cycles
                     |
                    [ L2 ]
              latency: 150-250 cycles
                     |
             [ HBM/GDDR VRAM ]
              latency: 300ns+ effective

Optimization lens: capacity vs latency

Scheduler context

diagram

WARP SCHEDULER VIEW — Instruction Issue & Scoreboard

cycle ->      0    1    2    3    4
eligible   [W1,W2,W5] [W2] [W2,W7] [W7] [W3,W7]
issued         W1      W2    W7      W7    W3
stall reason    -    dep wait  -   mem wait  -

Scheduler objective: keep issue slots non-empty.
Focus: eligible warp quality

Ownership layers

diagram

GPU OWNERSHIP LAYERS — Instruction Issue & Scoreboard

artifact area     owner
----------------  ----------------------------
architecture    SM RTL owner
RTL/microarch   verification owner
software/tools  performance lead

Rule: each metric needs a named owner before signoff.

Evidence to collect

Primary metric: issue slot utilization, dependency stall ratio, and scoreboard wait depth.
Primary artifact: scoreboard state timeline, issue reason breakdown, and stall attribution report.
Owners to include: SM RTL owner, verification owner, performance lead.
One reproducible failing workload and one stable comparator workload.
One counter capture that separates compute issue from memory/interconnect pressure.

Roofline lens

diagram

BANDWIDTH ROOFLINE — Instruction Issue & Scoreboard

performance
   ^
   |                compute ceiling
   |               /
   |              /
   |-------------/------------------ memory ceiling
   +------------------------------------------> operational intensity
      memory-bound             compute-bound

Interpretation: identify compute vs memory bound

Coalescing lens

diagram

COALESCING PATTERN — Instruction Issue & Scoreboard

WARP ADDRESSES
lane: 0 1 2 3 4 5 6 7
addr: 0 4 8 C 10 14 18 1C    -> contiguous -> 1 transaction segment

lane: 0 1 2 3 4 5 6 7
addr: 0 40 8 48 10 50 18 58  -> strided/scatter -> many segments

Effect: fewer coalesced segments => better bandwidth efficiency.
Focus: transaction inflation from scatter

Subpages in this topic

Each topic includes mechanism, inputs/outputs, reports, debug, worked example, pitfalls, interview, checklist, theory deep dive, design space, case study, walkthrough, matrix, software view, and silicon impact.

Key takeaways

Always connect warp behavior to measured counters before proposing fixes.
Treat memory transaction quality as equal priority to compute utilization.
Close decisions with explicit owners and reproducible benchmark evidence.

Common pitfalls

Copying tuning patterns from unrelated workloads or scenes.
Using occupancy as a success metric without stall classification.
Declaring closure without end-to-end frame or kernel validation.

GPU deep dive

Shader-core throughput is gated by issue policy, register-bank access, and pipeline hazard behavior.

Concept diagram

diagram

SM CORE LOOP

warp schedulers -> issue ports -> ALU/FPU/Tensor pipelines
scoreboard + register file gate progress

Metric graph

diagram

SM BOTTLENECK MIX

dependency stalls   ███████
bank conflicts      ████
pipeline bubbles    ███

Reports and artifacts

SM IPC dashboard
issue stall taxonomy
register-bank conflict log
shader unit utilization

Mini case study

Compiler register allocation shifted operand banking, doubling RF conflicts and causing a 14% shader regression.

Debug branches

Inspect scoreboard wait-depth trends
Track RF conflicts by instruction class
Separate front-end issue loss from backend saturation

Senior review question

Ask: which metric and benchmark pairing proves this topic is truly closed in production context?

Key takeaways

Always pair micro-kernel metrics with end-to-end workload impact.
Lock toolchain, driver, and launch metadata before comparing performance results.

Common pitfalls

Optimizing occupancy without checking memory-system saturation.
Comparing profiler captures from different driver or compiler builds.
Declaring wins without reproducible accuracy and performance gates.

Practice this lesson