ROCm
PerfXpert
AI Performance Analysis
%%HEADER_BADGES%%
☀ Light
Runtime:
%%TOTAL_MS%% ms
Kernels:
%%NUM_HOTSPOTS%%
Tier:
%%TIER_LABEL%%
Generated:
%%ANALYSIS_DATE%%
%%DB_PILL%%
📊
Overview
Tier %%TIER%%
%%ASSESSMENT%%
%%BN_ICON%%
Bottleneck
Primary Bottleneck
%%BN_DISPLAY%%
Confidence: %%CONFIDENCE%%%
⏱
Duration
Total Runtime
%%TOTAL_MS_VALUE%%
milliseconds • %%NUM_HOTSPOTS%% kernels
💻
%%KPI_KERNEL_LBL%%
Kernel Execution
%%KERNEL_PCT%%
%%KERNEL_MS%% ms active compute
%%TIER_ICON%%
%%TIER_STATUS_LBL%%
Analysis Tier
%%TIER%%
%%HAS_COUNTERS_SUB%%
%%FINDINGS_HTML%%
⏰
Execution Breakdown
Kernel
%%KERNEL_PCT_1F%%%
Memory Copies
%%MEMCPY_PCT_1F%%%
API Overhead
%%OVERHEAD_PCT_1F%%%
GPU Idle
%%IDLE_PCT_1F%%%
Kernel Execution
%%KERNEL_PCT_1F%%%
%%KERNEL_MS%% ms
Memory Copies
%%MEMCPY_PCT_1F%%%
%%MEMCPY_MS%% ms
API Overhead
%%OVERHEAD_PCT_1F%%%
%%OVERHEAD_MS%% ms
GPU Idle
%%IDLE_PCT_1F%%%
%%IDLE_MS%% ms
%%HOTSPOTS_HTML%% %%MEM_HTML%%
🔬
Hardware Counters
%%HW_BADGE%%
%%HW_INNER%%
%%ROOFLINE_SECTION%%
💡
Optimization Recommendations
%%RECS_BADGE%%
%%RECS_HTML%%
⇧