报告 #1591 跑分
已认证，设备已锁定 有效
- GPU
- Adreno (TM) 660
- 驱动
- turnip_faux123_v1.2.2 Mesa 26.3.0-devel (git-dd45fd0baa)
- 设备
- Samsung Galaxy S21+ 5G
- 芯片
- SM8350
- CPU
- 4 x 1.80 GHz + 3 x 2.42 GHz + 1 x 2.84 GHz, arm64-v8a
- 内存
- 7.2 GB
- 上报为
- samsung / SM-G996U1
- 收到时间
- Android
- 15
- 库文件 SHA-256
- cd5450905ed410ffa6987aef3023f18c5b0d0606f5bf6095a356dc513106fbe1
测试条件
- seconds_per_scene
- 6
- warmup_frames
- 30
- warmup_max_ms
- 2000
- warmup_min_frames
- 5
- render_width
- 1920
- render_height
- 1080
- heavy_width
- 3840
- heavy_height
- 2160
- heavy_below_ms
- 8
- heavy_particles
- 2
- cool_rule
- status_below_2_and_full_gpu_limit
- cooling_pauses
- 1
- cooling_s
- 15
- cooling_capped
- 0
- test_budget_s
- 180
- test_s
- 150
- system_log_unreadable
- False
- display_width
- 1080
- display_height
- 2400
- display_refresh_hz
- 120
- rounds
- 3
- scene_set
- 6
- hdr_format
- R16G16B16A16_SFLOAT
- vsync
- OFF
Every scene draws at 1920 by 1080, whatever the screen. A scene under 8.0 ms there is too light for the chip and runs again at a heavier load: at 3840 by 2160, as the row named with _2160p, or, for compute_nbody, with 2 times the particles, as the row named with _2x. hdr runs at the larger picture only. Compare drivers on one device; across devices, compare only rows that kept the GPU busy.
测试期间显卡芯片频率低于峰值的 90%（GPU 频率）：oit_linked_lists, compute_nbody。可能是设备压住了显卡芯片（发热或省电模式），也可能是场景太轻，用不着更高频率。这些数据只适合和频率相近的测试比较。
场景
| 场景 | 中位数 ms | 平均 ms | p95 ms | 轮次 | 开测时已发热 | GPU 频率 |
|---|---|---|---|---|---|---|
| oit_linked_lists | 29.64 | 29.56 | 29.86 | 3 | 0 | 752 / 840 MHz |
| subpasses | 5.83 | 5.89 | 6.86 | 3 | 0 | 830 / 840 MHz |
| compute_nbody | 12.33 | 12.19 | 14.16 | 3 | 0 | 625 / 840 MHz |
| pipeline_barriers | 3.56 | 3.61 | 4.10 | 3 | 0 | 819 / 840 MHz |
| subpasses_2160p | 22.55 | 22.33 | 23.87 | 3 | 0 | 799 / 840 MHz |
| pipeline_barriers_2160p | 13.70 | 13.64 | 14.37 | 3 | 0 | 768 / 840 MHz |
| hdr_2160p | 33.66 | 33.57 | 33.83 | 3 | 0 | 840 / 840 MHz |
单次运行（21）
| 场景 | 轮 | 状态 | 中位数 ms | 帧数 |
|---|---|---|---|---|
| oit_linked_lists | 0 | ok | 29.68 | 169 |
| subpasses | 0 | ok | 5.83 | 971 |
| compute_nbody | 0 | ok | 12.75 | 461 |
| pipeline_barriers | 0 | ok | 3.56 | 1594 |
| oit_linked_lists | 1 | ok | 29.64 | 168 |
| subpasses | 1 | ok | 6.14 | 940 |
| compute_nbody | 1 | ok | 13.52 | 425 |
| pipeline_barriers | 1 | ok | 3.65 | 1555 |
| oit_linked_lists | 2 | ok | 29.66 | 168 |
| subpasses | 2 | ok | 6.20 | 921 |
| compute_nbody | 2 | ok | 12.33 | 468 |
| pipeline_barriers | 2 | ok | 3.74 | 1541 |
| subpasses_2160p | 0 | ok | 22.55 | 239 |
| pipeline_barriers_2160p | 0 | ok | 13.70 | 389 |
| hdr_2160p | 0 | ok | 33.66 | 144 |
| subpasses_2160p | 1 | ok | 24.10 | 223 |
| pipeline_barriers_2160p | 1 | ok | 14.05 | 385 |
| hdr_2160p | 1 | ok | 33.72 | 144 |
| subpasses_2160p | 2 | ok | 24.19 | 218 |
| pipeline_barriers_2160p | 2 | ok | 14.06 | 375 |
| hdr_2160p | 2 | ok | 33.69 | 144 |