Commit 386b81

2025-07-05 01:18:37 deseven: gaming benchmarks, formatting
Guides/Power-Modes-and-Performance.md ..
@@ 19,20 19,22 @@
| Metric | [55W](https://www.passmark.com/baselines/V11/display.php?id=509787657486) | [85W](https://www.passmark.com/baselines/V11/display.php?id=509787842837) | [120W](https://www.passmark.com/baselines/V11/display.php?id=509787902430) |
| ----------------------------------- | ------- | ------- | ------- |
- | **CPU Mark** | **49635** | **58042<sup style="color: #1fea00; font-size: 0.8em;">+16.9%</sup>** | **62805<sup style="color: #1fea00; font-size: 0.8em;">+8.2%</sup>** |
- | Integer Math (MOps/s) | 193409 | 222270<sup style="color: #1fea00; font-size: 0.8em;">+14.9%</sup> | 233483<sup style="color: #1fea00; font-size: 0.8em;">+5.0%</sup> |
- | Floating Point Math (MOps/s) | 107679 | 127114<sup style="color: #1fea00; font-size: 0.8em;">+18.0%</sup> | 138255<sup style="color: #1fea00; font-size: 0.8em;">+8.8%</sup> |
- | Find Prime Numbers (MPrimes/s) | 263 | 286<sup style="color: #1fea00; font-size: 0.8em;">+8.7%</sup> | 302<sup style="color: #1fea00; font-size: 0.8em;">+5.6%</sup> |
- | Random String Sorting (kStrings/s) | 71735 | 88253<sup style="color: #1fea00; font-size: 0.8em;">+23.0%</sup> | 98842<sup style="color: #1fea00; font-size: 0.8em;">+12.0%</sup> |
- | Data Encryption (MB/s) | 33582 | 40310<sup style="color: #1fea00; font-size: 0.8em;">+20.0%</sup> | 44246<sup style="color: #1fea00; font-size: 0.8em;">+9.8%</sup> |
- | Data Compression (KB/s) | 539857 | 683287<sup style="color: #1fea00; font-size: 0.8em;">+26.6%</sup> | 777048<sup style="color: #1fea00; font-size: 0.8em;">+13.7%</sup> |
- | Physics (Frames/s) | 4802 | 5477<sup style="color: #1fea00; font-size: 0.8em;">+14.0%</sup> | 5806<sup style="color: #1fea00; font-size: 0.8em;">+6.0%</sup> |
- | Extended Instructions (MMatrices/s) | 38464 | 49404<sup style="color: #1fea00; font-size: 0.8em;">+28.4%</sup> | 58761<sup style="color: #1fea00; font-size: 0.8em;">+18.9%</sup> |
+ | **CPU Mark** | **49 635** | **58 042<sup style="color: #1fea00; font-size: 0.8em;">+16.9%</sup>** | **62 805<sup style="color: #1fea00; font-size: 0.8em;">+26.5%</sup>** |
+ | Integer Math (MOps/s) | 193 409 | 222 270<sup style="color: #1fea00; font-size: 0.8em;">+14.9%</sup> | 233 483<sup style="color: #1fea00; font-size: 0.8em;">+20.7%</sup> |
+ | Floating Point Math (MOps/s) | 107 679 | 127 114<sup style="color: #1fea00; font-size: 0.8em;">+18.0%</sup> | 138 255<sup style="color: #1fea00; font-size: 0.8em;">+28.4%</sup> |
+ | Find Prime Numbers (MPrimes/s) | 263 | 286<sup style="color: #1fea00; font-size: 0.8em;">+8.7%</sup> | 302<sup style="color: #1fea00; font-size: 0.8em;">+14.8%</sup> |
+ | Random String Sorting (kStrings/s) | 71 735 | 88 253<sup style="color: #1fea00; font-size: 0.8em;">+23.0%</sup> | 98 842<sup style="color: #1fea00; font-size: 0.8em;">+37.8%</sup> |
+ | Data Encryption (MB/s) | 33 582 | 40 310<sup style="color: #1fea00; font-size: 0.8em;">+20.0%</sup> | 44 246<sup style="color: #1fea00; font-size: 0.8em;">+31.8%</sup> |
+ | Data Compression (KB/s) | 539 857 | 683 287<sup style="color: #1fea00; font-size: 0.8em;">+26.6%</sup> | 777 048<sup style="color: #1fea00; font-size: 0.8em;">+43.9%</sup> |
+ | Physics (Frames/s) | 4 802 | 5 477<sup style="color: #1fea00; font-size: 0.8em;">+14.1%</sup> | 5 806<sup style="color: #1fea00; font-size: 0.8em;">+20.9%</sup> |
+ | Extended Instructions (MMatrices/s) | 38 464 | 49 404<sup style="color: #1fea00; font-size: 0.8em;">+28.4%</sup> | 58 761<sup style="color: #1fea00; font-size: 0.8em;">+52.8%</sup> |
+
+ \* **All percentage values are relative to the base 55W result**
### CPU - Geekbench
:::info
# Notice
- Geekbench does NOT test the system under full load, a lot of the tests don't even utilize all of the available cores, consider these results to be a reflection of an average daily lite usage with no extreme load
+ Geekbench does NOT test the system under full load, a lot of the tests don't even utilize all of the available cores, consider these results to be a reflection of an average daily lite usage with no substantial load
:::
Details:
@@ 41,44 43,58 @@
| Metric | [55W](https://browser.geekbench.com/v6/cpu/12718643) | [85W](https://browser.geekbench.com/v6/cpu/12718729) | [120W](https://browser.geekbench.com/v6/cpu/12719008) |
| --------------------- | ------- | ------- | ------- |
- | **Multi-Core Score** | **17328** | **17544<sup style="color: #1fea00; font-size: 0.8em;">+1.2%</sup>** | **17516<sup style="color: red; font-size: 0.8em;">-0.2%</sup>** |
- | File Compression | 11184 | 12063<sup style="color: #1fea00; font-size: 0.8em;">+7.9%</sup> | 12197<sup style="color: #1fea00; font-size: 0.8em;">+1.1%</sup> |
- | Navigation | 13578 | 13050<sup style="color: red; font-size: 0.8em;">-3.9%</sup> | 13322<sup style="color: #1fea00; font-size: 0.8em;">+2.1%</sup> |
- | HTML5 Browser | 16030 | 17040<sup style="color: #1fea00; font-size: 0.8em;">+6.3%</sup> | 15610<sup style="color: red; font-size: 0.8em;">-8.4%</sup> |
- | PDF Renderer | 25044 | 24893<sup style="color: red; font-size: 0.8em;">-0.6%</sup> | 26726<sup style="color: #1fea00; font-size: 0.8em;">+7.4%</sup> |
- | Photo Library | 25989 | 27260<sup style="color: #1fea00; font-size: 0.8em;">+4.9%</sup> | 27099<sup style="color: red; font-size: 0.8em;">-0.6%</sup> |
- | Clang | 42312 | 44101<sup style="color: #1fea00; font-size: 0.8em;">+4.2%</sup> | 44315<sup style="color: #1fea00; font-size: 0.8em;">+0.5%</sup> |
- | Text Processing | 3723 | 3714<sup style="color: red; font-size: 0.8em;">-0.2%</sup> | 3667<sup style="color: red; font-size: 0.8em;">-1.3%</sup> |
- | Asset Compression | 30848 | 31534<sup style="color: #1fea00; font-size: 0.8em;">+2.2%</sup> | 32421<sup style="color: #1fea00; font-size: 0.8em;">+2.8%</sup> |
- | Object Detection | 13032 | 13075<sup style="color: #1fea00; font-size: 0.8em;">+0.3%</sup> | 12545<sup style="color: red; font-size: 0.8em;">-4.1%</sup> |
- | Background Blur | 12406 | 12308<sup style="color: red; font-size: 0.8em;">-0.8%</sup> | 12207<sup style="color: red; font-size: 0.8em;">-0.8%</sup> |
- | Horizon Detection | 20156 | 19947<sup style="color: red; font-size: 0.8em;">-1.0%</sup> | 19792<sup style="color: red; font-size: 0.8em;">-0.8%</sup> |
- | Object Remover | 18049 | 18382<sup style="color: #1fea00; font-size: 0.8em;">+1.8%</sup> | 18796<sup style="color: #1fea00; font-size: 0.8em;">+2.3%</sup> |
- | HDR | 13929 | 13775<sup style="color: red; font-size: 0.8em;">-1.1%</sup> | 13791<sup style="color: #1fea00; font-size: 0.8em;">+0.1%</sup> |
- | Photo Filter | 14335 | 13066<sup style="color: red; font-size: 0.8em;">-8.9%</sup> | 13849<sup style="color: #1fea00; font-size: 0.8em;">+6.0%</sup> |
- | Ray Tracer | 38300 | 40547<sup style="color: #1fea00; font-size: 0.8em;">+5.9%</sup> | 38962<sup style="color: red; font-size: 0.8em;">-3.9%</sup> |
- | Structure from Motion | 22188 | 22210<sup style="color: #1fea00; font-size: 0.8em;">+0.1%</sup> | 21352<sup style="color: red; font-size: 0.8em;">-3.9%</sup> |
+ | **Multi-Core Score** | **17 328** | **17 544<sup style="color: #1fea00; font-size: 0.8em;">+1.2%</sup>** | **17 516<sup style="color: #1fea00; font-size: 0.8em;">+1.1%</sup>** |
+ | File Compression | 11 184 | 12 063<sup style="color: #1fea00; font-size: 0.8em;">+7.9%</sup> | 12 197<sup style="color: #1fea00; font-size: 0.8em;">+9.1%</sup> |
+ | Navigation | 13 578 | 13 050<sup style="color: red; font-size: 0.8em;">-3.9%</sup> | 13 322<sup style="color: red; font-size: 0.8em;">-1.9%</sup> |
+ | HTML5 Browser | 16 030 | 17 040<sup style="color: #1fea00; font-size: 0.8em;">+6.3%</sup> | 15 610<sup style="color: red; font-size: 0.8em;">-2.6%</sup> |
+ | PDF Renderer | 25 044 | 24 893<sup style="color: red; font-size: 0.8em;">-0.6%</sup> | 26 726<sup style="color: #1fea00; font-size: 0.8em;">+6.7%</sup> |
+ | Photo Library | 25 989 | 27 260<sup style="color: #1fea00; font-size: 0.8em;">+4.9%</sup> | 27 099<sup style="color: #1fea00; font-size: 0.8em;">+4.3%</sup> |
+ | Clang | 42 312 | 44 101<sup style="color: #1fea00; font-size: 0.8em;">+4.2%</sup> | 44 315<sup style="color: #1fea00; font-size: 0.8em;">+4.7%</sup> |
+ | Text Processing | 3 723 | 3 714<sup style="color: red; font-size: 0.8em;">-0.2%</sup> | 3 667<sup style="color: red; font-size: 0.8em;">-1.5%</sup> |
+ | Asset Compression | 30 848 | 31 534<sup style="color: #1fea00; font-size: 0.8em;">+2.2%</sup> | 32 421<sup style="color: #1fea00; font-size: 0.8em;">+5.1%</sup> |
+ | Object Detection | 13 032 | 13 075<sup style="color: #1fea00; font-size: 0.8em;">+0.3%</sup> | 12 545<sup style="color: red; font-size: 0.8em;">-3.7%</sup> |
+ | Background Blur | 12 406 | 12 308<sup style="color: red; font-size: 0.8em;">-0.8%</sup> | 12 207<sup style="color: red; font-size: 0.8em;">-1.6%</sup> |
+ | Horizon Detection | 20 156 | 19 947<sup style="color: red; font-size: 0.8em;">-1.0%</sup> | 19 792<sup style="color: red; font-size: 0.8em;">-1.8%</sup> |
+ | Object Remover | 18 049 | 18 382<sup style="color: #1fea00; font-size: 0.8em;">+1.8%</sup> | 18 796<sup style="color: #1fea00; font-size: 0.8em;">+4.1%</sup> |
+ | HDR | 13 929 | 13 775<sup style="color: red; font-size: 0.8em;">-1.1%</sup> | 13 791<sup style="color: red; font-size: 0.8em;">-1.0%</sup> |
+ | Photo Filter | 14 335 | 13 066<sup style="color: red; font-size: 0.8em;">-8.9%</sup> | 13 849<sup style="color: red; font-size: 0.8em;">-3.4%</sup> |
+ | Ray Tracer | 38 300 | 40 547<sup style="color: #1fea00; font-size: 0.8em;">+5.9%</sup> | 38 962<sup style="color: #1fea00; font-size: 0.8em;">+1.7%</sup> |
+ | Structure from Motion | 22 188 | 22 210<sup style="color: #1fea00; font-size: 0.8em;">+0.1%</sup> | 21 352<sup style="color: red; font-size: 0.8em;">-3.8%</sup> |
+ \* **All percentage values are relative to the base 55W result**
### GPU - Various Benchmarks
Details:
- - UNIGINE Superposition was run with the DirectX 1080p Extreme preset.
- - 3DMARK Steel Nomad Light was run in Vulkan.
+ - UNIGINE Superposition - DirectX, 1080p Extreme preset.
+ - 3DMARK Steel Nomad Light - Vulkan, 1080p
+ - 3DMARK Time Spy - 1080p
| Benchmark | 55W | 85W | 120W |
| ------------------------ | --- | --- | ---- |
- | PassMark 2D Mark | 621 | 706<sup style="color: #1fea00; font-size: 0.8em;">+13.7%</sup> | 758<sup style="color: #1fea00; font-size: 0.8em;">+7.4%</sup> |
- | PassMark DirectX 10 | 59 | 65<sup style="color: #1fea00; font-size: 0.8em;">+10.2%</sup> | 78<sup style="color: #1fea00; font-size: 0.8em;">+20.0%</sup> |
- | PassMark DirectX 11 | 112 | 134<sup style="color: #1fea00; font-size: 0.8em;">+19.6%</sup> | 148<sup style="color: #1fea00; font-size: 0.8em;">+10.4%</sup> |
- | PassMark DirectX 12 | 63 | 67<sup style="color: #1fea00; font-size: 0.8em;">+6.3%</sup> | 80<sup style="color: #1fea00; font-size: 0.8em;">+19.4%</sup> |
- | PassMark GPU Compute | 6229 | 6777<sup style="color: #1fea00; font-size: 0.8em;">+8.8%</sup> | 7311<sup style="color: #1fea00; font-size: 0.8em;">+7.9%</sup> |
- | UNIGINE Superposition | 5247 | 6103<sup style="color: #1fea00; font-size: 0.8em;">+16.3%</sup> | 6522<sup style="color: #1fea00; font-size: 0.8em;">+6.9%</sup> |
- | 3DMARK Steel Nomad Light | 8494 | 9883<sup style="color: #1fea00; font-size: 0.8em;">+16.4%</sup> | 10608<sup style="color: #1fea00; font-size: 0.8em;">+7.3%</sup> |
- | 3DMARK Time Spy | 8891 | 10230<sup style="color: #1fea00; font-size: 0.8em;">+15.1%</sup> | 10710<sup style="color: #1fea00; font-size: 0.8em;">+4.7%</sup> |
+ | PassMark 2D Mark | 621 | 706<sup style="color: #1fea00; font-size: 0.8em;">+13.7%</sup> | 758<sup style="color: #1fea00; font-size: 0.8em;">+22.1%</sup> |
+ | PassMark DirectX 10 | 59 | 65<sup style="color: #1fea00; font-size: 0.8em;">+10.2%</sup> | 78<sup style="color: #1fea00; font-size: 0.8em;">+32.2%</sup> |
+ | PassMark DirectX 11 | 112 | 134<sup style="color: #1fea00; font-size: 0.8em;">+19.6%</sup> | 148<sup style="color: #1fea00; font-size: 0.8em;">+32.1%</sup> |
+ | PassMark DirectX 12 | 63 | 67<sup style="color: #1fea00; font-size: 0.8em;">+6.3%</sup> | 80<sup style="color: #1fea00; font-size: 0.8em;">+27.0%</sup> |
+ | PassMark GPU Compute | 6 229 | 6 777<sup style="color: #1fea00; font-size: 0.8em;">+8.8%</sup> | 7 311<sup style="color: #1fea00; font-size: 0.8em;">+17.4%</sup> |
+ | UNIGINE Superposition | 5 247 | 6 103<sup style="color: #1fea00; font-size: 0.8em;">+16.3%</sup> | 6 522<sup style="color: #1fea00; font-size: 0.8em;">+24.3%</sup> |
+ | 3DMARK Steel Nomad Light | 8 494 | 9 883<sup style="color: #1fea00; font-size: 0.8em;">+16.4%</sup> | 10 608<sup style="color: #1fea00; font-size: 0.8em;">+24.9%</sup> |
+ | 3DMARK Time Spy | 8 891 | 10 230<sup style="color: #1fea00; font-size: 0.8em;">+15.1%</sup> | 10 710<sup style="color: #1fea00; font-size: 0.8em;">+20.5%</sup> |
+ \* **All percentage values are relative to the base 55W result**
### GPU - Gaming
+ Details:
+ - Cyberpunk 2077 - high preset, 1080p, built-in benchmark
+ - The Talos Principle - ultra preset, 1080p, built-in benchmark, 60 seconds limit
+ - Metro Exodus Enhanced Edition - ultra preset (ray tracing set to normal), 1080p, built-in benchmark
+
+ | Game | 55W | 85W | 120W |
+ | ------------------------ | --- | --- | ---- |
+ | Cyberpunk 2077 | 53 | 62<sup style="color: #1fea00; font-size: 0.8em;">+17.0%</sup> | 66<sup style="color: #1fea00; font-size: 0.8em;">+24.5%</sup> |
+ | The Talos Principle | 102 | 113<sup style="color: #1fea00; font-size: 0.8em;">+10.8%</sup> | 117<sup style="color: #1fea00; font-size: 0.8em;">+14.7%</sup> |
+ | Metro Exodus Enhanced Edition | 51 | 59<sup style="color: #1fea00; font-size: 0.8em;">+15.7%</sup> | 63<sup style="color: #1fea00; font-size: 0.8em;">+23.5%</sup> |
+ \* **All percentage values are relative to the base 55W result**
+
+ ### GPU - LLM
`WIP`
### Conslusions
- **For CPU** we have a classic situation of diminishing returns, where **85W mode seems to be the sweet spot**. Going higher doesn't make much sense unless you want to squeeze the absolute maximum out of your system, and going lower is only suitable if you don't use your system resources fully (especially when you don't utilize all of the cores) or don't care about performance loss.
-
- **For GPU** the situation is a bit different, seems that it can utilise power more effectively, but more testing is needed, I'm on it.
+ A classic situation of diminishing returns, where **85W mode seems to be the sweet spot**. Going higher doesn't make much sense unless you want to squeeze the absolute maximum out of your system, and going lower is only suitable if you don't use your system resources fully (especially when you don't utilize all of the cores) or don't care about performance loss.
0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9