This is an old revision of the document!
Table of Contents
With the plethora of new devices the ability to benchmark them for numeric comparison seems to be useful. It will also help us document any future optimizations to code like multi-core utilization.
Summary
Time in seconds, lower is better
| Phones | Read2Render | IsoPach Build Time | Creating 3D |
|---|---|---|---|
| Galaxy S3 | 344.79 | 226.44 | 11.07 |
| Galaxy S4 | 244.41 | 173.23 | 12.99 |
| Galaxy S5 | 164.30 | 118.73 | 8.44 |
| HTC Evo | NC | NC | NC |
| Nexus 5 | 304.25 | 219.4 | 15.1 |
| Tablets | |||
| Nexus 7 2012 | 313.94 | 242.69 | 10.50 |
| Nexus 7 2013 | 237.42 | 167.86 | 10.7 |
| Nexus 7 2013 L | 186.95 | 116.37 | 4.862 |
| Nexus 10 | 242.47 | 174.38 | 7.96 |
| Fujitsu 522 | 323.97 | 237.48 | 25.23 |
| Asus TF300 | 411.93 | 303.16 | 26.93 |
| Panasonic FZ-A1 | 804.79 | 595.80 | 60.96 |
| Galaxy Tab | |||
| Galaxy Tab S 8.4 | 142.92 | 106.56 | 5.81 |
Test Explanation
- Read2Render is the most important and is the total time from reading the file to showing cut/fill color.
- IsoPach Build time is usually the most time consuming element. This is the process of creating surface meshes and the combined surfaces (ISOPach) that the volumes and cut/fill color relies upon. This result is saved in the ADF file after the first run.
- Creating 3D is the calculation of the 3D views and creating the cut/fill color from the ISOPach.
Nexus 2013 L is running a preview version of Android L
* the HTC EVO cannot complete this test.
Detailed
Times in Seconds
| Phones | Galaxy S3 | Galaxy S4 | Nexus 5 | Galaxy S5 |
|---|---|---|---|---|
| Read2Render | 344.79 | 244.41 | 304.30 | 164.28 |
| ADF Read | 0.14 | 0.15 | 0.06 | 0.14 |
| OG Build Time | 0.82 | 0.65 | 0.60 | 0.43 |
| DN Build Time | 50.15 | 6.50 | 5.50 | 3.99 |
| ISOpach Build Time | 226.44 | 173.23 | 219.40 | 118.74 |
| ISOpach Copy Design | 0.80 | 0.68 | 0.60 | 0.44 |
| ISOpach Clip Existing | 69.06 | 46.77 | 49.30 | 32.61 |
| ISOpach Drape | 69.95 | 68.10 | 85.10 | 48.19 |
| ISOpach setting elevations | 4.86 | 5.20 | 5.70 | 2.95 |
| ISOpach meshing | 64.45 | 38.79 | 57.20 | 24.63 |
| ISOpach decomposition | 17.24 | 14.60 | 21.50 | 9.87 |
| Creating 3D | 11.07 | 12.99 | 15.10 | 8.44 |
| Code Test Version | 1.06 Rogue | 1.06 1462.1387 | 1.06 1462.1387 | 1.06 1462.1387 |
What Timings are most important, What is saved after read.
Times in Seconds
| Tablets | Nexus 7 | Nexus 7 2013 | Nexus 10 | Asus TF300 | Fujitsu 522 | Panasonic FZ-A1 | Samsung Galaxy 7 |
|---|---|---|---|---|---|---|---|
| Read2Render | 313.94 | 237.42 | 242.47 | 411.93 | 323.97 | 804.79 | fail |
| ADF Read | 0.10 | 0.09 | 0.09 | 0.14 | 0.10 | .60 | 0.42 |
| OG Build Time | 0.72 | 0.63 | 0.63 | 0.86 | 0.85 | 4.96 | 1.228 |
| DN Build Time | 6.57 | 5.8 | 5.80 | 7.37 | 0.676 | 45.68 | 10.86 |
| ISOpach Build Time | 242.69 | 167.87 | 167.87 | 303.16 | 237.48 | 595.80 | 1146 |
| ISOpach Copy Design | 0.80 | 0.69 | 0.75 | 1.138 | 1.00 | 1.86 | 16.03 |
| ISOpach Clip Existing | 71.23 | 45.39 | 54.76 | 86.57 | 66.96 | 127.80 | 96.35 |
| ISOpach Drape | 106.77 | 72.33 | 77.46 | 131.27 | 96.74 | 209.71 | 939 |
| ISOpach setting elevations | 4.08 | 4.12 | 3.03 | 4.21 | 4.49 | 13.27 | 7.42 |
| ISOpach meshing | 41.843 | 31.17 | 26.99 | 52.22 | 46.85 | 202.83 | 55.92 |
| ISOpach decomposition | 17.89 | 14.11 | 11.32 | 26.65 | 21.36 | 40.31 | 31.43 |
| Creating 3D | 10.5 | 10.7 | 7.96 | 26.93 | 25.23 | 60.96 | fail |
| Code Test Version | 1.06 1462.1387 | 1.06 1462.1387 | 1.06 Rogue | 1.06 1462.1387 | 1.06 1462.1387 | 1.06 1462.1387 | 1.6 1665.1546 |
Times in Seconds
| Tablets cont. | Galaxy Tab S 8.4 | *Nexus 7 2013 L |
|---|---|---|
| Read2Render | 142.92 | 186.95 |
| ADF Read | 0.11 | .59 |
| OG Build Time | 0.70 | .63 |
| DN Build Time | 4.02 | 6.06 |
| ISOpach Build Time | 106.56 | 116.37 |
| ISOpach Copy Design | 0.55 | 4.17 |
| ISOpach Clip Existing | 31.92 | 31.18 |
| ISOpach Drape | 43.01 | 43.78 |
| ISOpach setting elevations | 1.88 | 4.69 |
| ISOpach meshing | 20.83 | 25.34 |
| ISOpach decomposition | 8.23 | 7.17 |
| Creating 3D | 5.81 | 4.86 |
| Code Test Version | 1.06 1462.1387 | 1.8.2806.2337 |
Note Samsung Galaxy 7 ran out of memory on making the 3D surface.
What Timings are most important, What is saved after read.
Setting up to Test
The program version used to test is 146.1387 https://agtek.s3.amazonaws.com/Agtek/U82FSwRG2ala
File
Using eggman
Statistics
Subjective Observations
The Fujitsu 522 is very slightly brighter than the Nexus 10 in full sunlight. Both are much brighter than the Nexus 7. The Nexus 10 screen is much higher resolution than either and therefore much crisper.
The Samsung S5 is noticeably brighter than the S4 and S3. The Samsung Galaxy S 8.4 AMOLED screen is very bright
Modern SmartDirt 1.8 Testing
The version of Smartdirt used for the original testing is getting quite old and newer versions are slightly slower in calculating ISOpachs but faster at 3D. Going forward I'll run the tests on both versions.
Summary
Time in seconds, lower is better
| Phones | Read2Render | IsoPach Build Time | Creating 3D |
|---|---|---|---|
| Galaxy S3 | xxx.xx | xxx.xx | xx.xx |
| Galaxy S4 | xxx.xx | xxx.xx | xx.xx |
| Galaxy S5 | xxx.xx | xxx.xx | xx.xx |
| HTC Evo | NC | NC | NC |
| Nexus 5 | xxx.xx | xxx.xx | xx.xx |
| Tablets | |||
| Nexus 7 2012 | xxx.xx | xxx.xx | xx.xx |
| Nexus 7 2013 | xxx.xx | xxx.xx | xx.xx |
| Nexus 10 | xxx.xx | xxx.xx | xx.xx |
| Fujitsu 522 | xxx.xx | xxx.xx | xx.xx |
| Asus TF300 | xxx.xx | xxx.xx | xx.xx |
| Panasonic FZ-A1 | xxx.xx | xxx.xx | xx.xx |
| Galaxy Tab | |||
| Galaxy Tab S 8.4 | 157.94 | 126.78 | 4.05 |
Test Explanation
- Read2Render is the most important and is the total time from reading the file to showing cut/fill color.
- IsoPach Build time is usually the most time consuming element. This is the process of creating surface meshes and the combined surfaces (ISOPach) that the volumes and cut/fill color relies upon. This result is saved in the ADF file after the first run.
- Creating 3D is the calculation of the 3D views and creating the cut/fill color from the ISOPach.
* the HTC EVO cannot complete this test.
