Flash Player 10.1 Performance: Part 5
Part five of this series on Flash Player 10.0 versus Flash Player 10.1 performance continues re-testing performance articles just like the first, second, third, and fourth articles did. Read on for more performance comparisons!
Introduction
This part’s performance methodology will be the same as
Sorting Vectors
| Method | Unsorted 10.0 | Pre-Sorted 10.0 | Unsorted 10.1 | Pre-Sorted 10.1 |
|---|---|---|---|---|
| Vector via sort() | 1603 | 1190 | 1110 | 814 |
| Array via sortOn() | 644 | 350 | 588 | 343 |
| Vector via quickSort() | 3960 | 3027 | 2932 | 2249 |
| Vector via shellSort() | 6142 | 3455 | 5262 | 2642 |
| Vector via Array.sortOn() | 925 | 468 | 879 | 462 |
Without exception, all of the sorting approaches are faster in Flash Player 10.0. Unfortunately, most of the performance gains are minor with the exceptions of quickSort and shellSort. Even so, these boosted manual sort functions are still vastly slower than the alternatives.
Increment and Decrement
| Environment | Post-Increment | Pre-increment | Add One | Post-Decrement | Pre-Decrement | Subtract One |
|---|---|---|---|---|---|---|
| Flash Player 10.0 | 236 | 236 | 236 | 236 | 236 | 236 |
| Flash Player 10.1 | 202 | 202 | 202 | 202 | 202 | 202 |
There are two findings to take away from this re-test: your choice of increment/decrement style still doesn’t matter and all increments and decrements are now about 17% faster.
Array vs. Vector
| Environment | Array (read) | Vector (fixed, read) | Vector (dynamic, read) | Array (write) | Vector (fixed, write) | Vector (dynamic, write) |
|---|---|---|---|---|---|---|
| Flash Player 10.0 | 835 | 599 | 595 | 1912 | 698 | 714 |
| Flash Player 10.1 | 1112 | 708 | 709 | 1925 | 808 | 810 |
Sadly, both Array and Vector got slower in Flash Player 10.1. While this is not really pronounced when writing to them, reading is about 19-33% slower. For more on this, see Map Performance in the previous article in this series.
Array vs. Vector: Part II
Read Performance
| Environment | Array | Vector (int) | Vector (uint) | Vector (Number) | Vector (Boolean) | Vector (String) | Vector (Object) |
|---|---|---|---|---|---|---|---|
| Flash Player 10.0 | 38 | 29 (31%) | 31 (23%) | 30 (27%) | 29 (31%) | 29 (31%) | 30 (27%) |
| Flash Player 10.1 | 38 | 30 (27%) | 29 (31%) | 31 (23%) | 30 (27%) | 29 (31%) | 30 (27%) |
Write Performance
| Environment | Array | Vector (int) | Vector (uint) | Vector (Number) | Vector (Boolean) | Vector (String) | Vector (Object) |
|---|---|---|---|---|---|---|---|
| Flash Player 10.0 | 96 | 37 (259%) | 37 (259%) | 39 (246%) | 72 (33%) | 90 (7%) | 131 (-36%) |
| Flash Player 10.1 | 97 | 37 (262%) | 40 (235%) | 38 (255%) | 75 (29%) | 90 (8%) | 125 (-22%) |
Delete Performance
| Environment | Array | Vector (int) | Vector (uint) | Vector (Number) | Vector (Boolean) | Vector (String) | Vector (Object) |
|---|---|---|---|---|---|---|---|
| Flash Player 10.0 | 4544 | 4429 (3%) | 4431 (3%) | 4473 (2%) | 4506 (1%) | 4510 (1%) | 4485 (1%) |
| Flash Player 10.1 | 4630 | 4596 (1%) | 4546 (2%) | 4591 (1%) | 4628 (0%) | 4592 (1%) | 4597 (1%) |
At these lower volumes, the performance difference in reading and writing are less apparent than in the re-test for the first part of this article. Deleting is still enormously expensive though, but relatively unchanged in proportion to Flash Player 10.0.
Inline Math.ceil() Part II
| Environment | Inline (positive) | Inline (positive and negative) | Math.ceil() |
|---|---|---|---|
| Flash Player 10.0 | 777 (3.5x) | 775 (3.5x) | 2741 |
| Flash Player 10.1 | 740 (1.8x) | 745 (1.8x) | 1360 |
It seems that Math.ceil() has received a much-needed performance boost in Flash Player 10.1; it is now twice as fast as it was in Flash Player 10.0! The inlined version, however, hasn’t been sped up at all and therefore sees its lead diminished to a mere 1.8x speedup. It is therefore still useful for performance-critical areas of code that need to take the mathematical ceiling.
One More To Come
With only three more performance articles to re-test, the next article in this series will be the last. Stay tuned for the conclusion!