February 27, 2013 at 11:33Comparison of css3 & html5 browsers (v 2.0)
(I'm afraid to run into a condemnation of the meaninglessness of such comparisons and frequent repetitions, but still try ...)
Based on a recent comparison, I decided to check for myself all the popular browsers and, if possible, catch test versions. But it was precisely the reason for writing the post that served up some strange points revealed during testing (about them below) and yesterday's release of the IE 10 release .
To begin, I will describe hardware and software, because this plays a role in performance tests.
The test suite I took is the same as Bo_bda :
The bulk of the comparison was carried out on the night of February 19-20. Just then, Firefox upgraded to version 19 and Firefox nightly to version 22, and I tested all 4 versions. The second part of the comparison I conducted yesterday (February 26) in the evening, in connection with the release of IE 10.
The Maze Solver and Peacekeeper tests ran 2 times in a row without clearing the cache, without restarting the browser and not refreshing the page.
P.S. Opera did not find a test version of version older than version 12.14 on their blog . Opera 12.14 is the latest public release, so it is the only one in the test.
UPD . Added the 7th item to the observation.
UPD 2 . Added test Opera 12.14 x64 .
Based on a recent comparison, I decided to check for myself all the popular browsers and, if possible, catch test versions. But it was precisely the reason for writing the post that served up some strange points revealed during testing (about them below) and yesterday's release of the IE 10 release .
To begin, I will describe hardware and software, because this plays a role in performance tests.
Iron
- percent - Intel Core i5-2500K 3.4 GHz
- mother - Asus P8Z77-V
- RAM - 8 GB (DDR3-1333)
- video - NVIDIA GeForce GTX 560 Ti
- ROM - SSD OCZ Vertex 4 SATA III
BY
- Microsoft Windows 7 x64 Home Premium SP1
- Java 7.13
Tested browsers
| Browser | User agent |
|---|---|
| Opera 12.14 1738 | Opera / 9.80 (Windows NT 6.1; WOW64) Presto / 2.12.388 Version / 12.14 |
| Opera 12.14 1738 x64 | Opera / 9.80 (Windows NT 6.1; Win64; x64) Presto / 2.12.388 Version / 12.14 |
| Chrome 24.0.1312.57 m | Mozilla / 5.0 (Windows NT 6.1; WOW64) AppleWebKit / 537.17 (KHTML, like Gecko) Chrome / 24.0.1312.57 Safari / 537.17 |
| Chrome 27.0.1417.0 canary | Mozilla / 5.0 (Windows NT 6.1; WOW64) AppleWebKit / 537.32 (KHTML, like Gecko) Chrome / 27.0.1417.0 Safari / 537.32 |
| IE 9.0.8112.16421 | Mozilla / 5.0 (compatible; MSIE 9.0; Windows NT 6.1; WOW64; Trident / 5.0) |
| IE 9.0.8112.16421 x64 | Mozilla / 5.0 (compatible; MSIE 9.0; Windows NT 6.1; WOW64; Trident / 5.0) |
| IE 10.0.9200.16441 Pre-Release | Mozilla / 5.0 (compatible; MSIE 10.0; Windows NT 6.1; WOW64; Trident / 6.0) |
| IE 10.0.9200.16521 | Mozilla / 5.0 (compatible; MSIE 10.0; Windows NT 6.1; WOW64; Trident / 6.0) |
| Firefox 18.0.2 | Mozilla / 5.0 (Windows NT 6.1; WOW64; rv: 18.0) Gecko / 20100101 Firefox / 18.0 |
| Firefox 19.0 | Mozilla / 5.0 (Windows NT 6.1; WOW64; rv: 19.0) Gecko / 20100101 Firefox / 19.0 |
| Firefox 21.0a1 (2013-02-13) | Mozilla / 5.0 (Windows NT 6.1; WOW64; rv: 21.0) Gecko / 20130213 Firefox / 21.0 |
| Firefox 22.0a1 (2013-02-19) en-US | Mozilla / 5.0 (Windows NT 6.1; WOW64; rv: 22.0) Gecko / 20130219 Firefox / 22.0 |
| Safari 5.1.7 (7534.57.2) | Mozilla / 5.0 (Windows NT 6.1; WOW64) AppleWebKit / 534.57.2 (KHTML, like Gecko) Version / 5.1.7 Safari / 534.57.2 |
Testing
The test suite I took is the same as Bo_bda :
- CSS tests
- HTML5 Tests
The bulk of the comparison was carried out on the night of February 19-20. Just then, Firefox upgraded to version 19 and Firefox nightly to version 22, and I tested all 4 versions. The second part of the comparison I conducted yesterday (February 26) in the evening, in connection with the release of IE 10.
The Maze Solver and Peacekeeper tests ran 2 times in a row without clearing the cache, without restarting the browser and not refreshing the page.
results
| Maze solver | The CSS3 Test | HTML5 test | HTML5 test (2) | Peacekeeper | ||||
|---|---|---|---|---|---|---|---|---|
| 1 | 2 | out of 937 | out of 500 | out of 160 | 1 | 2 | ||
| Opera | 12.14 1738 | 7.2 | 7.0 | 58% 451 | 404 +9 | 146 | 3819 2/7 | 3805 4/7 |
| Opera | 12.14 1738 x64 | 7.2 | 7.2 | 58% 451 | 404 +9 | 146 | 3952 2/7 | 3942 4/7 |
| Chrome | 24.0.1312.57 m | 4.4 | 1.4 | 63% 562 | 448 +13 | 148 | 4514 6/7 | 4525 6/7 |
| Chrome | 27.0.1417.0 canary | 4.2 | 1.3 | 65% 573 | 468 +13 | 151 | 4831 6/7 | 4837 6/7 |
| IE | 9.0.8112.16421 | 12 | 4.5 | 33% 274 | 138 +5 | 91 | 2297 3/7 | 2268 3/7 |
| IE | 9.0.8112.16421 x64 | thirteen | 5.0 | 33% 274 | 138 +5 | 91 | 1505 3/7 | 1488 3/7 |
| IE | 10.0.9200.16441 Pre-Release | thirteen | 5.5 | 54% 444 | 320 +6 | 127 | 2731 3/7 | 2715 3/7 |
| IE | 10.0.9200.16521 | thirteen | 5.5 | 54% 444 | 320 +6 | 127 | 2822 3/7 | 2855 3/7 |
| Firefox | 18.0.2 | 47 | 47 | 56% 500 | 393 +10 | 129 | 2447 5/7 | 2415 5/7 |
| Firefox | 19.0 | 5.2 | 5.2 | 58% 505 | 393 +10 | 129 | 2830 5/7 | 2827 5/7 |
| Firefox | 21.0a1 (2013-02-13) | 7.2 | 7.2 | 63% 539 | 399 +14 | 134 | 2888 6/7 | 2895 6/7 |
| Firefox | 22.0a1 (2013-02-19) en-US | 14 | 14 | 63% 539 | 399 +14 | 134 | 2513 6/7 | 2566 6/7 |
| Safari | 5.1.7 (7534.57.2) | 4.5 | 1.4 | 51% 462 | 278 +2 | 97 | 2411 1/7 | 2396 1/7 |
Notes
- CSS3 Selectors Test and Acid3 did not write tests to the table, because all browsers performed them 100% (by points 41/41 574/574 and 100/100, respectively).
- Links to Peacekeeper test results: February 19-20 (all browsers except IE 10 release) and February 26 (only IE 10 release). UPD 2. Test Opera x64 of February 27 .
- Safari could not be tested, as The version is already outdated. And there will be no new versions for Windows due to the fact that Apple stopped supporting Safari under Windows on July 26, 2012.
Observations
- When you restart the Maze Solver test, Chrome, IE, and Safari did not explore the labyrinth again, but walked along the well-known path. As a result of this, the transit time was approximately 3 times less than during the first passage. Where this behavior comes from and why Firefox and Opera didn't show it is not clear.
- But re-launching Peacekeeper in my case did not give any growth, only a statistical error. Although Bo_bda in its comparison Firefox and Chrome added almost 2 times.
- The strange behavior of Opera when restarting Peacekeeper is also noticed. It increases HTML5 Capabilities from 2/7 to 4/7.
- Firefox was very pleased with its update from the 18th to the 19th version. The speed of the Maze Solver maze increased 9 (!!!) times (in this test, the 19th version even outperformed Firefox nightly 21) and the number of parrots in Peacekeeper increased by 15%. And by the way, this is not just synthetics ... The increase in speed is noticeable at least in browser games.
- But Firefox nightly 22 slowed down in the Maze Solver test 2 times and missed 12% of the parrots in Peacekeeper relative to the 21st version, which is very strange.
- And the last one. IE 10 release showed exactly the same results for all tests as IE 10 Pre-Release, only in the Peacekeeper test it issued about 4% more parrots.
- IE 9 x64 in the Peacekeeper test showed a result 35% worse than IE 9.
- Opera x64 in the Peacekeeper test showed a result 3.5% better than Opera x32.
P.S. Opera did not find a test version of version older than version 12.14 on their blog . Opera 12.14 is the latest public release, so it is the only one in the test.
UPD . Added the 7th item to the observation.
UPD 2 . Added test Opera 12.14 x64 .