|
|
| 16-11-07 / 00:05 : Leopard G4 performance / Tiger (cjed) | Here is the results of comparative tests of Leopard and Tiger through Xbench on a 1,25 Ghz MDD G4 (base Radeon 9000 Pro) :
We see a slight performance hit on processor tests (notably on altivec tests, 50% less) on Leopard, a slight hit in thread tests (with a single core, should be compensated in machines that use multiple cores), UI performances two times worse (but here is room to improve it as last Tiger versions showed 3 three times better results than the first Tiger release). On the other hand memory performance is better and Quartz engine has been optimized, notably in text rendering (3 times faster, and that was previously a slowdown factor in windows content resizing). OpenGL and disk performances are equal. | | | Comments | Merci pour ce test vraiment objectif! Donc un tigre est plus costaud qu'un léopard, c'est bon à savoir ;)
(submitted at 11-16-07, 15:37 by ) | Oui, mais je crois avoir lu quelque part que certains doutent des résultats de xbench. Bref il faudrait comparer avec d'autres logiciels de benchs, et des tests de logiciels. Mais les premiers éléments semblent logiques : découpage plus important en threads (charge supplémentaire pour cette gesion, mais impact minimum si plusieurs core), UI plus gourmande et optimisée plutôt pour les cartes récentes, problème de l'affichage des textes sous Quartz enfin traité.
(submitted at 11-16-07, 20:31 by cjed) | On peut lire d'autres résultats concordants : http://lowendmac.com/mail/mb07/1106.html#4.
(submitted at 11-17-07, 00:31 by cjed) | | Write a comment | |
|
