turpit
Splendid
zenze589 :
Well I don't know if you noticed but the E6850 scored better than the Q6600 at supreme commander, which is supposed to be a game that utilizes multiple cores, as well as nearly all of the other games tested.
Correct me if I'm wrong, but I believe that the E6850 also has greater overclocking potential than the Q6600, which would also make it more appealing.
I bring this up because I am about to build myself a new computer because I am currently running a agonizingly slow Athlon XP 2000+. And one of the things that I have been researching heavily is the choice between the E6850 and Q6600. Based on what I have been reading on these forums over the past week I was fairly sure I was going to go with the Q6600, but after looking at those benchmarks I'm having doubts...
Correct me if I'm wrong, but I believe that the E6850 also has greater overclocking potential than the Q6600, which would also make it more appealing.
I bring this up because I am about to build myself a new computer because I am currently running a agonizingly slow Athlon XP 2000+. And one of the things that I have been researching heavily is the choice between the E6850 and Q6600. Based on what I have been reading on these forums over the past week I was fairly sure I was going to go with the Q6600, but after looking at those benchmarks I'm having doubts...
That quad cores will out perform a faster dual core is a common misconception, and one that has been promulgated relentlessly by the 'wide eyed' tennie bopper crowd that is overcome by the 'more is better' theory, but dont understand multi threading or advertising.
Saying that an app is multithreaded or optimized for multicore does not mean it will give better perfromance for a greater # of cores. There are at least 3 ways (that I know of) to write multithreaded programs; fine, course and hybrid. Depending on which method is used to write a given program, that application may or may not see improved performance by increasing core #.
In the case of course threading, it will give you the best overall performance increase of the 3 methods, but a thread must be writen for each core...meaning a course threaded app written for 2 cores will see no improvement by going to 4 cores. It will not scale.
Fine threading on the other hand will scale, but will not give the same performance boost that course threading will. So a fine threaded app will not run as fast on 2 cores as the same app would if it were written in course, but the fine threaded app will improve in perfromance with more cores, up to the point it runs out of individual loops to process.
Hybrid threading combines the two methods and like fine threading will make scale up on more cores, but also like fine threading, does not give as good a gain as course threading.
So, when an app is advertised as making use of, or optimized for mulitcore, it doesnt mean it will scale up in performance for every core added, only that it will use more than one core. Of course, most kiddies see 'optimized for multicore' and just assume it means the app will take as many cores as you can throw at it. Those same kiddies then proceed to run around the internet crying 'buy the quad, buy the quad, it rulz' but as the test results show, clockspeed is still important.
Something to consider. Dual core processors have been around for awhile now. Apps that make use of dual core (with the exception of the heavy data crunchers like 3d rendering, video editing, encoding ect) havent exactly been been 'raining' down upon us. There are a handfull of upcoming big title games that will make use of quad core, but these are only a fraction of apps currently or soon available. But because these games are highly anticipated, and highly hyped, they are getting a lot of attention, making it seem as if every new app to be released is going to be quad capable. Just not true. Trust the benchmarks, not the stuff you read in the forums. 😉