Seems to be an awsome cpu.......hes still testing but he got 3dm06 cpu test to run on air 4.4ghz http://www.xtremesystems.org/forums/showpost.php?p=4341374&postcount=66 heres the whole thread read throught it and see some decent results http://www.xtremesystems.org/forums/showthread.php?t=249530
Very interesting indeed http://valid.canardpc.com/show_oc.php?id=1132113 These chips are going to be nothing but WIN! Psychlone
after all the intel bunch were saying, at the price of a 1055T and a cheapo AM3 board with bios flash will get you same power as an intel rig costing £300 more lol
I don't get it.. I've got a Phenom 2 965 clocked at just over 4Ghz. My 1M SuperPI score was 17sec flat. Just slightly behind this 6-core Thuron at 4.2Ghz. I figured the Thuron would have left me behind.. like at 10sec.. What gives???
Actually it muddles across all cores, but one is predominant. it'll go from one to the other, but, doesn't use a single core and does not use 100%. share's the load, core 0 could be 80%, core 1 could be 12%, and so forth. Run AMD Overdrive and set to Status, CPU and see the cores run up and down. You can also use CoreTemp and set the intervals tightly so there's a slight delay or real time actual.
I dont let it run like that. I usually test each core on its own by setting affinity........theres usually cores that do better than others in single threaded tests, you can use the faster core to get a better result if thats what you are looking for in stuff like superpi 1m........Ive done that a few times for speed contests I entered. I like tests like cinebench-cinebench 11.5 better where it utilizes all the cores.
...which is why wPrime is a better quick-and-dirty test than SuperPi...it uses all cores simultaneously instead of ramping up one and partially using the rest like SuperPi does. The Thuban at STOCK speeds was pumping out 32M wPrime at 4.something seconds - that's a full 300(and something) seconds quicker than my 940 overclocked to 4GHz. Which corresponds to my prediction years ago that core speeds will fall by the wayside to more cores. Having more cores to run multithreaded apps (of which, there isn't may *yet*) is going to be what get us there quicker than having less cores but higher core speeds. I've been talking about this for years - seems that some software devs are just now catching on to the power of multi-core processing! Psychlone
