![]() Let me explain: when L1 or L2 got a cache miss they do not ask for one particular address… the ask for a chunk of address. Incrementing the size of the cache helps in multitasking (more program’s contexts can be stored in the cache), but still, an application is required to make a good use of the cache. The think you want to avoid is the monstruous “cache miss”. Is this address in the L2 cache table? Yes give L1. Is this address in the L1 cache table? Yes good. Now that work this way: the CPU request an address. Caching (L1/2) is here to reduce the frequency of access to the physical RAM, that is as slow as a snail from the CPU point of view. Should I wait for a second revision or an Intel Core 2 Duo or start shopping?ĭoes anyone here know if L2 cache generally is important for number crunching or rendering? My project supervisor suggested (!) that I buy a new laptop before the fall. * May be a snappier/less hungry 65nm revision around the corner. * Faster than the Turion and Sempron 3400+, slower than the heavyweight Pentium 4 and Athlon desktop CPUs. * separate L2 cache for the two cores, but now only 512KB per core – meaning lower latency, but less space for buffering. The power consumption of the dual core Turion X2 is perhaps slightly less than the single core Turion. * better power management than the Turion. In short they test the TL-52 1.6GHz 2x512KB L2 and point out that: ![]()
0 Comments
Leave a Reply. |
AuthorWrite something about yourself. No need to be fancy, just an overview. ArchivesCategories |