I've recently had a chance to compare my Core 2 Quad Q6600 2.4 GHz (4 cores, 4 threads) with DDR2-800 RAM against a Core i7 920 2.66 GHz (4 cores, 8 threads) with DDR3-1067 RAM.
I was surprised by the results.
In single-threaded CPU-bound tasks (e.g. SuperPiMod 1.5), the i7 920 was about 65% faster (51 seconds vs. 31 seconds).
In multithreaded CPU-bound tasks (e.g. wPrime 2.03), the i7 920 was almost twice as fast (20 seconds vs. 10 seconds).
I've confirmed it in a real world test case by taking a somewhat large C++ project (612 source files) and recompiling it from scratch in Visual Studio 2010. The project was stored on a RAM disk, so HDD was not a factor.
It consistently took 30 seconds to fully compile on i7 920, and 60 seconds on the Q6600.
Is this normal? I thought the Core i7s were only 20-30% faster clock-per-clock than Core 2 CPUs. In any case, I'm impressed. I wonder how a 2500K would stack up.
I was surprised by the results.
In single-threaded CPU-bound tasks (e.g. SuperPiMod 1.5), the i7 920 was about 65% faster (51 seconds vs. 31 seconds).
In multithreaded CPU-bound tasks (e.g. wPrime 2.03), the i7 920 was almost twice as fast (20 seconds vs. 10 seconds).
I've confirmed it in a real world test case by taking a somewhat large C++ project (612 source files) and recompiling it from scratch in Visual Studio 2010. The project was stored on a RAM disk, so HDD was not a factor.
It consistently took 30 seconds to fully compile on i7 920, and 60 seconds on the Q6600.
Is this normal? I thought the Core i7s were only 20-30% faster clock-per-clock than Core 2 CPUs. In any case, I'm impressed. I wonder how a 2500K would stack up.