Hi, Sad Clouds wrote:
300MHz x 2 = 600Mhz ? 20% faster?It really depends on what is being compared: CPU, memory, disk, etc. It is best to isolate each performance test by running very specific synthetic benchmarks. Benchmarking with real workloads like source code compiling is OK, but is not "apples to apples", as the disk and memory latency significantly affect CPU performance metrics for example.
I know it is not perfect Apples to Apples, but a significative comparision. I always felt the netra to be faster. It is newer, uses newer Memory, SCSI controller, Ethernet... The Ultra 2 is really the beginning of the UltraSPARC! For years I had a single CPU only and was running Solaris. Besides, amazing what was called "Desktop" back then! In today's modern offices it wouldn't even fit or it would bend the desk. Is wider, thicker and heavier than towers...
I got the second CPU and then used it little.Compilation is a good test for "actual usage", it involves hard disk, memory and CPU without resorting to FPU (I hope so, don't know what modern compilers do nowadays).
The Netra has faster disk, faster RAM, faster bus... and still is slower :)It is just an info, if someone has one around, still be useful for testing or building!
And as a comparison AMD Turion(tm) 64 X2 Mobile Technology TL-52 169.52 real 285.78 user 27.50 sysis about 10x faster using both cores, but it is not a 3Ghz cpu, a 1.6Ghz. So the despised Turion is actually more efficient per clock-cycle than my beoved UltraSPARC. This irritates me :)
Riccardo .