Performance within an architecture should scale linearly with clock rate unless
you hit a bottleneck such as FSB throughput - no bottleneck in this case:
T5750 Passmark = 1106:
http://www.cpubenchmark.net/cpu_lookup. ... 40+2.00GHz
T5500 Passmark = 881:
http://www.cpubenchmark.net/cpu_lookup. ... 40+1.66GHz
On the other hand, the CPU is so fast compared to disk and internet throughput
that unless you are running a CPU intensive task you probably will not see much
difference. Given the very low cost, and if you are planning to take it apart to
refresh the thermal paste or whatever, then why not?
Here is the T7200 with the larger 4M cache but same 2 GHz speed, Passmark = 1205:
http://www.cpubenchmark.net/cpu_lookup. ... 40+2.00GHz
The T7300 has an FSB of 800 but will not run in this chipset but can be compared to the
slower 667 FSB, Passmark = 1193 - this should not be lower than the T7200 but the difference
is in the noise anyway. Perhaps the memory ran single channel here instead of dual:
http://www.cpubenchmark.net/cpu_lookup. ... 40+2.00GHz
Here is the E2180 which is a 65 nm, 64 bit Core 2 based desktop processor, marketed under
the Pentium name, with 1M of cache and an 800 FSB, just to see the results with even a
smaller cache. Passmark = 1056:
http://www.cpubenchmark.net/cpu_lookup. ... 40+2.00GHz
Some details on the E2180 desktop processor:
http://en.wikipedia.org/wiki/List_of_In ... 2865_nm.29
I do not know the details of the Passmark benchmark, it does take advantage of multiple
cores but if it is so small that it fits in even a small cache then it will not demonstrate
cache size and FSB differences in a significant way.