It seems like performance scale with MHz on AMD.
Memory speed doesn't matter?
AMD XP seems as good as A64.
Not just MHz factor (installed memory is identical on all systems). Note that the XP1600+ is close to double the XP1700+ CS/hr because it has SSE capability, but the XP1700+ does not.
Wish we had a 3dnow app. ALL my systems have THAT capability and Akos apps for S4 used it to excellent advantage.
Memory speed shouldn't matter. The working set of the science app fits easly into the CPU cache. AMDs can fit it all into the fastest L1 area, P4's need to put some into the slower L2 cache which hurts their relative performance.
It seems like performance
)
It seems like performance scale with MHz on AMD.
Memory speed doesn't matter?
AMD XP seems as good as A64.
RE: It seems like
)
Not just MHz factor (installed memory is identical on all systems). Note that the XP1600+ is close to double the XP1700+ CS/hr because it has SSE capability, but the XP1700+ does not.
Wish we had a 3dnow app. ALL my systems have THAT capability and Akos apps for S4 used it to excellent advantage.
Seti Classic Final Total: 11446 WU.
Computer: Dual 800MHz P3
)
Computer: Dual 800MHz P3 Coppermine
Short units:
Processing time: Average for 14 results 9292.66 seconds
Credit claimed: 20.40
No long work units so far.
As a possible point of interest I see that the initial replication and minimum quorum have been reduced to 2.
Nothing travels faster than the speed of light with the possible exception of bad news, which obeys its own special laws.
Douglas Adams (1952 - 2001)
Memory speed shouldn't
)
Memory speed shouldn't matter. The working set of the science app fits easly into the CPU cache. AMDs can fit it all into the fastest L1 area, P4's need to put some into the slower L2 cache which hurts their relative performance.