My dual Xeon 2.4 @ 3.0GHz (hyperthreading enabled),w2k, is benchmarking at about half the cpu-capability.
My dual AthlonMP, w2k, is benchmarking at full cpu-capability.

Both machines crunch their WU's in roughly the same time (10k~12k seconds)

Both are running BoincManager 4.72
Both are running the 4.04 core application for beta.

Claimed credit for the Xeons is at about 50% of that of the Athlons.

Is this correct?

My dual Xeon 2.4 @ 3.0GHz (hyperthreading enabled),w2k, is benchmarking at about half the cpu-capability.
My dual AthlonMP, w2k, is benchmarking at full cpu-capability.

Both machines crunch their WU's in roughly the same time (10k~12k seconds)

Both are running BoincManager 4.72
Both are running the 4.04 core application for beta.

Claimed credit for the Xeons is at about 50% of that of the Athlons.

Is this correct?

And the correct answer is somewhat.

The athlons are two entirely seperate chips so you get double the performance of a single one. The xeons are also two seperate chips and you get about double the performance of a single one. However the OS and BOINC see them as 4 CPUs rather than two so the benchmark numbers reported by BOINC should be about half of what you might expect.

The BOINC benchmarks have known problems with hyperthreaded CPUs and to a lesser extent multiple CPUs. When the more pressing issues are taken care of they will be retuned.

PS. w2k does not really work with HT processors allthough it appears to. It takes XP or later to use them at full potential.

---

BOINC WIKI

BOINCing since 2002/12/8

PS. w2k does not really work with HT processors allthough it appears to. It takes XP or later to use them at full potential.

There's a great document to read on that. Too bad it's only about the server versions of 2000. For Professional Intel advises to disable the HT function of the CPU.

---