Looking for benchmark score for Pentium 4 651 LGA775

[SumGuy]

Is there a CPUmark score for a Pentium 4 651 LGA775 (Cedar Mill / 65 micron, 3.4 ghz SL9KE)?

I believe this is a single-core CPU - not sure if it has hyperthreading (like some socket 478 Pentiums had). I don't think this is a Pentium D.

[David (PassMark)]

The Pentium 4 range appears in the CPU mega list.
And on the low end CPU graph.

[SumGuy]

The cpu that you're refering to:
http://www.cpubenchmark.net/cpu.php?....40GHz&id=1077

Is a socket 478. The one I'm looking for is socket 775:
http://ark.intel.com/products/27483/...Hz-800-MHz-FSB

Now I don't know - did Intel take some socket 478 Pentium 4's and turn them directly into socket 775 with no change in performance?

And why would the CPUmark score (401) be almost half the single thread score (736) ?

[SumGuy]

The cpu score that you reference is for the socket-478 version of the Pentium 4. I'm looking for a cpu score for a socket-775 (LGA775) Pentium 4 651. Perhaps they would be the same (are there versions of the same core or die that Intel made in both socket 478/775 formats and hence would give identical performance scores?).

I note that the socket-478 cpu gives an overall score that's half the single-thread score. Does hyperthreading cause this? Is this performance effect real?

[David (PassMark)]

The problem with older CPUs is that they never identified themselves. They had no internal model number to reference. So when trying to identify them via software you had to look at the list of features available, then take a bit of a guess.

I haven't checked, but if there was two different sockets for the P4 at the same clock speed then they have likely been lumped together.

Single thread rating may be higher than the overall rating, thread performance is just one component of the CPU Mark. (i.e. the relative scale is different).

These CPUs are so old, does it really matter now in any case?