Sign in with
Sign up | Sign in

Benchmarking The EX475 With Sempron 3400+, LE-1640, And BE-2350

Hacking The HP EX470/475 MediaSmart Servers
By

Interestingly, the LE-1640 and BE-2350 bring different strengths to the contest so that the outcome is not as cut and dried as you might expect. Table 4 includes the values from the NASPT benchmarks for these three processors, where all were equipped with 2 GB of RAM.

Table 4: Benchmarks for Sempron 3400+, LE-1640, BE-2350 with 2.0 GB RAM

Task
3400+
LE-1640
BE-2350
Notes/Remarks
HD Video Playback
33.2
31.0
28.6
Weak inverse correlation to procstrength
2x HD Playback
11.7
10.9
9.8
Inverse correlation to procstrength
3x HD Playback
9.9
8.1
9.2
Weak inverse correlation to procstrength
4x HD Playback
9.8
7.8
8.6
Weak inverse correlation to memory size
HD Video Record
137.0
134.5
137.5
Little or no benefit to procstrength
HD Playback and Record
39.0
38.0
36.0
Weak inverse correlation to procstrength
HD 2x Playback 2x Record
20.9
21.2
21.9
Little or no benefit to procstrength
HD Playback with Office
36.0
34.0
30.5
Inverse correlation to procstrength
HD Playback with Backup
27.6
26.3
28.1
Mixed bag: no clear benefit
Content Creation
6.7
6.5
6.7
Mixed bag: No clear benefit
Office Productivity
23.4
22.9
23.8
Mixed bag: No clear benefit
Backup
48.7
54.5
52.3
Mild correlation to clock rate
Restore
18.4
19.2
17.7
Mixed bag: No clear benefit
File Copy to NAS
72.4
74.2
74.2
Mixed bag: No clear benefit
File Copy from NAS
38.3
40.6
35.3
Mixed bag: No clear benefit
Dir Copy to NAS
5.6
4.0
5.7
Mixed bag: No clear benefit
Dir Copy from NAS
34.1
34.8
31.5
Mixed bag: No clear benefit
Photo Album
23.3
20.4
19.3
Inverse correlation to procstrength


I went into this exercise expecting the dual-core BE-2350 to swamp both of the other single-core processors, but came away believing there’s no clear advantage here. Except for inveterate tinkerers, this tells me that a CPU upgrade isn’t really necessary. In fact, if you accept the sum of all benchmark values for each processor as a rough-and-ready basis for comparison, the 3400+ comes out on top (596), followed by the LE-1640 (588.9), trailed by the BE-2350 (576.7).

Of course, there’s only a 1.2% difference between the 3400+ and the LE-1640, and 2.1% between the LE-1640 and the BE-2350, so we’re talking about a total variation of 3.3%, which I have to believe is completely within the range of sampling error. With no clear advantage for upgrading, this becomes a completely discretionary task.

Check prices for HP's MediaSmart EX475

React To This Article