HP Pavilion a1640n & a1630n: Intel vs. AMD?

Complete|Mobile Systems
Viewing page 3 of 5 pages. Previous 1 2 3 4 5 Next

POWERING UP

The first order of the day was a quick examination of the BIOS. It was very quick indeed. As is typical of systems from Big Brands, the BIOS is extremely limited. There are almost no options to speak of in either system. They both have hardware monitoring, however, for temperature sensing of key components. Both systems also came equipped with standby/sleep functions, although there were no selectable options for them in the BIOS.

Both systems were preloaded with Windows Media Center Edition (MCE) 2005. They also came preloaded with some applications, including...

  • MS Works v.8 (a simpler alternative to MS Office)
  • HP Photosmart Premier 8.0 (which seems similar to one of Adobe's general purpose imaging packages)
  • Sonic DigitalMedia Plus v.7 (optical disc burning/management software suite)
  • MS Money 2006

As usual, there were numerous trial and limited time software products preloaded into both systems, and even links to "hot deals" from various HP partners. Most of these were just a nuisance, and most users would probably want to uninstall them as quickly as possible. But that's probably a long and tedious job as well. We've seen similar "partner" packaged software in other big brands as well and have the same opinion of the practice: It's mostly a nuisance.

Aside from such annoyances, first impressions were...

  • The systems are noisy but not as bad as we've heard. The buzzing and humming of the fans, and the chattering of the hard-mounted hard drive when seeking was about the same in both systems. Not great, but better than others we've encountered in the past.
  • The overall feel of both systems was pretty good. That is to say, speedy and responsive enough not to feel much different in casual Windows procedures and web surfing than higher performance systems in the lab.

TESTING

Before running any tests or comparisons, the OS and drivers on both systems were totally and completely updated using Windows Update and HP's own HP Update. This actually took a couple of hours. Out of curiosity, the most recent G965 chipset graphics driver (win2k_xp1427.exe) was downloaded directly from the Intel web site. During the installation, an error message appeared stating that the driver is not validated for this computer and that the appropriate driver should be obtained from the computer manufacturer. The same thing occurred with the AMD system. In other words, the drivers in these systems are optimized by or for HP, and maintained through the HP Update service.

The following tools were used during testing:

Software:

CPUBurn (CPU stress test)
Prime 95 (CPU & memory stress test)
Speedfan v. 4.32 (temperature monitoring)
3DMark06 (video performance benchmark)
3DMark05 (previous version)
PCMark05 (system performance benchmark)
HP Photosmart Premier 8.0
Command & Conquer: Generals (2003 early 3D real time strategy game)
Far Cry (2004 first person shooter game)
Web browsing, document reading/editing

Hardware:

BenQ FP991 19" LCD monitor
ADI 19" CRT monitor (many years old)
Linksys 2-port KVM switch
Seasonic Power Angel AC power meter
Brüel & Kjaer Type 2203 Sound Level Meter
SPCR's digital audio recording system

Ambient conditions at the time of testing were 18 dBA and 21°C.

Acoustic, Thermal and Power Measurements
HP Pavilion a1640n (Intel E6300)
Activity State
SPL (dBA@1m)
CPU
HDD
AC Power
Idle (SpeedStep)
29
44°C
39°C
76W
2 x CPUBurn
30
69°C
39°C
108W
2 x Prime 95
30
69°C
39°C
117W
HDD defragment
36
45°C
42°C
78W
HP Pavilion a1630n (AMD A64X2 4600+)
Activity State
SPL (dBA@1m)
CPU
HDD
AC Power
Idle (CoolnQuiet)
27
29°C
30°C
47W
2 x CPUBurn
27
64°C
30°C
135W
2 x Prime 95
27
62°C
30°C
120W
HDD defragment
28
30°C
32°C
49W

Acoustics

Despite the slightly lower Sound Pressure Levels (SPL) measured for the AMD-based PC, there wasn't much significant difference in the sound of these systems. Both had a generally rough, if subdued, machine-like character most of the time. The big difference was during hard drive activity, when the Intel-based system was much louder than the AMD one. But the two system were running different makes of hard drives, and the HDD in the AMD-based system just happened to be quieter. The Intel-based system also became a bit noisier under prolonged CPU load as its CPU fan seemed to speed up a little. In contrast, the AMD-based system sounded the same at idle or at maximum load. Noise caused by vibration of the chassis, especially by hard drive seek vibration, was one of the bigger annoyances in both systems. Of course, the AMD-based system had one nasty annoyance: Its CPU fan started with a shocking full speed blast that lasted about 4-5 seconds.

By SPCR standards, the level of noise was fairly high, but more important than that was the poor quality of the sound. Neither PC had the smooth quality needed to make a computer sound quiet. The Intel-based system was worse in this regard, but it was mostly caused by the louder, harsher sounding hard drive. This was just an accident of fate; the drives could easily have been swapped on another production run.

Thermals

The temperature of the CPU in both systems went up into the 60s, with the Intel CPU reading a bit higher. As the readings are from different CPUs made by different manufacturers, compare the data only with skepticism. The main thing to note is that neither system experienced any throttling or misbehavior at high CPU load testing. Ditto the hard drive temperatures, which remained quite low in both systems.

The much lower temperature of the AMD processor at idle (compared to the Intel processor) may be explained by this system's near-30W lower power consumption at idle. See next paragraph.

Power

47W AC at idle is mighty impressive for a system with such credentials. The only other systems examined by SPCR that idle below 50W are those built around mobile processors or the VIA EPIA min-ITX integrated motherboards. Remember, the system's AMD A64-X2 4600+ is a dual-core processor that runs at 2.4 GHz. Aside from the power supply not being equipped with Active PFC, the HP Pavilion a1630n is actually ready to meet the toughest <50W requirement for Category A: Desktops of the new EnergyStar Computer Spec 4.0 that becomes effective in just a few months on July 20, 2007. This, despite the fact that it actually falls into the less stringent Category B: Dual-Core Systems with 1GB RAM, which allows idle power up to 65W.

The single greatest power difference between the two systems was at idle. Intel Enhanced Speedstep (EIST) and the similar AMD Cool'n'Quiet (CNQ)was operating on both systems to keep CPU power down during idle, but no matter what, the Intel system never dropped below 76W.

In contrast, the AMD system idled at 47W, nearly 30W lower. At the highest loads (Prime 95 for the Intel system and CPUBurn for the AMD system), the advantage went to the Intel system: 117W versus 135W, a difference of 18W.

A couple of recent AC Plug Load research studies* found that typical home systems idle about 2/3 of the time they are on and run at high load about 1/3 of the time. If this formula is applied to the two systems on hand, and a typical daily power-on time of 4 hours is used, the results for energy usage would be as follows:

Home Energy Use Estimate
System
Daily
a1640n
(Intel)
358 Wh
a1630n
(AMD)
304 Wh
AMD advantage
15% less energy

If the systems were assumed to be typical office machines, the difference would be considerably bigger, because according to the Energy Star program of the EPA, most office PCs run at idle 90% of the time that they are on. Assuming an 8 hour day, this means the daily energy usage would be considerably different from the home computer. The AMD system's energy advantage then becomes 29%.

Office Energy Use Estimate
System
Daily Average
a1640n
(Intel)
664 Wh
a1630n
(AMD)
473 Wh
AMD advantage
29% less energy

These estimates assume that the high load in real office or home systems is the highest load we could apply in the lab. This is unrealistic. Most people don't run CPUBurn or Prime 95 for two hours every day. The typical high load in a home computer would be most likely to be games. In an office computer it might be manipulations with a very large database file, processing large images with apps like Photoshop, or perhaps applying 3D processes in an architectural program. The simple fact is that none of these programs apply as large and continuous a load as either of the CPU stress utilities used here. The lower the peak load, the greater the significance of the idle load... and the greater the AMD system's energy efficiency advantage.



Previous 1 2 3 4 5 Next

Complete|Mobile Systems - Article Index
Help support this site, buy the HP Pavilion Media Center a1630n Desktop from one of our affiliate retailers!