Liquid-cooled HP Z800 Workstation Test Drive

Figure 1. The HP Z800's liquid-cooling system radiator.

Figure 1. The HP Z800's liquid-cooling system radiator.

I produce a lot of screencams and other narration-type recordings, and workstation noise is a constant concern. I also have multiple computers around my office, most off testing some software program or rendering some project. While "cacophony" is definitely too strong a word to apply, less noise is always good. For this reason, I was excited when HP called to offer a quick spin with its new liquid-cooled Z800 workstation.

Related Links

Test Drive: Intel Nehalem, Part 1

A few months ago, I ran some Adobe Creative Suite 4 (CS4) benchmarks on different computers that isolated how CS4 performed with formats ranging from DV to Red. Now that Intel''s Nehalem processor is upon us, those numbers are obsolete...

Test Drive: Intel Nehalem, Part 2

Welcome back to our presentation of how HP''s new Intel Nehalem-based workstations compare to older workstations when rendering from Adobe Creative Suite 4 (CS4)...

Review: HP Z800

Sporting a completely redesigned case and Intel''s new Nehalem processor, the HP Z800 knocks the socks off HP''s existing workstation line...

If you''ve heard of liquid-cooled systems in the past, you''ve probably associated them with gamers, who deploy the added efficiency of liquid cooling to prevent their CPUs from burning out when they overclock them to within a hertz of their natural life. As you might guess, this isn''t HP''s focus. Rather, HP focuses on the reduced noise that liquid-cooling systems deliver as compared to the output of traditional fans.

Of course, multiple components contribute to overall workstation noise—CPU and other exhaust fans, spinning hard disk drives, and fans on graphics cards being the most prominent. However, since liquid is more efficient at dissipating heat than air (which is why a 78-degree pool seems freezing when you jump in, while a 78-degree air temperature feels very comfortable), the associated cooling fans don''t have to work as hard to remove the heat.

According to HP''s marketing documents, this translates to an 8dB difference in noise levels, from 38dB to 30dB. In perspective, this puts both noise levels below a faint whisper. However, since a 10dB delta is perceived as a 50 percent difference, the makes the traditional fan-cooled system about 70 percent to 80 percent louder than the liquid-cooled. That''s the theory, anyway; you can read about how I tested this below.

<i />Figure 2. Note the two cooling stations on the CPUs and the radiator fan on the left.

Figure 2. Note the two cooling stations on the CPUs and the radiator fan on the left.

Before that, though let''s understand some basics of the liquid-cooling system, which will be available for about $250 extra on the Z800 and $150 extra on the Z400. First, it''s a closed system, so you never have to add liquids—not even after six months or 3,000 computing hours (whichever occurs first). Second, the coolant is a mixture of water and glycol, which is both nontoxic and biodegradable.

Physically, there are two main components: the radiator in the back (I kid you not) and the cooling stations located on the CPUs themselves. The radiator is shown in Figure 1. Though it looks pretty bulky, it really doesn''t require more space than the traditional model because the power cord on the upper right typically comes out a bit further.

The internal cooling stations are shown in Figure 2, and you can also see the fans on the inside of the radiator on the left. Note that the liquid-cooled Z800 has the same internal cowlings—top and bottom—as the air-cooled Z800, though I''ve removed them for the photo.

Figure 3. Waveform for air-cooled system shown with 500 percent vertical zoom.

Figure 3. Waveform for air-cooled system shown with 500 percent vertical zoom.


To test the noise characteristics, I configured the two Z800s almost identically, with Intel Hyper-threading Technology (HTT), Intel Turbo Boost, and Enhanced Intel Turbo Boost Technology all enabled. Note the slight differences in the configurations of the two systems: The liquid-cooled system was running Intel''s new 3.33GHz Nehalem Xeons with 24GB of RAM, while the air-cooled system was running at 3.2GHz with 18GB of RAM. Both systems were configured with 64-bit Windows Vista.

I installed Adobe Creative Suite 4 (CS4) on both systems and, one at a time, started encoding a ballet project that consumed 100 percent of all available CPUs in the past. Then I hung a Shure SM57 microphone behind the system being tested to record 10 seconds of audio in Adobe Audition.

Then, in Audition, I grabbed a 10-second section of both recordings and used the Zoom-in Vertically tool to zoom in by 500 percent. The waveform for the air-cooled system— again, at 500 percent vertical zoom—is shown in Figure 3.

Figure 4. Waveform for liquid-cooled system shown with 500 percent vertical zoom.

Figure 4. Waveform for liquid-cooled system shown with 500 percent vertical zoom.

Figure 4 shows the waveform for the liquid-cooled system.

The waveforms obviously illustrate a significant difference in both volume and regularity. Lower volume is always good, and in a recording scenario, regularity makes noise much easier to remove via the noise-reduction modules found in most audio editing programs.

Beyond the waveforms, you can listen to the two recordings, though you''ll likely need headphones to really isolate the difference between the two. (Listen to the liquid-cooled system and the air-cooled system. Note that I boosted both recordings by 20dB to make them more easily audible.) In addition to volume, you''ll notice that the air-cooled system had a whinier, dentist-drill sound with lots of irregular noise as reflected in the waveform. In contrast, the liquid-cooled system was much more regular and lower-pitched.

One final point: Though the liquid-cooled system ran at a higher clock speed, which theoretically should generate more heat, it also ran slightly cooler than the air-cooled system as measured by Real Temp 3.0, a utility from Kevin Glynn. The difference wasn''t significant—about 1 degree Celsius cooler on all measured cores, but it''s nice to know that the liquid-cooling system delivered equal or better cooling than the noisier air-based system, since excessive heat can shorten processor life.

Real-world Experience

In addition to our tests, I wanted to hear from a user about why he opted for a liquid cooled system and his experience thus far. Dave Torres, an analysis engineer at Briggs & Stratton, was kind enough to answer my questions.

When did you start considering liquid-cooled computers?

Torres: I started to consider liquid-cooled computers when I was informed that it would be an option on HP's new lineup of workstations. I had previously seen gaming computers with liquid cooling but never really considered it as a serious option for workstations until HP introduced it.

When did you get them in?

We've had a liquid-cooled workstation for about one month.

What types of users are using them for which types of applications?

The workstation is being used by engineers to perform advanced computer simulations of engine components (CAD and CAE applications such as Pro/E, HyperWorks, Abaqus, and Fluent).

What''s been the experience so far?

The decrease in noise is a big benefit to users sitting next to a workstation all day. With workstations generating evermore heat, the fan noise has become significant.

Is there an ROI piece to this, or is the benefit improved user experience (or does an improved user experience deliver the ROI)?

I think that there could be a benefit of increased hardware longevity with better cooling, but that remains to be seen. The most immediate benefit is reduced noise for the user. In the end, I think that there is a ROI in having a more pleasant work environment, but that is subjective, and it seems that not every one notices the fan noise, or lack thereof.