Lab test: Four Dell and HP workstations

Lab test: Four Dell and HP workstations

Our system and graphics performance tests show terrific price-performance at the low end of the quad-core workstation spectrum, and awe-inspiring power at the top; HP takes the bantam belt, while Dell is heavyweight champ

There was a time when workstations occupied a highly competitive niche in the hardware market. In those days, some 10 years ago, companies such as Sun Microsystems, SGI, IBM, HP, and Dell competed fiercely to deliver the top desktop systems characterized by powerful graphics and processing engines. An added element to this competition was the vendors' reliance on vastly different processor architectures to deliver the knockout performance. A decade later, the market segment is significantly different.

Today, the RISC processors that characterized the Sun, SGI, IBM, and HP machines are mostly memories, and all leading workstations are built on x86 processors. The number of vendors has also shrunk dramatically. SGI abandoned workstations, IBM's workstation division has morphed into Lenovo, and Sun -- a once-dominant player -- occupies a minor niche. On the one hand, this evolution has led to a market that is homogeneous in its product delivery and devoid of the intense competition of days of yore. On the other, workstations today deliver unimaginably more power than a decade ago at undreamed of prices.

To get a good cross-section of the market, we contacted the four principal vendors of x86 workstations, Dell, HP, Lenovo, and Sun, and requested dual-processor, quad-core machines as a midrange baseline. Only Dell and HP could provide such systems. Sun was in a prolonged holding pattern waiting for AMD to ship quad-core Opterons (Barcelona). And Lenovo, which is just starting up its workstation line of business, could not respond with the configuration we wanted within our time frame. We expect to review machines from both vendors in the next few months.

Meanwhile, this lab review focuses on major players HP and Dell, who together own the lion's share of the market. We asked them each for a midrange system costing US$5,000 retail, and we required they use identical processors, the same amount of RAM, and the same releases of Windows XP so that we could compare what other magic they could add to this base and still stay at US$5,000. As we'll see shortly, the magic looked a lot alike. We also examined a value-oriented system from HP and a very high-end workstation from Dell. Curiously, the high and low ends appear to represent the best values, while the midrange serves as a refuge for those who don't want to spend for the high-end Dell, and whose needs are not met by the entry-level HP.

The benchmarks

The benchmarks I used in this evaluation represent a departure from the metrics traditionally used. I decided to rely solely on high-quality benchmarks available at no cost to readers. This approach gives excellent quantitative insight, while allowing you to reproduce our tests on your systems.

I used the Sandra XII benchmark suite from SiSoftware to measure component performance. This suite is probably the best and most comprehensive suite available for x86 desktops and servers. A free personal version is available from SiSoftware's Web site. Graphics were tested using the highly regarded and freely available ViewPerf 10.0 benchmark suite from SPEC (Standard Performance Evaluation Corporation). ViewPerf works as a Windows-only series of tests that assesses a system's video capabilities. Its final mark (the average of five benchmark results) is influenced by both the graphics card and the system processor. I ran the Quad version of ViewPerf 10.0 on all the workstations.

Follow Us

Join the newsletter!

Error: Please check your email address.
Show Comments