New Workstation Technology HP sent us its latest XW4100 workstation, and it's sporting some brand new technology inside. Revving under the hood is Intel's newest chip, the P4 3.0GHz C, running alongside a fast new 800MHz front-side bus. The new P4 C also uses Intel's tricky hyperthreading technology, making one chip act like two. All this comes in a pleasing package, a fine product that's the first new desktop from the newly-combined Compaq and HP. Let's take a look.First, let's just talk about the computer itself. The first thing to impress me? It's quiet. That's always an important consideration if you don't want that whooshing sound akin to a vacuum cleaner roaring inside your head all day. HP has achieved quiet nirvana with the XW4100 -- it's so silent, you can hardly tell it's on. Bravo. It also makes life easy for you with plenty of input-output accommodations, with 1394 (FireWire) inputs front and back, along with USB 2.0 front and back as well. It's easy to open, too, with a car door-like lever inviting you to open it up and take a look inside. It's not as easy to open as a Mac, though, but it's an improvement over the safe-cracking techniques needed to open lots of PCs. [an error occurred while processing this directive]
Anyway, there's a lot more nice technology inside this box, with the new 800 MHz front-side bus (FSB) being the star of the show. The fastest FSB available before was a 533MHz chip set, also offered on Intel boxes. This new 800 MHz version offers a wider data path, increasing the bandwidth of the machine so that it matches up well with a gigabyte of speedy PC 2700 DDR-333 RAM included with our test unit. Those numbers all sound impressive, but how much faster does all that technology make the computer run? Well, not much. Take a look at our benchmark tests (table below), the same tests we've been running on various workstations for about a year now, and compare this machine with others we’ve tested. It's faster, but its extra speed is bordering on statistical insignificance. We ran six After Effects benchmarks, each using real-world After Effects projects, three Photoshop benchmarks using smaller files that would be typical of those used by digital video editors, and a new benchmark we're using, CineBench 2003, which tests the computer's raw processor speed while rendering complex graphics. Surprisingly, there's not a whole lot of speed difference between this HP workstation and our previous king-of-the-hill, the Dell 3.06GHz P4 Precision Workstation 350 (which also includes Intel's hyperthreading) we tested a few months ago. I thought, because of the new 800 MHz front-side bus, we'd see a bigger improvement, but no. However, the HP was slightly faster than the Dell, somewhat significant because the new 3GHz P4 "C" chip is running 60MHz slower than the Dell's 3.06GHz P4.
The good news is, the hyperthreading (HT) inside these new Intel chips makes a notable speed difference. Hyperthreading is Intel's technique for making a single processor behave like two, splitting up the data into two parts and processing it separately and simultaneously. This machine, for some reason, arrived with its hyperthreading capability turned off, but a quick trip to the BIOS let us easily turn it on. I'm just wondering, though, how many new users will even know about this default-off state of hyperthreading? If you do end up with a P4 machine with hyperthreading -- that would be, at this writing, a box containing either the 3.06 P4 or this 3.0 P4 C -- be sure it's enabled in the BIOS, because I saw a nearly 20% speed increase with the feature turned on. For example, I ran our After Effects benchmarks with hyperthreading both on and off, and noticed an improvement on every test with hyperthreading enabled. This was especially noticeable in render-intensive routines like benchmark 6, which without hyperthreading completed in 5:37, but with hyperthreading enabled, shaved off more than a minute, at 4:15. 1 2 Next [an error occurred while processing this directive] ![]() |
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||