Yes, that's right, one week before the Athlon 64 processor is due for its introduction to the public, we have a 2GHz K8 chip ready to take on a Pentium 4 3.2GHz. One might think this could be an enlightening preview of things to come. And one might be right.
If you haven't read our workstation comparo, you'd best do so now, because I'm not going to rehash all of the Opteron's improvements over the Athlon XP it replaces. I won't be talking about its SSE2 support, or its slightly longer pipeline, or its AMD64 extensions for 64-bit computing. AMD's new 0.13-micron SOI fab processo won't be discussed here. And I especially won't be talking about the Opteron's on-chip memory controller, with dual channels of DDR memory. Nope, none of that. So you'd best read that other article.
I will, however, mention that we've tested the Opteron 146 with dual channels of DDR400 memory, upping the ante from the DDR333 speeds we used for our previous article. Registered DDR400 memory ain't easy to come by, but we managed to snag some for use with the 146. The move to 400MHz memory ought to put the Opteron 146 at rough parity, memory-wise, with our Pentium 4 3.2GHz test system, which also has dual-channel DDR400 memory.
So the stage is set. The Opteron's core and memory clock speeds have both been ratcheted up, and the Pentium 4 over in the corner is starting to look a little nervous. Let's see what happens next.
We are blatantly recycling the benchmark results from our workstation comparo, so you'll see a number of interesting workstation configs compared to the Opteron 146 today. Obviously, the most relevant competition is the Pentium 4. We'll also be intrigued to see how the Opteron 146a K8 at 2GHzmatches up against the Athlon XP 3200+a K7 at 2.2GHz. AMD's K8 core seems to have improved clock-for-clock performance in a range of scenarios, and we'll be watching to see how the K7 and K8 compare.
As we noted yesterday, one facet of our tests that may send market segmentation fetishists into a tizzy is our use of a non-workstation-class GeForce FX 5900 Ultra graphics card. Never mind that there's a Quadro FX based on the same chip to match it, the card we've used for testing doesn't fit the workstation mold, and we own up to it. Truth is, we blew our budget on the workstation-class processors, instead. The better-tuned and certified OpenGL drivers that come with a Quadro would only have made a difference in a couple of our benchmarks, anyhow: SPECviewperf and the OpenGL portions of Cinebench. Otherwise, most of the tests we've run don't use real-time 3D graphics.