Systems' setup and notes
We took a look at an Intel Tigerton system here and compared the quad-processor setup against a dual-processor Clovertown system. We're now using those Clovertown numbers by which to judge the Stoakley setup by.
Hardware
| System | Harpertown - 2 socket | Clovertown - 2 socket |
|---|---|---|
| Processor(s) | 2 x Intel Xeon E5462 (2.8GHz, 12MiB L2 cache, 1600FSB, LGA771) | 2 x Intel Xeon X5365 (3.0GHz, 8MiB L2 cache, 1333FSB, LGA771) |
| Motherboards | Supermicro X7WDA | Intel S5000XVN |
| Chipsets | Intel Seaburg | Intel 5000X |
| Memory | 8 x 2GiB FB-DIMMs (800MHz) Only 3GiB addressable due to OS |
4 x 1GiB Micron FB-DIMMs |
| Graphics card(s) | NVIDIA Quadro FX 4600 | ATI FireGL V3200 128MiB |
| Operating system | Windows XP SP2 (updated), 32-bit | Windows 2003 Server Enterprise Edition R2, 64-bit with Service Pack 2 |
Software
| 2D Benchmarks | SiSoftware Sandra XII LITE CPU arithmetic SiSoftware Sandra XII LITE CPU multimedia SiSoftware Sandra XII LITE memory latency POV-Ray 32-bit 3.7.0 BETA 21 - internal benchmark mode CINEBENCH R10 |
|---|
Testing notes
A few things to note here. The extremely limited time available to benchmark the Stoakley system dictated that only a few benchmarks could be run.
We're nonplussed by the Stoakley's operating system. We're not sure why a 32-bit OS was used in conjunction with a server/workstation-oriented machine, and benchmark results will be negatively impacted by it.
Please also note that we're looking at the performance of the fastest Clovertown processors and referencing it to the third-rung Harpertown CPUs.
Lastly, our benchmark suite does not include any SSE4.1-optimised applications. We'll have to wait until we can test a Harpertown system in a HEXUS-controlled environment and with more time at our disposal. In short, the following benchmark results may not be fully indicative of performance on a properly configured system.
