These leaked results are coming in the same manner as the months-early AMD Interlagos numbers, the yet-to-be-out Trinity, early Ivy Bridge numbers, and other pre-production benchmark results from engineering samples... An engineer -- either unintentionally or willingly (it may have been very well intentional this time seeing as it was just days after I was ranting about the poor Chrome driver support) -- within VIA China pushing out some early results when validating a system under Linux using the Phoronix Test Suite. While nearly every major IHV/ISV is using the Phoronix Test Suite (and/or related components like OpenBenchmarking.org and Phoromatic), it seems a select few companies/engineers always enjoy using it to spread early performance results to the world by uploading their Phoronix Test Suite benchmark results to the collaborative OpenBenchmarking.org platform, which is another Phoronix product.
New VIA/S3 results of unreleased hardware were last pushed up to OpenBenchmarking.org at the end of December. I hadn't noticed these early results since usually those VIA benchmarks tend to not be interesting, but it's been picked up by a reader in the Phoronix Forums and other blogs.
When searching for the Chrome 600 there are the first of these Linux numbers. These results are from a system that has a VIA Nano 1.2GHz Quad-Core, a motherboard exposed as VIA VT3456 VT8611BMB, 2GB of DDR3-1333MHz system memory, and S3 Chrome 600 graphics. Ubuntu 10.04 with the Linux 2.6.32 kernel was the system in use.
Initial benchmark results on the Chrome 600 are rather disappointing. Enemy Territory at 800 x 600 is running at about 24 FPS, Lightsmark at 800 x 600 is just at 11 FPS, OpenArena at 800 x 600 is short of 60 FPS, and World of Padman at the same resolution is around 43 FPS. Simply put, if you haven't run any of these Phoronix Test Suite tests previously, the latest-generation of AMD APUs can easily smack the hell out of the Chrome 600 graphics. Here are some AMD Llano and Brazos benchmark results for reference.
These graphics numbers are improved over the previous-generation Chrome graphics, but still less than what's needed to save VIA Technologies. On the low-end VIA not only has great competition now from the AMD APUs and Intel's Atom (including Medfield), but there's now some compelling ARM platforms too even with good ARM graphics.
As far as other details on the Chrome 600, OpenGL 3.2 is the exposed GL version (at least what the driver is supporting), and the S3 binary blob graphics are still in use. It appears that the PCI device ID on this new chip is 0x3a01 (the vendor ID is still 0x1106) with the s3g kernel module. The VIA VT3456 is believed to be what will be marketed as the VIA VX11.
The audio on this "VIA VT3456 VT8611BMB" platform is still with a 0x3288 device ID and the HDA Intel driver, which is the same ID for the VIA audio adapter found on other motherboards. The network support also doesn't appear to be new but its device ID is 0x3119.
Don't expect open-source graphics drivers for this upcoming hardware (their open-source strategy was defenestrated, but they have promised a few open-source gifts) and you can only expect their binary graphics drivers to be rarely updated.
Still working out more details on these results and system information out of VIA China, expect updates.