Gave it a quick go following along with the video intro tutorial in DSPro 220.127.116.11 on Win7 64 (bottom of barrel AMD quad core/6 G RAM).
Sounds like a more than adequate setup.
Due to the nature that hair and fur is a complex system to replicate, having a older and slower machine will slow things down.
Garibaldi needs a graphics card that supports: hardware vertex/fragment shaders, framebuffer objects, and multisample antialiasing.
This means things are unlikely to function well with ati/amd or nvdia cards older than 5 years old. Any pre intel 'HD' (i series cpus) integrated graphics will not function.
CPU speed can make a big difference in render performance. The interface is designed to be responsive with any reasonable cpu (turn down the preview percent/mask small test distribution textures if you have issues).
A dual core cpu (like an intel core duo) will be fine, anything faster should be great.
RAM is currently a biggest roadblock. 4GB on a 64bit machine without many other open applications is great. Anything less will get a bit more tight when tring to handle any large complex setups.
Realistically I can't support machines with 2GB for version 1 of garibaldi express. But basic scenes will function fine.
The target is that as the beta versions progress things will become more opitimised make better use of resources.
I only tested on a default sphere (1 meter), and only encountered one issue. Not sure if it is a Windows version bug, or not.
When I got to the CLUMP tab, and tried experimenting there, nothing happened. I tried all the sliders (following along with tut), and my hair model exhibited no changes at all.
Unfortunately the sphere primitive is not really a good test object with the default settings. A 1 meter sphere has a lot of surface area. The value of the distubution settings will create aproximely the amount set per square cm. So a distribution of 10 will put 10 hairs in every square centimeter.
The clump amount needs to be less than the main hair distribution amount. The distribution of clumps on the characters surface use the same scale as the distribution of follicles. So if the clump amount is set to 1 then they will be about one clump in every square centimeter. Set fractional values for the amount will create larger clumps.
Both the main follicle distribution amount and clump amount need to have a texture set. As default the hair density texture will be automatically selected in a new scene. Grey values can be used in these texture maps to vary the density of the distribution over the characters surface.
I did find one thing interesting though. I scaled down the sphere, and hid it to use the hair on Genesis. When I scaled down the sphere to about the size of Genesis' head the more I scaled it down the more the hair scaled up. This was also true for individual X/Y/Z scaling giving interesting, and wild results. I know that is not intended use, but it might come in handy some day.
I guess you mean the hair stayed the same size... the garibaldi hair will try to retain its original shape while staying attached to the underlying geometry/figure. This behaviour make more scene on a figure/character i guess, but seems a bit weird when it is applied to a simple scaling sphere.