Up to this point I think the only requirement known is that DirectX 9 is a must as rFactor 2 won't support DX7 or 8.
That is correct! Gjon / ISI always wants that rf / rf2 will be accessible to almost everyone. I think that even quiete old computers will be able to handle rf2 if they support dx9. They would just need to lower details and cannot drive with >20 opponents or something similar. Once the development is in the final stage i am sure they re able to offer minimum requerements. I would say a low end system which had already issues with rf1 will not work propably with rf2. A System which could handle rf1 with medium details should be able to run rf2 on lowest details. Minimum system reqiuerements will not be huge. Recommended requierements to run all stuff is another story .
I just hope they make it where it will make use of multi core processors. I upgraded to a dual core last year, and would like rf2 to use it to its full potential )
It should be possible. It is even possible to use rF1 with multi-core when you add the modifier +fullproc So I imagine 4,6 and even 8 core support SHOULD be implemented.
I don't think it is needed any longer. At some point there was a windows bug, perhaps in XP or Vista. don't recall which, but to circumvent this bug ISI added this option. So AFAIK there is no benefit nowadays on win7 to do that.
During recent tests, me and my friend noticed there is performance gain with "+fullproc" parameter. Not only in fps in general (about 20%) but also during huge collisions, when you loose less fps with that parameter added. Tests were done on i3 530 and i5 750. Tested GPUs were: Ati HD6950 (and modded to 6970), HD5750 and nVidia GF GTX 295. OS is Windows 7.
wow, find that very surprising, that much ? perhaps I should be putting it back in, given that rF1 was designed to use a single core CPUs by design. there is some threading but fairly minor I do believe. I wonder if it is the HT support as the core 2 duos, quads are not HT, I never noted much difference worth talking about.
Nothing is designed to use "single" core, there are just applications which are optimized to benefit from 2 or more cores. Its up to the os-scheduler to share the load to the available cores. But yeah rf1 cannot benefit much from multicore, but i always used the +fullproc switch.
True enough Mr Purist I was coming form the point if running multiple threads, rF1 does not do much of this, but worded it badly.
Of course the less fps you have, the less benefit you will be having from +fullproc (beside situations with heavy CPU load, like during massive collisions). So don't expect you will get 36 instead of 30fps, because at this point graphics card will be a bottleneck.
That's not entirely true. Code can benefit a lot when it can assume that it will only be run on a single core, not having to worry about writing data in CPU caches back to memory or synchronizing access to data structures that can be used by more cores concurrently. So in general a piece of code that is designed to run on only one core will be quicker than the same piece of code that can take advantage of multiple cores. Of course as soon as you have cores and there is work to distribute amongst them, you will still gain a lot. For rFactor (1 or 2) there is also the trade off that neither will (probably, nobody knows about rFactor 2 yet) require a multi core CPU, so the code still has to work on a single core as well.
There is also a patcher that allows the rFactor.exe to make use of more ram than it's original 32bit origins allow