I asked Kevin the same question and his reply was to NOT change anything in the acdvanced tab unless you have difficulties. He also refrained from mentioning the exact intention of these settings, so I'm still in the dark about these settings. I have no idea what can be achieved in terms of performance by altering some of these settings. I hope someone can fill us in.
Those are actually some very nice control settings especially with the current HDD's shipping with bad blocks. I may have to email the Intel raid engineers and request the same. Basically you have 2 sets of settings there. 1 Set controls the Timeout and error recovery options that controllers have with enterprise drives. Often times even though the enterprise drives have the timeout feature it does not kick on in time before the raid controller marks the drive out as bad. Those settings allow you to adjust that if that becomes a problem. The other set has to do with caching models related to AV streams. AV streams require constant communication that is consistent and mechanical drives have to handle those with the latency every mechanical drive has for seek time due to the way they work. Those settings help you optimize the caching model to best handle multiple consistent streams of data that have to remain consistent. You can always play with these settings and run some AV tests along with drive benchmark tests to see if there is some optimal manual configs. I would not adjust the Controller timeout settings unless you start having drives drop out of raids in a short period of time.