Greetings,
Not sure why the gentlemen in the post from 2012 suggests you will have "long term problems" connecting your boot drive to the 9128 controller. His claim is unsubstantiated and vague. That said, searching google shows a mixed bag of reviews for the controller.
What is the brand and model of your SSD?
For comparison (example)
Samsung 850 EVO 512GB 550/520 max R/W
Samsung 850 Pro 512GB 550/520 max R/W
There is always some overhead.
Suggest you do the following:
Do 3 or more consecutive tests with Crystal Disk Mark. Take the results off all tests and divide by 4. Use that as a baseline for the Marvel controller.
Try one other test. Use the largest file you have and copy it back and forth between the SSD and array. Note the time and throughput.
Now move the SSD to the Intel controller. It should operate as standalone SATA in AHCI mode, even with the operate mode set to RAID. You might have to use a certain port (refer to your manual) don't recall.
Perform another round of benchmarking and see which is better. Overall R/W.
Then test the file copy operation I mentioned above and compare. Just curious to see the disk I/O between different controller and with the disks on the same controller.
This won't hurt anything and you can move the SSD back depending on the results.
One other
important thing to note. Are you using microsoft's SATA driver for the Marvell Controller or the Marvell driver? Marvell's driver might be better.
https://www.gigabyte.com/Motherboard/GA-X58A-UD3R-rev-20#support-dlNote: You can use System Restore to test the various drivers without any negative impact to your system.
Also, please be sure to install Intel chipset .inf drivers as this can also improve overall system performance.
Please report your findings. I don't need images of the benchmarking results, just your written findings what worked best.
**Edit, just watched a few youtube videos from 2012, consensus seems to be the slower Intel controller out performs the Marvell, but I would do my own testing to confirm.