Just conducted a little test using CrystalDiskMark 3.0.1 x64 with a Corsair 64GB C300 on my ASUS P8P67 Deluxe. Here's the results for the MARVELL 6GB port... Code: Sequential Read : 344.473 MB/s Sequential Write : 55.078 MB/s Random Read 512KB : 315.742 MB/s Random Write 512KB : 53.995 MB/s Random Read 4KB (QD=1) : 26.197 MB/s [ 6395.8 IOPS] Random Write 4KB (QD=1) : 39.230 MB/s [ 9577.5 IOPS] Random Read 4KB (QD=32) : 192.378 MB/s [ 46967.4 IOPS] Random Write 4KB (QD=32) : 66.061 MB/s [ 16128.1 IOPS] Here's the results for the INTEL 6GB port... Code: Sequential Read : 361.130 MB/s Sequential Write : 54.011 MB/s Random Read 512KB : 317.750 MB/s Random Write 512KB : 64.234 MB/s Random Read 4KB (QD=1) : 29.853 MB/s [ 7288.4 IOPS] Random Write 4KB (QD=1) : 41.736 MB/s [ 10189.6 IOPS] Random Read 4KB (QD=32) : 216.583 MB/s [ 52876.6 IOPS] Random Write 4KB (QD=32) : 79.158 MB/s [ 19325.8 IOPS] INTEL wins! I could've swore I read in a Bit-Tech article they recommend the MARVELL ports over INTEL for 6GB drives? Never mind, at least I now know which to connect my top drives to in the future!
That could well have been with the X58 boards where the intel was/is only 3Gb/s... The current Marvell chipsets on most boards (the 9128) are just shonky for SSDs (inc 3Gb/s ones)... ...whereas the 9182 is better, but nowhere near as common or as good as the intel. [NB the 9182 will allow okay-ish lower cost pcie solutions at some point - though with a bit of added latency i'd assume.] in fact, unless you're going to go over something like 2x V3s in R0, the 6Gb/s intel is arguably better than expensive raid cards. At 2 SSDs, there can be slightly lower sequential speeds but you gain on the 4Ks & latency - whereas above 2 then, as the intel has reached it's limits for sequentials, the gains that can be made by using a 'proper' raid card are huge. [NB the same basically applied with the 3Gb/s intel vs a 'proper' raid card - there was a small gain with my 2x V Turbos on the lsi 9260 over the intel, but the gains when they were swapped for 4x V2s was huge vs the intel.]
they probably used a different method to bench too.. now you can load up big butts II a little faster!
Did you run the benchmark at least 3 times for each? Where you pick the one that gets twice the same results? Did you ensure that you have the latest firmware of the SATA controller and driver for both, and latest BIOS? My experience with the Intel X58 controller 3Gbps is that it ain't good. During intensive HDD usage, my audio on my dedicated sound card skips (no latency problem), and already caused problem where it broke Windows and corrupted a project of mine. (lucky I do had an hour backup done, and did a image of my disk using Windows 7 the night before). I had enough of it, and decided to use my Marvel SATA 6 one, and since, 0 problem of any kind. Just for kicks, I decided to move my HDD back to the Intel one (no need to re-install as my SATA controllers, both, are set to AHCI), and the audio skipping is back. I have confirmed this problem with a friend with the Xonar, instead of X-Fi. By audio "Skip" I don't mean continuously skips.. it just skips maybe once or twice when it loads a big game save or level, when music plays, or 2-3 times when I defrag my HDD, while playing music. In my book, this is unacceptable. So thanks, but no thanks, I'll stick to my Marvel SATA controller any day now over Intel's. If my old 2005 (same GPU, same sound card, Win7 64-bit, dual core (AMD Athlon X2 4400+ Socket 939) 3GB DDR1 RAM, nForce 4 32XSLI chip, system with it's SATA II controller, had no problem like that... than something is definitely wrong with the Intel controller.
I just did a single test for kicks really not an exhaustive testing session! Each port gives blistering speeds compared to my mechanical HDD so I'm not that fussed! Interesting info on the sound issue GoodBytes, I'll be picking up a Xonar myself so I'll be listening carefully for any sound distortions whilst using the INTEL 6GB port.
Unless there's something odd about the mobo, then that's unusual... Driver conflict perhaps - & assume you're using the latest RST intel drivers? Or maybe buying Gigabyte rather than Asus. As an odd additional thought, have you tried going to - Control Panel (All CP Items) -> Sound (new popup window) --> Communications ---> & change the setting to "Do Nothing" - just there's a chance that it 'might' solve the problem (or at least rule it out as the cause). Otherwise, whilst not exhaustive testing, it's in line with all the other tests that have been done by various people & sites.
I think I mentioned before but when Intel's controller is running at full tilt, it seems to take priority over most things. I don't get audio skipping with my i5 based Intel controller and X-FI Xtreme gamer, but I've noticed it makes my system less responsive than my old Nvidia IDE controller did.
Latest BIOS, firmware for both controllers, and drivers, even tried Gigabyte site drivers, and 1 version older. If Citizen (not the watch maker) ever made a motherboard at 10$, I'll prefer it over ASUS... it least they have 2 employee working there, over 1 at customer support. Already disabled since day 1, because I have multiple sources linked to the PC, and Skype screws everything up if it's running. Maybe Skype latest version fixed it's problem, I never tried.. but in case it's disabled already. I am not saying that the benchmark are wrong... I am saying I would prefer to get SLIGHTLY slower speed by going with the Marvel one, over the not so great Intel SATA controller (the controller itself or the drivers).
The i5 doesn't have the problem, which I can confirm with my brother Gigabyte powered motherboard, which feature the X-Fi Xtream Platinum and a Core i7 860.
We both have the same results. No, did not swamp cards. The issue is not problematic as it rare occurrence... no one defrag his HDD everyday and plays music at the time. I COULD live with it, but this problem is still unacceptable in my book. People might not care, but I do. I seek for the best computer experience possible (without breaking the bank account, and being smart and reasonable in purchasing things, of course). If the issue occurred with my old 2005 computer, I would not complain.., but it did not, and I think that in 2010 (or even 2009) it's unacceptable that a company doesn't pay attention to details and get away with it. Example: RIM PlayBook, great device, excellent device on paper... but when no attention is paid on any details... it's a crappy product at the end of the day, and no one buys it.
Have deleted the rest of your reply as god knows what's causing that - have never had any issues with sound cards & SSDs/HDDs on any controller so... ...though my last non-Asus board was probably back when i had either a x486 or an (original) Celeron - was obviously only joking about it buying one over the other... ...but then i don't need customer services particularly & buy to my needs (with the SB build for example, Gigabyte didn't have a board that did everything i needed - though did look). Also wasn't suggesting that you hadn't tried sensible things to rectify - hence suggesting the one odd thing i could think of that wasn't down to an unusual driver/software conflict... Well, it would have surprised me greatly if you'd not updated the bios, tried the latest drivers (from intel & whatnot rather than Gigabyte), set the pcie speed to 100 (as this can cause cards to become unstable), disabling any power saving modes (both in the bios & windows), etc, etc... [the vague list is if anyone else is reading as, whilst i actually do assume that you know what you're doing, other people might find it useful] Anyway, what i was picking up on was your orig comment to the OP - "Did you run the benchmark at least 3 times for each? Where you pick the one that gets twice the same results? Did you ensure that you have the latest firmware of the SATA controller and driver for both, and latest BIOS?" - where, whilst you're completely correct that all of this & more; for example - the drive should have been SE'd before & between b/ms so that every cell started clean, since the test is of the controller not the SSD (since it would appear that the Marvell controller was tested first, the intel scores 'may' be on the low side d.t. dirty nand skewing the results) & you'd choose the fastest possible SSD (&/or 2 in R0) since the Marvell controller has far lower max sequentials; rather than a (comparatively slow) 64GB C300 - should be done if the OP were a review site... ...the OP was simply showing the kind of results that you'd expect from that SSD in r.l. usage between the 2 controllers. So not perfect but illustrative - not least since the sound card issues that you're having appears to be very anomalous; there not appearing to be some obvious site/forum full of people who are having the same problem (on the basis of a very quick google search). Oh, & i do defrag semi-constantly - well forgetting all of it's snazzy preemptive stuff, Diskeeper# uses idle time to defrag, so... [#other defrag software is available]
