320 GB SATA2 drive slower than a 160 GB SATA2 drive ?

[Guest jahman]

>I don

[Guest baksteen33]

Thanks Jahman, appreciated. FWIW, my internal statics say, if stutters are reported etc, it's often either too little RAM, or (single) partitioned harddrives resp sub ideal ATA setups. The latter two are - at first - obviously a big hickup, but when the implications are understood and everything is in place again, this almost always leads to positive results. @ Nbouc, the HD should be able do better in the transfer dept, AFAIK, ca 80MB should be possible on the faster sectors. The operation might not be complete yet? You might be using a default Windows driver? Is there an 'ICH' mention in the device manager? Kind regards Jaap

[NBouc 54]

Hi Jaap,There is no ACH mention in Device Manager. The controllers are listed under "IDE ATA/ATAPI controllers" as:Intel® ICH9 2 port Serial ATA Storage Controller 2- 2620Intel® ICH9 4 port Serial ATA Storage Controller 1- 2626Since it is an X38 chipset, I was expecting to see "82801IR" somewhere in the name, even though I have not set AHCI nor RAID.The drivers for the controllers are identified as:- device provider: Intel- driver Date: 05/02/2007- driver version: 9.3.0.1011- digital signer: Microsoft WHCPThe drivers details show them from Microsoft:- atapi.sys Microsoft V5.1.2600.2180 (xpsp_sp2_rtm ...)- pciide.sys Microsoft V5.1.2600.0 (XPClient ...)- pciidex.sys Microsoft V5.1.2600.2180 (xpsp_sp2_rtm ...)Would I have better performance updating them with the Intel's F6 disk drivers, considering I am not using AHCI nor RAID ?

[Avcomware 1]

Hi, jahman.This thread is going off topic, but I want to clarify some of the misunderstandings about the

[Guest baksteen33]

Hi Nbouc, My suspicion is you're running standard Windows drivers or those from the 'inf'. The latter most likely. Also, Intel's 'HD chipsets' have some featuritis, I frankly couldn't tell whether it makes a difference or not in your case. The HD itself however, could do much better than ca 60MB/s on the fast sectors, more like 80. As you certainly know, the best and most flexible Intel interface comes with an ICHx plus a 'R' suffix --> i.e. ICH9R. What I say is about chipsets with this 'R' only. If you set it up via F6 it will look like the enclosed picture. It used to be possible to inject AHCI drivers in the aftermaths, but I don't know whether it is still possible? Maybe somebody else does? ACHI, Raid, ICH and software support is a pretty messy business, when Intel was under more pressure they were less rigorous sometimes. FSX means transferring much more data, you can't afford thrashing any longer. ;-) Particularly, IMHO, make sure FSX is on the fastest bits of the drives and not on slower secondary partitions. Best wishes and kind regards Jaap PS, don't forget to reduce the garbage bin with XP, particularly if you make a large partition. ;-) The old, XP garbage bin formula = 10% of the partition... Vista is more intelligent but I would recommend to lessen the bin there too. FWIW, I always use a ca 400MB. Edit: I forgot to mention, sorry Nbouc, if you go AHCI/Raid and F6, you should enable the interfaces in the Bios before going ahead. Make sure the controller is set to AHCI or whatever you choose. ATA drivers and bios settings must match each other.

[Guest Nick_N]

AgreedFSX is a file access monster.With RAID0 and FSX you must disable NCQ and other polling/optimizer settings or you will LOSE performance in RAID.. a massive loss in real world results too.FSX (especially with add-ons) incorportates a huge average file size and it would be considered a random access file system, therefore it requires a minimum of 256K STRIPE (higher is actually better) and all the network optimizing crap disabled to get the most out of it, and, a real RAID card not motherboard RAID and certainly not some cheap 100 dollar card solution.Geometry and the basic specs of rotation speed, access time and file placement are the best way to go in a FSX system. A properly spec'd single high performance drive will out perform 2xRAID0 in FSX on motherboard RAID, out-perform meaning less overhead, more CPU going to the application and less storage related file-call stuttersperiodThe right hardware RAID solution with the right configuration of it will overcome the limits of motherboard or cheap RAID and it will in fact allow a greatly increased performance curve to emerge in FSX. Otherwise a single drive properly configured and laid out is the best solution, with all the fancy optimizing in the BIOS shut down. My opinion is those are great for networks and BAD for FSX.Every test setup I have run in the last year and a half prove that. There is a huge difference between the theory, science, math, engineering, and the end result. Usually the 'experts' are dumbfounded after they are proven wrong and then after the fact, go figure out where their statement was in error or needed to be 'adjusted" for the application being run and in that they say.. 'well, I was not wrong.. I just did not understand the application"LOL

[Guest baksteen33]

Hi AV, We've had these discussions in the past and I'm not going to go any further than saying, your thoughts omit at least one very important factor IMsnsHO: HD-head travel instead of reading and writing, also called thrashing. Besides, secondary partitions are always slower and will get less MBs across... Please don't tell me you want to debate that as well, dear AV. Best wishes and kind regards Jaap

[Guest baksteen33]

>>Geometry and the basic specs of rotation speed, access time and file placement are the best way to go in a FSX system. A properly spec'd single high performance drive will out perform 2xRAID0 in FSX on motherboard RAID, out-perform meaning less overhead, more CPU going to the application and less storage related file-call stutters<< Ooohhh Nick, stutters, my almost daily day and night support nightmare! :-) I particularly liked the 'overhead' bit. Very true. Cheers and kind regards Jaap

[Guest Nick_N]

One thing I see very little of is stutters however the combination of my refusing to see less and push FSX as far as possible on my systems, and, the nature of the application being ahead of its time for the hardware (note: I also use a SS HDD RAID system too.. love it) I still see the occasional stutter here and there.. that will clear up over the next year, I am sure of that.I try to keep it simple and chuck out all the engineering jargonIt use to drive my staff engineers nuts because they would love to complicate the #### out explaining something as simple as replacing a light bulb. Nature of the beast and it took me about 20-30 years to get over it too. I use to let them get away with "I did not understand the application" then I got sick of that BS and started nailing them for it. Under me they either think, or sink.There is a time and place for going in-depth and working the problem out through engineering and theory.. then comes the real world and although one will directly apply to the other but I have found more often than not (and although not perfect) the approach of Occam's Razor is ususally the best way to attack a problem, figure out the issues, and solve it. Keep It Sweet and Simple.

[Avcomware 1]

Hi, Jaap.

[Guest baksteen33]

Hi AV, There's no need for science really. I advise people on an almost daily base to get rid of such setups. If the results weren't overwhelmingly positive, I wouldn't be here saying so. ;-) Partitioning drives is too much Vodoo, better to keep it simple and efficient. Feel free to partition a single drive, of course, but never run an app and the OS off two different partitions on the same HD. Telling it's fine to split app and OS like this is causing a lot of developers a couple of headaches. FSX is a support nightmare if you can't count on a couple of dependencies. Last week I had 3 out of 3 cases using non default paths which they - literally - couldn't manage any more. OTOH, you hardly ever hear anything from those keeping it simple. Strange? Not really IMHO. With FSX, 'relocating' requires a different level of expertise. Kind regards Jaap

[NBouc 54]

Hi Jaap,I am indeed running the drivers from the INF. I tried to switch to the F6 AHCI drivers but guess what ... my floppy controller also seems to be malfunctioning. Unable to read/write floppy disk.I replaced the floppy drive with another but same result, thus my suspicion towards the floppy controller. Faulty SATA port #1 plus faulty floppy controller, this definitely calls for a Mobo replacement.BTW, I did once a driver update from IDE to AHCI on my previous system after Windows was installed. Not too complicated under Windows XP 32-bit, but I don't know about other OSs. In short:- update the driver via Device manager pointing it to the F6 disk.- select the correct controller.- reboot and enter BIOS.- set SATA as AHCI, save change and exit.- boot into Windows.Intel Matrix Storage manager can then be installed if one whishes to confirm NCQ is active.Take care, I must pay a visit to my Motherboard vendor ;-)

[NBouc 54]

Your level of discussion is way past my competence, but one statement got my attention. When you said;"Usually the 'experts' are dumbfounded after they are proven wrong and then after the fact, go figure out where their statement was in error or needed to be 'adjusted" for the application being run and in that they say.. 'well, I was not wrong.. I just did not understand the application".Have you been working with my ex-wife ?LoL

[Guest jahman]

Jaap,I've been following your arguments and I could't have said it better meeself! :-)Cheers,-jahman.

[Guest baksteen33]

Hi Nbouc, I was looking for some information and looked on a bit. It still seems possible to implement AHCI in the aftermaths. Here are a couple of links with general information, first, MS themselves on enabling the AHCI driver in Vista post install: http://support.microsoft.com/kb/922976Or this one: http://forums.pcperspective.com/showthread.php?t=444831Another one with links inside: http://www.omgili.com/omgili.search?q=inst...vista+ahci+ich8I've always done these things for scratch, so please excuse my 'injection negligence'. :-) Kind regards Jaap