getting ready for windows 7

[hyster]

i forgot to mention u can uninstall IE8 in windows features

[MrRiplEy[H]]

I would think you would want to reformat a drive in the case a nasty virus hits. Its like going into orbit and niking the planet its the only real way to be sure. 

Skuzzy points out that reading and writing data at the exact same time is impossible but compare accessing two seperate folders on a single drive versus two folders on seperate drives for a moment. Particularly in the case of a dual partition the head of the drive has to move a great distance when compared to two seperate drives accessing folders that 'by design of zone filing' will be at the fastest read area of the drive. I will say that this too is very difficult for humans to see and realize the difference but with programs so slaved to I/O as we have today this can become a severe data flow problem that results in video stuttering and I have myself seen where using more than a single drive can really clear things up and smooth things out.

What's your point? The ability to format easily at will is the benefit I'm looking for with partitioning.

You're comparing apples to oranges with your multiple drive configuration. If the user has just one large drive he should partition it. Adding drives doubles the cost, partitioning costs nothing. I'm not disagreeing with having multiple harddrives - the grim reality is, however, that many people can't afford them. I have 5 harddrives in my gaming rig and 3 on my sons gaming rig.

[Chalenge]

Sometimes you can be so dense but I certainly understand being stubborn. 

The point is that using the small drive he ALREADY HAS for the OS is only going to be slow during boot and he can always use the other drive for programs.

You can keep spreading this partition foolishness all you want but IMHO it is just a poor mans way of having extra drives and YES it slows the system performance. I understand minke says he has enough invested in the computer already and does not want to spend more money but partitions are in my way of thinking a poor way to use a hard drive. The worse scenario I can think of is some process requiring a disk read while you are playing a game and suddenly you are stuck seeing warps and freeze frames and major stutters... might not be so bad if you are willing to put up with that but I am not.

[MrRiplEy[H]]

Sometimes you can be so dense but I certainly understand being stubborn.  

The point is that using the small drive he ALREADY HAS for the OS is only going to be slow during boot and he can always use the other drive for programs.

You can keep spreading this partition foolishness all you want but IMHO it is just a poor mans way of having extra drives and YES it slows the system performance. I understand minke says he has enough invested in the computer already and does not want to spend more money but partitions are in my way of thinking a poor way to use a hard drive. The worse scenario I can think of is some process requiring a disk read while you are playing a game and suddenly you are stuck seeing warps and freeze frames and major stutters... might not be so bad if you are willing to put up with that but I am not.

It's not foolishness. Well, except if you call Microsoft expert articles foolish.

If he plans to keep the OS on the old worn out drive then naturally he won't have to partition the new drive. Doing so will give him a penalty in OS responsiveness though. Especially if he doesn't move his swap file to the new harddrive.

Windows7 will do agressive caching and search indexing that will be negatively hit by using the old slower drive. Now that's a way to ruin your system performance and you're worried about partitioning? LOL!

[Chalenge]

Not too long ago a certain someone named Ripley told me that even the older 5200rpm drives have just as good a performance as 7200 drives with large memory caches. Different day different story I guess.

I know for a fact that multiple drives will allow for faster access of data especially over partitioning.

I went to school with some people that work at Microsoft Ripley. They are not infallible.

[Ghastly]

Multiple drives can help improve overall performance, if they are configured correctly to spread accesses evenly across them.  The reduction in seek times would be the improvement made.

While Skuzzy is quite correct that the actual transfer from the drives cannot occur simultaneously (and he knows more about drives that I do by a long shot) - I do believe that another design aspect that improves performance at times is that the drives can also simultaneously read data into their onboard cache, such that while the DMA transfer is in effect for one drive the others can be retrieving the data blocks requested until execution of the command queue is suspended due to the full cache - and thereby imparting a certain amount of parallelism into the slowest portion of the data read phase (the actual access of the platters).   You can't transfer the data off of the drives simultanously with standard controller hardware - but you can ready it simultanously.

What I don't know is how well Windows schedules the command queues to take advantage of this. I suspect it's quite poorly - especially since a Raid 1 device in Windows seems to perform marginally better at best and often no better than a single drive.  

What I also know is that in our testing and evaluation of the openfiler SAN (caveat and disclaimer - this was not a scientific test evaluation, but an in the field eval intended to resolve another issue), using the openfiler SAN (i..e a Linux distro based on IET) configured with similar drives and controller as an existing Windows Server with DAS outperformed the existing server by about 80% - despite the fact that the overall memory use of the NAS was in the neighborhood of 1 GB of the 4GB installed vs 4 GB for the Windows Server, leading me to suspect that there is a lot of unexplained overhead in the Windows 2003 disk subsystem on this particular system.  I was hoping for on par at best, and was very pleasantly surprised (I had expected the overhead of ISCSI to degrade performance slightly - I wasn't expecting better than the existing performance).  Whether there are instances where Windows performs better I can't say either.

All YMMV, IMHO, etc.

<S>

[Skuzzy]

There area number of esoteric ways to get some level of performance gains in the Windows disk subsystem.  True, multiple drives can fill thier cache simultaneously.  Unfortunately, Windows takes a very serial approach to reads from a disk drive.

Due to the above, many drives in the market make use of a segmented cache architecture and they continue to fill that cache after satisfying the Windows request.  The hope being they will be able to provide faster access to read data.

By using multiple hard drives, you statistically increase the chances of a cache hit occurring within the disk subsystem.

Of course, how effective this is, remains a hit and miss proposition as the Windows OS also provides read caching.

[MrRiplEy[H]]

Not too long ago a certain someone named Ripley told me that even the older 5200rpm drives have just as good a performance as 7200 drives with large memory caches. Different day different story I guess.

I know for a fact that multiple drives will allow for faster access of data especially over partitioning.

I went to school with some people that work at Microsoft Ripley. They are not infallible.

Please show me where I said that because I have never said anything such. This is the second time in a row that you make claims that are exactly reverse to the ones I made in reality. Either you're trolling or you simply do not understand what you read.

What I said was that a benchmark test showed a new 2Tb harddrive @ 5400rpm beating a smaller older 7200rpm drive. Which was and is a fact. It seems you still cannot understand the concept of data density and what it does to read/write performance. I'll give you a clue: it increases it proportionally with increase of density.

This is because the tracking head can pass more data per revolution compared to a smaller, lower density drive. If the data density difference is large enough (+ larger cache effect etc. taken into account) even the slower spinning drive can beat the faster spinning one in throughput performance.

As it is very clear that you cannot comprehend the technical details we're discussing here IMO it would be best to avoid commenting.

I'll give you a benefit of a doubt with Microsoft experts being fallible. Please prove the article I posted as being wrong from a professional, reliable source.