Disk Partitioning Advice

Once again, Walter Mossberg has offered incomplete and potentially dangerous computer advice in the Wall Street Journal. The December 6, 2007 edition of Mossberg's Mailbox had a question from someone whose lone hard disk was divided into two partitions; a small C disk that was almost full and a large D disk with lots of available space. The questioner asked about merging the two partitions together. Mr. Mossberg said that Partition Magic can be used for this purpose and that it "works well".

It is malpractice to suggest changing partitions in any way shape or form without first making a disk image backup. When things go wrong, as they inevitably do, you can lose access to all the files in a partition.

I jumped on the Partition Magic bandwagon early. In the late 1990s, before the availability of virtual machines on PCs, we used it in an R&D lab to run multiple operating systems on a single computer. For years I have used it on my personal machines for a host of reasons.

Partition Magic has it's fair share of quirks and problems, not the least of which is that it appears to have been abandoned by Symantec. The Partition Magic gripes at my computergripes.com site are consistently the most popular topic on the site.

Among the operations that can be performed on partitions, combining two of them is perhaps the most dangerous. It is more complex than resizing a single partition and is a relatively new feature. Personally, I never attempted it, both because of the risk and because there are other ways to accomplish the same thing.

In this case, I would shrink the D partition to the minimum allowable size (plus a small fudge factor for good luck), then enlarge the C partition to include the space just given up by the D partition.* Next, I would copy all the files from D to C, then wipe out the D partition and, finally, expand the C partition so that it takes up the whole hard disk.

But, before combining partitions, I would look to avoid the whole thing by moving files from the C disk/partition to the D disk/partition.

Some of the poorly chosen Windows defaults that I mentioned last time, can be tweaked to free up space. For example, the Recycle Bin defaults to 10% of the partition in Windows XP and System Restore claims 12% by default. The minimum for System Restore in XP is 200 megabytes, give it 300 or 400 and you will probably reclaim many gigabytes. Internet Explorer also consumes large quantities of hard disk space. I doubt you will notice any change if you limit the IE cache to 30 or 40 megabytes.

Windows Update creates folders in the C:\Windows folder with names like $NtUninstallKBxxxxxx$. The total uncompressed size of these folders was 245MB, 285MB and 536MB on three different Windows XP machines that I checked. These folders can be moved out of the C disk/partition, as they are only used to un-install bug fixes. If there is a large collection of pictures, music and/or videos, they can certainly be moved to free up space. Finally, there is the Disk Cleanup feature of XP that exists for just this purpose (get the Properties of the C disk, it's a button on the General tab).

Partition Magic is also expensive. Similar software, GParted, is available for free in Linux (see sample screenshots). You can boot your computer using a Linux Live CD and run GParted that way. I have done this with Ubuntu and Knoppix but many other Linux versions/distributions also include partitioning software.

You don't read PC magazine for mutual fund advice and you shouldn't read the Wall Street Journal for computer advice.

* I'm simplifying things a bit. There is actually another necessary step - after shrinking the D partition, it has to be moved to the right before the C partition can be be expanded. Also, if after this shrink/re-size operation all the files from the D partition don't fit onto the C partition, then another round of shrink/resize would be needed. Backup, backup, backup.

Six top hard disk drives for speed and capacity

You have to ask yourself whether you want speed, battery life or capacity?

Over the past few months, Computerworld has reviewed the latest in hard disk drive technology, from large-capacity 1TB models and energy-efficient "green" drives to solid-state disks that outpace their spinning disk counterparts on every dimension of speed and power consumption. For your perusal are six of the best drives we could find that you might want to consider for your computer or a friend's this holiday season.

Review: Samsung's 64GB flash hard drive

The Samsung 64GB SSD drive

The drive has the capacity to be your workhorse and is heartier than others in that it can withstand a 1,500G shock, and it has a mean time between failure of over 2 million hours, versus under 500,000 hours for the company's other drives. It sips power at just 1 watt when active, 0.1 watt when idle, and 0.06 watt in standby mode.

Review: Samsung Electronics Co.'s 32GB, 1.8-in. SSD drive -- the fastest yet

The Samsung 32GB SSD drive

At 32GB, its capacity hardly competes with the 1.8-in. hard drives inside current-generation 60GB or 80GB iPods or even the little 40GB drive inside the iRiver H340. As well, Samsung's SSD commands pricing that hovers around $500 for just its 32GB model. But this is one fast drive with random data access times of 16.3 milliseconds and an average read of 30.6Mbit/sec. It's also quite miserly with a 10% CPU utilization rate.

Review: Seagate's whopper of a drive -- the 1TB Barracuda

The Seagate Barracuda 1TB drive

The odds are excellent that Windows will never again tell you that you're running low on hard disk space with this 1TB drive, and that alone might be worth the price of admission. Seagate Technology LLC's latest data monster uses a Serial Advanced Technology Attachment NCQ interface and carries a rather sizable 32MB of cache.

Review: Western Digital's 'green' RE2-GP 1TB hard drive

The Western Digital 1TB RE2-GP drive

Western Digital Corp.'s RE2-GP drive uses less power during startup, meaning that on average, you save $10 annually on your electrical bill per drive. The drive is rated at a relatively common 8.9 milliseconds seek time and has a 16MB cache -- half the size of other comparable drives. But Western Digitals's drive has an average operational power consumption of just 7.4 watts, which makes it between 22% and 33% more efficient than its three primary competitors.

Review: Hitachi's old-school 1TB DeskStar 7K1000

The Hitachi 1TB DeskStar 7K1000

This drive doesn't boast efficiency, but its slightly lower platter density allows it to achieve better error-checking without the need for sophisticated firmware, and that translates into fast read speeds that rival other high-capacity disk drives that we compared it with. The drive's burst speed even tops that of Western Digital's speedy RE2-GP by a hair, and the 7K1000's average read is about 10% faster than the tested results of the RE2-GP as well.

Review: Sony's flash-based notebook -- a road warrior's dream

The Sony Vaio TZ191N notebook with 32GB SSD drive

Although this drive comes attached to a Sony Vaio TZ191N notebook, we thought it would be good to include on a comparison basis. Sony Corp.'s 32GB hard drive capacity is tiny by comparison, especially when 6GB of that space is taken up by a hidden partition, this flash drive afforded us with 5 hours, 35 minutes of laptop battery use. Data-transfer rates were 48.7MB/sec. for power saver mode, 49.5MB/sec. for high-performance mode, and 50.2MB/sec. for balanced mode, all with the same 0.3-millisecond access time.

Hacking a Graphic Card (VGA)

In the bad old days, you'd buy a video card that ran at some standard clock frequency. The core clock was the core clock, the memory clock was the memory clock, and that was that.

Then a couple of companies decided to try their hands at selling and supporting video cards clocked higher than the reference clocks. In a commodity market, selling an overclocked card was one way to eke out a little more profit margin and distinguish yourself from the competition. At first, the GPU makers frowned on this, but not for long. In fact, Nvidia and ATI have begun embracing the idea. They allow companies that resell cards, that are often indistinguishable from each other, to compete on something other than software bundles and prices.

When Nvidia launched the 8800 GT, they told the press they left lots of headroom in the card, so card makers could, if they so desired, pump up clock rates and give buyers a little more value. Of course, that added value comes at an extra cost. As we noted in our recent 8800 GT roundup, the factory overclocked cards cost quite a bit more than the standard cards. These higher clock rates can make a big difference in some games, and less so in others.

Of course, you can simply take control of the situation by overclocking the card yourself. This article shows you just how to do that.

The Software Approach

You can, of course, take a stock video card and overclock it. Both Nvidia and ATI offer overclocking features in their respective video control panels, though in Nvidia's case, you have to download and install the additional nTune software package. Alternatively, you can use 3rd party packages, such as RivaTuner.

Since overclocking an Nvidia-based card requires installing additional software—whether Nvidia sanctioned or not—we generally use RivaTuner. It offers more flexibility with a slick user interface.

click on image for full view

One problem we've run into with RivaTuner is that Windows Vista prior to service pack 1 won't let you auto-load the utility on bootup. You get an error telling you that Windows has prevented software from running on startup, so you manually have to start it. After SP1, this no longer seems to be a problem—at least with RC1 of the Vista service pack.

ATI's Overdrive utility is built into the ATI Catalyst Control Center, though you do have to press the padlock button to enable the capability, which is grayed out by default.

click on image for full view

In fact, lots of people tweak their graphics hardware by using software utilities. Both Nvidia and ATI control panels offer graceful fallbacks in case you push the card too aggressively. However, once you find a clock rate that works consistently, and is stable in all your games, loading a software hack at boot time seems a little wasteful. What if you could simply update the BIOS on your video card so the new clocks become a permanent fixture of the card? You can.

Danger, Will Robinson!

Hacking your video card BIOS is fraught with peril. While using a software utility to overclock the card doesn't void your warranty, hacking your video card's BIOS will certainly throw any warranty support out the window. If you load a BIOS into your card that bricks it, you can't simply upload a new BIOS, nor can you return it for an RMA. So if you want to muck around in your video card BIOS, you should do so with the clear idea that you may end up with a $300 paperweight if you take too many risks.

If you understand and accept the risks, editing and uploading a custom BIOS allows you to add permanent customizations to your card.

The BIOS Hacking Community
As with most technology these days, a whole community of hackers has emerged around video cards. These range from hacked drivers, like the famous Omega drivers, to laptop video drivers more current than most OEM systems. In fact, we recently updated 8600 GT drivers on a Macbook Pro running Windows Vista to the much more recent 163.75 drivers. The original Nvidia drivers Apple shipped with the Leopard version of Boot Camp don't even support Aero.

Drivers are one thing, though. You can uninstall hacked drivers and install officially supported versions if need be. Hacking a BIOS requires a deeper level of commitment than simply moving a slider on a software utility. Once you load the BIOS, that card runs at the parameters built into that BIOS. If the card can't boot, there's generally no going back—though some users have successfully reflashed an apparently bricked card by using a secondary graphics card to actually boot the PC. But that's not something you can count on.

The primary source we've found for graphics card BIOS resources is MVKTech.net. You can find individual BIOS files uploaded by users. Want to turn your generic 8800 GT into an ASUS EN8800GT TOP? You can do that if you want. But be aware that not every 8800 GT is created equal, as we'll see shortly. Installing an aggressively clocked BIOS onto a reference-clocked board may simply result in severe graphics instability, game issues or worse.

Note that there are many resources for Nvidia cards, including BIOS flash utilities, BIOS files and a real gem: the NiBiTor BIOS editor. We'll get to the BIOS editor shortly.

The resources for ATI cards aren't as readily available, alas. The two utilities we've found, ATI BIOS Editor and RaBiT, don't seem to correctly read or understand BIOS files from the latest 38XX series of cards.

One use we've found for BIOS editing is to burn into the card overclocking settings we know to be stable through software means. Let's take a quick look at how to edit and burn a BIOS. The card we'll use is the Gibabyte GV-NX88T512H-B, which we reviewed in our GeForce 8800 GT roundup.

Editing the 8800 GT BIOS

We used the NiBiTor BIOS editor to tweak the settings on the Gigabyte 8800 GT. After some trial and error, we determined that the Gigabyte card was stable at a core clock of 700MHz and a memory clock of 940MHz. At higher memory clocks, we began seeing image artifacts.

Step One: Back Up Your Existing BIOS!

The first thing we did was extract the existing Gigabyte BIOS. If you're running Windows XP, you can do this in Windows with the NiBiTor BIOS editor. However, this doesn't work under Windows Vista. An easier way is to boot to the DOS prompt, either from a floppy or optical drive and use the nvflash BIOS flashing utility. In our case, we booted to the DOS prompt from a floppy, but used a USB flash key, which was the B: drive, to perform all the BIOS flashing.

Using nvflash 5.57, we backed up the existing BIOS on the Gigabyte card with the following command:

nvflash --backup gigabyte.rom

This created a file called "gigabyte.rom" on the flash key that contained the original BIOS. We then copied that BIOS onto our main system and used NiBiTor to tweak the settings.

Editing the clock rates in NiBiTor is quite easy. You can even set up different clock rates for different classes of performance (2D, 3D, etc.). Curiously, the Gigabyte BIOS only had clock speeds set in the "Extra" fields, which is how we made our changes.

The default core clock is 600MHz, with the shader clock set to 2.5x the core clock, or 1500MHz. The memory clock is 900MHz. We've determined through using RivaTuner that the Gigabyte card is stable at 700/1750/940, so that's what we set up the BIOS to do.

click on image for full view

Note that changing clock speeds isn't the only reason to edit the BIOS. Let's say, for example, that you're building a quiet system, and don't really plan on running high performance 3D applications. You may want to edit the fan settings to tweak fan speeds down in 2D mode. That's done in the Temperatures tab.

click on image for full view

Of course, we'd like to know what BIOS we're running. So you can edit the BIOS bootup string. That's done under the Boot Settings tab.

click on image for full view

In our case, we altered the BIOS bootup message to read:

ExtremeTech Experimental BIOS 8800 GT by Gigabyte

That way, we know we're running a hacked BIOS when we boot the system.

Once the hacked BIOS is edited, we saved it off to a file, copied it to the USB memory key, then booted from the DOS floppy. Using nvflash 5.57, we updated the BIOS on the card with the following command:

nvflash -5 etbios.rom

Now we have an 8800 GT running at a 700MHz core clock, 1750MHz shader clock and 940MHz memory clock.

Final Thoughts: Does It Matter?

The card has been quite stable running with our hacked BIOS. Let's take a quick peek at performance of our 700/1750/940 card versus the original 600/1600/900 card.

click on image for full view

While the card seems to scale nicely with 3DMark06, it scales less well with games. In fact, Crysis at 1680x1050 with AA and AF enabled seems to run a tad slower on our hacked card. It could just be an anomaly. Or it could be that we've created a card that's a little unbalanced in certain cases, and Crysis with AA and AF simply uncovered this problem. For example, if changing the ratio of memory to core clock rates results in unintended stalls somewhere in the pipeline, then seeing performance go down a notch might not be surprising.

Read our roundup of Four Hot 8800 GT Graphics Cards.

Of course, Nvidia and the card makers spend hundreds of man hours finding the right mix of core, shader, and memory clocks, so our few hours tweaking the card is nothing compared to that. Still, if you go into this with the right mindset, you can end up with a card that's just as good as those pricey overclocked ones. But remember that you won't be able to return the card for repair if something goes wrong. If that thought gives you the willies, we recommend paying a few bucks more for a factory overclocked card.