Short-stroking A Hard Drive

Has any been able to do this.Any tips on how to do this option.

 

If you want a partition on that drive that keeps the transfer rate above ~100MB/sec, create only a 400GB partition (that's ~40% used, see your graph).

 

What happens if i go any higher or lower.

Will i get a gain in speed transfer rate if i go lower than 400gigs.

Be cause i really don't need that much space.That is what i got the backup drive for.

Edit: and i am only using about 85 gigs on my hard drive.

 

Dropping the size to ~200GB would increase the average speed and minimum speeds a little more, it would also improve the latency by a small margin.

If you've only had 85GB max. on your drive, short-stroking it is not likely to improve it by much and if it does, it's unlikely to be noticeable outside benchmarks.

Do whatever you want, it's your hardware and your data.

 

I seem to see a difference in speed.

I put it at 200Gig for the main drive.And every thing seems to work faster.

 

Good result then.

 

Frankly, I bet that is due to the placebo effect and other factors and not really the result of short stroking. I say that because we are still talking milliseconds. And the majority of computer performance comes from the CPU, RAM and graphics solution, not the drive. Write times are not really user perceptible. Read times are, but the reality is, once the file is loaded into RAM, the drive steps out of the picture. And most of the time needed to retrieve a file is in seeking out the first file segment. Unless the drive is heavily fragmented, the vast majority of the subsequent file segments are right next door to previous segment.

Years and years ago, short stroking was beneficial. This is because hard drives were slow and they had tiny buffers making them even slower. In fact, most defragging programs would move your files to the beginning of the disks just to take advantage of the more data per track in those longer out rings. But today, just defragging the files is more important than moving the files to the beginning of the disk (or consolidating free space). This is partly because Windows and program updates are constantly updating files so the updated files are constantly being stored in new locations with a big hole appearing once the old version is deleted.

The best performance gains on a hard drive are achieved by keeping file fragmentation minimized (and Windows does a decent job of that automatically) and by keeping a large chunk of disk space free for the Page File (Windows managed, of course) and Windows temp files to operate in freely.

Now you if really want to actually "see" a genuine over all boost in performance from your drives, replace them with SSDs.

 

This the results after a smaller partition.

 

I see nothing significant there. Access Time and Maximum Transfer Rate are identical and the average transfer rate actually got worse!

 

Here's a comparison for ya in a Acer 605 with a j1900 quad core

 

How is that a comparison? We would need a before and after shot. As it is now, two different drives in two different computers don't tell us anything.

But if you want to compare, check out a SSD:

 

I was looking at his cutting a 1 tb disk in half specs to an original wd500 of similar size. Tho I don't know his cpu or ram or graphics as was brought out earlier. Just a look to see. I ran the HDTune on a usb external in a #2 port and then a #3 port.
Just by eyesight I can't tell the difference. Probably can't tell the difference between the usb or internal either. I probably could tell some difference with your ssd example.

 

Here are my specs if you want to know.

Operating System
Windows 7 Professional 64-bit SP1
CPU
AMD A10-8700P
Carrizo 28nm Technology
RAM
8.00GB Dual-Channel DDR3 @ 798MHz (11-11-11-28)
Motherboard
HP 80B0 (P0) 55 °C
Graphics
Generic PnP Monitor (1600x900@60Hz)
MLT1921 (1440x900@60Hz)
512MB ATI AMD Radeon R6 Graphics (HP) 49 °C
Storage
931GB Western Digital WDC WD10 JPVX-11JC3T0 SATA Disk Device (SATA) 31 °C
931GB Western Digital WD My Passport 0827 USB Device (USB (SATA)) 31 °C
3GB SDHC Card (SCSI)
Optical Drives
hp DVDRW SU208GB SATA CdRom Device
Audio
AMD High Definition Audio Device

The 3GB SDHC Card (SCSI) helps with my Ready boost.FYI this is a laptop.

 

My thoughts are actually on the other half of the disk.. Does it now lose speed that was given up in the short stroke? Unless that is going to left as unallocated space, which to me is a waste. If it is used for a different os or as a data partition does it suffer? The same might be true of any partitioning scheme under normal conditions ie ; installing Linux or moving data files to the second or third partition - is there a loss of speed?

 

So far i don't see any issue.

Even my mmo games seem to load 3 times faster with a smaller drive.I just kept a smaller partition for my Backup data.

And left the rest alone.

 

Any time you have more than one partition on a disk, there is some (albeit very small) overhead.

What do you mean "smaller drive"? A separate physical drive? Or a separate partition on the boot drive?

Any time you have two physical drives, you can see a performance gain (depending on the task being performed) because Windows can access both drives at the same time. But if Windows is on one partition and your games are on the other partition on the same drive, there are no gains.

No way would there really be 3 times faster load times. Sorry. Note that when you change the partitions on a drive, the files that were moved and saved in the new partition are automatically defragmented. If you see improved load times, that might explain why (besides the placebo affect).

 

A separate physical drive? No i mean a smaller partition.Sorry for the confusion.

And when i moved it to a smaller partition i did a defrag on the HDD.I did see a better improvement instantly when i reinstalled every thing.

 

What did you reinstall?

 

And that is why - not because of the short stroking. When you reinstalled everything, all the file segments were laid down in sequential clusters and tracks. So when you clicked on a program to load, the drive's R/W head just needed to "hunt" for the first file segment, then all the rest of the segments were right there in order.

If you disable defragging, fail to keep your disk decluttered (with regular use of CCleaner or Windows Disk Cleanup), and fail to keep a large chunk of free disk space available (I like at least 30GB) those load times will degrade again. That is just the natural course of things.

 

Exactly. Windows is 'faster', not the HDD.

 

With the Samsung Magician doing it's thing, the rates go through the roof.

 

In my experience, Samsung Magician either did nothing, or dragged things down with additional overhead. And it caused sleep problems for me and others have reported the same.

If you note, we have the same 850 Pro series. I would not call yours "through the roof" compared to mine. Our results are pretty similar but I would say my maximum is "through the roof" compared to yours.

Windows knows how to manage SSDs just fine without any 3rd party software.

 

Fortunately, the Magician software works quite well on both my machines. On the ThinkPad, it has allowed the machine to overcome the limitations of being SATA II. And yes, I ran it without Magician in the beginning just to see if there was a visible difference. There is.

I left the PC idle for the HD Tune benchmark. What caused your transfer rate to spike?

Do you have an ATTO benchmark to reference?

 

I don't know what caused the spike - I just sat back and watched the test run. I just ran it again (though with Pandora streaming this time, if that matters) and this time, no huge spike but my max was 349.3, average 333.3 MB/sec, and access time .1 ms. So again without Magician very similar times (if not slightly better) than yours with Magician.

No, I don't have ATTO. FTR, I don't normally run benchmark programs because I think people put too much and undeserving faith in them. This is actually why Microsoft removed the Windows Index Experience from Windows. People were obsessing over it unnecessarily. Benchmarks are synthetic scenarios that abuse our hardware. They are great in test and comparison labs but normal users don't need them.

Years ago, I did. But in spite of what many want to believe, parts of Microsoft have some pretty sharp folks working there. I can't say much for the bean counters and marketing departments, but the Windows development team are top notch and know how to make Windows take optimal advantage of the hardware it runs on. This is why I say leave the defaults alone. They work just fine.

 

The HDTune doesn't pick up the Magician transfer rate, just the base rate. ATTO and CrystalDiskMark will pick up the increased rate. CrystalDiskMark will also show you the compressed data rate, which tends to be slower.

I tend not to run benchmarks on a regular basis. They're just a curiosity tool. If I'm shelling out dough for an alleged performance benefit, I like to see proof.