HDD speed tester?

[keeperos]

I just built a new system and am trying to see how to best make use of my two HDDs (I can no longer use my 3rd IDE HDD).

The one I currently use for my system is the newer Western Digital Caviar SE16 WD2500KS.
250GB, SATA II, 16MB cache.
The 2nd one is an older Seagate Barracuda ST3200822AS. 200GB SATA I, 8MB cache.

What I need is the most basic of programs to test the read/write speed of both as to decide which I should use for the various applications.

Any ideas?

 

[SomeJoe7777]

Try HDTach or HDTune.

 

[Zorg]

Try HDTach or HDTune.

I like HDTune. Another good benchmark, and all around information, program is SiSoftware Sandra Lite

 

[azimuth40]

Try HDTach or HDTune.

And another which is small and does not need to be installed and free is DiskSpeed

You will need to run it as admin under vista like most things that interrogate hardware

 

[MrCommunistGen]

I prefer HDTach because of its speed, but HDTune gives more detailed latency info. Plus, HDTach won't run on Vista.

Oh yeah, was gonna mention this but forgot. In terms of computer responsiveness faster latency will trump higher sequential reads. Some examples: my friend has RAID0 using old 36GB Raptors. The sequential read is actually lower than my 2x160GB WD 7200RPM drives in RAID0 (I get at least 100MB/s over the majority of the 320GB partition) but his computer felt SO much more responsive especially when loading apps. I guess this is also why using my RAID0 as a boot drive doesn't really feel any faster than my 80GB 7200RPM Hitachi which has reasonably fast seeks even though its a low end hard drive and only has 2MB cache.

-mcg

 

[Zorg]

And another which is small and does not need to be installed and free is DiskSpeed

You will need to run it as admin under vista like most things that interrogate hardware

http://www.roadkil.net/images/ScreenShots/dskspeed1.JPG

I tried that and it doesn't jibe with the other benches, oh well.

 

[azimuth40]

I tried that and it doesn't jibe with the other benches, oh well.

As with most tests it depends on the parameters used by the programmer. Recent versions of Sandra go berserk on Vista, Everest does not have a disk test, HDTach just doesn't run correctly. So I only have HDTune and DISK SPEED. HDTune does a lengthy exhaustive test and Diskspeed does a 5 second test. However they both compare at the beginning of HDTunes tests with Diskspeed showing only the fastest values. A empty 500gig Seagate internal shows around 74 MB/sec on both programs and access around 8 ms.

I did find that both programs can not be open at the same time. I also found that DiskSpeed has a problem under Vista associating the correct drive letter with the correct physical drives sometimes so I used the physical selections in the pull down. It could have something to do with me having one of those flash card readers installed, several virtual drives install and two external 500 gig Seagate Freeagents on USB ports.

Anyway for a quick test which is what the original poster asked for when testing drives on the same system Diskspeed is probably useful since it can just be carried on a pen drive. When comparing drives on different systems of unknown type then as usual everyones mileage will vary no matter what the benchmark program is.

 

[Zorg]

As with most tests it depends on the parameters used by the programmer. Recent versions of Sandra go berserk on Vista, Everest does not have a disk test, HDTach just doesn't run correctly. So I only have HDTune and DISK SPEED. HDTune does a lengthy exhaustive test and Diskspeed does a 5 second test. However they both compare at the beginning of HDTunes tests with Diskspeed showing only the fastest values. A empty 500gig Seagate internal shows around 74 MB/sec on both programs and access around 8 ms.

I did find that both programs can not be open at the same time. I also found that DiskSpeed has a problem under Vista associating the correct drive letter with the correct physical drives sometimes so I used the physical selections in the pull down. It could have something to do with me having one of those flash card readers installed, several virtual drives install and two external 500 gig Seagate Freeagents on USB ports.

Anyway for a quick test which is what the original poster asked for when testing drives on the same system Diskspeed is probably useful since it can just be carried on a pen drive. When comparing drives on different systems of unknown type then as usual everyones mileage will vary no matter what the benchmark program is.

All I said is that it did not agree with the other benches. To make it simple I'll post screens of HDTune and Disk speed. I realized when comparing it to HDTune that it called out my "C" drive, which is the Raptor, Physical 1 not physical 0. That is the reason for the initial confusion. Maybe they should give it an update with the drive name, I don't want to guess if I have the right drive. The sequential read is in line but the random read is 4.826MB/sec. I have a hard time accepting that figure, although, I haven't got a bench handy to compare it to, so maybe they are randomly seeking very small numbers of bits that don't represent even moderately fragmented files. The access time is a blistering 5.89ms. now I want to believe that, but the 8.4ms in HDTune is closer to reality. I could go on about what the OP wanted as far as the size of the file and time spent on the bench, but that is all moot. I had to download the prog to run it again because I don't trust it and so I deleted it. Feel free to let me know where I went wrong.

P.S. The fact that Vista doesn't work with Sandra, and many many other programs, is one of the many many reasons I refuse to use it.

HdTune

Disk speed

 

[keeperos]

Let me elaborate. Both HDs are fairly full. About 40-50 GB free on either one.

At this point I want to see if better results would be yielded if I were to install all the games/heavy appz on the slower Seagate or if it'd be better to keep them at the faster WD which also has windows XP on it and move all the more permanent data (my documents, downloads photos, mp3s etc) to the Seagate. The answer will come depending on how slower the Seagate will be really...

 

[Zorg]

Let me elaborate. Both HDs are fairly full. About 40-50 GB free on either one.

At this point I want to see if better results would be yielded if I were to install all the games/heavy appz on the slower Seagate or if it'd be better to keep them at the faster WD which also has windows XP on it and move all the more permanent data (my documents, downloads photos, mp3s etc) to the Seagate. The answer will come depending on how slower the Seagate will be really...

Well I would use HDTune. Why do you have any more questions? Sandra, which is linked above, will give you read speeds at each location of the drives and HDTune will give you a graph similar to the one above. Why don't you try downloading them and running some benches?

 

[keeperos]

Erm, ok, sorry!

Ok, I used both HD Tune and DskSpeed and had similar results, that is both HDDs performed almost exactly the same with the OLDER Seagate being a tad faster while the WD (because of the double cache) having a more stable graph... Both HDs were 20% free so they were being tested on the same grounds too.
.
(click to enlarge)

Any idea why this is happening? Is that the Seagate's win or the WD's loss in performance that I'm seeing?
I can't believe there's no difference in performance between a SATA 1 and a SATA 2 hdd...

 

[choirbass]

the interface bandwidth alone wont affect hdd performance if the hdds themselves arent fast enough to take advantage of it. for instance, sata150 (sata 1) allows up to 150MB/s transfer rates. sata300 (sata 2) allows up to 300MB/s transfer rates. since neither of your hdds are capable of transferring faster than 150MB/s, it wont matter which sata interface is being used. its not much to be concerned about though, as no single hdd is capable yet of taking advantage of what sata300 has to offer... its just available for when it eventually would happen, a way of futureproofing.

imo, the seagate is marginally faster for regular use, but only due to its slightly faster access times. the transfer rates in general are less important for regular uses though.

either way, the performance and capacity between them is close enough to where it wont really matter too much which you use for which purpose.

 

[onestar]

DiskSpeed is small and quick...Personally, for me, that is a huge plus. It is great for comparing disk performance for the various interfaces I use on various platforms.

 

[geofelt]

Both drives at 7200rpm have about the same capability. You might activate the performance monitor to see where all the activity is usually done. Spread out your data to the other drive.

 

[keeperos]

Great, for this small difference I think it'll be safer and faster if the system is in one HDD and the heavy apps in the other. Case closed

Thanks guys!

 

[Zorg]

Based on the graphs it's in the wind, they are very close.

 

[sp88]

A reliable method to speed test your disk subsystem's is by using SQLIO. Although its not officially supported by microsoft, it produces pretty nice output's to analyse.

You can do some very intersting tests against your hardisk. Like random and sequential reads with multiple threads with multiple IO operations etc.

http://www.slashroot.in/microsoft-sqlio-disk-performance-test-and-benchmark-tool