Testing HDD/SSD Speeds

[cubies1982]

Could someone recommend me some good software that will accurately test my drives read/write speeds?

[Enderman]

Crystaldiskmark.

[Electronics Wizardy]

crystal disk info

 

atto disk mark

 

dd

[Bradl79]

AS SSD

[Captain_WD]

15 hours ago, cubies1982 said:

~snip~

Hey there

 

The guys suggested some good tools. I'd say use multiple tools and see what all of them will bring up as a result and post it here if you are not sure if the results are normal or not.

If you doubt your storage devices' health I'd suggest to run diagnostic tools from the manufacturers' websites and see if the drives pass their tests. Also, checking the raw values of the S.M.A.R.T. status is a great way of getting a more detailed picture of the drives' condition. 

 

Let me know if you have any questions! 

 

Captain_WD. 

[cubies1982]

On 10/26/2016 at 1:07 PM, Captain_WD said:

Hey there

 

The guys suggested some good tools. I'd say use multiple tools and see what all of them will bring up as a result and post it here if you are not sure if the results are normal or not.

If you doubt your storage devices' health I'd suggest to run diagnostic tools from the manufacturers' websites and see if the drives pass their tests. Also, checking the raw values of the S.M.A.R.T. status is a great way of getting a more detailed picture of the drives' condition. 

 

Let me know if you have any questions! 

 

Captain_WD. 

I want to know if the speeds listed on the website are the same speeds I am getting inside my terminal.

[Captain_WD]

1 hour ago, cubies1982 said:

~snip~

What are your SSD's brand and model? If you check out the spec sheet you should see the speeds there and when you run a benchmark tool you can compare those results with the results on the spec sheet. 

 

Captain_WD. 

[corrado33]

Speeds reported by companies are from synthetic benchmarks. You will probably never reach those speeds transferring normal files to or from the device. 

 

Personally, I find synthetic benchmarks, (Crystal Disk Mark, Atto Disk Benchmark) useless for determining how quickly you will be able to transfer files to or from a drive.

 

If you want to know more, check out my flash drive comparison thread.

 

[cubies1982]

7 minutes ago, corrado33 said:

Speeds reported by companies are from synthetic benchmarks. You will probably never reach those speeds transferring normal files to or from the device. 

 

Personally, I find synthetic benchmarks, (Crystal Disk Mark, Atto Disk Benchmark) useless for determining how quickly you will be able to transfer files to or from a drive.

 

If you want to know more, check out my flash drive comparison thread.

 

Maybe I could purchase a harddrive and say a USB pen drive and then transfer a 1GB file from one to the other and then look at the speed between the two devices. Do my own little bench marking.

[corrado33]

4 minutes ago, cubies1982 said:

Maybe I could purchase a harddrive and say a USB pen drive and then transfer a 1GB file from one to the other and then look at the speed between the two devices. Do my own little bench marking.

Your idea is correct, your execution is wrong.

 

You'd have to buy a really nice (aka not budget) usb 3.0/3.1 flash drive to outperform even a slow 5400 RPM HDD.

 

My sandisk extreme drive managed it, but only barely.

 

I did that test except I used my internal NVME M.2 SSD which easily outperforms anything else that I have. 

[cubies1982]

1 minute ago, corrado33 said:

Your idea is correct, your execution is wrong.

 

You'd have to buy a really nice (aka not budget) usb 3.0/3.1 flash drive to outperform even a slow 5400 RPM HDD.

 

I did that test except I used my internal NVME M.2 SSD which easily outperforms anything else that I have. 

Would also work with two SSDs? Move a 1GB file from one SSD to the other and Windows will measure the transfer speed?

[corrado33]

6 minutes ago, cubies1982 said:

Would also work with two SSDs? Move a 1GB file from one SSD to the other and Windows will measure the transfer speed?

I'd say no. While you would get good transfer speeds, you could never be sure which drive was bottlenecking the other, so you won't get the true speeds. 

 

In general, writing is slower than reading, but that wasn't ALWAYS the case in my tests for some reason. (Check out the crystal disk mark results for my m.2 drive, writes were faster than reads in some instances.) I couldn't do real world benchmarks on the m.2 because I had nothing faster than it.

 

The only think you could probably do is create a RAM disk (literally a writeable disk made up of your memory) then transfer a file to/from that. 

 

Quick google search for creating a ram disk showed me that you can download a simple program to do it for you.

 

https://www.tekrevue.com/tip/create-10-gbs-ram-disk-windows/

[Captain_WD]

15 hours ago, cubies1982 said:

~snip~

Transferring a file from one drive to another is a tricky way of determining the performance of a drive. You will have to have a faster target drive in order not to bottleneck the drive you are benchmarking. One thing you could do, as @corrado33 suggested, is to create a small RAM drive from your system's memory and transfer data from and to it and see how fast can your storage drive perform. RAM drives have higher performance than most (if not all) storage drives. Your testing storage drive needs to be the slowest point in your transfer in order to truly benchmark it. 

 

Here's a link with more info on RAM drives:

 

Captain_WD.