How to know if drive A (SSD) is really faster than drive B (NVME)

orlando1974 · 08 Mar 2020

Hi, I've got 2 inertnal SSD hard drives being used as part of my music studio, where I'm trying to get the quickest load/read times possible. These drives are not being written to - just read. Upon loading audio samples into my digital instruments, I'm not sure if I can tell a difference between the regular SSD drive and the NMVE one.

Is there a way to see some data/numbers rather than just rely on what I'm perceiving as "not much difference"?

Bree · 08 Mar 2020

orlando1974 said:

Is there a way to see some data/numbers rather than just rely on what I'm perceiving as "not much difference"?

I use CrystalDiskMark, the portable apps version is here:

CrystalDiskMark Portable (disk benchmark utility) | PortableApps.com

orlando1974 · 08 Mar 2020

Thanks!

Pejole2165 · 08 Mar 2020

The thing with SSDs is that the manufacturers sell them based on the sequential read/ write speeds, the truth is that most users will be reading/ writing much smaller files and the important spec of SSDs to be aware of is the random read/ write speeds as this has the most effect on average use cases. If your samples are less than say 500 Mb in size, you won't necessarily see a difference as 2.5" SATA SSDs top out about 540 Mb/s. If your samples are over say 1 Gb or greater in size then you might notice a difference with the NVME pulling ahead.

orlando1974 · 08 Mar 2020

Pejole2165 said:

If your samples are less than say 500 Mb in size, you won't necessarily see a difference as 2.5" SATA SSDs top out about 540 Mb/s. If your samples are over say 1 Gb or greater in size then you might notice a difference with the NVME pulling ahead.

So glad you chimed in! Looking at my sample libraries, I've got samples between 500kb (wav & ncw) and 1.95GB (nkx) - the latter number appearing to be the max size allowed by Native Instruments, though it could just be a coincidence. Regardless, I never knew there could be such a massive difference in the file size of the samples used by my libraries!

Sounds to me then that if I really want to optimize (especially if I'm noticing glitches), I should put those libraries that have samples of 1GB + on the NVM if possible, with the understanding that it may or may not make a "significant" difference - depending on the "random read" specs of each drive.

sygnus21 · 08 Mar 2020

orlando1974 said:

Sounds to me then that if I really want to optimize (especially if I'm noticing glitches), I should put those libraries that have samples of 1GB + on the NVM if possible, with the understanding that it may or may not make a "significant" difference - depending on the "random read" specs of each drive.

Honestly in real world use it won't make a huge difference. I'd personally leave the all the files on one drive. Which drive is up to you but keep the files together.

As a photographer processing files anywhere from 70meg to 2Gig I keep all my files on a second NVME drive in my system. The first MVME drive is the OS drive. But honestly I could have put them on any one of my other 3 SATA SSD in the system probably wouldn't notice a significant difference. BTW before I had dual NVME drives my photo files were on my fastest SATA SSD drive, and NOT on the OS NVMe drive as that would have eaten a lot of space there.

That said, understand not all NVME or SATA SSD drives offer top performance. Thus there is the possibility that a high end SATA SSD outperforms a low level NVME drive.

Yes, NVME drives offer higher throughput speeds than SATA SSD's but other factors in the system could equalize that difference between them. And again, the quality of the drives also factor.

At the end of the day there are way too many factors that can play with speeds and thus why vendors use the term "up to" when discussing speed performance.

My two cents.

orlando1974 · 08 Mar 2020

Ok, I did measured random reads for 1GB using their default values and these are the results for the 3 drives I use in the studio. Does anything look off?

My 7200 RPM HDD:

How to know if drive A (SSD) is really faster than drive B (NVME)-2020-03-08_random-read_samplelibraries-hdd.jpg

My Sata SSD:

How to know if drive A (SSD) is really faster than drive B (NVME)-2020-03-08_random-read_samplelibraries-ssd.jpg

My NVME

How to know if drive A (SSD) is really faster than drive B (NVME)-2020-03-08_random-read_samplelibraries-nvme.jpg

sygnus21 · 08 Mar 2020

I'm not sure those are valid test as it looks like none of them completed. Someone with HDD's confirm this?

Bree · 08 Mar 2020

sygnus21 said:

I'm not sure those are valid test as it looks like none of them completed. Someone with HDD's confirm this?

No, that looks like the OP just ran the last two tests. Each can be run individually by clicking its named button, only by clicking the 'All' button will all 4 tests be run.

orlando1974 · 08 Mar 2020

Bree said:

No, that looks like the OP just ran the last two tests. Each can be run individually by clicking its named button, only by clicking the 'All' button will all 4 tests be run.

Indeed, I wanted to run tests only for the value that was most relevant for my personal application.