SSD read speed with Synchron Pianos.

[johnstaf]

Hi all,

I have a Samsung T5 SSD. and a 15 inch MacBook Pro (2018). The Black Magic speed test shows read speeds of 500 MB/s. Synchron Pianos shows read speeds of 190. Is this normal?

I wonder if this is because Synchron isn't simply reading a big file from start to finish, unlike a speed test.

Also, I have the HFS journaled file system. Should I switch over to APFS? I have APFS on my other drives, so this was an oversight on my part.

Incidentally, the Activity Monitor on my system doesn't show any activity when playing the Synchron Pianos. It doesn't matter, but I found it interesting.

[Paul]

Hi, 

The big files make a difference, but HFS is just as good as APFS. Just make sure you don't use ExFat on SSD drives for our Synchron Series. 

Best, 
Paul

[johnstaf]

Thanks Paul.

[mushanga]

@johnstaf said:

Hi all,

I have a Samsung T5 SSD. and a 15 inch MacBook Pro (2018). The Black Magic speed test shows read speeds of 500 MB/s. Synchron Pianos shows read speeds of 190. Is this normal?

I wonder if this is because Synchron isn't simply reading a big file from start to finish, unlike a speed test.

Also, I have the HFS journaled file system. Should I switch over to APFS? I have APFS on my other drives, so this was an oversight on my part.

Incidentally, the Activity Monitor on my system doesn't show any activity when playing the Synchron Pianos. It doesn't matter, but I found it interesting.

Hi John,

I had a similar question for Ben from VSL, and he confirmed that the Synchron Player measures the 64k random read speed, whereas the Blackmagic disk speed test measures the sequential read speed (giving a different, far greater value).

I have been looking for a decent disk speed tester for macOS which measures the 64k random read speed so that I can verify this independently from the Synchron Player, but am yet to come across one.

Hope this helps,

Matt

[johnstaf]

Thanks Matt. That's great to know. I wasn't sure if I was getting a reasonable speed or not.

I'm looking at getting a new SSD, and I might try the T7. Maybe a 2nd T5 with half of the mics on it would work well.

[David B.]

@johnstaf said:

Thanks Matt. That's great to know. I wasn't sure if I was getting a reasonable speed or not.

I'm looking at getting a new SSD, and I might try the T7. Maybe a 2nd T5 with half of the mics on it would work well.

Synchron Pianos gives me a read speed 380 with my Samsung X5 (thunderbolt 3/NVNe) external SSD. I just ordered a Glyph Atom Pro (Thunderbolt 3/NVNe) from Guitar Center. I should receive it next week.

Reviews indicate that the Glyph Atom Pro manages heat very well with sustained advertised speeds. The reason I bought it is because I need more SSD space and I wanted the best performing T3 SSD on the market. The Samsung X5 has been sufficient (although it does get pretty hot). 

I'd like to know why your activity monitor (in Synchron Pianos) is not registering any CPU activity. Can any VSL techs answer that?

God Bless,

David

[johnstaf]

The Glyph seems particularly impressive. 

Parallel drives work well for other demanding instruments such as Dimension Strings. For Synchron Pianos, the Glyph seems like a great choice. It will be interesting to hear how you get on with it.

The Activity Monitor mentioned in my other post is the built-in MacOS utility. It doesn't show any extra activity on the drive when using Synchron Pianos. It usually shows external SSD activity as well as that of the system disk. The Synchron CPU meter works.

[David B.]

@johnstaf said:

The Activity Monitor mentioned in my other post is the built-in MacOS utility. It doesn't show any extra activity on the drive when using Synchron Pianos. It usually shows external SSD activity as well as that of the system disk. The Synchron CPU meter works.

I see. Thanks for the clarification.

[johnstaf]

I just thought I'd add that the read speed (as shown in the Synchron players) for the T7 is slower than the T5. If I run other tests, the read speed for small files is again below that of the T5.

In actual usage the T7 seems to perform very well.

[Alkan]

@johnstaf said:

I just thought I'd add that the read speed (as shown in the Synchron players) for the T7 is slower than the T5. If I run other tests, the read speed for small files is again below that of the T5.

In actual usage the T7 seems to perform very well.

I've been eyeing a 2TB T5 or T7 but am wondering whether the theoretical speed improvement of the T7 will translate to faster real-world performance. I'd be using the USB-C port on a 2019 27" Core i9 iMac. Should I just get another T5 then?

[Macker]

I'm also very puzzled by the Synchron Player's speed readouts.

I recently added a 1TB T7(Touch) to one of the USB-C ports on my 2017 5K iMac (4.2GHz 4-core i7 7700K). This computer came with High Sierra on its internal 2.12TB Fusion Drive formatted as Journaled HFS+, and the only change I've made is expanding RAM to the max 64GB. Synchron Player had previously been showing storage drive read speeds of around 200MB/s to 220MB/s, which seemed ok to me as that's about the best read speed of the internal spinning-platter hard disk.

Initially I formatted the T7 to APFS. Then when, to my horror, I saw that after loading a big SY Dimension Strings preset from the T7 into Synchron Player, the speed readouts were only around 190MB/s, I reformatted the T7 to Journaled HFS+. But Synchron Player speed readouts were still pretty much the same. I've kept Journaled HFS+ in the T7.

I tested the T7's sequential speed with the Black Magic app's 1GB stress test:– 907MB/s read, 833MB/s write. Quite a bit less than Samsung's quoted max speeds as I expected, but not bad.

Then I compared the times taken to load the SYDimStr Violins1/Full/All Players preset (3.1GB) from internal HDD, and then from external T7 (4 threads assigned to loading, 4096 preload). I didn't play anything, just wanted the Synchron Player to measure speed while loading the preset. Each time I ran the test after a reboot so nothing was cached in RAM, and I told Spotlight not to index the T7. The tests were run in LPX10.4.8 which was completely empty except for a single Synchron Player plugged into an instrument strip; audio buffer set at 256.

Timed-test results: Internal HDD:- 9 mins 59 sec; External T7:- 53 sec. Wow! Over 11 times faster!

After loading from the internal HDD Synchron speed showed 259MB/s. But ... after loading from T7, Synchron Player speed readout showed only 202.7 MB/s. How is that logical??

So, Synchron speed readout says the T7 SSD is a bit slower than the iMac's hard disk drive ... excuse me??? Well the nice big fat difference in Synchron preset download times tells a very different story.

I've been building a new VEP7 template for over a week now - Synchron mostly, several Kontakt 6 and a few VIPros, all fed from the T7 - and I'm running increasingly large chunks of score for testing. So far I'm happy with the T7, but dang it, what's up with the Synchron speed test?

[johnstaf]

The performance of the T7 isn't good for small file access.

Accessing small non-sequential chunks of data is similar to accessing millions of small files. The smaller the chunks, the more non-sequential reads must be carried out per gigabyte, which slows things down.

Of course none of this explains why the T7 is slower than the T5 in this regard.

I hope it's a fixable bug rather than a design limitation. 

[johnstaf]

I hooked my 2013 Samsung SSD via USB 3 to my 2012 Samsung i5 dual-core laptop. According to the Windows version of Synchron Pianos I'm getting 600 MB/s.

The same drive shows around 150 MB/s in the Mac version with my 2018 MBP 15.

The performance is better on the Mac though with the T7 with a 64 sample buffer. I can still manage 2 mic pairs with a 128 buffer on my old laptop with the older SSD. 😀

For my next trick, I'll try to get everything up and running on a 1970s calculator! 🤡

[Alkan]

Is there a consensus on whether the Samsung T7 would be a better choice for VSL than the T5? From what I've read, apparently the older T5 is faster at random reads. Am I right to assume that VSL libraries are stored as tons of small files, so when a preset is loaded random reads determine the loading speed? If so, this would imply that not only is the T7 not worth the price difference, but the older T5 would actually perform faster for this application. (I'm using only Synchron Pianos.) Thanks in advance for any insights.

[Macker]

Johnstaf, your Windows computer + T5(?) Synchron speed result is most interesting - many thanks for that.

Now we just need someone who streams VSL libraries from a T7 to a Windows computer (with the correct type of USB port for the T7) to tell us what the PC-version Synchron speed meter says about the T7.

Alkan, I really wouldn't get too attached yet to the idea that the T5 is better than the T7 for sample libraries. If it turns out to be true - and there seems to be only one piece of highly dubious evidence so far - then Samsung would lose face and probably be hit by several class actions from thousands of miffed customers claiming they've been scammed. But I seriously doubt it's true, and we don't want anyone here ending up looking like a myth-loving chump, do we, Alkan.

I'd love to know, for starters, what sector size is used by the T5 and T7 internal controllers. I doubt that they have the variable sector option, and I seriously doubt that they're each different. Anyway, both these products seem aimed at the fiercely competitive mass market rather than enterprise and high-end professionals, and so Samsung understandably keep a lot of the engineering design detail out of the public eye.

[johnstaf]

@Macker said:

Johnstaf, your Windows computer + T5(?) Synchron speed result is most interesting - many thanks for that.

It wasn't even the T5. It was a drive from 2013! 😀

[johnstaf]

@Alkan said:

Am I right to assume that VSL libraries are stored as tons of small files, so when a preset is loaded random reads determine the loading speed?

The actual files are several gigabytes each, but the parts of the file are not necessarily accessed sequentially, so it's similar to reading small files.

I'd imagine (but I don't know) that the player and files are designed in such a way that some large chunks can be read sequentially. 

[johnstaf]

I changed the file system on my T5 to NTFS,  and the Windows version of Synchron Pianos shows a read speed of 600 MB/s.

[Macker]

Thank you and well done, John. We sorely needed that crucial piece of evidence. I think we can rest our case there.

I've been playing various full Sy Dimension Strings presets live with two big handfuls of notes (playing all 5 sections together - 32 individual instrument players), without the slightest hint of any streaming starvation from my T7. Previously trying to do that while streaming from the iMac's internal HDD was pretty hopeless - sample streams plopping-out all over the place - yet Synchron Player's speedometer says my T7 is slower than the HDD.

Dear VSL, I hate to be a thorn in your side in these difficult times, but in this case I wonder if you would very kindly acknowledge that it appears the ball is now in your court.

[MS]

Measuring speeds with a fusion drive is not so easy. The drive will most likely have everything cached in the SSD part, and will thus measure as an SSD. When reading more data randomly, from several libraries, it will perform a lot worse than a proper SSD.