M5Pro /w 1TB slow SSD Speeds

[vatter69]

So I keep seeing those 14XXX read/write benches on the new MacBooks. I got the new M5pro with 18/20 cores 1TB SSD and 48 GB ram. Closed everything but the BlackMagic Bench and used 5GB option.

This is what I get. Not saying it's slow but certainly noway near the reported 14GB/sec I keep seeing. Anyone see the same?

Edit - not sure if it matters but running the 26.4 Release Candidate.

 

[Macintosh101]

I get the same speeds as you on my 1TB M5 MacBook Pro. My understanding is the higher read + write speeds you referenced are on models with higher storage capacities(e.g. 2TB upwards)

 

[turbineseaplane]

Good LORD those are fast.

That’s insane.

 

[torana355]

Never thought id see the day where 10 gigabytes SSD speeds were called slow haha

On topic i'm assuming the really fast speeds were for a 2tb or 4tb model.

 

[vatter69]

Yea like I wrote, not really slow heh - but still 40% less than the reported results 😉

 

[1BadManVan]

try downloading Amorphous disk mark

a more direct type of test most use with windows as its based of crystal disembark and its touted as being more accurate and comprehensive test. And for me it seems more accurate as my 990 pro drive reads at the rated speeds on that test.

 

[Non-Euclidean]

First world problems.

 

[Alameda]

So I keep seeing those 14XXX read/write benches on the new MacBooks. I got the new M5pro with 18/20 cores 1TB SSD and 48 GB ram. Closed everything but the BlackMagic Bench and used 5GB option.

This is what I get. Not saying it's slow but certainly noway near the reported 14GB/sec I keep seeing. Anyone see the same?

Edit - not sure if it matters but running the 26.4 Release Candidate.

View attachment 2616053

The Blackmagic tool looks really cool but it isn't very accurate. As another poster said, use Amorphous.

 

[1BadManVan]

For example, black magic is about 6500 read/write on my m5 air. Amorphous is over 7000 and I believe that to be more accurate since it's in line with the ssd spec when testing my 990 pro desktop on it.

 

[bzgnyc2]

I am curious if these speeds are sustained or just peak? For example, can these new Macs sustain 10 or 14GB/sec for a minute at a time (e.g. 500GB or 1TB test file size)?

 

[1BadManVan]

I am curious if these speeds are sustained or just peak? For example, can these new Macs sustain 10 or 14GB/sec for a minute at a time (e.g. 500GB or 1TB test file size)?

That's very dependant on the drive you're transferring it from more so than the internal ssd itself.

 

[ManuCH]

I'm not sure if that's a thing, but could it be that the drive gets slower (so to say) if it's fuller? Are the 14 GB/s speeds maybe only referred to an empty drive?

 

[Non-Euclidean]

I am seeing some very impressive numbers

MBP 15/16/16 24G/1TB

vs my Studio M4 Max (512 GB)

 

[vatter69]

Yes - managed to find a copy on archive.org (in Germany and only available in the US App Store).

Numbers fluctuate a lot though, re-running just the first SEQ1MQD8 test gives an avg of 13 and 12-16 on write. So point being, I have a normal SSD and people were just running it 30 times and posted the biggest number in the reviews I guess 😉

 

[Non-Euclidean]

Yes - managed to find a copy on archive.org (in Germany and only available in the US App Store).

Numbers fluctuate a lot though, re-running just the first SEQ1MQD8 test gives an avg of 13 and 12-16 on write. So point being, I have a normal SSD and people were just running it 30 times and posted the biggest number in the reviews I guess 😉

Which means they are putting unnecessary wear on their SSDs.

 

[Alameda]

Yes - managed to find a copy on archive.org (in Germany and only available in the US App Store).

Numbers fluctuate a lot though, re-running just the first SEQ1MQD8 test gives an avg of 13 and 12-16 on write. So point being, I have a normal SSD and people were just running it 30 times and posted the biggest number in the reviews I guess 😉

View attachment 2616959

I ran the test again using the same settings as you and got the same results, except for the SEQ1M QD8.

It may be that the System could be reading/writing in the background. Could be writing a log file, purging a temp file, taking a Time Machine snapshot. Difficult to say. *BUT* What we can say is this: This 1TB SSD is a lot faster than the SSD on the M1 MacBook Pro, and that SSD was already wicked fast.

I don't know how to read these benchmarks exactly and what is the most similar to real-world performance, but the M5 Pro SSD is benchmarking 2x to 3x faster than the M1 Pro SSD. Very impressive boost.

 

[bzgnyc2]

For those with the new M5 Pro, etc, can you try running with a larger test file (e.g. 32GB) and longer duration (perhaps even 10 min)? Even if just the sequential QD8 and/or QD1 tests.