2018 MacBook Pro Features 'Fastest SSD Ever' in a Laptop According to Benchmarks

[pl1984]

Yes, that result being an accurate picture of the performance of the disk. You know, a benchmark. BlackMagic and Crystal are long-established, credible benchmarks that are used by credible publications (though laptopmag does not count as a credible publication) All the mental hoops you’re jumping through to deny their results because you don’t like them are really quite remarkable. If you have nothing factual to add, we have nothing to discuss. Have a good one.

You just don't like the benchmark results I provided so you want to alter parameters to change the outcome to suite your desire.

I'll tell you what, I'll run my benchmark with the same parameters as those used in the Laptop Magazine benchmarks. Just let me know what they are.

 

[brilliantthings]

Seriously, the disk test is fine as-is. "Fine", in that it's a synthetic best-case scenario that doesn't even try to mirror a real-world situation. If you want an accurate disk benchmark you'd need something like a recorded disk I/O stream captured from genuine applications that actually stresses the disk subsystem. Database or fileserver apps, and so on. Anandtech and other serious review sites run tests like that.

So you don't think this is just a function of APFS's new file "copy" function?

 

[Lennyvalentin]

So you don't think this is just a function of APFS's new file "copy" function?

No, as mentioned repeatedly that's not how the test works.

 

[Feenician]

You just don't like the benchmark results I provided so you want to alter parameters to change the outcome to suite your desire.

I'll tell you what, I'll run my benchmark with the same parameters as those used in the Laptop Magazine benchmarks. Just let me know what they are.

I don’t like or dislike the results you provided. I simply said they did not reflect the disk performance. The file is too small and is cached. The commands I gave you were absolutely, honestly good a good faith attempt to produce real results. The write test (the dd command) produces, as near as I know how, a realist reproduction. The read test (the hdparm command) much less so because I simply don’t know a way offhand to do. Feel free to discard the read test - the write test seems more relevant to the discussion anyway. You appear to believe I’m trying to deceive you or stack the cards. I am not.

 

[pl1984]

I don’t like or dislike the results you provided. I simply said they did not reflect the disk performance. The file is too small and is cached. The commands I gave you were absolutely, honestly good a good faith attempt to produce real results. The write test (the dd command) produces, as near as I know how, a realist reproduction. The read test (the hdparm command) much less so because I simply don’t know a way offhand to do. Feel free to discard the read test - the write test seems more relevant to the discussion anyway. You appear to believe I’m trying to deceive you or stack the cards. I am not.

You do not feel my benchmark is reflective of disk performance? Hmm, I wonder where we've heard that before? Any ideas?

 

[Feenician]

You do not feel my benchmark is reflective of disk performance? Hmm, I wonder where we've heard that before? Any ideas?

It’s not about “feel”. The test you ran, factually, does not reflect the disks performance.

Once again, it’s clear you want to argue. I’m not interested in doing so. Have a good day.

 

[chucker23n1]

No idea but these should give you a reasonable(ish) idea.

For write (please ensure you have 5 Gb free first)

dd if=/dev/urandom of=5gigfile.bin bs=1m count=5120 oflag=direct

I see about 144 Mb/s from a spinner on a 3 year old Dell running Mint

Maybe Linux urandom is a lot faster than macOS urandom. oflag doesn't exist on macOS dd, and without it, I get rather pathetic results of 11.52 MiB/s on my 2013's built-in SSD, and 10.23 MiB/s on my 2010's Samsung 840.

With /dev/zero instead, that changes to 395.22 MiB/s and 43.61 MiB/s.

 

[pl1984]

It’s not about “feel”. The test you ran, factually, does not reflect the disks performance.

Once again, it’s clear you want to argue. I’m not interested in doing so. Have a good day.

The reality is you have doubts about the benchmark I ran because you do not feel it accurately reflects what is possible. I have the same issue with the Laptop Magazine benchmark. I question their results. How is your doubt any more valid than mine?

 

[Feenician]

Maybe Linux urandom is a lot faster than macOS urandom. oflag doesn't exist on macOS dd, and without it, I get rather pathetic results of 11.52 MiB/s on my 2013's built-in SSD, and 10.23 MiB/s on my 2010's Samsung 840.

With /dev/zero instead, that changes to 395.22 MiB/s and 43.61 MiB/s.

No idea why you’d see such low results on macOS for urandom. Looks like on Mac you’re seeing more realistic results for /dev/zero. That was giving me impossibly high number for a hybrid spinner. Don’t know if there’s an optimization at the kernel level, or maybe at the filesystem level (ext 4 in this case) that sees it’s just a stream of zeros and uses a different path.

My test was carried out on a 3 year old Dell precision under Linux Mint. The drive is a hybrid spinner (I actually forgot it was a hybrid earlier. I wasn’t being disingenuous though ). The SSD cache is tiny and without it, I’d expect to see 50-60Mb/s read/write. That’s pretty standard speed for a laptop spinner of the era.

 

[sunman42]

Flash storage speed compares impressively with ~ 1600 Mbyte/s on my 2017 iMac 27-inch with NVMe 500 Gbyte SSD.

 

[brilliantthings]

Why are so many posts on here sarcastic?
[doublepost=1531607841][/doublepost]Like this one:

You do not feel my benchmark is reflective of disk performance? Hmm, I wonder where we've heard that before? Any ideas?

I don't know what you're referring to. Why don't you just say it directly?

 

[chucker23n1]

No idea why you’d see such low results on macOS for urandom. Looks like on Mac you’re seeing more realistic results for /dev/zero. That was giving me impossibly high number for a hybrid spinner. Don’t know if there’s an optimization at the kernel level, or maybe at the filesystem level (ext 4 in this case) that sees it’s just a stream of zeros and uses a different path.

I don't know, but with a 4 GiB file with actual data (from VMWare), I also get 342.74 MiB/s, so I don't think there's a special optimization for /dev/zero going on.

 

[rodrigo_pereira]

never wanted to go back to Windows until I heard some good things about 10.

It's actually a pretty decent interface, and the recent fluent design language is a very welcome addition, but:

I had issues like after waking up my windows machine from sleep, the start menu stops working.

I have a ThinkPad (and I sorta-kinda love this laptop for it's brutalist design) for dotnet development, and Visual Studio is my favorite IDE, but recently I have this unsolvable problem with screen brightness. It's always set to full brightness which makes this laptop usable only in well-lit environments. Another thing - my external screen stopped working after one of the updates. Screen brightness and external screen works with no problem under GNU+Linux, so it's some drivers-or-whatever related problem.

It's far from the experience I've expected.

 

[Feenician]

I don't know, but with a 4 GiB file with actual data (from VMWare), I also get 342.74 MiB/s, so I don't think there's a special optimization for /dev/zero going on.

Yeah, seemed like something happening on the Linux side of the house there. Your results sound on the money.

 

[VictorTango777]

I use two regularly.

And while I don't use the z, q, ^, and ` keys very often, I'm glad they are there.

Most ironic comment so far

https://www.macrumors.com/2018/07/12/apple-stops-selling-2015-macbook-pro/

 

[pl1984]

Why are so many posts on here sarcastic?
[doublepost=1531607841][/doublepost]Like this one:


I don't know what you're referring to. Why don't you just say it directly?

Already have. I have questions about the Laptop Magaizine benchmarks. Without the configuration of the test systems or details of the methodology I cannot put any weight in them.

 

[stylinexpat]

Just placed my order for the new MacBook Pro. Look forward to getting it this month. None of the previous MacBook Pro laptops I have bought have disappointed me. This will be the most powerful one I have bought and owned.
[doublepost=1531609030][/doublepost]

 

[Feenician]

Just placed my order for the new MacBook Pro. Look forward to getting it this month. None of the previous MacBook Pro laptops I have bought have disappointed me. This will be the most powerful one I have bought and owned.
[doublepost=1531609030][/doublepost]Just placed my order for the new MacBook Pro. Look forward to getting it this month. None of the previous MacBook Pro laptops I have bought have disappointed me. This will be the most powerful one I have bought and owned.

Enjoy! 32Gb refurbs can’t come soon enough for me.

 

[Internet Enzyme]

It's absurd to suggest that a dedicated GPU is required for smooth UI.

Not really. running a display thats high resolution such as the 4k 60hz display my macbook pro is connected to does require a decent amount of performance. My macbook pro with iris 6100 graphics regularly stutters when resizing windows. Having a dedicated gpu is an across the board improvement requisite for being marketed as a ‘pro, machine

 

[stylinexpat]

Enjoy! 32Gb refurbs can’t come soon enough for me.

32Gb of RAM must be sweet. I ordered the 15” model with 512,16 and the i7 8th gen. This should be enough for me for the time being.

 

[Internet Enzyme]

It's absurd to suggest that a dedicated GPU is required for smooth UI.

try switching spaces or doing expose or mission control with a mac from a few years ago with an integrated gpu. They age terribly when compared to the dedicated gpus that these very expensive machines really should come with

 

[Michael Goff]

try switching spaces or doing expose or mission control with a mac from a few years ago with an integrated gpu. They age terribly when compared to the dedicated gpus that these very expensive machines really should come with

That just tells me macOS isn't coded well.

 

[-BigMac-]

Completely useless comparison...

Install Windows 10 with NTFS on it, and then report results pls.
This is a Mac device vs 5 Windows Devices.

We need Apples for Apples

 

[Tech198]

Isn't that SSD soldered in? Meaning... maybe it's fast now, but in a short while when it gets surpassed by something even faster, too bad -- no swapping it out for you! Chuck the whole machine and buy a new one!

And that will be an expensive loss..... Something you wouldn't even wanna DIY yourself

I'd feel safer replacing an SSD card on my 2015 Macbook Pro, in fact i've done that several times.. I'd feel allot less confident tearing apart a 15" (AUS)$9,219.00 laptop.. with 4TB.

I'd probably be starting to sweat a bit

 

[OlliFlamme]

What does this have to do with OS? Apple is forced to preinstall non-essential utilities on Macs because free soft for OSX is hard to find.

Or the other way around: Apple pre-installs utilities and thus the need of free software is less important?
Don’t we all prefer to print to pdf out of the box, rather than having to look on the web and shift shady ware from really ‘free’ programs?
And what about Apple keeping these stock apps tidy in those OS updates that cost you nothing?
Just some thoughts...