Can anyone please explain whether this even makes a difference in real life usage. My laptop has an inferior SSD to my top tier gaming desktop but there is absolutely zero difference in real world usage/responsiveness.
Got a tip for us?
Let us know
Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.
2018 MacBook Pro Features 'Fastest SSD Ever' in a Laptop According to Benchmarks
- Thread starter MacRumors
- Start date
- Sort by reaction score
Probably the same MLC Samsung 64 Layer V-NAND seen in the 970 Pro. Not sure why you two are fighting, misunderstandings happen. Can't we just have a nice discourse.
Probably not, unless you are doing a lot of video exporting, IO intensive stuff like heavy database work or the likes. Not much difference for a standard user or gamer. Still nice to see technology move forward though.
Last edited:
What facts? The testing methodology along with comparison system configuration is completely lacking. As such we have nothing to go on other than some numbers for which we have no idea how they were achieved.
[doublepost=1531586269][/doublepost]
I would say for the majority of users sequential read / write performance is not really important. IMO the majority of people benefit from SSDs because of their much faster random read / write performance. It's nice to see SSD performance continuing to increase but, IMO, I'd like to see more emphasis on random performance.
What ROOKIE at Laptop Mag did this review? Any techie person would know (and SHOULD MENTION in the review and comparisons to be FAIR) that the slower Windows notebooks had SATA SSD's in them (and even then, those speeds listed are too slow, so what did they do wrong?), and the new Macbook Pro has NVMe PCI-e SSD (on a Proprietary APPLE stick that you cannot upgrade).
What wasn't mentioned is that anyone can add a m.2 NVMe SSD to a Windows notebook. Cheaply and there is a large selection to choose from. But you'd never know that from reading the article! I picked up an HP EX920 1TB m.2 NVMe SDD (3200MB/s read and 1800MB/s write) for $229. No $3500 Macbook Pro needed.
What wasn't mentioned is that anyone can add a m.2 NVMe SSD to a Windows notebook. Cheaply and there is a large selection to choose from. But you'd never know that from reading the article! I picked up an HP EX920 1TB m.2 NVMe SDD (3200MB/s read and 1800MB/s write) for $229. No $3500 Macbook Pro needed.
They didn’t compare NVMe with SATA.
https://forums.macrumors.com/thread...ng-to-benchmarks.2127397/page-6#post-26243488
Mid-tier drives plus only two PCIExpress channels will result in SATA tier performance though.
I was reading something on the Dell forums where someone upgraded to the new and fast VNMe drive but still could not get fast read and write speeds which had something to do with the Dell 13 XPS only having two channels.
Yep. For the XPS in particular putting a 970 in would be a waste of money. It’s not like 400-500Mb/s write is objectively bad or anything but it is what it is.
You have no idea what they compared as they didn't provide any details of what they compared nor did they provide any idea of the methodology they used to perform the comparison.
Here's my benchmark of a 2012 Mac Mini with its stock, non-SSD hard disk:
dd if=temp of=temp2 bs=4M
51+1 records in
51+1 records out
214329773 bytes (214 MB, 204 MiB) copied, 0.187166 s, 1.1 GB/s
1.1GB/s is an outstanding speed for anything other than a PCIe based SSD. By this metric Apple really knows how to make non-SSD hard disks.
I’ll test my sister’s XPS 9350 when she comes it has a PM961 because I recall getting over 1GB/s writes. Even if it only had 2 lanes 300-400MB/s only makes sense if it saturated the dynamic SLC cache and was writing directly to TLC portion after 10-40GB if continued writes depending on drive size. The Toshiba XG5 I know needed an update as it was hitting only 300-400MB/s even without saturating the dynamic SLC cache, it can hit around 1GB/s normally. Pretty sure the two lane limit was on the TB3 port on the 9350.
I wonder if that person had the laptop on RAID-ON or AHCI mode in bios as some are picky that way, I found stock Microsoft nvstore (nVME driver) to be more reliable than Rapid Storage and until recently even Samsung’s on nVME driver.
Last edited:
Which is why I did my own research. Their numbers stack up with Crystal benchmark on notebookcheck.net
Dd is a horrible test on Mac since it has no oflag=direct option. You’re testing the buffer. Also, 214Mb is not a useful file size to stress an SSD. Go ahead and run the 5Gb stress test on BlackMagic and report the speed.
Except those lame speeds for Windows notebooks (obviously chosen to show magnitudes faster speeds for Macbook Pro) are slower than even SATA III. More like SATA II.
That post you referenced...claims they tested a 2017 XPS 13, but where does it say the Dell XPS 13 in this test was a 2017 model? It doesn't. It could have been ANY model year XPS 13 and certainly one with a SATA SSD in it. Again, most likely cherry-picked to make Windows notebooks look magnitudes slower than a model 2018 Macbook Pro.
Just a biased rookie doing the testing at Laptop Mag.
Which does nothing to change what I said: You have no idea what they compared. The fact you may have done your own research doesn't say anything about the comparison systems nor does it say anything about the test methodology.
The benchmark is what it is: 1.1GB/s. I think you'd be accepting of this benchmark as it shows that Apple's traditional hard disks really stomp the competition...just like this SSD test you're championing.
[doublepost=1531590674][/doublepost]
His argument is this one XPS, for which we know utilizes an NVMe SSD, achieved SATA SSD levels of speed and therefore is representative of all non Macintosh base NVMe SSDs.
I agree...The article doesn't even say what they used (or how the did the testing) for the Windows notebooks. It's ridiculous. I'd run a similar test on my 2017 XPS 13 except I have no idea what they did to test!
I'm wondering how a SATA III hard drive can do 1.1GB/s. It'd be a miracle since SATA III is saturated at about 560MB/s.
Well, you know, if you actually wanted to check, instead of just complain you could do that.
2017 Dell XPS with Toshiba XG4 NVMe
https://www.notebookcheck.net/Dell-XPS-13-i7-8550U-QHD-Laptop-Review.257650.0.html
2018 Dell XPS 13 with Lite-On CA3-8D256-Q11
Those figures are on around the ones in the OP. Now what?
It can't, here's the exact same test on the exact same hardware:
dd if=temp of=temp2 bs=4M
51+1 records in
51+1 records out
214329773 bytes (214 MB, 204 MiB) copied, 0.828976 s, 259 MB/s
I think we would all agree these results are more acceptable and, ironically, outperforms the three slowest SSD comparison PCs used in the Laptop Magazine comparison. We have two different sets of results despite the exact same test.
The “benchmark” showed nothing but buffer speed. You went and ran BlackMagic, didn’t like the real number and buried your head in the sand.
[doublepost=1531591240][/doublepost]
dd on a small file shows buffer performance and nothing more.
Yeah definitely capable of speeds over 1Gb/s in that very same NC article.
"very slow for an NVMe SSD at only around 331 MB/s compared to 1347 MB/s on the Samsung SM951 in the 256 GB XPS 13 SKU. The slow speeds seem to be a recurring attribute amongst XPS 13 units equipped with Toshiba SSDs as shown by our table below. At the time of writing, neither Dell nor Toshiba have addressed the issue." -Notebookcheck
Pretty much the experience I stated with the XG5 I stated above pre firmware update.
"very slow for an NVMe SSD at only around 331 MB/s compared to 1347 MB/s on the Samsung SM951 in the 256 GB XPS 13 SKU. The slow speeds seem to be a recurring attribute amongst XPS 13 units equipped with Toshiba SSDs as shown by our table below. At the time of writing, neither Dell nor Toshiba have addressed the issue." -Notebookcheck
Pretty much the experience I stated with the XG5 I stated above pre firmware update.
That's why I like Notebookcheck they mention caveats in different SKU of the same model regarding display and SSDs, and are more in depth about their testing methods.
It only outperforms TWO that I can see. And really only one because there is this test that includes the Surface Book 2 and shows MUCH HIGHER speeds for it.
With a lack of testing methodology on the Windows notebooks (yet they say how they tested the Macbooks), the numbers from this review are questionable, at the least.
Regardless, the lack of any mention of faster NVMe drives for Windows at a much lower cost is inexcusable. Take any of those low cost Windows notebooks and change the SSD for $150 (something you'll never be able to do on this Macbook Pro) and voila! The same "double take" speeds.
I believe that currently there is an issue with the drivers of these for sleep mode (at least on the new X1 Carbon models). Something conflicting between Windows drivers and bios from manufacturer of laptop and actual NVMe bios. That is where things get complicated. An issue arises then people don’t know whom to blame. Blame manufacturer of laptop,manufacturer of hardware device or the Windows operating system. This is why at the end I like Apple devices. You only have one person to deal with and complain to. Where as with a Dell laptop running Windows it could be Dell,Windows or the manufacturer of the hardware component inside. Each is guessing and blaming the other party.
[doublepost=1531591986][/doublepost]
That is a huge difference between Read and Write speeds. On most Samsung manufacturers website for NVMe storage the difference written for their Read and Write speeds is not that big.
Yes, Dell are doing their customers a disservice putting those SSDs in. Though, with only two PCI Express lanes, they wouldn’t be able to utilize the full speed better units anyway.
Which truth is that? So far, I've read that the XPS 13 *is* capable of >1GB/s speed, using a different stock SSD and that the slow benchmark is attributed to a particularly slow SSD model (Toshiba) that needs a firmware update.
Could you please link to somewhere credible actually showing that?
Even then, in 2018, 1Gb/s is pretty pedestrian. The 2016 MBP is faster than that, and my 2015 pulls around 760Mb/s.