Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.
Given that in SSD benchmarks, drives of the same brand and design are fastest in their 512Gb or above sizes it would seem the larger capacity flash performs faster than lower capacity flash. I'd imagine with SSDs a 512Gb drive uses 4 x 128Gb chips, a 256Gb uses 4 x 64Gb chips (because performance is similar but the 512Gb drives have faster IOPs and read/write times) and 128Gb SSDs likely use 4 x 32Gb chips and offer the lowest performance.
 
Given that in SSD benchmarks, drives of the same brand and design are fastest in their 512Gb or above sizes it would seem the larger capacity flash performs faster than lower capacity flash. I'd imagine with SSDs a 512Gb drive uses 4 x 128Gb chips, a 256Gb uses 4 x 64Gb chips (because performance is similar but the 512Gb drives have faster IOPs and read/write times) and 128Gb SSDs likely use 4 x 32Gb chips and offer the lowest performance.

Thanks for the insight. Purpose of thread was just to see if there are any significant benefits in terms of performance gains by getting a 128GB over 16GB
 
I knew there was another reason I chose 128GB! ;) Waiting for the 6 vs 6+ crowd to argue over memory configuration now.
 
Thanks for the insight. Purpose of thread was just to see if there are any significant benefits in terms of performance gains by getting a 128GB over 16GB

Only that iOS appears to slow down as the storage gets very close to full. 128GB will obviously stave that off longer.

----------

What makes people think that higher capacity will be faster? There is zero evidence to suggest this.

there's been raw IO benchmark evidence that shows that some larger capacity NAND Flash storage systems perform faster than smaller capacity models. But that sort of performance will barely make a difference, and only then in iO heavy tasks. It's unlikely you'll ever notice it in an iOS device.
 
So why did u get the 128GB?

I store all my media locally and had about 15GB free on my 5S before I upgraded. It feels nice to have the extra cap room, notwithstanding the 64GB config wasn't available when I pre-ordered.
 
Anyone having a 128GB/64GB that can vouch for the performance?

Will having 128GB/64Gb allow tabs and programs to refresh less often?
 
Given that in SSD benchmarks, drives of the same brand and design are fastest in their 512Gb or above sizes it would seem the larger capacity flash performs faster than lower capacity flash. I'd imagine with SSDs a 512Gb drive uses 4 x 128Gb chips, a 256Gb uses 4 x 64Gb chips (because performance is similar but the 512Gb drives have faster IOPs and read/write times) and 128Gb SSDs likely use 4 x 32Gb chips and offer the lowest performance.

From what I understand, that's because larger SSDs have more physical NAND chips, and they do a sort of "internal" raid 0 to increase speed. I don't know if that would apply to our iPhones which only have a single NAND chip
 
From what I understand, that's because larger SSDs have more physical NAND chips, and they do a sort of "internal" raid 0 to increase speed. I don't know if that would apply to our iPhones which only have a single NAND chip
It does apply to a certain extent. There's internal parallelization within the NAND, too. However, top speeds can usually be had on drives one step lower than largest possible. That's because on the largest SSDs, the bandwidth/channel becomes a bottleneck so the further increase in parallelization becomes somewhat detrimental.
 
It depends on the controller. The NAND employed in phones isn't all that fast, so if the phone is able to write to and read from, say, 4 32GB chips at the same time, it's going to be faster than a single 16Gb chip while loading a several hundred megabyte game for example. Especially since storage is the bottleneck, and not the CPU. I find it strange that people dismiss this offhand, as we simply don't know if iPhones work like that or not. Common USB 3.0 sticks do this, so I don't regard it as unthinkable.
 
Given that in SSD benchmarks, drives of the same brand and design are fastest in their 512Gb or above sizes it would seem the larger capacity flash performs faster than lower capacity flash. I'd imagine with SSDs a 512Gb drive uses 4 x 128Gb chips, a 256Gb uses 4 x 64Gb chips (because performance is similar but the 512Gb drives have faster IOPs and read/write times) and 128Gb SSDs likely use 4 x 32Gb chips and offer the lowest performance.
A lot of SSD manufacturers are cheap. One big NAND package is usually cheaper than multiple smaller NAND packages. You're more likely to find:
1TB: 8x 128GB
512GB: 4x 128GB
256GB: 2x 128GB
128GB: 2x 64GB

The increase in channels is usually where you get the big RAID-0 like performance boost. I reckon the only reason they can't max out the NAND on smaller SSDs is because then, performance will be atrocious.
 
Ah

Anecdotally my iPhone 6+ 128gig runs geekbench faster than my old 2.5 ghz G5 tower. That's what's really impressive in how processor speed and whatnot has increased over the last 5-7 years. No liquid cooling needed.
 
It depends on the controller. The NAND employed in phones isn't all that fast, so if the phone is able to write to and read from, say, 4 32GB chips at the same time, it's going to be faster than a single 16Gb chip while loading a several hundred megabyte game for example. Especially since storage is the bottleneck, and not the CPU. I find it strange that people dismiss this offhand, as we simply don't know if iPhones work like that or not. Common USB 3.0 sticks do this, so I don't regard it as unthinkable.
The NAND die itself is fast (or at least likely just as fast as the ones used in NGFF/mSATA/2.5" SSDs). The issue is there's not as many dies on which to divide the load. On a 512GB SSD, assuming 128Gb (16GB) dies, you've got a total of 32 dies in parallel. On a 128GB iPhone, you've only got 8 dies. You're also limited in terms of power consumption on an iPhone so it's not like you can use the same powerful controllers found in SSDs.

By the way, my iPhone 6 and 6 Plus 64GB bench at ~65MB/s for sequential write and ~220MB/s for sequential read. My iPad 4 128GB does ~60MB/s sequential write and ~160MB/s sequential read. :)
 
Last edited:
Having some free storage is important, having a full 16GB phone would make the phone perform worse than one with 2-4GB free, but only slightly.

It's the same with computers.
 
You're also limited in terms of power consumption on an iPhone so it's not like you can use the same powerful controllers found in SSDs

True, I hadn't even considered that. You can't go too crazy or it will eat into the battery. On the other hand, storage operations would be finished quicker, so..
But I was just speculating anyway. It would be interesting to see if a 16GB iPhone 6 will match your benchmark numbers.
 
The NAND die itself is fast (or at least likely just as fast as the ones used in NGFF/mSATA/2.5" SSDs). The issue is there's not as many dies on which to divide the load. On a 512GB SSD, assuming 128Gb (16GB) dies, you've got a total of 32 dies in parallel. On a 128GB iPhone, you've only got 8 dies. You're also limited in terms of power consumption on an iPhone so it's not like you can use the same powerful controllers found in SSDs.

By the way, my iPhone 6 and 6 Plus 64GB bench at ~65MB/s for sequential write and ~220MB/s for sequential read. My iPad 4 128GB does ~60MB/s sequential write and ~160MB/s sequential read. :)

What did you use to measure the read and write speeds?
 
I feel the exact opposite of this :S I feel like my iPhone 6 128gb is slower than the 64/16gb models.
 
True, I hadn't even considered that. You can't go too crazy or it will eat into the battery. On the other hand, storage operations would be finished quicker, so..
But I was just speculating anyway. It would be interesting to see if a 16GB iPhone 6 will match your benchmark numbers.
It wouldn't. However, the performance difference is unlikely to actually make an appreciable difference in normal day-to-day usage. The 128GB will probably bench around the same, though. Again, these are sequentials. For running iOS and apps, random IO matters more and there probably isn't that large a gap there between 16GB and 64/128GB.

What did you use to measure the read and write speeds?
PerformanceTest by PassMark.
 
Last edited:
Register on MacRumors! This sidebar will go away, and you'll see fewer ads.