[swift-users] Help! Slicing an array is very expensive

Howard Lovatt howard.lovatt at gmail.com
Tue May 9 02:59:45 CDT 2017


My mistake. If I create a new array I get the problem. EG:

import Foundation

func elapsed(s: DispatchTime, e: DispatchTime) -> Double {
    return Double(e.uptimeNanoseconds - s.uptimeNanoseconds) / 1_000_000_000
}

let s = 250_000_000
var a = [UInt8]()
a.reserveCapacity(s)

let sa = DispatchTime.now()
for i in 1 ... s {
    a.append(0)
}
let ea = DispatchTime.now()
print("Append time: \(elapsed(s: sa, e: ea)) s")

let st = DispatchTime.now()
a = Array(a[0 ..< (s >> 1)])
let et = DispatchTime.now()
print("Trim time: \(elapsed(s: st, e: et)) s")
print("a count: \(a.count), capacity: \(a.capacity)")


Prints:

Append time: 2.65726525 s
Trim time: 36.954417356 s
a count: 125000000, capacity: 125001696


  -- Howard.

On 9 May 2017 at 17:53, Howard Lovatt <howard.lovatt at gmail.com> wrote:

> I find trimming relative to appending OK on my 6 year old MacBook Pro. EG:
>
> import Foundation
>
> func elapsed(s: DispatchTime, e: DispatchTime) -> Double {
>     return Double(e.uptimeNanoseconds - s.uptimeNanoseconds) /
> 1_000_000_000
> }
>
> let s = 250_000_000
> var a = [UInt8]()
> a.reserveCapacity(s)
>
> let sa = DispatchTime.now()
> for i in 1 ... s {
>     a.append(0)
> }
> let ea = DispatchTime.now()
> print("Append time: \(elapsed(s: sa, e: ea)) s")
>
> let st = DispatchTime.now()
> let ha = a[0 ..< (s >> 1)]
> let et = DispatchTime.now()
> print("Trim time: \(elapsed(s: st, e: et)) s")
>
> print("ha count: \(ha.count), capacity: \(ha.capacity)")
>
>
> Prints:
>
> Append time: 2.612397389 s
> Trim time: 0.000444132 s
> ha count: 125000000, capacity: 125000000
>
>
>   -- Howard.
>
> On 9 May 2017 at 12:56, Kelvin Ma via swift-users <swift-users at swift.org>
> wrote:
>
>> Depending on what you’re trying to do with the data, you might be better
>> off using an UnsafeBufferPointer and allocating and reallocating that,
>> C-style.
>>
>> On Mon, May 8, 2017 at 7:01 PM, Philippe Hausler via swift-users <
>> swift-users at swift.org> wrote:
>>
>>> I wonder if Data might be a better tool for the job here since it is
>>> it’s own slice type and would avoid copying large amounts of data into
>>> temporary buffers.
>>>
>>> > On May 8, 2017, at 16:47, Rick Mann via swift-users <
>>> swift-users at swift.org> wrote:
>>> >
>>> > I have this C library that interacts with some hardware over the
>>> network that produces a ton of data. It tells me up front the maximum size
>>> the data might be so I can allocate a buffer for it, then does a bunch of
>>> network requests downloading that data into the buffer, then tells me when
>>> it's done and what the final, smaller size is.
>>> >
>>> > Thanks to previous discussions on the list, I settled on using a
>>> [UInt8] as the buffer, because it let me avoid various .withUnsafePointer{}
>>> calls (I need the unsafe buffer pointer to live outside the scope of the
>>> closures). Unfortunately, When I go to shrink the buffer to its final size
>>> with:
>>> >
>>> >    self.dataBuffer = Array(self.dataBuffer![0 ..< finalBufferSize])
>>> >
>>> > This ends up taking over 2 minutes to complete (on an iPad Pro).
>>> finalBufferSize is very large, 240 MB, but I think it's doing a very naive
>>> copy.
>>> >
>>> > I've since worked around this problem, but is there any way to improve
>>> on this?
>>> >
>>> > Thanks,
>>> >
>>> > --
>>> > Rick Mann
>>> > rmann at latencyzero.com
>>> >
>>> >
>>> > _______________________________________________
>>> > swift-users mailing list
>>> > swift-users at swift.org
>>> > https://lists.swift.org/mailman/listinfo/swift-users
>>>
>>> _______________________________________________
>>> swift-users mailing list
>>> swift-users at swift.org
>>> https://lists.swift.org/mailman/listinfo/swift-users
>>>
>>
>>
>> _______________________________________________
>> swift-users mailing list
>> swift-users at swift.org
>> https://lists.swift.org/mailman/listinfo/swift-users
>>
>>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.swift.org/pipermail/swift-users/attachments/20170509/1dddda27/attachment.html>


More information about the swift-users mailing list