Re: Curl's network performance is slower than IDM.

On Thu, Jun 18, 2020 at 13:34 Patrick Schlangen via curl-library <
curl-library_at_cool.haxx.se> wrote:

> Hi,
>
> interesting thread.
>
> Am 18.06.2020 um 12:09 schrieb Daniel Stenberg via curl-library <
> curl-library_at_cool.haxx.se>:
>
> The question is then of course still what IDM does that makes it that much
> faster or perhaps what curl should do to improve. But without more details
> and knowledge of IDM's set of tricks, I'm not sure we will achieve that
> right now.
>
>
> I’ve downloaded that program out of curiosity and compared it against curl
> on macOS.
>
> curl is running natively on macOS, IDM in a Windows 10 VMware Fusion
> virtual machine on the same machine. I only tested a couple of times with
> the file (http://speedtest-ny.turnkeyinternet.net/1000mb.bin) and timed
> the total download time (using `time` for curl and using a stopwatch for
> IDM).
>
> Interestingly, IDM showed max 13.5 MB/s with an average total download
> time of 93 seconds, while curl showed max 12.0 MB/s and took 86 seconds for
> the same file on average. So actually curl was faster but displayed the
> lower max transfer rate.
>
> Important: Start your stopwatch already when you are in the dialog shown
> in my attached screenshot. They work with a little trick here: They start
> downloading in the background before you even press "Start Download"! (See
> network activity in my screenshot...)
>

That is a clever technical/psychological trick. Not too different than what
Apple does/did w DHCP[0]. An example of “real world” speculative
execution.[1]

[0]
https://news.ycombinator.com/item?id=2755461

[1]
https://en.wikipedia.org/wiki/Speculative_execution?wprov=sfti1

> @Hongyi: I’d recommend to retry the tests while measuring the time it
> takes to download the file instead of relying on the displayed download
> speed. Either curl could be too pessimistic in the displayed speed or IDM
> could be too optimistic. (In the end the selling point of their product is
> to be faster than the competitors.)
>
> Disclaimer: I didn’t spend much time, and from a statistical point it’s
> probably not valid to draw any conclusions from my experiments.
>
> Best,
>
> Patrick
>
> -------------------------------------------------------------------
> Unsubscribe: https://cool.haxx.se/list/listinfo/curl-library
> Etiquette: https://curl.haxx.se/mail/etiquette.html

-------------------------------------------------------------------
Unsubscribe: https://cool.haxx.se/list/listinfo/curl-library
Etiquette: https://curl.haxx.se/mail/etiquette.html