Buy commercial curl support from WolfSSL. We help you work
out your issues, debug your libcurl applications, use the API, port to new
platforms, add new features and more. With a team lead by the curl founder
himself.
Re: Response time appears to be much higher on 7.78 version of libcurl (vs 7.60 version)
- Contemporary messages sorted: [ by date ] [ by thread ] [ by subject ] [ by author ] [ by messages with attachments ]
From: Daniel Stenberg via curl-library <curl-library_at_lists.haxx.se>
Date: Sat, 18 Sep 2021 12:16:49 +0200 (CEST)
On Fri, 17 Sep 2021, Question Throwaway via curl-library wrote:
> With 7.60 I regularly will get an average of 450-500MS response time
>
> With 7.78 I regularly get between 4000-5500MS response time
Your showed code doesn't seem to be complete so we can't run that locally, but
I would suggest you try that test code on another platform first as I think
you'll see that there won't be any such extreme delays there.
Getting a ~10 times slower performance is not just a minor degradation or
wrong option. It's a sign of something seriously wrong. I would urge you to
get a full DEBUGFUNCTION log with timed entries to spot where all that time is
spent. Alternatively doing a full wireshark dump, but that might be tricker
for HTTPS traffic from that target.
Did you use the same TLS backend in both your libcurl tests? Does it make any
difference if you use another?
Date: Sat, 18 Sep 2021 12:16:49 +0200 (CEST)
On Fri, 17 Sep 2021, Question Throwaway via curl-library wrote:
> With 7.60 I regularly will get an average of 450-500MS response time
>
> With 7.78 I regularly get between 4000-5500MS response time
Your showed code doesn't seem to be complete so we can't run that locally, but
I would suggest you try that test code on another platform first as I think
you'll see that there won't be any such extreme delays there.
Getting a ~10 times slower performance is not just a minor degradation or
wrong option. It's a sign of something seriously wrong. I would urge you to
get a full DEBUGFUNCTION log with timed entries to spot where all that time is
spent. Alternatively doing a full wireshark dump, but that might be tricker
for HTTPS traffic from that target.
Did you use the same TLS backend in both your libcurl tests? Does it make any
difference if you use another?
-- / daniel.haxx.se | Commercial curl support up to 24x7 is available! | Private help, bug fixes, support, ports, new features | https://curl.se/support.html -- Unsubscribe: https://lists.haxx.se/listinfo/curl-library Etiquette: https://curl.haxx.se/mail/etiquette.htmlReceived on 2021-09-18