Hi,

I’ve recently started trying out Csound 7 which has prompted me to use Core Audio instead of Port Audio and I’ve noticed that regardless of how simple my code is or what I set ksmps to or what CsOptions flags I use, elapsed time at end of performance always shows almost exactly the same number for real and CPU and Activity Monitor shows Csound using a fairly constant 100.3%. 
I’m not experiencing any problems with audio performance so I’m wondering if this is the expected behaviour with Core Audio or if there is a problem somewhere in my system?
The exact same thing happens with Csound 6.18 and the last file I tested was only using around 6% CPU when I tried it with Port Audio in 6.18.
I wondered if it might be related to my MOTU M4 USB audio interface but it happens even when that is disconnected and I use the built in speakers. I deleted the drivers for the MOTU M4 and that didn’t change anything either.
I’m using a MacBook Air M3 running macOS Sequoia 15.5. 

Thanks,
Adam
Csound mailing list
Csound@listserv.heanet.ie
https://listserv.heanet.ie/cgi-bin/wa?A0=CSOUND
Send bugs reports to
        https://github.com/csound/csound/issues
Discussions of bugs and features can be posted here

That’s not your system, the coreaudio backend does that alright. What happens is that while
Csound is waiting for samples to be played, it runs a loop that keeps the CPU busy. The
reason you see no impact on performance is that if Csound needs to do more computation,
it then takes less time in that loop, and the result is a steady 100% cpu stat. However if
we look at the overall system resource use, we will see that this is not actually having an
impact on user cpu (well it’s more than portaudio), and I think it’s because it’s not doing
much computation except for looping.

There is code there commented out to give back the cpu to the system. I tested it and
it brings the stats back to normal. I will mainline that code and monitor its performance.

Thanks
========================
Prof. Victor Lazzarini
Maynooth University
Ireland






> On 4 Jul 2025, at 00:57, Adam Willetts  wrote:
>
> [You don't often get email from adam@ethermap.org. Learn why this is important at https://aka.ms/LearnAboutSenderIdentification ]
>
> *Warning*
>
> This email originated from outside of Maynooth University's Mail System. Do not reply, click links or open attachments unless you recognise the sender and know the content is safe.
>
> Hi,
>
> I’ve recently started trying out Csound 7 which has prompted me to use Core Audio instead of Port Audio and I’ve noticed that regardless of how simple my code is or what I set ksmps to or what CsOptions flags I use, elapsed time at end of performance always shows almost exactly the same number for real and CPU and Activity Monitor shows Csound using a fairly constant 100.3%.
> I’m not experiencing any problems with audio performance so I’m wondering if this is the expected behaviour with Core Audio or if there is a problem somewhere in my system?
> The exact same thing happens with Csound 6.18 and the last file I tested was only using around 6% CPU when I tried it with Port Audio in 6.18.
> I wondered if it might be related to my MOTU M4 USB audio interface but it happens even when that is disconnected and I use the built in speakers. I deleted the drivers for the MOTU M4 and that didn’t change anything either.
> I’m using a MacBook Air M3 running macOS Sequoia 15.5.
>
> Thanks,
> Adam
> Csound mailing list
> Csound@listserv.heanet.ie
> https://listserv.heanet.ie/cgi-bin/wa?A0=CSOUND
> Send bugs reports to
>        https://github.com/csound/csound/issues
> Discussions of bugs and features can be posted here


Csound mailing list
Csound@listserv.heanet.ie
https://listserv.heanet.ie/cgi-bin/wa?A0=CSOUND
Send bugs reports to
        https://github.com/csound/csound/issues
Discussions of bugs and features can be posted here

Hi Victor

Thanks for the detailed explanation. It’s a relief to know that everything is working as intended and I really appreciate the insight into how Csound works with the core audio backend.
I’m excited to get on with exploring the new features of Csound 7 now that I know that 100% CPU stat is nothing to worry about.

Thanks
Adam

> On 4 Jul 2025, at 19:57, Victor Lazzarini <000010b17ddd988e-dmarc-request@LISTSERV.HEANET.IE> wrote:
> 
> That’s not your system, the coreaudio backend does that alright. What happens is that while
> Csound is waiting for samples to be played, it runs a loop that keeps the CPU busy. The
> reason you see no impact on performance is that if Csound needs to do more computation,
> it then takes less time in that loop, and the result is a steady 100% cpu stat. However if
> we look at the overall system resource use, we will see that this is not actually having an
> impact on user cpu (well it’s more than portaudio), and I think it’s because it’s not doing
> much computation except for looping.
> 
> There is code there commented out to give back the cpu to the system. I tested it and
> it brings the stats back to normal. I will mainline that code and monitor its performance.
> 
> Thanks
> ========================
> Prof. Victor Lazzarini
> Maynooth University
> Ireland
> 
> 
> 
> 
> 
> 
>> On 4 Jul 2025, at 00:57, Adam Willetts  wrote:
>> 
>> [You don't often get email from adam@ethermap.org. Learn why this is important at https://aka.ms/LearnAboutSenderIdentification ]
>> 
>> *Warning*
>> 
>> This email originated from outside of Maynooth University's Mail System. Do not reply, click links or open attachments unless you recognise the sender and know the content is safe.
>> 
>> Hi,
>> 
>> I’ve recently started trying out Csound 7 which has prompted me to use Core Audio instead of Port Audio and I’ve noticed that regardless of how simple my code is or what I set ksmps to or what CsOptions flags I use, elapsed time at end of performance always shows almost exactly the same number for real and CPU and Activity Monitor shows Csound using a fairly constant 100.3%.
>> I’m not experiencing any problems with audio performance so I’m wondering if this is the expected behaviour with Core Audio or if there is a problem somewhere in my system?
>> The exact same thing happens with Csound 6.18 and the last file I tested was only using around 6% CPU when I tried it with Port Audio in 6.18.
>> I wondered if it might be related to my MOTU M4 USB audio interface but it happens even when that is disconnected and I use the built in speakers. I deleted the drivers for the MOTU M4 and that didn’t change anything either.
>> I’m using a MacBook Air M3 running macOS Sequoia 15.5.
>> 
>> Thanks,
>> Adam
>> Csound mailing list
>> Csound@listserv.heanet.ie
>> https://listserv.heanet.ie/cgi-bin/wa?A0=CSOUND
>> Send bugs reports to
>>       https://github.com/csound/csound/issues
>> Discussions of bugs and features can be posted here
> 
> 
> Csound mailing list
> Csound@listserv.heanet.ie
> https://listserv.heanet.ie/cgi-bin/wa?A0=CSOUND
> Send bugs reports to
>        https://github.com/csound/csound/issues
> Discussions of bugs and features can be posted here

Csound mailing list
Csound@listserv.heanet.ie
https://listserv.heanet.ie/cgi-bin/wa?A0=CSOUND
Send bugs reports to
        https://github.com/csound/csound/issues
Discussions of bugs and features can be posted here

Thx. I have some changes now that give cpu back, combined with other changes that also improve its performance. Hopefully they will be merged soon.

Prof. Victor Lazzarini
Maynooth University
Ireland

> On 5 Jul 2025, at 02:12, Adam Willetts  wrote:
> 
> Hi Victor
> 
> Thanks for the detailed explanation. It’s a relief to know that everything is working as intended and I really appreciate the insight into how Csound works with the core audio backend.
> I’m excited to get on with exploring the new features of Csound 7 now that I know that 100% CPU stat is nothing to worry about.
> 
> Thanks
> Adam
> 
>> On 4 Jul 2025, at 19:57, Victor Lazzarini <000010b17ddd988e-dmarc-request@LISTSERV.HEANET.IE> wrote:
>> 
>> That’s not your system, the coreaudio backend does that alright. What happens is that while
>> Csound is waiting for samples to be played, it runs a loop that keeps the CPU busy. The
>> reason you see no impact on performance is that if Csound needs to do more computation,
>> it then takes less time in that loop, and the result is a steady 100% cpu stat. However if
>> we look at the overall system resource use, we will see that this is not actually having an
>> impact on user cpu (well it’s more than portaudio), and I think it’s because it’s not doing
>> much computation except for looping.
>> 
>> There is code there commented out to give back the cpu to the system. I tested it and
>> it brings the stats back to normal. I will mainline that code and monitor its performance.
>> 
>> Thanks
>> ========================
>> Prof. Victor Lazzarini
>> Maynooth University
>> Ireland
>> 
>> 
>> 
>> 
>> 
>> 
>>>> On 4 Jul 2025, at 00:57, Adam Willetts  wrote:
>>> 
>>> [You don't often get email from adam@ethermap.org. Learn why this is important at https://aka.ms/LearnAboutSenderIdentification ]
>>> 
>>> *Warning*
>>> 
>>> This email originated from outside of Maynooth University's Mail System. Do not reply, click links or open attachments unless you recognise the sender and know the content is safe.
>>> 
>>> Hi,
>>> 
>>> I’ve recently started trying out Csound 7 which has prompted me to use Core Audio instead of Port Audio and I’ve noticed that regardless of how simple my code is or what I set ksmps to or what CsOptions flags I use, elapsed time at end of performance always shows almost exactly the same number for real and CPU and Activity Monitor shows Csound using a fairly constant 100.3%.
>>> I’m not experiencing any problems with audio performance so I’m wondering if this is the expected behaviour with Core Audio or if there is a problem somewhere in my system?
>>> The exact same thing happens with Csound 6.18 and the last file I tested was only using around 6% CPU when I tried it with Port Audio in 6.18.
>>> I wondered if it might be related to my MOTU M4 USB audio interface but it happens even when that is disconnected and I use the built in speakers. I deleted the drivers for the MOTU M4 and that didn’t change anything either.
>>> I’m using a MacBook Air M3 running macOS Sequoia 15.5.
>>> 
>>> Thanks,
>>> Adam
>>> Csound mailing list
>>> Csound@listserv.heanet.ie
>>> https://listserv.heanet.ie/cgi-bin/wa?A0=CSOUND
>>> Send bugs reports to
>>>      https://github.com/csound/csound/issues
>>> Discussions of bugs and features can be posted here
>> 
>> 
>> Csound mailing list
>> Csound@listserv.heanet.ie
>> https://listserv.heanet.ie/cgi-bin/wa?A0=CSOUND
>> Send bugs reports to
>>       https://github.com/csound/csound/issues
>> Discussions of bugs and features can be posted here
> 
> Csound mailing list
> Csound@listserv.heanet.ie
> https://listserv.heanet.ie/cgi-bin/wa?A0=CSOUND
> Send bugs reports to
>        https://github.com/csound/csound/issues
> Discussions of bugs and features can be posted here

Csound mailing list
Csound@listserv.heanet.ie
https://listserv.heanet.ie/cgi-bin/wa?A0=CSOUND
Send bugs reports to
        https://github.com/csound/csound/issues
Discussions of bugs and features can be posted here

Thanks for this Adam, it’s good to have a confirmation. I also have a PR improving the responsiveness of portaudio on *nix,
which should bring its performance to a similar level as rtauhal. Once that is approved and merged, there will be a new
updated release.

For a short intro to a number of new features in 7.0, we have

https://csound.com/manual/intro/whats-new-in-csound-7/

there are other things which have not yet been added to that introduction, hopefully I can do some more
documenting now that we’re almost settled on the features.

best regards
========================
Prof. Victor Lazzarini
Maynooth University
Ireland






> On 26 Jul 2025, at 05:29, Adam Willetts  wrote:
> 
> You don't often get email from adam@ethermap.org. Learn why this is important I just tried the new build that has the core audio improvements merged and I’m completely blown away by the performance.  
> As before, core audio seems to be able to cope with smaller buffer sizes and has subjectively slightly nicer sound quality (at least to my ears) compared to port audio but the difference in CPU performance is now huge. The piece I’m currently working on is using 10 - 12% CPU with port audio but only 2.5 - 2.8% with core audio!!! 
> Admittedly, there are only a small number of short duration voices playing simultaneously so this might not be indicative of potential performance under more demanding conditions but I am using ksmps = 1 so I’m extremely impressed.
> 
> With Csound 7 now reaching a beta release, I’d like to offer a massive thank you and congratulations to everyone who has contributed to its development, especially the core development team. All of your hard work and dedicated commitment to both innovation and backwards compatibility ensure that Csound continues to evolve and help shape the future of computer music while also being unrivalled for its active preservation of computer music history. I’ve barely scratched the surface of the new features since I started moving from Csound 6 to Csound 7 a few weeks ago but I’ve been having a wonderful time and I am absolutely loving the new syntax.
> 
> Thanks!
> Adam
> 
> 
> 
>> On 5 Jul 2025, at 19:39, vlz  wrote:
>> 
>> Thx. I have some changes now that give cpu back, combined with other changes that also improve its performance. Hopefully they will be merged soon.
>> 
>> Prof. Victor Lazzarini
>> Maynooth University
>> Ireland
>> 
>>> On 5 Jul 2025, at 02:12, Adam Willetts  wrote:
>>> 
>>> Hi Victor
>>> 
>>> Thanks for the detailed explanation. It’s a relief to know that everything is working as intended and I really appreciate the insight into how Csound works with the core audio backend.
>>> I’m excited to get on with exploring the new features of Csound 7 now that I know that 100% CPU stat is nothing to worry about.
>>> 
>>> Thanks
>>> Adam
>>> 
>>>> On 4 Jul 2025, at 19:57, Victor Lazzarini <000010b17ddd988e-dmarc-request@LISTSERV.HEANET.IE> wrote:
>>>> 
>>>> That’s not your system, the coreaudio backend does that alright. What happens is that while
>>>> Csound is waiting for samples to be played, it runs a loop that keeps the CPU busy. The
>>>> reason you see no impact on performance is that if Csound needs to do more computation,
>>>> it then takes less time in that loop, and the result is a steady 100% cpu stat. However if
>>>> we look at the overall system resource use, we will see that this is not actually having an
>>>> impact on user cpu (well it’s more than portaudio), and I think it’s because it’s not doing
>>>> much computation except for looping.
>>>> 
>>>> There is code there commented out to give back the cpu to the system. I tested it and
>>>> it brings the stats back to normal. I will mainline that code and monitor its performance.
>>>> 
>>>> Thanks
>>>> ========================
>>>> Prof. Victor Lazzarini
>>>> Maynooth University
>>>> Ireland
>>>> 
>>>> 
>>>> 
>>>> 
>>>> 
>>>> 
>>>>>> On 4 Jul 2025, at 00:57, Adam Willetts  wrote:
>>>>> 
>>>>> [You don't often get email from adam@ethermap.org. Learn why this is important at https://aka.ms/LearnAboutSenderIdentification ]
>>>>> 
>>>>> *Warning*
>>>>> 
>>>>> This email originated from outside of Maynooth University's Mail System. Do not reply, click links or open attachments unless you recognise the sender and know the content is safe.
>>>>> 
>>>>> Hi,
>>>>> 
>>>>> I’ve recently started trying out Csound 7 which has prompted me to use Core Audio instead of Port Audio and I’ve noticed that regardless of how simple my code is or what I set ksmps to or what CsOptions flags I use, elapsed time at end of performance always shows almost exactly the same number for real and CPU and Activity Monitor shows Csound using a fairly constant 100.3%.
>>>>> I’m not experiencing any problems with audio performance so I’m wondering if this is the expected behaviour with Core Audio or if there is a problem somewhere in my system?
>>>>> The exact same thing happens with Csound 6.18 and the last file I tested was only using around 6% CPU when I tried it with Port Audio in 6.18.
>>>>> I wondered if it might be related to my MOTU M4 USB audio interface but it happens even when that is disconnected and I use the built in speakers. I deleted the drivers for the MOTU M4 and that didn’t change anything either.
>>>>> I’m using a MacBook Air M3 running macOS Sequoia 15.5.
>>>>> 
>>>>> Thanks,
>>>>> Adam
>>>>> Csound mailing list
>>>>> Csound@listserv.heanet.ie
>>>>> https://listserv.heanet.ie/cgi-bin/wa?A0=CSOUND
>>>>> Send bugs reports to
>>>>>     https://github.com/csound/csound/issues
>>>>> Discussions of bugs and features can be posted here
>>>> 
>>>> 
>>>> Csound mailing list
>>>> Csound@listserv.heanet.ie
>>>> https://listserv.heanet.ie/cgi-bin/wa?A0=CSOUND
>>>> Send bugs reports to
>>>>      https://github.com/csound/csound/issues
>>>> Discussions of bugs and features can be posted here
>>> 
>>> Csound mailing list
>>> Csound@listserv.heanet.ie
>>> https://listserv.heanet.ie/cgi-bin/wa?A0=CSOUND
>>> Send bugs reports to
>>>       https://github.com/csound/csound/issues
>>> Discussions of bugs and features can be posted here
>> 
>> Csound mailing list
>> Csound@listserv.heanet.ie
>> https://listserv.heanet.ie/cgi-bin/wa?A0=CSOUND
>> Send bugs reports to
>>        https://github.com/csound/csound/issues
>> Discussions of bugs and features can be posted here
> 
> Csound mailing list Csound@listserv.heanet.ie https://listserv.heanet.ie/cgi-bin/wa?A0=CSOUND Send bugs reports to https://github.com/csound/csound/issues Discussions of bugs and features can be posted here


Csound mailing list
Csound@listserv.heanet.ie
https://listserv.heanet.ie/cgi-bin/wa?A0=CSOUND
Send bugs reports to
        https://github.com/csound/csound/issues
Discussions of bugs and features can be posted here