Free program to maximise or change your cpu parameters - Quick CPU

[Mikeb]

Just came across this free software program which maybe useful in setting up your pc for audio playback. Its very easy to use and as far as I can tell clean. Direct link below.

https://coderbag.com/product/quickcpu

 

[audio_PHIL_e]

The web site makes it clear that the program is targetted at Intel CPUs. It won't do anything for you is your CPU is ARM or something else that's non-Intel.

 

[Chumpchops]

Not sure why this might be needed for any audio. Audio playback requires very little overhead from any modern Intel cpu, and certainly no reqs formultithreading or core reservations parking.

…might be useful for more IO intensive transaction stuff Or to minimise power consumption by closing cores when playing music. Or just if you like fiddling about.

TBH the inbuilt Intel and AMD auto optimisation tools already perform most utility function.

 

[tuga]

Not sure why this might be needed for any audio. Audio playback requires very little overhead from any modern Intel cpu, and certainly no reqs formultithreading or core reservations parking.

…might be useful for more IO intensive transaction stuff Or to minimise power consumption by closing cores when playing music. Or just if you like fiddling about.

TBH the inbuilt Intel and AMD auto optimisation tools already perform most utility function.

Upconverting with HQPlayer can max out even top-level i7s and i9s, which is why some people combine CPU and GPU power.

https://audiophilestyle.com/forums/topic/37775-best-cpu-for-hqplayer/

 

[Chumpchops]

Upconverting with HQPlayer can max out even top-level i7s and i9s, which is why some people combine CPU and GPU power.

https://audiophilestyle.com/forums/topic/37775-best-cpu-for-hqplayer/

Ha ha. That thread is funny. Only one person experiencing problems with his diy pc build. And that appears similar situation to saying that ‘my hi fi distorts when I turn volume to 11! ‘.

The wisdom (or otherwise) of upsampling to DSD 512, and using potentially flakey (in system resource management terms) one man band SW to do so aside, most i5 and i7 chipsets and their AMD Ryzen oppos, are more than capable of running with little or no latency using there own on board utilities. Majority of posts on that thread concur with this.

That said, and contentious fraudiophile jokes aside, you can never have too much compute power. Keep on buying.

I work for largest PC and HPC global market share company btw ;-)

 

[tuga]

Ha ha. That thread is funny. Only one person experiencing problems with his diy pc build. And that appears similar situation to saying that ‘my hi fi distorts when I turn volume to 11! ‘.

The wisdom (or otherwise) of upsampling to DSD 512, and using potentially flakey (in system resource management terms) one man band SW to do so aside, most i5 and i7 chipsets and their AMD Ryzen oppos, are more than capable of running with little or no latency using there own on board utilities. Majority of posts on that thread concur with this.

That said, and contentious fraudiophile jokes aside, you can never have too much compute power. Keep on buying.

I work for largest PC and HPC global market share company btw ;-)

Just out of curiosity, what's your experience with HQPlayer?
You could download the free demo if you have a DSD-able DAC and see how your PC performs.

I have a 2018 i3. It can't do the more resource-hungry filters or noise-shapers at DSD256, even if I turn of DRC. Even people using the M1 Mac mini are struglling with those.

My older 2012 i7 laptop can't even do the lighter ones. I have yet to test my wife's new M1 laptop.

 

[tuga]

Ha ha. That thread is funny. Only one person experiencing problems with his diy pc build. And that appears similar situation to saying that ‘my hi fi distorts when I turn volume to 11! ‘.

The wisdom (or otherwise) of upsampling to DSD 512, and using potentially flakey (in system resource management terms) one man band SW to do so aside, most i5 and i7 chipsets and their AMD Ryzen oppos, are more than capable of running with little or no latency using there own on board utilities. Majority of posts on that thread concur with this.

That said, and contentious fraudiophile jokes aside, you can never have too much compute power. Keep on buying.

I work for largest PC and HPC global market share company btw ;-)

The issue is not latency by the way but drop-outs, if you're lucky, or dop-ins in most cases.

 

[Chumpchops]

Just out of curiosity, what's your experience with HQPlayer?
You could download the free demo if you have a DSD-able DAC and see how your PC performs.

I have a 2018 i3. It can't do the more resource-hungry filters or noise-shapers at DSD256, even if I turn of DRC. Even people using the M1 Mac mini are struglling with those.

My older 2012 i7 laptop can't even do the lighter ones. I have yet to test my wife's new M1 laptop.

None. I have played with J River and Roon, but they were not for me.

I have few chunky PCs, but no longer use a PC for streaming or music playing SW. LMS runs on a Pi. If it’s your thing the 256 settings should run ok on those above. I suspect it is the developers code that may not make most efficient use of multicore loads and threading within using concurrencies. A surfeit of digital sonic knowledge over CPU usage optimisation code.it takes a huge development effort to compile for the varios cpu iterations on the market

…reading around HQP, I’d be intrigued to fiddle with filters in hpq, but ultimately that’s shag n hassle when a well resolved DAC can already do this with adjustments. Having moved to integrated 3 DACs per active box, with RC and mastertunings in the Buchardts, it’s all a bit academic tbh. But still have my Benchmark and Dyn actives, so may fiddle in my study at some point.

but, if I was looking at blowing £6k on an M Scaler, I’d certainly look at the SW option. I just don’t think I need ‘rounder wheels’

 

[Chumpchops]

The issue is not latency by the way but drop-outs, if you're lucky, or dop-ins in most cases.

Drop outs are often cause by latency - cpu times out as it overloads increasing latency, filling buffers then ditching digits, even if their are spare cores and threads. Back to processor optimisation coding in SW.
cheers

 

[tuga]

Drop outs are often cause by latency - cpu times out as it overloads increasing latency, filling buffers then ditching digits, even if their are spare cores and threads. Back to processor optimisation coding in SW.
cheers

According to Linkedin, HQPlayer's developer seems to have a bit of experience in the field:

Senior Linux Software Engineer 
Intel Corporation 
2011 - 2018 (7 years)
Principal Specialist 
Nokia 
2004 - 2011 (7 years) 




System Engineer 
Image Soft Oy 
1997 - 2004 (7 years)

And the activity monitor shows the process evenly distributed over all cores.

 

[tuga]

But still have my Benchmark

The Benchmark up or downsamples all inputs to 110kHz or 210kHz depending on the model.
And the ESS chip will not allow the bypassing of filters.
You won't be able to fully explore the benefits.

 

[Chumpchops]

The Benchmark up or downsamples all inputs to 110kHz or 210kHz depending on the model.
And the ESS chip will not allow the bypassing of filters.
You won't be able to fully explore the benefits.

Like I said, I am ambivalent to the whole rounder wheel upscaling thang

 

[tuga]

Like I said, I am ambivalent to the whole rounder wheel upscaling thang

Fair enough.

 

[Chumpchops]

According to Linkedin, HQPlayer's developer seems to have a bit of experience in the field:

Senior Linux Software Engineer 
Intel Corporation 
2011 - 2018 (7 years)
Principal Specialist 
Nokia 
2004 - 2011 (7 years) 




System Engineer 
Image Soft Oy 
1997 - 2004 (7 years)

And the activity monitor shows the process evenly d

But no silicon design or implementation experience. Activity may be distributed evenly across cores, but not with intra-core threading. Intel has teams of multiple discipline engineers from the silicon to the surrounding firmware and Sw levels. One man can’t do all of these in the time that this sw has come to market. It takes 1000s of man hours. But hey buy it. Personall, I buy products not projects, no matter how laudable their intent.
ymmv

 

[tuga]

But no silicon design or implementation experience. Activity may be distributed evenly across cores, but not with intra-core threading. Intel has teams of multiple discipline engineers from the silicon to the surrounding firmware and Sw levels. One man can’t do all of these in the time that this sw has come to market. It takes 1000s of man hours. But hey buy it. Personall, I buy products not projects, no matter how laudable their intent.
ymmv

I suggest you inform yourself a bit more before commenting.

 

[Chumpchops]

I suggest you inform yourself a bit more before commenting.

You are free to suggest all you like. I could likewise, but would not be that rude.

 

[tuga]

You are free to suggest all you like. I could likewise, but would not be that rude.

Your opinion would be different.

 

[Chumpchops]

Your opinion would be different.

You are clueless to even make that statement. And dull.

 

[Cable Monkey]

Try to be objective and lay off the personal insults chaps. It makes for better reading.