Which CPU properties are best for using GP

I would like to give an update.
My new machine is currently working absolutely problemless,
without any cracks as well loaded with many CPU stressing
pluginns.

Following Hardware is in use:
Intel Core i5-12600K / ASUS PRIME Z690-P WIFI D4 DDR4 Bundle

Prozessor
Intel Core i5-12600K, 10 Core (6+4), 6x 3700 MHz
overclocked to 4.900 MHz

Prozessor Cooler
be quiet! Shadow Rock LP

Grafik Card
Intel U/HD-Grafik OnChip

Mainboard
ASUS PRIME Z690-P WIFI D4

RAM
32 GB DDR4-RAM, Dual Channel (2x 16 GB), 3600 MHz, Corsair Vengeance LPX Black (+109,90 €)

Maindrive / SSD / m.2 SSD
1 x 500 GB M.2 PCIe 3.0 SSD Samsung 980 (Lesen/Schreiben: max. 3100 MB/s | 2600 MB/s)
1 x 1000 GB M.2 PCIe 4.0 SSD Samsung 980 Pro (Lesen/Schreiben: max. 7000 MB/s | 5000 MB/s)

Housing
be quiet! Pure Base 500DX schwarz, mit Glas window,

power adapter
600 Watt be quiet! Pure Power 11, 92 % Effizienz, 80 Plus Gold certified

WLAN/Bluetooth
2400 MBit/s Wireless LAN, Intel® AX 201, WiFi 6 802.11 ax WLAN, 2.4 GHz & 5 GHz Dual Band, Bluetooth 5.1

Netzweork Card
onBoard LAN
Sound Card
onBoard HD Sound

Boost (not overclocked) frequency, I assume?

Nice setup :+1:t2:

don’t know…I thought it calls overclocking :blush:

The recent Intel desktop processors have a “base” frequency and a “max turbo” frequency. On the i5-12600k I believe the base is 3.7 and the max turbo default is 4.9.

With various settings you can force them to always run at the “max turbo” frequency, and with many motherboards (including the Asus Prime Z690) you can pretty reliably overclock them higher to 5.1 or 5.2 gHz if desired.

The system I’m on right now is an i5-12600k that I have running at 5.0gHz all the time.

My experience with audio on Windows PCs is that you get much more benefit from forcing them to always run at the “max turbo” speed than by pushing the overclocking.

The simple reason is that in the name of saving power Windows will slow your CPU speed down to the “base” frequency under light load, and even shut down cores if you let it. The nature of running “real time” audio is that the loads will on average tend to be fairly light, so your CPU will tend to run at “base frequency” (or at least drop to it frequently) if you let it, then you’ll get glitching if the base frequency wasn’t quite enough to handle one sample buffer.

In the gaming world the tradeoffs tend to be different. There you might be better off letting the CPU run at 3.7GHz most of the time but overclocking the “max turbo” to 5.5GHz, but most systems will only be able to sustain the thermal load at that speed briefly before thermal throttling happens. In my experience that kind of thing doesn’t help audio. Your glitches will happen when the CPU is running below top speed, and the system won’t decide to ramp up the speed until after the glitch has happened. Then it’s too late.

Bottom line, I’m “overclocking” mine from 4.9GHz to 5.0GHz, but the overclocking part is less important than forcing it to not drop down to 3.7. You can force windows to run at “max turbo” all the time without getting into more elaborate “overclocking”.

3 Likes

Good explanation! This is exactly what I did (at least I guess that I set the it correctly in the bios setting of the motherboard).
Beside of the basis or general performance of the CPU and other performance relevant hardware it is important that “all” components “fit together” for live audio machines.
For that reason I listet my complete hardware setup.

As mentioned before, it is helpful to eliminate all not needed software (and it’s update function) in Windows to avoid any possible interruption during the performance.

What audio interface are you using and at what buffer size/latency?