• We would like to remind our members that this is a privately owned, run and supported forum. You are here at the invitation and discretion of the owners. As such, rules and standards of conduct will be applied that help keep this forum functioning as the owners desire. These include, but are not limited to, removing content and even access to the forum.

    Please give yourself a refresher on the forum rules you agreed to follow when you signed up.

FM9 CPU Utilization Question

Deadpool_25

Power User
Just curious how the CPU usage differs between the two units. This is my very general, probably-not-accurate understanding.

FM3 (2 cores)
  • Amp and Delays (including all delay types? MTD, Plex, Etc?)
  • Everything else (including cab?)

FM9 (4 cores)
  • Amps
  • Reverbs
  • Delays (all types?)
  • Everything else (?)
Is that correct-ish?
 

DLC86

Fractal Fanatic
FM9 (4 cores)
  • Amps
  • Reverbs
  • Delays (all types?)
  • Everything else (?)
Since an fm3 core can only run one amp (and a delay) and the cores in the FM9 should be roughly the same, I'd say it's more likely like this:
  • Amp 1
  • Amp 2
  • Delays and reverbs
  • Everything else
 

Sixstring

Axe-Master
I've started running out of CPU on my Axe-Fx II and am wondering how the FM9 compares? Does the FM9 have more processing power?
Axe FX II if I'm not mistaking has 2 shark processors one of amp and cab sims and one for the effects. Per the WiKi, The FM9 uses the same processor type (“Griffin” DSP: one ARM and two SHARC+ cores) (4 cores) as previously posted

I'm going to go out on a limb here and say that it probably has more power to process than the Axe FX II, someone correct me if I'm wrong.
 

mr_fender

Axe-Master
I've started running out of CPU on my Axe-Fx II and am wondering how the FM9 compares? Does the FM9 have more processing power?

I believe the Axe II has about 40% of the processing power of the Axe III, while the FM9 is supposed to have about 50%. So it looks like the FM9's main DSPs might be roughly 25% more powerful than the Axe II. However, the Axe II's main processors handled a lot of system function overhead like USB and user interface processing. The FM9, like the Axe III, uses a separate dedicated processor for those functions, so it has even more main DSP power freed up for just audio processing. As a result, using Axe Edit or USB has zero impact on system performance. Also, no need to pause AxeEdit anytime you mess with the front panel or foot controller either. It all works together seamlessly.
 
Last edited:

unix-guy

Legend!
Since an fm3 core can only run one amp (and a delay) and the cores in the FM9 should be roughly the same, I'd say it's more likely like this:
  • Amp 1
  • Amp 2
  • Delays and reverbs
  • Everything else
No, that's not correct as confirmed by Cliff in another thread.

The OP is correct except I have not seen confirmation that all types of delays are using the same core... But if there's a single core just for delays then it's a pretty good assumption.

As far as 2 Amp blocks on a single core it's a good point... But it's what Cliff said.
 

DLC86

Fractal Fanatic
No, that's not correct as confirmed by Cliff in another thread.

The OP is correct except I have not seen confirmation that all types of delays are using the same core... But if there's a single core just for delays then it's a pretty good assumption.

As far as 2 Amp blocks on a single core it's a good point... But it's what Cliff said.
Could you point me to the post? I can't seem to find where Cliff said it
 

Deadpool_25

Power User
I’m not sure if/where Cliff actually said it but this is in the Wiki:

“The delays and reverbs on the FM9 each have a dedicated core DSP.”

I’m not sure the source of that information but the Wiki is usually pretty accurate. That implies that reverbs are on one core and delays on another.

Because the amp block and some of the delay blocks (perhaps just Delay and not MTD, etc.—idk) run on a single core in the FM3, and because they seem to have offloaded the delays, perhaps they can now fit two amp blocks on one core. I’m just speculating and making uneducated guesses based on incomplete information though.

Maybe when mine arrives I can add an amp block to a preset (or even use a blank preset) and start loading it up with delay, MTD, Plex, etc blocks—all the delay type blocks—and see what effect that has on the displayed CPU usage. If I understand it correctly, adding Reverb1 and Reverb2 have a negligible effect, presumably because they’re on their own core. If the same applies to any/all of the delay types that might indicate they’re also on their own core.

Additionally, both the FM3 and FM9 have only two Delay blocks. If the FM3 had them sharing with an Amp block and the FM9 moved them to their own block, it sure seems possible that more than just those two delay blocks could use that core, hence why I’m curious if other (delay types) blocks that were (we think) included in the “everything else” core may have also been moved to the FM9’s “delay core.”
 
Last edited:

@danserrano

Experienced
The reverbs having its own dedicated processor is huge. I saw the video posted by chris and adding two reverb blocks at hi res raised the overall patch cpu by 1%. That’s fantastic
 

Deadpool_25

Power User
FM9 Manual:

"Some blocks such as Amp, Delay, and Reverb have hardly any effect on the meter at all because they run on dedicated CPU cores."

I think that "Delay" applies only to the regular Delay blocks, not to the Multitap, Megatap etc.

Thanks Yek! I missed that line in the manual.
 

javajunkie

Moderator
Moderator
FM9 Manual:

"Some blocks such as Amp, Delay, and Reverb have hardly any effect on the meter at all because they run on dedicated CPU cores."

I think that "Delay" applies only to the regular Delay blocks, not to the Multitap, Megatap etc.
The Plex delay definitely is not on the dedicated delay core. It consumes quite a bit of processing power, especially when set to Plex Delay.
Delays hardly add any thing, Reverbs hardly add anything (doesn't atter in if you set to ultra high or economy - the dsp meter does not significantly change)..

Drive, Pitch, and Plex Verb are the blocks that eat up the most processing for me. I usually stack drive pedals. On the III , I use 3 drive blocks just for easy of thought. boost, overdrive, distortion. I cut down to 2 on the FM9.

Getting rid of unneeded shunts can help too. The don't use much, but they ad up.

With the dedicate cores, I am able to get a lot more than I thought ported over.
 

@danserrano

Experienced
The Plex delay definitely is not on the dedicated delay core. It consumes quite a bit of processing power, especially when set to Plex Delay.
Delays hardly add any thing, Reverbs hardly add anything (doesn't atter in if you set to ultra high or economy - the dsp meter does not significantly change)..

Drive, Pitch, and Plex Verb are the blocks that eat up the most processing for me. I usually stack drive pedals. On the III , I use 3 drive blocks just for easy of thought. boost, overdrive, distortion. I cut down to 2 on the FM9.

Getting rid of unneeded shunts can help too. The don't use much, but they ad up.

With the dedicate cores, I am able to get a lot more than I thought ported over.
Whats your cpu on the fm9 give or take with that type of patch?
 

javajunkie

Moderator
Moderator
1631912943596.png
With Cabs set to Ultra Res, I kept the plex delay on the Plex Delay Channel until I need it for the Plex verb, the cpu goes to 85% once I change it to Plex Delay. That is right on the verge of muting out.
 
Top Bottom