Have either of you adjusted (or even looked at) the settings for the frequency display (FFT display)? If you set the window size large, it will definitely suck up CPU. What are the defaults for n V4?
IIRC, in V3, window size is 2048, and “FFT size” and “Compute every # samples” are both set to “Window Size”. If any of these are set differently (bigger for window & fft, smaller for Compute), it will cause CPU usage to increase.
This won’t affect showing the mixer, unless the mixer now shows an FFT by default. (I don’t run N v4 yet.)
Jeff, it’s definately not an FFT thing. You don’t even need the analyzer up for this to occur. With 5-6 tracks running no plugs my CPU usage is at something like 20-25%. Now, I am still using ver 3.3, so any fix that 4.0 might have wouldn’t apply to me, but it seems that same batch of code is causing it in both version. I mean, I don’t think there was any change to the deep guts of the code from 3.3 to 4.0 .
This is indeed odd. For comparison sake with Bubba’s example above…Last night I had a 6-Track song all 24/48 .wavs loaded. Playing back with SIR reverb on Aux 1 and the Kjaerhaus Classic compressor on the master effects, CPU use was around 28%. Dropping the SIR, the CPU went to about 12%. No effects at all save for EQ adjustments in n’s mixer yielded 8-10% CPU. All with n-Track V3.3.
I run an Athlon XP 2500 with 512 meg RAM. You would think that Bubba’s P4 2.6 Ghz machine would KILL results like mine?
TG - curious 
double post
AFAIK, I am just running the default setting. I haven’t changed anything on a new install. I know in the past I did mess with the settings on this and had all kinds of problems until I did a clean reinstall, but that was with ver 3.3.
Blessings, Terry
| Quote (theshaggyfreak @ Oct. 10 2004,07:32) |
| My machine runs a p4 1.7ghz, and I tested both the mixer and the built in EQ plugin. The mixer makes the CPU usage jump about 6-8%. The EQ jumps about 20-30%. |
Btw, these jumps only happen when the plugin is being viewed. If it is hidden, I barely notice a hit on the CPU.
TG, that’s what I am saying. My Athlon 1800+ smokes my P4 2.66Ghz. Doesn’t make a whole lot of sense. Ver 3.3 seems to have the narmalization bug at near 0 levels on the P4 and not on the Athlon as well. Drove me nuts mixin g for the Italian Festival a few weeks back. Some tracks would go silent for parts of songs and bump the CPU to near max and get all poppy and clicky and until the track hit a non silent part I couldn’t get the system to respond to Ctl-Alt-Del or any other command. Can you tell I am frustrated with my P4 and N?
| Quote (gtr4him @ Oct. 06 2004,13:38) |
| ....my newer DAW, Athlon 2500+ (1.8ghz) SMOKES with n-Track |
Okay, yeah, this is OT,
but are you sure the 2500+ runs at 1.8ghz? Cause I'm pretty sure that's what my 2200+ ran at before I lightly overclocked it to 1.93ghz. No way I made a jump from a 2200+ to a 2600+ or so with what little I did.
I'll second that emotion on Athlons and audio/Ntrack. No problems for me, and I've even got a via KT333 chipset.
j
I know my Athlon XP 2500+ Barton-core is at 1.83 GHz, but AMD rates it higher than the Thoroughbred-core 2200+ @ 1.8, and even the 2 GHz 2400+, because it supposedly performs better due to the extra cache, 333 FSB, and whatever else. Remember: it’s a “performance rating”…confusion is part of the game 
Tony
I upgraded my MD a few weeks ago and am now getting it back in working order. It’s basically a new machine considering everything I changed. There is a new hard drive in there, but it’s a 7200 RPM WD – the same as my main recording drive with the old config. All the old drivers are still in the machine, including the old boot C: drive. It’s now E:.
P4 - 3.2 ghz - 1 gig memory (Soyo Dragon2 V 1.0)
WinXp Sp2
n-Tracks V3.3-1516
10 - 24 bit stereo tracks
5 - 24 bit mono tracks
13 plug-ins - not counting built-in EQ, which is being used on most tracks
2 - Aux returns being used
4 subgroups
Output to the Layla20 in MME 16 bit mode (Gina24 and onboard C-Media – MAJOR YUCKO – are enabled and working, except in the case of the C-Media working is relative. It audibly clicks when a CD is placed in the CD or DVD drive.)
All track output meters showing in the mixer and output VU - CPU: ~8% to ~9%total
This is MUCH better than the old PIII 750 mhz with 512 megs ram. That config was up around ~60-70% all the time.
I’m WAY amazed at the difference in this new MB and CPU.
I haven’t been able to get these songs to work in V4. so I can’t compare this to it. I’ve been in touch with Flavio, so he’s working on finding the problem. It’s not related to the new MB.
Now the one thing you have different in your setup, Phoo, is that you are using MME drivers. I am using ASIO. Do you suppose that could cause a difference? I am on build 1516 as well and a song like that would totally cramp up my P4. I was running 6-8 stereo tracks with about 4 effects on a recent project and the CPU was hitting 80-90% at sections where one track was silent. Very strange. I might just grab the demo of ver 4.0 and see what happens. Flavio supposedly squished a similar bug a few weeks back. Let’s see what happens even though some have reported the bug came back in later builds.
Maybe…I’ll switch drivers and see what it does. I never used ASIO because I didn’t need low latency and ASIO didn’t do well for a long time. The latest Echo drivers fixed that, but I still don’t need low latency. (I bet a lot of folks that don’t need it think they do…different thread discussion).
Does yours get better when switching to MME?
The silent track thing should be fixed, but that doesn’t mean CPU usage won’t still go up when things get quiet (math done on very small floating point numbers)…wouldn’t a blank track do the same thing since it will simply send zeros to be processed?
Does yours get better when switching to MME?
I haveone of the new EMUs and they are an ASIO or WDM only beast. The WDM is there only for game and monitoring support, so for multitrack, ASIO is it.
The silent track thing should be fixed, but that doesn't mean CPU usage won't still go up when things get quiet (math done on very small floating point numbers)....wouldn't a blank track do the same thing since it will simply send zeros to be processed?
I don't know. The CPU goes nuts when the track has not wav information at all. So in this case it was a 30 second wav, had maybe 2 minutes of nothing in the track, and then another wav. As soon as it hit the empty parts, things went crazy so I assume the track is treated as a bunch of 0 samples rather than ignored.
I haveone of the new EMUs and they are an ASIO or WDM only beast. The WDM is there only for game and monitoring support, so for multitrack, ASIO is it.
The silent track thing should be fixed, but that doesn't mean CPU usage won't still go up when things get quiet (math done on very small floating point numbers)....wouldn't a blank track do the same thing since it will simply send zeros to be processed?
I don't know. The CPU goes nuts when the track has not wav information at all. So in this case it was a 30 second wav, had maybe 2 minutes of nothing in the track, and then another wav. As soon as it hit the empty parts, things went crazy so I assume the track is treated as a bunch of 0 samples rather than ignored.
Bubba,
Does it make any difference if you playback with only the WDM driver selected? This is a weird one. I’m going to be building a DAW for our church soon and briefly thought about a second-hand P4 machine I can get fairly cheap. NOW I’m thinking I should stick to what I know and that is that n works well on my AMD box. A guy can build a nice AMD box for about 600 clams these days…
TG
It seems to make a difference, but very slight. I need to go home and play with hyperthreading and see if that changes anything. It seems HT screws with a lot of DAWs out there and I may just be b$&*#ing about nothing. I am almost cetain I have HT disabled, but I had better pop into the BIOS and be 110% sure.
I have HT turned on for all machines I have that support it (work, and now at home). It might screw some things up, but it shouldn’t, and it should make some things a lot better. No harm in toggling it to see which works best.
I haven’t see any problems related to HT on any of my machines, but I also have two multi-processor machines (one is multi-proc AND HT – an Xeon), and those things have major problems with some things - so far all driver related.
I’ll try using MME and WDM tonight and see if that makes any difference.
Blessings, Terry
Well, I didn’t have time to record anything last night, and all the songs that I already have recorded are 24 bit. MME wouldn’t play 24 bit so I couldn’t try it last night. WDM did the same thing. The song I loaded was running about 36% CPU until I opened the EQ on one of the tracks. It then shot to 98%.
Blessings, Terry
Has anyone heard anything from Flavio recently on this issue?
I’m getting ready to build a new computer and I was originally going to go Intel all the way including a P4. But now I’m wondering if I should consider an Athalon.
…gary
Has this changed for anyone?