ASN SP4 makes P3D stutters after 15-20 min of usage.

[G-YMML1 582]

Not an issue here. 100/100 Comcast cable. AVSIM downloads ~14Mb/sec

[G-YMML1 582]

First turn off all AA, and default the NVCP profile. Anti-aliasing simply caps the fps at some point blurring the results while trying to set up for performance. Once set up, bring on AA until it starts to dent the performance then back off a notch.

 

With the 4 core, enable HT and use the Affinity Mask 85 and test. If you can make ASN run above core zero, on say cores 1,2,3, then continue to use AM=85. If you can only force ASN to start on core zero, then try AM=116. This uses 3 cores like AM=14 with HT Off, leaving core zero free for ASN. However with HT on the AM=116 can utilise 4 logical processors in a way that uses those three cores more efficiently. And with HT enabled, process switching time is halved overall. HT on will increase the heat if the proper AM is not used, and the sim will be unable to perform any better at 4GHz with more than 4 jobs running on 4 LPs.

 

When making performance tests use for example the stock B58, since the stock A36 requires a different mode of AA and can confuse the results while finding a setup. Beware some planes (like the A36) and scenery objects, can look poorly anti-aliased unless NI and Sparse grid sampling is utilised, halving your fps.

 

So far, my communication with HiFi support solves nothing, but we're still digging. However, as we already established that in my case ASN starts to put a lot of load on CPU after 15-20 minutes of activity, one of the suggestions I received is to assign ASN to the specific core and see what happens with the CPU load. Meanwhile, in the light of previous discussion in this thread I kinda confused if it's ok to play around with specific core assignment for ASN with HT=OFF. My understanding was that messing with cores is justifiable with HT = ON only.

 

Steve?

[jrs182p 23]

I've noticed ASN uses hardly any cpu when you first run it but after a while it starts showing about 15-20% CPU usage in task manager.

[G-YMML1 582]

I've noticed ASN uses hardly any cpu when you first run it but after a while it starts showing about 15-20% CPU usage in task manager.

 

In my case, >45% in some cases. ASN SP4 here

[G-YMML1 582]

What if my move ASN to another laptop and purchase WideFS to connect it with the sim on main PC? Will it solve CPU load issue?

[be77solo 125]

You've got me curios now, ASN shouldn't use anywhere near that much CPU.

 

Now I've got to go check mine, will report back.

[jrs182p 23]

I can confirm I've been flying for about two hours now and my ASN is using 15-20% CPU. and on longer flights specifically if you keep loading in new flight plans if your doing short hops the CPU usage just keeps rising.

[funkyj4ever 2]

how to put asn on core 0?

[G-YMML1 582]

how to put asn on core 0?

 

Going to test this theory later tonight even though I have HT=Off, AF=14 and OC-ed to 4.6. SteveW mentioned that any messing with the cores should be done with HT=ON

[be77solo 125]

Ran P3Dv3.1 this morning for a little over an hour.  Took off from Milford Sound, NZ (Orbx Region) in the JF DH 104 Dove with latest ASN running using live weather and flew NE at about 10K'. Had a few scattered storms, plenty of clouds and some spotty rain... bumpy ride over all ha.

 

Anyway, I kept task manager up as seen in the picture and ASN was almost always at ~2%, occasionally up to 8-10% but only when fiddling with the ASN weather map, and then it would drop right back down to ~1.5-2%.

This is running Windows 10 64, no affinity tweaks or masks, HT on using a mobile i7-4710HQ @3.7Ghz and 970m GPU, framerate locked and stable at 27fps.

 

 

[G-YMML1 582]

Arrived home, beginning my test in 5m

Here is what happened after 20 minutes of sitting at KJFK, updates are disabled, weather is snow + overcast. FPS are stable, stuttering is present.

https://www.dropbox.com/s/jnvdi57bf6m6nlq/123.PNG?dl=0

...almost 50% now and I'm not flying anywhere!

[vgbaron 3,626]

Does this happen with live weather disabled too?  I've flown for hours (5 or 6) with absolutely NO sign ificant change in cpu usage with ASN. I haven't a clue WHAT it is but there is SOMETHING on your system that is not playing nice with ASN.

 

What happens if you disable weather updates completely?

 

Look really close at any anti virus checker you have to be sure it is not scanning the ASN folder periodically.

 

Vic

[G-YMML1 582]

Update: 

 

1. Disabled cores 1,2 and 3 for ASN, leaving core 0 only, The stuttering has gone IMMEDIATELY and CPU usage dropped from 53% to 13% in a second. 

 

2. Checked via PE and found out that P3D uses cores 1-2-3, but not core 0. Have no idea why, but it is what it is.

 

3. Tested ASN with each individual core and found out that cores 0,2 and 3 are fine, Core 1 creates HEAVY stuttering immediately.

 

4. Will be continuing to test single core theory for ASN for a next couple days, but hope that my problems is finally solved (keep my fingers crossed). 

 

5. I still want someone to confirm that moving ASN exclusively to core 0 with HT = Off in BIOS and my modest 4.6 OC will not hurt my hardware.

 

...And big kudos for HiFi support - Kostas, David and Lisa

[SteveW 1,792]

My understanding was that messing with cores is justifiable with HT = ON only.

 

Steve?

HT On or Off we can still mess with cores However, with HT On we're messing with cores that each emulate two cores. It's just that when HT is on we might want to ensure an app does not compete with itself on the same core.

 

I have HT=Off, AF=14 and OC-ed to 4.6.

2. Checked via PE and found out that P3D uses cores 1-2-3, but not core 0. Have no idea why, but it is what it is.

With HT Off and an AM=14=(1110) that would be correct. Reason I put (1110), this is the binary representation of your HT Off cores, core zero on the right (3210).

[warbirds 555]

I put a ticket in with HiFi about this and asked if there was a way to get the previous version. I did have the previous version but lost it in a pc crash and not really wanting the new version with this problem.