What Could Cause PCI Overflow

A place to talk about whatever Scope music/gear related stuff you want.

Moderators: valis, garyb

Post Reply
dawman
Posts: 13275
Joined: Sun Jul 24, 2005 4:00 pm
Location: PROJECT WINDOW

What Could Cause PCI Overflow

Post by dawman » Wed Aug 22, 2018 8:10 pm

Is it possible that too many voices being streamed on a PCI-e 1X can cause this.
If so would using a PCI-e 16X/ PCI 3.0 slot maybe not get these messages?
I understand a 1X in a 16X defaults to 1X speeds, but perhaps being so close to the CPU could help?

I’m ocassionally getting PCI Error messages before the end of the gig.
Not used to that, but I am blending three electric Pianos together for the sheer power and tone (awesome) but I’m just trying to figure out what causes this so I can avoid it.

Any ideas’s are 🙏

DragonSF
Posts: 389
Joined: Sun Nov 08, 2015 4:28 pm

Re: What Could Cause PCI Overflow

Post by DragonSF » Wed Aug 22, 2018 9:39 pm

You should tell us, what are you streaming over PCI?

User avatar
garyb
Posts: 21707
Joined: Sun Apr 15, 2001 4:00 pm
Location: ghetto by the sea

Re: What Could Cause PCI Overflow

Post by garyb » Wed Aug 22, 2018 10:56 pm

i think it's a matter of actual timing of data transfer, rather than a total amount of data over a longer period of time. controller efficiency, cpu cycles, a lot of things might cause it. check for an updated bios. what else has changed? if it's at the end of gigs, heat? things are overclocked?

User avatar
valis
Posts: 4704
Joined: Sun Sep 23, 2001 4:00 pm
Location: West Coast USA
Contact:

Re: What Could Cause PCI Overflow

Post by valis » Wed Aug 22, 2018 11:51 pm

Is the PCI-e slot one of the two GPU slots? PLX chips abound...and this might indeed cause issues as this is a switch for ganging PCI-e devices on a single data path.

List your motherboard model, revision & which devices are in which PCI-e slots in complete please, so we can eliminate that as a possible issue.

dawman
Posts: 13275
Joined: Sun Jul 24, 2005 4:00 pm
Location: PROJECT WINDOW

Re: What Could Cause PCI Overflow

Post by dawman » Thu Aug 23, 2018 5:36 am

ASRock H97m WS
Ribbon Extender coming off of PCI-e 4X as there’s no 1X, totally forgot about that.

No overclocking, fans are full tilt.
But I will go to the hardware monitor next time and check.

Weird because you can still play.
When you close down Scope to re boot it, usually takes a minute, it just vanishes.

I’m going to just play this sucker for 8 hours today, got lots to do.
Then maybe get the message.
My luck it will be like yesterday and work great.

dawman
Posts: 13275
Joined: Sun Jul 24, 2005 4:00 pm
Location: PROJECT WINDOW

Re: What Could Cause PCI Overflow

Post by dawman » Thu Aug 23, 2018 11:51 am

59-60C on all Cores, no troubles, but that’s with triple 22k barrel fans on silent mode.
Went into the BIOS and put them on full tilt. CPU was on silent, I wasn’t aware of that.
Now it’s loud as a vacuum cleaner but 49-51C.

It’s been on for 3 hours I’ll keep it on all day since I’m learning songs for another group.
I’m in 3 bands now.
One is Disco Night at a house gig at a Casino, then a jazz group with a Monster guitar player.
The dance band is fun too.
But everybody’s loading me up with material.
I like challenges.

Pressure turns Coal into Diamonds, guess we’ll see if I shine or not.

User avatar
valis
Posts: 4704
Joined: Sun Sep 23, 2001 4:00 pm
Location: West Coast USA
Contact:

Re: What Could Cause PCI Overflow

Post by valis » Thu Aug 23, 2018 1:28 pm

Coal into diamonds indeed =]

User avatar
Bud Weiser
Posts: 1443
Joined: Tue Sep 14, 2010 5:29 am
Location: nowhere land

Re: What Could Cause PCI Overflow

Post by Bud Weiser » Thu Aug 23, 2018 5:52 pm

garyb wrote:
Wed Aug 22, 2018 10:56 pm
i think it's a matter of actual timing of data transfer, rather than a total amount of data over a longer period of time. controller efficiency, cpu cycles, a lot of things might cause it. check for an updated bios. what else has changed? if it's at the end of gigs, heat? things are overclocked?
Do different PCIe standards (PCIe 1.0 / 1.5 / 2.0 / 3.0) make any difference w/ XITE-1 ?
When buying a motherboard we know what we get,- but what´s up w/ XITE-1 PCIe card itself ?

We have different card revisionss in the ballpark already,- are they different in standard and data thruput ?

(tried to submit pics, but it seems to be impossible uploading 2 pics (jpg, size about 100kb each) in only one post.


Bud

User avatar
garyb
Posts: 21707
Joined: Sun Apr 15, 2001 4:00 pm
Location: ghetto by the sea

Re: What Could Cause PCI Overflow

Post by garyb » Thu Aug 23, 2018 6:31 pm

no. the first standard is sufficient. the problems are simply motherboard implementations of the standard. realtime audio is not a priority.

User avatar
garyb
Posts: 21707
Joined: Sun Apr 15, 2001 4:00 pm
Location: ghetto by the sea

Re: What Could Cause PCI Overflow

Post by garyb » Thu Aug 23, 2018 6:34 pm

this is true of everything in the computer world. the current tech never gets fully developed as newer and newer is always the solution. because of this, system theoretical limits are never met. buyers are suckers.

User avatar
valis
Posts: 4704
Joined: Sun Sep 23, 2001 4:00 pm
Location: West Coast USA
Contact:

Re: What Could Cause PCI Overflow

Post by valis » Thu Aug 23, 2018 8:33 pm

60C isn't all that hot, 140F. It could be that you had a hot spot on one of the ring links that connected to a PCIe lane set though. It also could be that there was heat building up affecting ram or one of the onboard motherboard chips, and now all that extra airflow has helped.

You will never see peak theoretical bandwidth used in a real world usage scenario, as you can only hit those peaks when running artificial workloads that stress that specific component (ram, a cpu core instruction set, PCIe bandwidth etc) without utilizing other parts of the system. When you get into real-world workloads, the way that the data moves around the system bus and occupies different portions of the CPU, memory, storage systems and such is almost never going to match that artificial scenario that's designed to stress a single component, whether it's PCIe bandwidth or a cpu pipeline. Still, gains do bring real-world improvements, at least until we move ahead with features that absorb all that extra power and then begin to push things further driving us up to our next upgrade cycle... what fun!

Way back in the day we had tech sites that went to such depths to discuss things, Arstechnica had excellent whitepapers and Anand actually had wonderfully in-depth coverage with their Computer Shopper articles and then the first decade and a half of his website. Alas, you're unlikely to find such in-depth coverage about modern Train engines (the kind that run on tracks) as that was hitting its peak almost a century and a half ago, and hardly anyone sits around dreaming about railroad adventures anymore (at least compared to its heyday). The same with modern computing tech, as long as your graphics card features matches up with the latest hot game, the gamers are happy. And the rest of us are currently just along for the ride... that includes the Xite PCIe cards.

dawman
Posts: 13275
Joined: Sun Jul 24, 2005 4:00 pm
Location: PROJECT WINDOW

Re: What Could Cause PCI Overflow

Post by dawman » Sat Aug 25, 2018 1:23 am

I was hitting 74C on 4 x Cores using a 3 layer EPiano Sound.
Cantabury Rhodes w/ Phaser, PianoTeq Rhodes and Omnisphere Crystal EPno.
With fans at maximum it drops to 54C on all Cores.

It’s heat related.
I had this PC on for 13 hours playing the big CPU eating 3 way layer.
I thought the fans were on high but not until I fixed the BIOS settings.
They’re so damn loud I thought silent was the full max.
Now this things holds a piece of paper in place they’re sucking so much air.

Pretty sure this was heat.
Haven’t seen the message for days now.

User avatar
valis
Posts: 4704
Joined: Sun Sep 23, 2001 4:00 pm
Location: West Coast USA
Contact:

Re: What Could Cause PCI Overflow

Post by valis » Sat Aug 25, 2018 8:15 am

Probably right then, 74C is almost as hot as some of the RAM sticks get in my Xeon rig under a 3D rendering workload, and when they hit 78C they die shortly after. This is because of the buffer chop that enables more than 2 sticks per channel, and my response was a fan strapped over them running at speeds that are annoyingly loud under load but I’ve not had any fail since. I hope the same for you

User avatar
valis
Posts: 4704
Joined: Sun Sep 23, 2001 4:00 pm
Location: West Coast USA
Contact:

Re: What Could Cause PCI Overflow

Post by valis » Sat Aug 25, 2018 8:15 am

Probably right then, 74C is almost as hot as some of the RAM sticks get in my Xeon rig under a 3D rendering workload, and when they hit 78C they die shortly after. This is because of the buffer chip that enables more than 2 sticks per channel, and my response was a fan strapped over them running at speeds that are annoyingly loud under load but I’ve not had any fail since. I hope the same for you

dawman
Posts: 13275
Joined: Sun Jul 24, 2005 4:00 pm
Location: PROJECT WINDOW

Re: What Could Cause PCI Overflow

Post by dawman » Sun Aug 26, 2018 5:00 pm

Thank you sir.
So it is written, so it shall be done

I never had any issues until I gigged outside, then I put the fans on full tilt
But me thinks this rig is 3 years old and is starting to need all the cooling it can get
I’m just leaving these suckers on annoyingly loud from no on.

Since I put them on full tilt, suppose to be 22k but I’m reading 14k max, haven’t had a problem.
In a few months I’ll have a Dexibel Module with all the sampled sounds I need.
Think the Hammond Module, racked synths and Dexibell will sound fine.

XITE-1 and PCs can retire to studio work.
Must say though Scope served me well live for 14 years. 9 of those years were XITE-1.

User avatar
valis
Posts: 4704
Joined: Sun Sep 23, 2001 4:00 pm
Location: West Coast USA
Contact:

Re: What Could Cause PCI Overflow

Post by valis » Sun Aug 26, 2018 6:09 pm

And so it is!

If you ever sell that Xite... =]

dawman
Posts: 13275
Joined: Sun Jul 24, 2005 4:00 pm
Location: PROJECT WINDOW

Re: What Could Cause PCI Overflow

Post by dawman » Wed Sep 05, 2018 3:31 am

Well the XITE-1 and PC will still be used.
The new SX7 Module isn’t ready for prime time.
But new pedalboard (48 bucks) is great, as are the Strymon Pedals.
If SpaceF convinces me his Integra 7 has great EPiano and acoustic instruments, the SX7 will be strings and Horn Sections, and Ill try again after New Years.

Brought back the TC Electronics Fireworx and eliminated Scopes delays with Strymon Timeline (hardware version of RoyTs BB Delay w/ extras/) and use the Fireworx for Reverbs. Fans are at 12,000 rpm.

Haven’t seen the message since.

User avatar
valis
Posts: 4704
Joined: Sun Sep 23, 2001 4:00 pm
Location: West Coast USA
Contact:

Re: What Could Cause PCI Overflow

Post by valis » Wed Sep 05, 2018 4:04 am

Single-handedly flipping the script on the meaning of the term "Jimmy-Rig" 8)

User avatar
Bud Weiser
Posts: 1443
Joined: Tue Sep 14, 2010 5:29 am
Location: nowhere land

Re: What Could Cause PCI Overflow

Post by Bud Weiser » Wed Sep 05, 2018 4:56 am

dawman wrote:
Wed Sep 05, 2018 3:31 am
...
The new SX7 Module isn’t ready for prime time.
...
If SpaceF convinces me his Integra 7 has great EPiano and acoustic instruments,

Integra 7:

EPs ...
https://www.youtube.com/watch?v=vXRzPp5Zlt0

Brass ...
https://www.youtube.com/watch?v=Z65Z-Opdaoo

Strings ...
https://www.youtube.com/watch?v=6mwAM9bAZkk

Synthpad ...
https://www.youtube.com/watch?v=Wj4-SjrMUjU

Polysynth ...
https://www.youtube.com/watch?v=QXGllvTrRfU

Choir ...
https://www.youtube.com/watch?v=L-HM_MK8rzk

Classic Guitar ...
https://www.youtube.com/watch?v=IbNqEwGJbcg

Violin ...
https://www.youtube.com/watch?v=Imy2Mxm_7TE


I stop here,- that guy has many more demo-vids for the Integra-7 and I´d say most people can do a complete gig/show w/ just only that module until it runs out of polyphony.
There´s at least 1 drawback w/ this toy,- there are 64 "studio setups" only (which is, in theory, still enough for 64 tunes though).
But,- what you have in each "studio setup" slot,- there are 16 per setup, each using a dedicated MIDI channel,- is what we know as "tones" in Roland world.
So, when single patches are composed from several "tones", that patch occupies more slots in a "studio setup" where each slot consumates voices.
Because there aren´t "performances" existing anymore and "tones" are what they were before,- a so called "Studio-Setup" is what a performance was before w/ the exception/advantage it now can handle 16 tones.
But when you use it fully MIDI multi-timbral, each slot on a different MIDI channel, you´ll hear "tones" only w/ all the limitations.
It´s a top notch module, but I always wished it had 256 voices instead of 127 only as also MUCH more "Studio-Setup" memory available.

In a studio, that all doesn´t matter much, but for live gigging it does.

Finaly,- this Dexibel and whatever hardware discussion should go to "off topic",- no ?

:)

Bud

Post Reply