Thanks for your send.http://forum.notebookreview.com/showthread.php?t=213664 :





_______________________________________
Blue Ray DVD Copy

--------------------
___________________________
DVD to Zune Converter

Hi ,

Im using :

An ancient by very reliable Dell Inspiron 9100 3.2Ghz P4 2G RAM 333 (800Mhz FSB)
M-Audio Fast Track Ultra USB 2.0

2. My DPC values (after lots of analysis and tweeks, including battery removal) are always 10us -> 30us idle
and 70us -> 80us (max) when recording 8 tracks of audio at 24/48 in Ableton 7.0.2. This is very consistent and low.

Despite the continous low DPC latency I get intermittent and apparently random audio dropouts (between 2 and 10 minutes and sometimes in pairs) only when recording in Ableton 7.0.2. at 768 sample latency. Also I dont appear to have CPU spikes in conjunction with these dropouts. I cannot generate dropouts by running the test signal at full
with the lowest sample buffer latency or playback of multiple tracks of audio at low buffer settings.

Can anyone suggest another tool that can record the OS processes data so I can correlate it against the drop outs.

Any help appreciated...

-Iain

Here's my results with my newly custom built PC:


1. PC hardware: CPU: Intel i7920
Mobo: Gigabyte EX58-UD5
OS: Windows XP32 3GB Switch

2. DPC Latency typical and peak values: 5-6

3. RME Multiface II audio interface. Enabled Lan. No antivirus. Powercore Firewire Hyper-Threading:ON

[image][/image]

--------------------
Cubase 5.5 RME UFX, Asus P6T, 12Gb Corsair Ram, Intel i7930, Powercore Firewire, Asus 9400GT Silent

Here is my new build:

Gigabyte GA-EP45-UD3LR
Quad 9400
4gig Corsair TWIN2X4096-8500C5C (5, 6, 6, 18; 2.1v)
Gigabyte NVidia G Force 8400 GS
Corsiar 520W Powersupply
RME Fireface 800
Adaptec Fireconnect PCI FW card (Texas instruments chip)
XP pro 32bit


DPC latency results idling are 3us constant!!!!

The mainbordas onboard audio and LAN are disabled and the Bios CPU options C1E, TM2 and EIST are disabled too.

Sid

Quote Remeniz Productions:

---

Previous...

Motherboard: ASUS A8N-SLI Premium
CPU: AMD Opteron 165. 1800MHz factory. Overclocked to 2558MHz
RAM: Crucial Ballistix Tracers DDR500 PC4000
Audio Interface: Creative Sound Blaster Audigy 2 Platinum EX
OS: Windows XP Pro 32BIT Service Pack 2

Absolute Maximum 160 microseconds
Typical 7 microseconds

---

Now...

Motherboard: Gigabyte GA-EX58-UD5
CPU: Intel Core i7 920 (D0) @ 2.66Mhz
RAM: 3GB (3x1GB) Corsair Dominator DDR3, PC3-12800 (1600)
Audio Interface: M-Audio Delta 1010
OS: Windows XP Pro 32BIT Service Pack 2

Absolute Maximum 6 microseconds
Typical 5 microseconds

Hi all, just posting in the hope that it helps someone somewhere! I was having nightmares with pop and crackle on my Acer 5630 and after running DPC latency checker discovered peaks all over the place. I disabled all the usual suspects, nothing doing. I resigned myself to buying a new laptop. Finally I read on this very forum of a guy who, in a last gasp attempt to cure the same problem, (not on the same laptop model) disasbled his DvD combo drive. It worked for him, so of course I immediately tried it. EUREKA!! No peaks whatsoever latency stays constant between 30 - 70. I can simply re-enable the drive for general use when not using the DAW (in my case Pro Tools M-Powered 7.4). Best of all, no more crackle or pop. Up to now I had tried everything (reinstalls, older bios, different USB cables etc, different drivers) but never had anyone recommended disabling the combo drive.
Of course it may not work for everyone - guess it all depends on chipsets/hardware etc - but hope it helps at least one of you!

Acer 5630, 2.5 gb ram, core 2 duo T5500, M-Audio FastTrack Pro USB, via firewire PCMCIA card with Seagate firewire external HDD for audio.

Welcome to the SOS Forums Bob!

And we always like folk who register and then post useful advice straight away

Some of these DVD combo drives can be a right pain for musicians can't they?


Martin

--------------------
YewTreeMagic

Hi Martin,

great work with your survey topics and the effort you put in.

I am about to build a new i7 system. The 920 seems a no brainer for the price point, but I am dead scared about Motherboard and DPC.

I know you have another survey on working i7 systems, but that doesn't seem to go near DPC.

I went Quad Core last year and got burned with the X48-DS4 board from Gigabyte and Vista 32. I always get DPC spikes around 1000us (every 5-10s) and spent a long long time trying to get to the bottom of it. Never did.

For various reasons I need a new setup and am committed to going for:

Windows 7 64bit
i7 920

I think the choice of MB is critical. You would think it would be easy to just find people who have working systems with that CPU and OS and with low DPC figures, but I just cant find any, Ive been googling for weeks. There are Tons of i7 systems with low DPC running XP though.

The closest I have come to an answer is this link
http://forums.tweaktown.com/f69/dpc-latency-problems-x58-ud5-30547/

I got a message from the author who got the amazingly low DPC (second screen shot, not the first) and he was running Vista 64. (not quite W7 64, but hey maybe close enough).

Do you know of any website or have come across anywhere that recommends an X58 MB with i7 and Windows 7 64bit?

If you or anyone could steer me in the right direction that would be amazing.

Thanks.
Paul.

Welcome to the SOS Forums Paul!

And thanks for the kind words about my efforts

In a way I'm not surprised that you haven't found reports of i920 systems running audio software under Windows 7 64-bit, since the official release date of this operating system was only a few weeks ago, and musicians are still trying out the possibilities and discovering whether or not their Windows XP audio interface drivers run well under Windows 7 in the absence of any official feedback from the manufacturer

Some specialist system builders are also holding back from Windows 7, or even if they are offering it as an option I'm been told by at least one builder who posts here that the majority of their customers still ask for Windows XP to be safe.

I also get the impression that some COre i7-related components are beginning to be in short supply for 'political' reasons, although I'll let the system builders fill you in on that one from their personal experience


Martin

--------------------
YewTreeMagic

Just a quick report on my brief experience with Windows 7 64 bit and DPC.

I assembled the following system this weekend:

i7 920
Asus P6T SE
3 X 2 GB OCZ gold 1333Mhz DDR3 RAM
2 sweex firewire PCI cards (VIA chipset)
Some kind of Gforce PCI-e graphics card (I think a 7300 or something like that)

I am currently using Windows 7 64 bit enterprise 90 day trial I downloaded from Microsoft, while I am waiting on my copy of windows 7 home premium 64 bit.

All these worked fine with minimum fuss:

RME FF800
Powercore 6000
Liquid mix
BCF 2000


DPC latency is on average about 10 times higher than XP sp3 32 bit, with all hardware enabled. It usually stays between 80 and 100 us, while XP was between 5-10us. I did get occasional very high spikes while using the DVD drive, but it doesn't happen if the DVD drive is idle.

I will try to disable unnecessary hardware to see what effect it has.

Hi Nuno,

Thanks for posting your results. I do recall someone else with a similar pattern to you (ie DPC much higher in Win7 compared to other OS)...

But the same person also reported that the Win7 machine could actually do more advanced audio work (more instances of plug-ins) than XP on the same machine.

This particular point is very interesting and should probably show us that aiming for 'the lowest' DPC might be irrelevant. Maybe we dont even know if that DPC tool is accurate across all the different OS in terms of its absolute readings. But I think it can report the spikes OK and they are the problem.

Could you also list some other specs on your machine (amount of RAM and type for example). I would be happy to follow your spec if thats OK with you.

Thanks again for taking the time to report your findings.
Paul.

Quote paul 440:

---

Hi Nuno,

Thanks for posting your results. I do recall someone else with a similar pattern to you (ie DPC much higher in Win7 compared to other OS)...

But the same person also reported that the Win7 machine could actually do more advanced audio work (more instances of plug-ins) than XP on the same machine.

This particular point is very interesting and should probably show us that aiming for 'the lowest' DPC might be irrelevant. Maybe we dont even know if that DPC tool is accurate across all the different OS in terms of its absolute readings. But I think it can report the spikes OK and they are the problem.

---

Hi Paul,

II think the DPC latency will always be strong indicator of at least how low a latency you will be able to achieve, because it is directly related to the probability that windows attends an interrupt issued by your soundcard before the audio buffer expires. That being said, I have changed way too many variables to be able to access the gain/loss in performance attributable to the operating system in this case. All the hardware changed.

A quick test yesterday seems to suggest that this machine is considerably more powerful than the Q6600 machine I had previously, but until I install everything again and manage to open one of my old projects it will be hard to quantify.

I have to wait for my Windows 7 home Premium OEM before I install all my IK and NI stuff as I am not going trough the hassle of copying/pasting serial and activation codes twice.

Quote paul 440:

---

Could you also list some other specs on your machine (amount of RAM and type for example). I would be happy to follow your spec if thats OK with you.

---

I am not sure what else I can provide that may be of use.

I use an Icute power supply which I think is 750W (completely over specified) and a Antec P182 case (which is great). The coller is a Zalman CPN9500 for which I unnecessarily bought an adapter bracket from socket 775 to 1336, because the Asus P6T actually takes coolers for both sockets.

The RAM type I wrote above. It's OCZ Gold DDR3 1333 MHz and I think the latency is CL9. I have 3 X 2 GB, but I will double that at some point.

Quote paul 440:

---

Thanks again for taking the time to report your findings.
Paul.

---

No Problem

Tomorrow I should have the definite copy of Windows here and will start installing everything. I will report back when I have more info.

Thanks Nuno,

thats already very helpful. I think MB, OS, CPU and RAM are fairly fundamental and youve provided all of those now!

Is there a chance you could just list the exact model of graphics card? (you said maybe a GEForce 7300?)

Thanks again!

Paul.

Quote paul 440:

---

Thanks Nuno,

thats already very helpful. I think MB, OS, CPU and RAM are fairly fundamental and youve provided all of those now!

Is there a chance you could just list the exact model of graphics card? (you said maybe a GEForce 7300?)

Thanks again!

Paul.

---

I checked that now. It's actually a NVidia Geforce 6200 LE. Just a regular one with passive cooling, so no noisy fans. I don't think graphics card choice is that important. I just picked the first one I found at the place I bought the parts.

CPU - Intel i7 860, HT and Turbo on, no overclocking

MotherBoard - GigaByte GA-P55M-UD2

Memory - 2x2GB DDR3 Kingston ValueRam at 1333MHz

OS - W7 64bit (One boot yet for software compatibility testing, but I want to make a DAW boot. Maybe there will be no need)

Soundcard - RME Fireface 400

VGA - Sapphire ATI HD3450 512MB (fanless, VGA, HDMI, DVI outs)


After the OS installation the DPC tool showed around 250-300us with occasional 500-600us peaks, which was shocking, even my second PC (Core2Duo E6400, 2.13Ghz and the PC at work (Core2Duo E6700, 2.53Ghz) had 30-60us with peaks at 127 and VERY rare - 300us.

Then I read the Motherboard manual and turned off the energy saving stuff in BIOS (C1, C3/C4/C5, i suppose). Huge improvement - 127us, peaks at 200-300us.
Then I switched on the "always maximum cpu power" in W7 power management, turned of all useless stuff in BIOS (COM port, LAN.)

Result: 30-40us average, 100-120 peaks.
Much better, but still not an i7 level (according to the results, posted here). It's quite sad, I don't have ideas anymore how to decrease the latency.

The good thing actually is that I've made a sick Ableton Live 8.1. -project with 15 tracks playing simultaneously (all MIDI tracks with several softsynths: Massive, Absynth, Zebra, Jupiter 8, Prophet 5, uhe ACE, AAS Tassman, AdmiralQuality Polyana, D16 Drumazon, Korg MS-20, Korg Monopoly, Korg Polysix.) ca2-3 effects on each (some quite CPU consuming: Guitar Rig 4, Revalver).

test results (according to the CPU usage meter in Live 8.1):
- 44100khz, 256 samples buffer: around 22-25%
- 44100, 512S: 17-20%
- 44100, 128S: 25-27%
- 44100, 64S: 29%
- 44100, 48S: 33%
- 96000, 128S: 51-53%
- 96000, 256S: 47-50%

no clicks or gliches in any case

actually I could never even think off running such a project on my E6400 XP Pro System (even with a super stripped down DAW boot) at 44100, 512S.

One more interesting thing: adding a new instrument didn't cause a raise of CPU usage. I had almost the same CPU usage results with 7 instances in this project.

Just finished building my new DAW and this is my spec:

CPU - Intel Core i7 920 2.66Ghz

MotherBoard - GigaByte GA-EX58-UD5

Memory - 6GB kit (2GBx3), Ballistix Tracer 240-pin DIMM (with LEDs) Tri-Channel

OS - Win XP sp3, 32 Bit – will upgrade to Win 7 later

Soundcard - M-Audio Delta 66+ omni


Currently running DPC on average of around 11µs with an occasional peak of 112µs. I get this result by disabling the Gigabyte GBB386x SCSI & RAID Controller as well as the Texas Inst. IE1394 host controller. Both of these cause peaks in my DPC of 4000µs.

I use a USB/wireless keyboard & Mouse and could reduce my latency further by changing it for a wired setup, but I do find it useful to control my DAW from some distance away when I am playing guitar etc. For me, this DPC score is good enough. However, if ever I connect a Firewire device that may change the situation!

Interestingly, as I playback from Sonar at the same time that I progressively disable items in Device manager, if I Print Screen, this action seems to settle the system down faster than if I left things to sort themselves out. Weird or what?

And I do leave the DPC checker running for 5-15 mins and reset after each Device Manager change, just to iron out any stray peaks.

Edited by Paul881 (10/01/10 06:43 AM)

I just upgraded my CPU to a quad core, and in case anyone is still building lga775 PCs, i'll post my results.

Q8200
Asus P5KC
4 Gigs DDR2 667 Mhz RAM
ASUS GS8400 video card (passive)
Seagate 7200.12 500Gig SATA system drive
2 x 7,200 SATA HDDs for recording, samples, backup

TC Konnekt 24D (on a Siig PCIe firewire card)
2 x UAD1
2 Creamware Pulsar II

Using Windows XP 32-bit, in audio mode (ie. wireless network adapter, firewall and related services disabled) i get DPC Latency Checker average of 10 uS, w/ peaks up to 30 or 40 uS. Before installing the DSP cards i was getting 5 uS on average.

The GS8400 is an old vid card; i originally went w/ an EVGA GeForce GT 240, but DPC results averaged 200+/- uS, 2/ peaks near 1,000 uS. Rather than wait around for new and better drivers, i went back to the trusty 8400.

I'm happy w/ this system, although i thought the quad core CPU would mean a smaller buffer size on the Konnekt 24D...unfortunately that's not the case. Best i can do is 384 samples = 4 ms (at 96Khz). I know that the Konnekt driver are widely unpopular...i may ditch the Konnekt, and just put an ADAT front-end on my Creamware cards (though i like the TC pre-amps). When Creamware (now called Sonic Core) finally release 64-bit drivers, i'll try XP64 right away.

The PK5C gets alot of bad press, and based on benchmarks tests i have seen recently, i'm certain i would get noticable improvements w/ a better mobo (CPU, memory and HDD performance). But i'm reluctant to put even more $$ into this system now that i7's are becoming affordable. Haven't driven this system hard yet in terms of large numbers of VSTi's and plugins, but early indications are good.

Tjere is one minor frustration; i thought i could finally do some major polyphonic stacking w/ the u-He ACE, but in fact i'm getting only slightly more polyphony (in accurate mode) than i did w/ my previous, dual core CPU. ACE is far and away the most CPU-hungry VSTi i've ever seen...i'm thankful for Cubase's 'freeze' function.

Regards
Glenn

--------------------
"The path of excess leads also to the palace of wisdom....but they let you in through the back door."

Samsung R780

DPC latency is high a lot of the time

The NVIDIA GeForce GT330M is the culprit. With the driver disabled the DPC latency drops to around 300 consistently. However, then you're stuck with Microsoft VGA drivers which are rather a strain on the eye.

Cheers,
Matt

you are my las resosrrt!

Before upgrading my windows vista to windows 7 my Toshiba P 300 laptop (4 GB ram + dual core T 5800 processor) had no problems with audio latency. And i could uise my midi controller (Reloop digital jockey with build in soundcard + Traktor Pro). WIthout any problems.
But after upgrading, and using de DPC checker every few minutes i have a value of more than 160000! De rest of the bars are green , this is not a controller problem. Also when disconnected i have this high latency every few minutes. I have disabled everything possible in the hardware configuration and used every single thing to rule out the cuase but cannot find it. SOmeone can help me?

I am using Windows 7 on a Lenovo T500. Just got a driver for my FW1804 so I am trying to move up to Win7.

Latencies are running about 500us with an occasional peak of up to 1600us. Mind you that has been while I have been hacking about but even quiescent I saw 1450.

Surprisingly audio performance is ok. I am running 3ms latency. The problems I see seem to be independent of latency. That simply means that I have other issues right now. I suspect the on-board Firewire interface is only capable of running half duplex because I get crackles when I hit my interface with lots of MIDI (glissing for example or moving a controller), which is not going to burn bandwidth but could be causing the interface to be "turned around" a lot and that might be slow. This is pure guesswork at present.

Have not disabled much yet. Just the battery ACPI control. That has not made any noticeable difference. Still have autorun on, for example. Did turn off screen candy.

The is no Hardware Profile support on Win 7. On XP update 3 it clearly was not working either. So if I disabled something in one profile it would disable it in both profiles.

There is supposed to be a way to set up powershell scripts to enable and disable stuff on Win 7. I am messing around with it but so far its not clear I have it working. It appears only a few things can be enabled and disabled in this way (like network adapters) since some other things don't have the Enable/Disable methods.

Also device Manager on Win 7 fails to indicate what is and is not enabled. Only way to find out is to right click and see if the menu contains "enable" or "disable".

More when I know more.

--- more ---

Turned things off until I got down to about 250us. Than started turning stuff back on. The following numbers show the latency I get with just one extra device on.

DVD 712. (This is a big one)
USB drive 396 (Just disconnected it)
Wifi 658 (Initial connect took longer)
Wired Network 725 (While talking to SoundonSound - insignificant
if not communicating)
ACPI Battery Control None discernible
Laptop Display 692

This one was a surprise. As the USB drive was fairly slow even though I was not accessing it I hypothesized that as I am using a USB mouse and keyboard undocking the laptop might improve stuff, but actually it got a lot worse. Have not actually isolated this to the display - its possible other things are turning on. It wasn't being on battery versus wall power - eliminated that.

BTW Win 7 does mark disabled devices. Its just not as easy to see as on XP.

Hello!

This possibly sounds like a DPC/ISR interrupt issue brought about by your new O/S and a possible outdated driver. Together, these related Windows 'events' essentially introduce unacceptable and all-but-uncontrollable system latency (as a consequence of extended FLIH/SLIH calls from h/w to a low-level s/w kernel).

[SLIHs - aka soft or slow interrupt handlers ... or Deferred Procedure Call (DPCs);
First-Level Interrupt Handler - aka hard or fast interrupt handlers].

Because ISRs have to execute as soon as possible, drivers typically postpone the completion of servicing an interrupt until after the Interrupt Service Routine returns control to an application's thread. Windows systems provide support for deferred procedure calls (DPCs) - which are queued by ISRs to execute at a later time and at a lower IRQL (IRQ Level) than the requesting ISR - however, please note that some BIOS code implementations, as opposed to driver code, per se, can have a massive impact on these interrupts too.

See here for more info:

Microsoft Developer Network Blog on DPCs and ISRs

DPC issues on previously 'stable' systems usually come about because of driver 'upgrades', new O/S implementations and BIOS alterations (or upgrades, where you have little or no control of such low-level calls). Dodgily over-engineered video drivers, et al, where the coders try to make their h/w implementation 'the best' of its kind by hogging / overusing interrupts and internal buses, have been a BIG cause of problems too.

The best (if most complicated) approach is to try to identify the faulty process/driver:

ISR and DPC Monitor

and then install a 'replacement' driver or software (DLLs, EXEs, etc.) to remedy any negative effects on audio performance.

Or you can simply adopt a brute force 'trial and error' approach by starting to update ALL your driver software (and/or apps) one at a time in order to make sure you have the latest versions for your newly-installed O/S. This goes some way to providing for overall system-wide 'compatibility' (and ensuring you have the 'right' drivers for that particular O/S), whilst allowing you to note what effect each change has on overall system latency / audio performance. Unfortunately, there are no guarantees that each constituent element will work together as a cohesive whole!

(Be aware, however, that the latest driver s/w does not always mean 'best' or 'optimal' in terms of compatibility with music systems and/or latency matters! There may well be some trial and error necessary ... and you may never get there, wherever "there" is! And "there" may not be as a consequence of a single change: low-level performance is a combination of extremely complicated interactions).



The aforementioned DPCs/ISRs can also be isolated and authorship confirmed by Microsoft's RATTV3 and System Performance Monitor utilities (XPerf - Trace Capture, Processing, and Command-Line Analysis Tool / XPerfView - Visual Trace Analysis Tool) which will likely indicate that your problem is related to one of the following:

• i) the manner in which the BIOS and ACPI interface; or
  
• ii) how your video card stifles processing to improve its own performance (or polls to check if another monitor is attached); or
  
• iii) a network-related driver constantly polling resources; or
  
• iv) A. N. Other badly written device driver; or
  
• v) a combination thereof (the worst kind to diagnose).
  

Processor Explorer is detailed here:

Microsoft SysInternals Process Explorer

As an example, given Windows 7 uses ACPI services to call into the BIOS (via acpi.sys) - the kernel queries temperature sensors, battery state, special keys and such features in this way - it all-too-often appears that the O/S kernel employs 'questionable' ACPI routines (as far as audio performance is concerned, e.g. employing a busy loop to poll hardware state) to API into a similarly dodgy BIOS implementation and this kinda thing tends to be the principle cause of near constant CPU-hogging ISRs.

For instance, I found that when a laptop's fan starts up, the volume of DPC interrupts (on a system I owned) increased massively and hence DPC latency rises, from a paltry background 'noise' of ~40-50us latency, to a high quantity of 4,000-5,000us (4-5ms) latency calls. This usually resulted in near-constant audio noise - crackles and complete audio drop-outs - during that phase, quite independent of any other system-, software-, hardware- or user-configurable parameter(s). (I could, of course, reduce the CPU speed so it didn't get too hot, removing the requirement for a fan ... but I like to use the plug-ins I've bought!!!)

Even stopping unnecessary Windows O/S services, disabling all possible hardware, ensuring no conflicting or shared IRQ's, having firmware and drivers installed at their best possible version, hard-drive DMA and interface method (IDE) being optimal, DEP off, all unnecessary background processes stopped, system start-up routines culled, etc., etc., - and even running a graphics adapter in VGA-compatible mode (to reduce the impact of video demands on the CPU and Chipset) - often does scant little to help with such low-level issues. All-in-all, even machines optimised to within an inch of their lives can have little (or no) control on optimising these low-level calls if it's a kernel-BIOS issue.

When MS's Driver Development Kit, "RATTV3" and "SPM" tools, coupled to SysInternals' "Process Explorer", points to the ACPI-BIOS implementation (or similar), be aware that this is something neither the PC manufacturer, audio-driver nor DAW s/w developers have any control over whatsoever. If this is the case, you'll probably have little choice but to go back to the 'working' O/S unless there is a 'better' - and this does not necessarily mean later - BIOS for your system ... and, of course, you are prepared to take the risk implementing it!

Hope some of this helps / makes sense ... and that it's all as correct as possible: I'm really no expert - I just read a lot of geeky stuff and pretend I understand it!!!

So, good luck!

--------------------
Steve
E: SteveECrane at GMail dotty com

Original post detailing issues HERE

Hi, I'm pleased to report that I've finally managed to sort out my issue with my new Gigabyte GA-G31M-ES2L with Core 2 Duo E6750 and DPC spikes of >8000 microseconds.

One of the culprits was, as stated in the OP, the NIC. However, disabling this failed to completely solve the problem, and I would still get the dropout-inducing spikes at random intervals.

After much tinkering and monitoring it became apparent that the CPU was still adjusting its frequency despite the W7 power management being set to High Performance mode which should disable it. I made a quick visit to the BIOS and after disabling the C1E and EIST functions the problem went away!

Could someone in the know please confirm what the real effects of disabling these settings are? Are the consequences just increased power consumption?

I would really prefer to have the options disabled at an OS level so that my main internet partition could benefit from EIST. I used to use CPU Rightmark to set the CPU to full power, but that doesn't appear to have been updated for a while and doesn't seem to work properly with my system. Are there any other methods of forcing the OS to do this?

Cheers,
Bod.

Hello!

My results:



My system:

Intel i7 920
6GB DDR3
Gigabyte EX58-UD4P
Asus NVIDIA Silent graphics card
Windows 7 Pro 64 bits
Echo Audiofire 4

--------------------
Intel i7 920 6GB DDR3, Gigabyte EX58-UD4P, Windows 7 Pro 64 bits, Cubase 5.1.1 64 bits, Echo Audiofire 4, DAV BG1 U, Rode NT1A, Rode NT5 MP, Yamaha HS50M.

Nice!

That’s what we like to see Metheny!


Martin

--------------------
YewTreeMagic

Thanks!
I saw Nuno´s results (similar system as me), and I see that Win XP achieves less latency, ten times lower as Win 7...
I also tested my system with a RME Multiface 2 and I had the same result through the PCI.

I have my LAN and antivirus (Norton 360) enabled, C1E and EIST disabled, HT enabled. The onboard audio card is disabled.

Regards!

--------------------
Intel i7 920 6GB DDR3, Gigabyte EX58-UD4P, Windows 7 Pro 64 bits, Cubase 5.1.1 64 bits, Echo Audiofire 4, DAV BG1 U, Rode NT1A, Rode NT5 MP, Yamaha HS50M.

Hi,
I am pleased to say that I have just solved my DPC latency issues which were making my PC in its current form almost unusable.

I'm running a Giga-Byte EP35C-DS3R motherboard, Q6600 CPU, 4GB of Crucial RAM, Giga-Byte 9600GT silent graphics, a Line6 UX2 interface and Windows 7x64 and I was getting an average of ~400ms with spikes of over 8000ms which meant that I couldn't even listen to the SOS podcast without it glitching never mind anything else! However, after following a random thread on the Cakewalk forum I tried going into the BIOS and disabling the following:

EIST (Speedstep)
C1E
Virtualisation

and now I have a workstation that is actually usable with an average of ~140ms and absolute maximum of 346ms. I can't tell you the relief! I was about to throw in the towel and buy a Mac I was that fed up!

Anyway, I hope that this may help someone else who's at their wits end like I was.

CJ

--------------------
It's a long way...

Hi SeeJay - glad you're finally sorted!

I thought disabling EIST and C1E in the BIOS were common knowledge though - whoops!

I certainly mentioned the disabling of ‘C1E CPU Enhanced Halt state’ in the Advanced CPU Features section of the BIOS way back in PC Notes December 2009 to cure strange squealing noises from Gigabyte UD5 motherboards

Perhaps I’d better put a timely reminder into my next PC notes column in case anyone else is still suffering


Martin

--------------------
YewTreeMagic

It would definitely be worth doing again Martin. It caught me out recently too after upgrading from a single core AMD cpu to an intel cpu.

OK - a reminder is going in the next column then


Martin

--------------------
YewTreeMagic

Hello,

I have an i7 920 which works really well as my main DAW. However, I recently upgraded and old Acer laptop, which is a gigging machine, with a Packard Bell Core 2 duo laptop which seems ok running Tracktor but pretty poor for apps like Ableton. See image below.

Thing is, the Packard Bell came with Vista Home Premium and I'm wondering if that could be something to do with its poor performance for streaming audio. I wonder this because when I got my Acer P 1.7 several years ago it came with Windows 2000. The first thing I did was install XP on it. To my horror, all audio, and I mean all, including ordinary streaming from media players like Winamp, was glitchy and unusable, both with its on board audio and my usb sound device.

But then, when I reinstalled Win 2000 onto a new partition (Bearing in mind the Acer only had a 40 gig HD) all audio worked fine on Windows 2000, smooth and glitch free. Same hardware, different OS. One worked smooth the other was absolutely unusable.

Now, I don't pretend to understand why this is but I'm prepared to try different OS's on my Packard Bell, starting with XP, but before I do I'm wondering if anyone here knows something I don't and could possibly throw some light my problem. Just a thought.

Cheers





By the way I've tried disabling all on board devices one at a time, to no avail.

Hello,

Regarding my post, directly above this one, and in case anyone is interested, I decided to take the plunge and replace my Vista OS by formatting and installing Windows XP instead. I had some trouble getting XP drivers for this machine as they were not available from Packard Bell's support page but found them on another site.

Anyway, the results. Although not perfect much, much better. I now get a steady current latency of about 200us or less with the very occasional peak into the yellow. I can live with these very usable results.

So if anyone else has been having similar problems, especially with Vista, it's well worth trying XP. Worked for me.

This just confirms what I already believed. Vista stinks

Just to add:

The machine is a Packard Bell EasyNote ARC 21 with an Intel Mobile 2 Core T5500 processor with 1 Gig of ram which it shares with the ATI Graphic. I'll soon be increasing the RAM to 1.5 gig in the hope of improving things further.

Cheers.

Edited by dbm1949 (25/08/10 02:38 PM)

I have
Asus M2N68-AM PLUS
AMD Athlon +4400
2 gb RAM
M-Audio 24/96 Sound card

Idle was 3
Cubase SX3 and Reason 4 24
Cubase 5 and reason 4 45

Hello again,

If anyone is interested, after replacing Vista with XP and performing further tweaks, including disabling wLan and turning off the "ACPI compliant battery method" via Device-Manager I now get the following latency results:



Result, no?

My conclusion is that all these problems are driver related.

Cheers

Edited by dbm1949 (26/08/10 10:48 PM)

Ok. I've been trying out Windows 7 Ultimate 64 bit on my internet drive and love it.



Here's my DPC results. I basically started the checker loaded up my browser, moused around a bit, cut and pasted the image then uploaded it to Photobucket. And I have my AV software running too. And the results above are what I got.

I have to now build up the courage to install Windows 7 on my studio drive...

i7 920 @ 3.2Ghz
GA-EX58-UD5
3GB Corsair XMS RAM
Windows 7 Ultimate 64

Hi,
I recently purchased a new Dell Latitude E6510 laptop, pre-installed with Windows 7 professional 64-bit and as mentioned here and on other online forums, the out-of-the-box DPC latencies were pretty terrible. While trying to find the cause of my DPC spikes, I came across Windows Performance Tools (WPT) Kit which is part of the Windows 7 SDK, that can pinpoint the driver that is causing DPC spikes over a specified time period. (I haven't seem this mentioned before in this thread, or in SOS - but apologies if I've missed it and and just repeating old news).

I found the detailed instructions on how to install and run here:

http://www.msfn.org/board/topic/140263-how-to-get-the-cause-of-high-cpu-us age-by-dpc-interrupt/

After following all the usual steps, disabling speed step, all configurable power management, DVD player, uninstalling Intel rapid storage etc., I managed to get my DPC spikes down to just under 2000us.




At this point I started using the Windows Performance Tools as detailed in the link above. After creating a trace file, the Windows Performance Analyzer provides graphs for CPU Usage, Disk I/O, Disk Utilization and also DPC CPU Usage and Interrupt CPU Usage for the duration of the trace:



A detailed summary can the be generated that shows a worst offenders list for DPC CPU Usage:



In my case, the main culprit was the NVIDIA nvlddmkm.sys driver. I updated the driver and made sure it's power management features were turned off, but it didn't make any difference. After a bit of googling, I found the 'NVIDIA PowerMizer Switch' on the notebookreview forum. After disabling the PowerMizer(brilliant name) functionality using this application I managed to get my DPC spikes down to a fairly reasonable 500us:



I've still a bit of work to try to get it lower, but at least it's a start.

Again, hope this is useful and not just a rehash of old information.

cheers,

david

--------------------
Dell Latitude E6510, Intel i7 M640@2.80GHz, 8GB DDR3, Windows 7 Pro 64 bits, Cubase 5.5 64 bits, Steinberg MR816 CSX

Hi david, and welcome to the SOS Forums!

It’s always good to hear of a new tool to track down DPC spikes - have a read of this thread as well about LatencyMon, which runs on Windows Vista and 7:

www.soundonsound.com/forum/showflat.php?Cat=&Board=PCMus&Number=860880

Hope it helps!


Martin

--------------------
YewTreeMagic

Thanks Martin!,
I had missed that thread. Using the LatencyMon application I managed to get rid of the remaining occasional DPC and hard page fault spikes. However, I'm still getting audio dropouts in Cubase when using my Steinberg MR816 CSX interface (which worked perfectly on my old Lenovo 3000 N100 laptop running XP). After scouring the web for the last few days, it sounds like the on-board Ricoh 1394 chipset may be causing the issue. I've purchased a FireWire express card with the TI Chipset, although from what I've read, I'm not sure if this will even help. Hope to have a chance to test it out tonight.

thanks again,

David

--------------------
Dell Latitude E6510, Intel i7 M640@2.80GHz, 8GB DDR3, Windows 7 Pro 64 bits, Cubase 5.5 64 bits, Steinberg MR816 CSX

Anyone noticed how different synths cause latency issues? I just installed Wizoo's Darbuka and Latigo that I had bought in 2005 and hadn't used or even installed on my new DAW. But the pops and clicks were terrible.

So in testing the Latency with it playing and without the difference was significant. So do I presume that this synth is now out of bounds for me to use?

As you were - and an important lesson (well, for me at least )! So when I first built my DAW built around a GA-EX58 UD5 Gigabyte MoBo, I checked my DPC and found it was rock solid around 9uSecs.

Then today, I got a load of clicks and pops so assumed it was an old synth. However, I decided to update my Graphics card driver and that did the trick!

So the lessons are - regularly check your graphics card drivers as well as your DPC! Simples!!!