New, Updated SonarTest

OK here are mine with all plugs:

Sonar 5 PE, without 64-bit:

2ms-38
4ms-26
7ms-21
10ms-19
16ms-17
20ms-16
49.9ms-14

With 64-bit on:

2ms-41
4ms-29
7ms-23
10ms-21
16ms-19
20ms-18
49.9ms-16

I am running a Gigabyte K8N Ultra-939nForce3 with an Opteron 165 Dual Core OC'd to 2250 per core. I run 1GB of G. Skill DDR550. I have a Matrox P650 Graphics Adapter. I have some more adjustments to make. This is a brand new rig. I just dumped my nForce4 board...but that would have done better, but it won't work with UAD-1, so away it goes!

This is a stock Gateway GT5032 Media Center PC -

Athlon 64 4200 X2
1 Gig RAM (upgrading to 2 or 3 gigs soon)
onboard Geforce 6100 video (upgrading to a real videocard soon)

25msec - %18
18msec - %20
12msec - %24
6msec - %29
3msec - %35 - %45
1.5msec - %62 - %79

It looks like I'm not sacrificing too much by going with an off the shelf system. At $850, I think it was a bargain.

ORIGINAL: saturdaysaint

This is a stock Gateway GT5032 Media Center PC -

Athlon 64 4200 X2
1 Gig RAM (upgrading to 2 or 3 gigs soon)
onboard Geforce 6100 video (upgrading to a real videocard soon)

25msec - %18
18msec - %20
12msec - %24
6msec - %29
3msec - %35 - %45
1.5msec - %62 - %79

It looks like I'm not sacrificing too much by going with an off the shelf system. At $850, I think it was a bargain.

No, that's quite respectable, but it IS a great CPU to begin with. Try that with a stock Dell with the disk drives hobbled in "quiet mode" LOL. Anyway, CAREFUL ABOUT "REAL" VIDEO CARDS!!! I just dumped a nice N6600 for a Matrox P650 for better compatibility. Got another machine (my wife's) with a rockin' 3D card; I don't want one on my DAW.

I'm currently setting up a new DAW running two separate SATA RAID-0 configurations using WD 10k/150GB/16MB, the Intel P940, 975x chipset mobo and 4GB of Corsair XMS2 Twin-X DDR2 800. Dual-boot system with XP-Pro and XP-64.

It'll be a couple of weeks before it's built and tuned but LMK if you're interested and I'll PM the numbers once available.


Brendan.

ORIGINAL: brendantownsend


I'm currently setting up a new DAW running two separate SATA RAID-0 configurations using WD 10k/150GB/16MB, the Intel P940, 975x chipset mobo and 4GB of Corsair XMS2 Twin-X DDR2 800. Dual-boot system with XP-Pro and XP-64.

It'll be a couple of weeks before it's built and tuned but LMK if you're interested and I'll PM the numbers once available.


Brendan.

This looks great on paper, and I wish you luck with it, but have you researched these components as far as compatibility with some of the top audio cards and DSP devices and mixers?

Isn't the 975 a PCI-E chipset? Suppose you want to add a UAD-1 down the road. Are folks having better luck with the 975 PCI-E than with other PCI-E chipsets??? Check out the rigs on the UAD forum.

Also, I just dumped my RAID 0. Unless you're a gamer, it doesn't do anything but give you even more trouble. The Raid drivers and chipsets will take up extra PCI cycles and will often conflict with sound card drivers. I have had three different top of the line MOBO with RAID, and I didnt see any benefit to audio rendering with any of them.

Now, I have 2 S-ATAs, one each for samples and production rendering, but I now record my projects to a 7500 RPM, 250GB ATA133 Samsung which is ultra-quiet and has lower access time than a RAID 0 setup. No offense, but are you about to spend a lot of money for what COULD end up, in practice, as a frustrating, sub-standard DAW?

The 940 probably has some pretty impressive floating point capability, still, if the rest of the rig is non-supportive, it could be frustrating.

Best,

Mark

Hello Tombo.

No offense, but...

Absolutely none taken, and I'm always keen to listen to those that have more experience than myself at DAW's, quirks and foybles. It's a hugely complex subject with lots of interaction and interdependencies. Certainly the bottom line is "Does it do the job well", and I've no particular pre-conceived notions as to the "best" arrangement or technology - if indeed there will ever be one. Your feedback is therefore very much appreciated. I'm not a professional and I know my limitations, though I love music composition and I just wanted a really good platform to move on.

Now without elaborating, suffice to say that I did particularly well on the stock market and this was an exploratory venture - though with serious intentions of-course. From the start I considered that if I just go with what I felt comfortable with then anything that I wished to switch out later would just get transferred to my (soon-to-be non-DAW) other workstation. I did start to look at building an AMD-based system, but I have never used AMD before and quite frankly so much AMD/Intel tribalistic talk put me right off. Refreshingly, I'm not reading this in your post and I fully accept that AMD architecture and potential looks particularly bright. I'll be watching developments very closely.

The 975 chipset? - well, a bit of a venture into the unknown with regard to future chipset possibilities so I figured that I would just give the latest a shot. Initially I don't plan installing a UAD-1, though this would certainly free up considerable CPU bandwidth and I may just do this yet. If the 975x gives me grief then the mobo goes. Certainly I would be interested to know of other experiences, both AMD and Intel, so I'll be monitoring the forums so I can make better informed decisions about future changes - or perhaps just a new DAW for the rack. For the immediate time it's more a case of get it built and get cracking on those bigger projects.

Very interesting f/b regarding RAID-0 as-well. I initially looked at RAID-1 for samples but in practice the reports suggested nothing even approaching x2 theoretical read (these dedicated to samples). It was then that I looked at RAID-0, though was not aware of PCI-hogging being a particular issue at that time. My third considered option was simply to mirror the drives and manually/arbitrarily select the drive and source when configuring the instruments in Kompakt/Kontakt - thereby distributing resource demand. I also purchased a plain ol' Seagate 400GB/8Mb/7200/ATA133 for regimented backups and recordings (in anticipation of using a RAID-0 configuration) though I'm sure this drive wouldn't touch your Samsung for performance. The 10K/16MB WD SATA-II's however offer 4.5ms seek and I probably wouldn't have to RAID them for necessary speed .

At this stage I think that I have sufficient flexibility in the setup to trial various configuration modes, though I will fully take on-board your advice and considerable experience with DAW's and RAID-0. Who could knock that?

Anyways - Better finish there as this is all a bit off-topic (sorry all).

Many thanks Tombo, all the best.

Brendan.

Oh yeah, no tribal stuff here. If Intel makes a great set, I will consider it. What I DO like is at least TWO succcessful chip companies keeping each other honest. And, I would love to see AMD get back in the Northbridge business.

Yeah, what I like about my CURRENT rig, is that I paid US$330 for a chip that, in its extremely stable, OC'd state, is outperforming chips from $600-$1,000+...now that is fun...

My big question really centered around AGP vs. PCI-E. I just haven't seen many happy campers with PCI-E, and that is why I sold my PCI-E, and went BACKWARDS to an AGP.

My reason for using UAD-1 is NOT about CPU savings. I have CPU to spare! It is about a quantum-level rendering of the Fairchild compressor, which its designer couldn't tell apart from the $30,000 "gold" instance that it was modeled on...and the same with the 1176...and the plate 140...and the LA2A...the rest I really don't care about. My engineering ability, I would say is intermediate, but I know what I like, and I like the sound of the above mentioned plugs.

Now, UAD-1 will have to get with the program at some point, but until they do, I am using AGP.

Good luck to you!

And one more thing...if Xeons were faster, I might have bought a dual Xeon setup...and I am seriously eyeing a dual CPU, Opteron 2** series Dual Core setup...think about it: four cores in two CPUs...more cowbell! More power!

Thanks Tombo.

That dual Opteron was something that I had seriously considered myself - maybe for my next DAW? Hehehe.

Your comments regarding PCI-E also reflect those of many others elsewhere, though with my limited current setup to-date I haven't experienced any particular problems that I'm aware relate to PCI-E (Intel D915GAV mobo, GeForce 6600GT PCI-E). I'm guessing it's therefore just a resource hog, or is this more a combinational thing with certain cards and bus mastering? Maybe I can disable bus matering in the driver and/or interrupt assignation in BIOS or something - if that's it?

Now, has anyone got a legacy PCI card?

Going the Intel route, I originally looked at Xeons (purely speculatively) but ruled them out due to the inescapable need for ECC RAM, and a Dual-Core would operate much more efficiently than separate single-CPU's dotted around a mobo. I went for the Pentium D940 in the end - primarily due to the 2MB cache for each core which itself is reported to yield significant performance gains for a DAW.

...think about it: four cores in two CPUs...more cowbell! More power!

Mmmm...yes, that sounds like an excellent way to go . When the price of the D955 drops considerably then I'll be right up there with you - and I do like cowbells (LOL).

<respect!>

Cheers mate.
Brendan.

bump

Athlon 64 X2 3800+ 2.0GHz
ASUS A8V Deluxe
Presonus Firepod ASIO driver version 2.14.25
Sonar 4.0.4 Producer Edition

Multiprocessor On:
25 ms - 19%
12 ms - 22%
6 ms - 28%
3 ms - 40%
1.5 ms - 51%

Mutliprocessor Off:
25 ms - 31%
12 ms - 36%
6 ms - 44%
3 ms - 61%
1.5 ms - 77%

JP

You need to get the AMD X2 driver. THose results are either backwards...or you need some updates. I have your config....with a 4600 X2....and it should certainly improve with enabling dual proc.

I have the nForce4 Asus SLI board..with a 4600 X2 in it. Also have a GeForce 6800 GT in it...it smokes for everything. As for UAD...I don't use it anyway. Took a while to figure out all the updates I needed to make her fly...but that she does.

ORIGINAL: Tombo

ORIGINAL: saturdaysaint

This is a stock Gateway GT5032 Media Center PC -

Athlon 64 4200 X2
1 Gig RAM (upgrading to 2 or 3 gigs soon)
onboard Geforce 6100 video (upgrading to a real videocard soon)

25msec - %18
18msec - %20
12msec - %24
6msec - %29
3msec - %35 - %45
1.5msec - %62 - %79

It looks like I'm not sacrificing too much by going with an off the shelf system. At $850, I think it was a bargain.

No, that's quite respectable, but it IS a great CPU to begin with. Try that with a stock Dell with the disk drives hobbled in "quiet mode" LOL. Anyway, CAREFUL ABOUT "REAL" VIDEO CARDS!!! I just dumped a nice N6600 for a Matrox P650 for better compatibility. Got another machine (my wife's) with a rockin' 3D card; I don't want one on my DAW.

I saw the results between an AMD X2 3800 and AMD X2 4400 processor and i saw a little difference in performance.Here in mexico the price of X2 3800 is about 400 dlls and X2 4400 is about 700dlls. My question is if 1Mb vs 512kb of cache have more stability??? i want to know because i like to buy a new DAW. THX

ORIGINAL: Elvenking
You need to get the AMD X2 driver. THose results are either backwards...or you need some updates. I have your config....with a 4600 X2....and it should certainly improve with enabling dual proc.

Huh? I think those numbers are great. I have the AMD X2 driver. Why in the world do you think they're backwards?

JP

OK.

Here's some numbers for my machine.

4400 x2 with 1gig high performance corsair RAM.

1.5ms - ~55%
2.9ms - ~35%
5.8ms - ~27%
11.6ms - 22% static.
23.2ms - 19% static.

all with multi-processing enabled and the RME fireface. Latest drivers and firmware.

oh, and no overclocking anywhere - and using the 64bit mixing engine.

Kind regards

Dave Rich

Hi Scott.

FWIW...

Having now completed my new DAW I revisited this thread to run a test but then noticed that previously I had overlooked the fact that this was for Sonar 3 users. Having arbitrarily tried the provided file (noting that I only started use of Cakewalk at Version 4) it would not run, though all instrument tracks appear but without content.

Although the results would not be representative, but if you wanted to generate a later compatible file (I have Sonar 5) then I would be happy to run a few tests for you.

For the record, DAW specs as follows:

CPU: Pentium D940 (2 x 3.2GHz, 2+2Mb cache)
Mobo: Asus P5WD2-E Premium (975X chipset)
Boot HDD's: 2 x 74GB 10k/8MB SATA Raptors in striped (RAID-0) array on PCI-based SiI-3112 h/w controller card
Data HDD's: 2 x 150GB 10k/16MB SATA Raptors in striped (RAID-0) array on Intel ICH7 h/w controller chip (on-board)
Memory: 2 x 2GB paired Corsair matched XMS2 PC6400 DDR2 800MHz FSB (i.e. 4GB total)
Graphics: GeForce 6600 running dual-monitors
Audio: Creative X-Fi Elite Pro ("un-professional" I know but with scant info I just had to try it)
OS: Win XP-Pro SP2 (soon to be dual-boot with XP-64)

Incidentally, in the light of other comments the two RAID-0 configurations are really just an experiment at the moment - both being hardware-driven implementations on the latest chipsets as opposed to software RAID. A few good audio tests should reveal whether anything like this might give me grief. The current graphics card is also PCI-E - guess I'm really tempting fate here , but there's nothing quite like the acid test and I can always switch at this early stage and relegate the h/w to my other systems if necessary.

All the best,

Brendan

Uhhm...I do not know what I was thinking....or where my brain was at the time....but I totally agree with you now. Apparantly....I was backwards...cause I thought I was reading worse performance with Multiprocessing enabled.....LOL. Forget I said anything.

ORIGINAL: jpkeys

ORIGINAL: Elvenking
You need to get the AMD X2 driver. THose results are either backwards...or you need some updates. I have your config....with a 4600 X2....and it should certainly improve with enabling dual proc.

Huh? I think those numbers are great. I have the AMD X2 driver. Why in the world do you think they're backwards?

JP

ORIGINAL: Elvenking
Forget I said anything.

Forgotten!

JP

Nice looking rig. Just one question/comment:

ORIGINAL: brendantownsend

For the record, DAW specs as follows:

CPU: Pentium D940 (2 x 3.2GHz, 2+2Mb cache)
Mobo: Asus P5WD2-E Premium (975X chipset)
Boot HDD's: 2 x 74GB 10k/8MB SATA Raptors in striped (RAID-0) array on PCI-based SiI-3112 h/w controller card
Data HDD's: 2 x 150GB 10k/16MB SATA Raptors in striped (RAID-0) array on Intel ICH7 h/w controller chip (on-board)
Memory: 2 x 2GB paired Corsair matched XMS2 PC6400 DDR2 800MHz FSB (i.e. 4GB total)
Graphics: GeForce 6600 running dual-monitors
Audio: Creative X-Fi Elite Pro ("un-professional" I know but with scant info I just had to try it)
OS: Win XP-Pro SP2 (soon to be dual-boot with XP-64)


Incidentally, in the light of other comments the two RAID-0 configurations are really just an experiment at the moment - both being hardware-driven implementations on the latest chipsets as opposed to software RAID.

I'm wondering if you wouldn't be better off using the 2x 74G disks seperately. One for boot and one for sampler data. Like that you keep the other RAID for audio data and have samples seperate, the OS seperate and maybe put the PictureCache on the OS disk too. Like that nothing gets in the way of anything else.

UnderTow (who is very happy he just ordered a Tyan Thunder K8W to go with the two Opteron 246s he got cheap. )

I am not sure if I am doing something incorrectly, but I am not able to get the results on the sonartest that I have been seeing posted on this forum. I am running an opteron 180 dc with a gigabyte k8nf9 ultra mobo, 4 gigs of corsair CL2 ram, and a tascam 1884. With the latency set to 2.9ms, i am seeing a fluctuation of between 50-60% on the cpu meter. Anybody think it's a function of the audio interface or have I not optimised the system enough? I'm running SP2 with the AMD X2 fix installed. Any help would be greatly appreciated.

Ben

Ooops

Here is mine

AMD Athlon 64 3200+ Venice
ASUS A8N-E NVIDIA nForce4 Ultra Motherboard
CORSAIR ValueSelect 1GB (2 x 512MB) DDR 400
HIS Hightech Radeon X800GTO IceQ II Turbo 256MB
3 Western Digital 160GB 7200rpm ATA100 Audio Drives
WindowsXP Professional SP2
Digi 002r 6.7


46.4ms........30%
23.2ms........34%
11.6ms........41%
5.8ms..........53%
2.9ms..........77%
1.5ms..........90%

Just updated my systems a bit from my initial post. Previously:

System Specs
Motherboard: Asus A7N8X deluxe (nForce2 dual channel)
RAM: 512Mb Corsir TwinX Timed at 2-2-2-5
Processor: AthlonXP 2700+ T-Bred B core
FSB: 333 MHz
CPU: 2.17 GHz
Multiplier: 13
Audio Device: WDM M-Audio Delta 0.29 driver set to 64 sample buffer

46.4ms: 33%
23.2ms: 33-36%
11.6ms: 39-43%
5.8ms: 51-58%
2.9ms: 73-76%
1.5ms: Warning shown but doesn't drop out

Same motherboard.. new chip. I got a mobile AthlonXP 2600+ barton core chip, widely regarded as a great overclocker, and a 1G pair of Corsair value select. The motherboard doesn't know what the mobile chip is, but the chip is completely unlocked so I could clock it however I want. I brought it up slowly and ended up matching a 3200+ setup exactly. New Results:

RAM: 1Gb Corsair Value Select Timed at 2.5-3-3-8
Processor: Mobile AthlonXP 2600+ Barton core
FSB: 400 MHz
CPU: 2.2 GHz
Multiplier: 11
Audio Device: ASIO M-Audio Delta, I forget which version

5.8ms = 46%
2.9ms = 65%
1.5ms = won't engage

Edit... with the exception of the 1.5ms not engaging, it's interesting that I'm beating the Athlon64 3200+ just posted... interesting.

MacBook Pro 17" (With XPSP2 under Boot Camp), 2.16 GHz Intel Core Duo, 2 GB DDR2 SDRAM, 667 MHz Bus Speed (according to "About This Mac...")

Onboard Sigmatel audio, ASIO4ALL 2.7, Sonar Producer 5.2, dither off, multiprocessing on

32 bit engine:

1.5ms dropout
2.2ms 36-76% (fluctuates a lot)
2.9ms 30-62%
4.4ms 50-55%
5.8ms 44-47%
11.6ms 37%
23.2ms 32%

64 bit engine:

1.5ms dropout
2.2ms 36-76%
2.9ms 35-68%
4.4ms 28-55%
5.8ms 49%
11.6ms 41%
23.2ms 36%

The CPU usage finally stabilizes at a buffer size of 256 and latency of 5.8ms. I think these are good numbers for a laptop, but I'm a bit disappointed; I was expecting better results. Of course, it could just be the onboard audio. I need to test with a better interface and see what happens.

specs :
17" Imac Core Duo 1.83Ghz, running Windows XP SP2 under Bootcamp
667Mhz buss
2 GB RAM
Stanton Final Scratch 2 Firewire audio interface

first column = Sonartest3 CPU load
second column = ADK Sonartest CPU load
third column = latency setting soundcard

19% 48% 25ms
20% 51% 18ms
22% 54% 12ms
24% 56% 10ms
25% 59% 8ms
27% 65% 6ms
34% 75% 4ms
40% 83% 3ms
50% 91% 2ms
66% audio engine stopped 1.5ms

to give you an idea : my dual Pentium 4 system (Asus P5 motherboard, 2GB RAM, 800Mhz buss, Terratec EWS 88 MT PCI audio interface) reached 75% with Sonartest 3 @ 1.5ms (11% more than the iMac), and the lowest latency I could get with the ADK sonartest was 7.6ms (93% CPU load), any value lower than that stopped the audio engine ...

the iMac cost me 800 EUR, and an additional 230 for the RAM ... 1030 EUR for a top notch system that's dead quiet ... I can recommend it to anyone. Especially the absolute absence of fan noise is a relief, my Pentium was stored in a separate closet with cable extensions for monitors and keyboard, no more need for that now.

in addition, it apparently even scores better than the MacBookPro with a faster processor ...

1.5ms no go
2.9ms no go
5.8ms 45%
8.7ms 39%
11.6ms 36%
17.4ms 33%
23.2ms 31%
29ms 30%
34.8ms 29%
40.6ms 28%
46.4ms 28%

P4 640 3.2GHz HT enabled 2048MB DDR2 533 RAM SONAR6PE m-audio 2496

@ Twigman : try disabling HT, it made my system way more reliable and responsive !

Hello,
Here are my results...

64bit Double Prec. Engine : on
Mulitproc Engines : on
Dither : Pow-r 1

ASIO 24bit 44100Hz
1.1ms 048 samples 27-40 %
1.5ms 064 samples 22-30 %
2.2ms 096 samples 18-24 %
2.9ms 128 samples 15-20 %
5.8ms 256 samples 11-14 %

WDM 24bit 44100Hz
1.1ms 29-40 %
2.2ms 18-24 %
4.4ms 12-17 %
5.4ms 11-16 %


Here is what got them...

3.5GHz @ 1.4V 7x500 1:1 RAM 4-4-4-12 @ 2.3v
P5B-Dlx || E6300 || Noctua NH-U12F sink || EVGA 7900GTO
Team X 2X1GB DDR2-800 (D9) || OCZ XTC Memory Cooler
Thermaltake VA8003BWS Case || Zalman ZM600-HP 600W
150GB Raptor + 2 300s || Dell 24 wide || RME Fireface 400



Performace looks similar between drivers.
Tomek.

-edit ASIO seems to have less audio problem sub 2.2ms...
and this was w/ Sonar 6.2

ORIGINAL: cappy

I am not able to get the results on the sonartest that I have been seeing posted on this forum. I am running an opteron 180 dc with a gigabyte k8nf9 ultra mobo, 4 gigs of corsair CL2 ram, and a tascam 1884.

I have an Opteron DC 180 as well on a Gigabyte K8-U 939, but with only 1gb Ram and a PCI Audiophile and am getting competitive numbers. I think possibly your problem lies in the Tascams drivers as I hear their not very efficient, but I could be wrong. Aside from AMD's DC fix, I also installed Microsofts Hotfix with the registry edit, and made sure the multiprocessor engine option was on in Sonar, plus the 64 bit engine option, other than that I can only guess what the problem might be.

Well, here's one to chuckle over

Asus M2V AMD 3800 X 2

Sonar 3

Soundcard: er, the builtin 'crap' realtek sound device on the motherboard, driven through ASIO4ALL 2.7

Only took one data point: at 5.8ms latency, CPU was 30%.

Not bad I thought for the crappiest possible output device you could pick.

Interestingly, on my laptop (Pentium M 1.5 IBM T40) with UA25, I cannot get this test to run even at max latency setting (48mS reported). What actually happens is that every time I try loading the test, Sonar crashes catastrophically with an addressing exception. This is worrying - but checking, it looks like there's a new driver for the UA25 - will grab that and retry in a moment.

Update: No, version 2.0.1 of the UA25 driver does not fix this problem. Sonar will crash every time the project is loaded, even with latency set to maximum. The crash is a 0xc000005 addressing exception. This is interesting because the laptop has been quite stable for a very long time indeed. (WinXP SP2, S3P 3.1.1)

If I disable the audio driver i.e deselect the UA25, then the project will load. This looks like we may have a driver issue for Edirol to look at. Can anyone else run this test against a UA25 - would be interesting to see if the problem can be reproduced.

In some cases, the speedup came out to more
than 5%, presumably because the measurement isn't really exact.