DAW Digi-001 Benchmarks Info

[Drastikpup1]

The XP-1900 at a phat price range is a waste of money for the 2 extra d-verbs you will get VS. an 1800. I noticed only a very weak increase in plug in power using D-verb with T-bird 1.4, XP-1500, and XP1600. Between the three, only an increase in 2 D-verbs 12, 14, 16 ECT.

With a full bandwidth of 24 tracks playing back in loop mode 1024 buffer with loop set to 30-minute sessions.

Prospective up-graders and new system buyers may be interested to know that the very in-expensive XP-1500 is about the same performance as the 1600 and 1700. After reading AMD’s white papers and having a 30-minute discussion with AMD they say the XP-CPU architecture is not written for 98SE or ME. It is optimized for XP.

So you will be seeing a slightly noticeable difference in the added instructions of the XP-CPU running 98SE or ME.

XP Beta Now!!!

Of course it would nice if Digi would bust the dam stuff before we all fly down to California and demand for a beta!!

I think the torture test is good because everyone is performing under the same rules but beyond that it’s not real world by any means.

Some of the recent testing we have done is to run two digi-001s into each other with a snake and the light-pipe going at the same time. Giving you the ability of really testing actual live tracks playing out of one digi-001’s I/O through a snake and into another Digi-001 at full volume 16 tracks at once, 8 analog and 8 ADAT.

The test we ran was in increments of 45 minutes of 16 tracks at once set at 1024 buffer and then at a 128 buffer.

After those tracks are recorded you’re left with 8, so we enable those and hit record at 1024 with low latency monitoring enabled for 45 minutes.

When all the tracks are full (24x 24bit 48khz,) we set loop playback and drag the markers so that the 24 tracks play back for 45 minutes and then loop on and on.

After this is set up in real time we start adding Digi rack plug-ins from the top of the list starting with compression, limiter, d-esser, 4 band eq, and delay.

We then remove all of those plug-ins and just load up straight D-verbs until the system hangs. We then back the D-verbs off until it can loop for three hours without a crash.

This to me is pretty real world. There are a lot of other ways to test like writing a lot of automation on 24 full audio tracks in addition to the above mentioned.

I was not seeing a lot of difference in XP-1500, XP-1600, XP-1700, between the 1500, 1600 one or two d-verbs, and the same for the 1700.

So what you are basically looking at here are 4-5 D-verbs gained by going XP-1700 VS XP-1500. The model is based off of a full 24 tracks, playing back with 85% CPU and a 1024 buffer speed. If you were doing fewer tracks like 12 or 16 you would see more plug-ins.

Why use the D-verbs??? Because they are hogs, but not even half as bad as the Bob Moog Bomb factory stuff. No XP-any thing will give you a lot of those.

Basically the Digi rack plug-ins are very weak CPU users. The easy way to tell is to load up a full 24x5 of them and you are not even at the 45% mark for System usage with a T-Bird 1.4.

D-verb is a better indicator of CPU performance because when you load enough of them up they hang the system.

Cutting and pasting wave files is not a very good measuring bar unless you are only doing two-track over dubs to the 24-track point.

I hope our DAWBOX test helps you better understand what the Digi-001 is capable of from a recording standpoint as well as a playback standpoint.

And after doing this for the better part of 16 months now with the PC for the Digi-001 I think this is a very reliable way to test the Digi.

Am I telling to dump the torture test?

No but if you can run some inputs into your Digi drom a drum machine or sampler or even a old tascam 8 track whatever as long as it is real time and actual audio going in not just phantom tracks recording hiss.

We did this same test with T-bird 1.4, XP-1500-2000 as well as the P-IV stuff CPU.

PIV was not as fast yet, but that may change.

Conclusion??

XP-CPU is a very stable system with a Via-KT266a or the 761 chipset. With 256 or 512..

Ram has doubled in price so and I have heard through the market that Samsung is merging with another company.

This should bring some stability in ram pricing.

A bit of a wash though becuase drives and cpu have dropped a bit.



Justin Lynch www.dawbox.com

System Tested


http://www.dawbox.com/amd-digi-ddr_producer.htm



[img]images/icons/grin.gif[/img] [img]images/icons/grin.gif[/img] [img]images/icons/grin.gif[/img] [img]images/icons/grin.gif[/img]

[QuikDraw]

<BLOCKQUOTE><font size="1" face="Verdana, Arial">quote:<HR>Originally posted by Drastikpup1:
...as long as it is real time and actual audio going in not just phantom tracks recording hiss.
<HR></BLOCKQUOTE>

So, what you are saying here is that you'll get different results depending on the audio that is being recorded? Recording "hiss" gives measurably different results than recording say, a bass line? Or guitar part?

Please explain the need to use "real audio" when testing a PTLE setup.

Thanks for the explanation.

Mike

[bstaley]

I am dual booting Win98SE/Win2000 with my XP1900 and I notice a pretty big difference when doing CPU intensive stuff like encoding video (VirtualDub, TMPGEnc, FlaskMPEG, etc.). I get many more FPS (frames per second). I wonder if Windows XP would be that much better than Win2000? I also notice Sound Forge and Vegas Video seem to run much better in Win2000 versus Win98SE. That's why I'm so looking forward to PT LE coming out for Windows XP. Not to mention that it will be more stable.

Interesting reading. Thanks.

[Allen Hallada]

The other thing that is important when upgrading processors in the future is to wait until a new series has been released. AMD is already releasing info on the new Clawhammer that will outperform the current XP2000+ by 100% (double). If one upgraded their system every year or every two years you would be better off than upgrading every 4-6 months. So if you went from the Thunderbird to the XP to the Clawhammer it would save you money over upgrading in the same series. Or even upgrading every other series like I am going to do and skip the XP series. My Thunderbird 1.33 Ghz is still outperforming all of the new XP1800+ and above, except for three, on the A7M266.
24 Audio Tracks + 15 Aux. tracks with 5 plugins each on the Davec test.
Sometimes it is good to learn how to tweak your current system before upgrading and wasting all that processing power.
Allen
[img]images/icons/smile.gif[/img]
[img]images/icons/smile.gif[/img]

[Drastikpup1]

The XP 1700 and the 1600 right now are 2$ apart the 1800 jumps a bit in price but the 1900 is alot more and I am sorry to say not alot more juice for the money.

I am not ripping anyone who got one just be aware that 100mhz in performance does not do much.

Allen is on the money the config has everything to do with it.

Yes a real track of audio will test the system better than recording his.

If you can plug whatever you can into the thing not matter what it is as long as it is loud and peaks the meters.

Than arm as many tracks as you can and make a 30 minute pass or a 2 hour pass it is up to you.

When your done with those 8 arm another 8 whislt those play back. Then arm 17-24 while 1-16 play back.

If your system can handle all three passes the first two at 128 buffer and the last one at 512 or 1024 for at least 30 minutes that will give you a good indication of it's recording power.

If DMA is not on or your OS is not tweaked you will possibly hang and crash as the test proceeds.

When you get the 24 done start tacking on plugs and automation.

(hint) add automation to each track on top of plug ins.

If you system makes it with all of the Digi-racks and automation it is running great.

Now let it loop over night...

If it is still looping in the morning without a dae error or a blue screen...

It is now a pretty solid machine worthy of trusting it when you need it the most.

Defrag after every session like religion.

PS. lose alot of sleep and drink some coffee because you going to be up late.

Justin
http://www.dawbox.com

[img]images/icons/smile.gif[/img] [img]images/icons/smile.gif[/img] [img]images/icons/smile.gif[/img] [img]images/icons/smile.gif[/img] [img]images/icons/smile.gif[/img]