Page 1 of 2

MSI B75A-G43 (intel B75 chipset) testing and results.

Posted: Fri Apr 12, 2013 11:45 am
by Polarity
It will be long series of posts, sorry.
But I had a lot of details to write down. I tested heavily the new mainboard and CPU,
and updating this text file during the various steps of my tests for 2 days.

MSI B75A-G43 intel B75 chipset, 3 legacy PCI slots
CUPU i7 3770K at default 3.5Ghz (intel turbo boost enabled)
Corsair 16gb RAM at 1600 Hz
Asus Radeon HD7850 2gb RAM
WD Raptor 74Gb (with WIN7) connected to SATA3 port.
+ WD Black 1Tb + WD Blue 1Tb + WD 320Gb SATA2 conencted to SATA2 ports
PSU Tagan U22-580W
Windows 7 64bit Home Premium
Scope 5.1 - 48KHz - 23ms latency (it has always been my usual use)

Consider that I didnt' reinstall Windows 7 or Scope 5.1...
I just connected my actual hard disks to the new motherboard (outside of the DAW case) and just messed a bit with drivers updating.

This system seems more stable and quicker than my previous P965 chipset DFI mobo with Q6600 CPU...
Cubase now is playing my projects without any glitches/crackles/pops anymore
and this alone is making me really happy! :)
Actually the Masterverb test is disappointing... a crap I dare to say...
but during the practical use it's performing very fine for my usual needs.

I tried to increase PCI latency timer from 64 to 128 and updated to latest BIOS version but didn't change the Masterverb test PCI limits warning.
Changing the slot from the 3rd (the more external) to the 2nd (central) PCI slot
seems to not bring big difference.
But on the 3rd slot the Pulsar has IRQ 18 all for itself alone,
while on 2nd slot gets IRQ 17 and so you can see with what share it from the image I attach.

I didn't tried slot 1 for now (cause the HD7850 videocard is two slot wide).
I will try another day perhaps removing that and using the onboard videocard.

Re: MSI B75A-G43 (intel B75 chipset) testing and results.

Posted: Fri Apr 12, 2013 11:46 am
by Polarity
the IRQs sharing image

Re: MSI B75A-G43 (intel B75 chipset) testing and results.

Posted: Fri Apr 12, 2013 11:53 am
by Polarity
These were the Reverb test results:
Everytime I always verified if the sound passed through the reverb units, placed in parallel to the mixer, without any artifacts or dropouts. At 48 KHz:

Pulsar2 in PCI slot3 (IRQ 18 alone)
3 Masterverb (reaching PCI capacity limits loading the 4th one)
1 Masterverb PRO (reaching PCI capacity limits loading the 2nd one)
but working with two Masterverb PRO sends/returns with Mixer STM 2448 and a ASIO-Cubase VSTsynths/plugins only (no Scope synths or external hardware expanders) with satisfying performance for my usual work.

Then I changed the slot and went on on that one.
Pulsar2 in PCI slot2 (IRQ 17 shared)
(with Hyperthreading ON)
3 masterverb (reaching PCI capacity limits loading the 4th one): 6 playing fine the sound file from Wavelab.
Not able loading the 7th Masterverb unit to get over the PCI overflow alert, Scope became instable and got out of sync. Anyway once resynced the Scope rate it played fine the file through the existing 6 units.

(with Hyperthreading OFF)
3 Masterverb without first PCI limits warning but I went go on adding till 7 and play fine Wavelab file through all of them.
After that I couldn't try 'cause I exhausted the 6 DSPs power.
I saved the project, closed it and tried to reload it: you know what? he didn't give me the PCI limits warning, only the DSPs exhausting. The yellow DSP bar was empty, so I tried to resync the master clock.
It worked and Scope played normally the audio file through the 7 Masterverb units without PCI overflow warning.
I was a bit surprised. It's a weird behaving.

PS: HyperThreading OFF seemed to make some difference in better at first try, but the day after when I retried the test with Hyperthreading ON again I got the PCI limits warning at the 5th unit.
I played sound through all 7 units (the maximum for the DSPs I have), so 1 unit more than the first time I tried.

Re: MSI B75A-G43 (intel B75 chipset) testing and results.

Posted: Fri Apr 12, 2013 11:54 am
by Polarity
B75 Masterverb Test 7units image

Re: MSI B75A-G43 (intel B75 chipset) testing and results.

Posted: Fri Apr 12, 2013 11:56 am
by Polarity
6 MasterVerb Classic without PCI limits warning. (what's the difference with the other Masterverb? I see it uses less DSP power).
With difficulties I made to connect to the mixer also the 8th but after some retries I made it play with all the others.
Of course I reached the available DSPs power, so couldn't go on.
Closed it and reloaded the saved project: it opened without any PCI limits warning and the audiofile played through all 8 units without problems.
I think it's a bit weird it doesn't give warnings at all.

Re: MSI B75A-G43 (intel B75 chipset) testing and results.

Posted: Fri Apr 12, 2013 11:57 am
by Polarity
B75 Masterverb Classic Test 8units image

Re: MSI B75A-G43 (intel B75 chipset) testing and results.

Posted: Fri Apr 12, 2013 11:59 am
by Polarity
2 Masterverb PRO without getting PCI capacity limits warning.
Got it loading the 3rd one, but didn't get it loading the 4th one.
Got again the warning loading the 5th unit.
Anyway it plays fine the audiofile through all the 5 Masterverb PRO units.
Then I had no more DSPs available.
Again saving and reloadind the project I got DSPs warning and empty bar and no sound: resyncing the Master brought Scope on service again and played soundfile fine through all 5 units.

Reloading again and again these two projects (even the day after), with Hypertherading off or on, I got same behaviour.
Usually I work with one MasterVerb PRO only connected to my STM2448 project, just as a glue for sounds coming from different sources.

Re: MSI B75A-G43 (intel B75 chipset) testing and results.

Posted: Fri Apr 12, 2013 12:01 pm
by Polarity
B75 Masterverb PRO Test 5units image

Re: MSI B75A-G43 (intel B75 chipset) testing and results.

Posted: Fri Apr 12, 2013 12:01 pm
by next to nothing
first thing to do (apart from the obvious reinstall) is to disable onboard audio. also disable whatever other devices you dont use. Unless you have special needs i dont see why you have two hd audio devices running. "Active management technology"? try turning it off if you do not have state secrets on your pc!

Re: MSI B75A-G43 (intel B75 chipset) testing and results.

Posted: Fri Apr 12, 2013 12:04 pm
by Polarity
1 SC-RMX160 S (reaching PCI capacity limits loading the 2nd one)
but after hitting RETRY button I loaded 2 other more and played sound through dynamic mixer without artifacts.
Not able loading the 4th SC-RMX160 to get over the PCI overflow alert (hitting retry button) when linking its outputs to mixer.
It played fine the audio through 3 SC-RMX160 units.

2 SC-Plate without warnings.

As last test yesterday night I opened and played while recording to a wav file a 5 years old Cubase project that was using real time also 4 Hardware Expanders (through 2 ADAT A/D converters) together VST instruments and drums (Stylus RMX) that were using 7 ASIO busses.
Mixer STM2448 with 1 MasterVerb PRO in send on almost all channels used (ASIO and Hardware Expanders).
All perfect: no glitches, no crackles, no pops, no dropouts.
For now I'm satisfied, I put away an annoying issue that in these 2 years I couldn't be able to get rid off whatever I tried!!

Finally today I tried also Scope synths:
Minimax alone: played the basic default patch without any problems.
5 full chords plays fine.
Instead with 6 notes some distortion, but with 6 notes of polyphony I exhausted all DSP power and got red led in the bar. However absolutely no delay with playing all the notes.

No problems also playing 4 Scope synths (Minimax, Prophit5, SB4044 and Uknow007) together at same time, each one monophonically.

For me the system is actually usable and workable, despite the not good (at least weird) results in masterverb test, so I think I'll reassemble all inside my DAW case and replace the DFI mobo.

Thank you for reading through all, and hope it can be useful.
I will update eventually if something changes.

All the best.

Re: MSI B75A-G43 (intel B75 chipset) testing and results.

Posted: Fri Apr 12, 2013 12:04 pm
by next to nothing
for masterverb test, use classic, and remove every other device including mixers, in and outs etc. use a completly empty project. not for practical reasons but for true comparison.

Re: MSI B75A-G43 (intel B75 chipset) testing and results.

Posted: Fri Apr 12, 2013 12:11 pm
by next to nothing
oh, also use 44.1 khz for the test. from what i see from your screenshots you are able to max out the p2, so i think you are ok :)

Re: MSI B75A-G43 (intel B75 chipset) testing and results.

Posted: Fri Apr 12, 2013 12:15 pm
by Polarity
OK, I will do.
But I think won't change much.

Usually I load Scope only when I want to so some music work (compose, recording or mastering), that's why I always kept enabled the onboard mobo audiocard.
I use it for everyday multimedia: browser, winamp, video editing and VLC and videogames.
I experienced audio delay with Pulsar and videogames on previous mainboards in past, so I didn't retry to use it for gaming.

Re: MSI B75A-G43 (intel B75 chipset) testing and results.

Posted: Fri Apr 12, 2013 12:21 pm
by Polarity
next to nothing wrote:oh, also use 44.1 khz for the test. from what i see from your screenshots you are able to max out the p2, so i think you are ok :)
I tried it, but for some strange reason Wavelab refused to play audiofile every time I clocked Scope at 44.1 Khz.
It happens also on my other system: perhaps this Win7 install has messed up something among them.

Re: MSI B75A-G43 (intel B75 chipset) testing and results.

Posted: Fri Apr 12, 2013 12:27 pm
by next to nothing
I totally understand. It's just some hints to show how the actual performance is compared to other systems. but as i said it seems you are maxing out the p2, so you are probably fine :)

ps. i use scope for gaming as well these days and it behaves perfectly.

Re: MSI B75A-G43 (intel B75 chipset) testing and results.

Posted: Fri Apr 12, 2013 12:30 pm
by Polarity
Here we go.
MasterVerb Classic test at 44.1 Khz in a complete empty project.

7 MasterVerb Classic without PCI limits warning.
Added the 8th one with retry button..
but can't get over the PCI warning with the 9th one (I hit the retry button for 4 times before giving up).

Re: MSI B75A-G43 (intel B75 chipset) testing and results.

Posted: Fri Apr 12, 2013 12:30 pm
by next to nothing
Polarity wrote:
next to nothing wrote:oh, also use 44.1 khz for the test. from what i see from your screenshots you are able to max out the p2, so i think you are ok :)
I tried it, but for some strange reason Wavelab refused to play audiofile every time I clocked Scope at 44.1 Khz.
It happens also on my other system: perhaps this Win7 install has messed up something among them.
yup, but for the "standard" masterverb test you dont have any audio going thru the system. its not a practical use test if you know what i mean. so if you want to compare it with other results you dont run any audio, you do 44.1 khz and with an empty project :)

Re: MSI B75A-G43 (intel B75 chipset) testing and results.

Posted: Fri Apr 12, 2013 12:31 pm
by next to nothing
ah, posted before i read the other post :)

Re: MSI B75A-G43 (intel B75 chipset) testing and results.

Posted: Fri Apr 12, 2013 8:54 pm
by dawman
Do you have the S|C 5.1 Room...?
It is a fine Reverb and running a pair of these, one in each AUX covers most needs.
One Verb for spaces that are smaller, like a wooden stage, etc. and then a medium to large Hall.
Personally this Reverb test has nothing to do with the system having limited resources.

I will be building a new Haswell 1U build using the i7 4770R/ and the MSI Military Grade 4 spec'd board, which also has legacy support.
From my experience with the i7 3770S/MSI-Z77 I am so impressed, the upgrade will be the first time I really jumped on an untested platform, but I won't use it until it becomes stable.
But it will be a spare 4U w/ 3 x Scope Cards. I believe there was so much stink from the user base about piss poor Legacy PCI support they came back in Spades.

I can tell you one thing though, the Ivy Bridges' GFX on die is advanced and very efficent, and the Haswell is 50% better than the Ivy Bridge, and 15-20% lower TDP.
Intel is really making some moves here lately when they don;t have to as AMD is barely making a dent in the server or consumer market, but thankfully enough to stay alive and keep Intel cheap.

But I wouldn't let the Reverb test disappoint you. I used 3 x Pro's and 9 MVPs was all I got , but I had Solaris, B2003, Modular, effects and Prowave rocking with headroom left.

Personally I actually prefer an external Hardware Reverb, they just seen to be more overwhelming in encapsulating the audio source.
A great Reverb that would be easy to run using 50 dollars for conversion tools in ADAT, and it's the Alesis Q20.
If one had an A16 you could run a pair of these, then a pair of hardware pedals like the Moog 12 Stage, Moogerfogger, or Rind Modulator.

Just talking out loud, and obviosuly another addicted Scope lover.

Cheers.

Re: MSI B75A-G43 (intel B75 chipset) testing and results.

Posted: Sat Apr 13, 2013 6:53 am
by Bud Weiser
dawman wrote:
I will be building a new Haswell 1U build using the i7 4770R/ and the MSI Military Grade 4 spec'd board, which also has legacy support.
... and the Haswell is 50% better than the Ivy Bridge, and 15-20% lower TDP.
May I ask which MSI military grade 4 board you have in mind exactly ?
And yes,- I´m interested in lowering TDP specs for DAW rackmounts too.

Bud