jcschild
""""See, when Scott asserted that certain USB and FW AIs performed better than some other PCI AIs, and I'd already come across a post previously in RME's forum by an RME sales rep/forum mod stating that RME's own PCI/PCIe AIs performed better latency-wise than their USB and FW AIs, and I happened to have a pretty good idea that RME sales rep/forum mod had worked closely with Scott, well, it was like, kinda too ironic, """"
do NOT twist my words.. i have never said that RME USB/FW works better than RME PCI/PCIe. (pretty dang close however)
Never said you had. I didn't twist anything you said.
Actually, I had earlier said:
Goddard
PCI can still offer better performance (esp. lower latency) than USB or FW.
jschild
i said RMEs USB/FW can work better than some OTHER PCI cards.. EG: Echo/M-Audio/Emu and other low budget older cards
Yep, that's pretty much how you responded to what I'd said above. Here is exactly what you said:
jschild
PCI does not nessesarily offer lower latency than USB/FW it depends on the interfaces you are comparing.
i would put an RME USB up again an M-Audio/Echo/Emu PCI all day long..
Now, I never disputed what you wrote there about it depending upon the interface (because I completely agreed - it definitely depends upon the particular interface).
I had never stated that
all PCI interfaces
absolutely performed better than
all USB or FW interfaces. I said "PCI
can still offer..." just as you've just said that "RME's USB/FW
can work better than some OTHER PCI".
Btw, perhaps your "all day long" comparison between an RME USB against an M-Audio, Echo, E-Mu PCI needs to be viewed in light of the newer RME USB's much higher cost compared to those "old low budget PCI" interfaces, rather than as proving any inherent performance advantage of USB against PCI (which there isn't - quite the opposite). Apples-to-apples, and dollars to donuts.
So, in response to what you'd pointed out (namely, about it depending upon the interfaces under comparison), I simply pointed to Chris' RME forum post, as an "apples-to-apples" performance comparison of PCI with FW/USB, noting that even with RME (the example you'd just given for your USB vs. PCI comparison), their PCI performed better than their USB/FW as had been stated by Chris (who I knew was one of the better DAW guys around and had worked with you) in RME's forum.
Seemed to me like a pretty authoritative source for a fair PCI vs FW/USB performance comparison on equal terms, rather than your "all day long" RME USB vs. "old low budget PCI" example. So what's the problem?
jschild i
have also clearly stated many times than a PCI/PCIe is only .75ms better low latency RME vs RME and even Matthis will say this.
Are you talking about RME's PCI compared to RME's PCIe? Or about RME's PCI/PCIe compared to RME's FW/USB?
I don't believe I ever asserted that PCI performance was superior to PCIe, but if anything I said came off that way it was never intended as such.
However, I may have stated that some PCIe cards offer little or no performance advantage or increased functionality over the PCI versions beyond PCIe compatibility while costing rather more. RME's HDSP and HDSPe PCI host cards being just one example (perhaps this was dictated by a need to remain compatible with their outboard gear), and MOTU's 424 cards being another.
jschild i
have also stated that there is NO difference in PCI to PCIe benchmark wise (low latency)
And I don't believe I have ever asserted any such difference, have I? Nor would I expect there to be much if any difference latency-wise. Any difference would (hopefully) be in the form of improved bandwidth and lower system loading and DPC latency for PCIe over PCI,
Which of course raises the question, for people using still-supported and/or -workable PCI cards, whether it is advantageous (or even necessary) to dump their PCI card and get a PCIe card. I mean, if one can still use one's PCI interface in a new mobo, and a new PCIe version won't offer much if any improvement, why go to the expense?
But I was talking about PCI vs. USB/FW performance, not PCI vs. PCIe.
So let's look at USB and FW. In this regard, even with a USB interface having a superior USB implementation (as RME claim to have) and drivers, the host PC still has to do all of the heavy lifting for USB i/o (due to the way USB is implemented on the host/root side), which loads the cpu and increases DPC latency. And there still arise incompatibilities between certain USB interfaces and certain USB controller chips (and system chipsets) in PCs, most recently involving PC USB3 ports.
FW is better than USB in regard to system loading and DPC latency, because FW controller chips can offload most of the i/o processing and thus don't load the cpu like USB does. However, FW performance and compatibility is still often a concern, and can be a real "chipshoot" gamble both on the interface end (DICE=bad & slow) and PC side, as may be seen from the compatibility info given on most interface manufacturers' website and the benchmarks on the DAWBench site.
jschild
lastly. even on an X79 there are still some PCI that simply do not work right... whilst RME will work fine on Z77 when others will not.
Yes, there have always been compatibility issues, for various reasons, on various platforms. But that does not change the fact that the X79 does in fact offer native PCI support, and in any case, such compatibility issues may not necessarily be down to the X79's PCI. A BIOS which does not correctly detect or configure PCI devices, or the different way Windows (since Vista iirc) enumerates and configures PCI devices may be involved. Or it could be drivers which don't play well with 6 cores and 4 channel memory on X79.
PCI compatibility on bridged PCI seems to be a big chipshoot. Even RME appear to have problems with bridged PCI on some Z- mobo's (Z68) judging from posts I've come across on their and other forums. Personally, if I were going with a non-native PCI mobo, I would just as soon get a PCIe interface card. if I could find one suitable to purpose.
If anything, compatibility is card-by-card and mobo-by-mobo, pretty much as it has been for a long time. IOW, it depends upon the interface, and upon the mobo, same as it ever was.
jschild
keep it stright.
Hey, I've been straight up about everything, not shouting misinformation about all 1155 and 2011 having bridged PCI (really, you of all people ought to know better).
Lessee now...
OP posts in forum seeking advice on what platform to build for new DAW, advantages/disadvantages and PCI connectivity.
You reply, posting misinformation about non-availability of native PCI support on 1155 and 2011 platforms, then state that PCI may or may not work depending on interface, and that RME does work but other interfaces probably won't or maybe might, and also state that PCI is pretty much dead.
OP then asks (quoting your earlier post) whether, since PCIe interface choice is limited and FW seems to be on the way out, he has to settle for a USB2 interface now.
I post, pointing out that X79 and some 1155 do still have native PCI, so that native PCI is not dead quite yet, and that PCI can still offer better performance, especially latency, than FW or USB.
You post, asserting once more mis-info about X79 PCI being bridged and pointing to Intel's incomplete X79 block diagram, state that PCI performance advantantage over FW/USB depends on the interface, that you would put up RME USB against M-Audio, Echo and E-Mu PCI all day long, and recommend a number of FW and USB interfaces to the OP.
I post, pointing out again that X79 does support native PCI and referencing Intel's X79 datasheet (linked on same webpage as the block diagram you linked to), and, in reply to your statement about PCI vs. USB/FW performance depending upon the interface, suggest an apples-to-apples comparison of RME PCI vs. RME FW/USB, pointing to Chris's post in RME's forum, and note that PCI compatibility also depends upon the drivers and OS and application, and mention about WaveRT.
You post, questioning my qualification to speak to audio interface performance, bragging about being an RME' beta tester, and dismissing WaveRT and calling anything but ASIO (btw, no surprise, it's, the only driver mode RME support) a joke.
I post, responding with an explanation of where I'm coming from, suggesting that I might in fact be as qualified to speak to technical matters regarding DAWs and audio interfaces, and addressing your assertion regarding WaveRT. I also suggest (a bit tongue-in-cheek, as in oh, how ironic!) that the reason why it is being found that PCI interfaces can somehow work on X79 mobo's might be that x79 mobo's support native PCI (a technical fact about which you were still completely wrong, were still spreading incorrect info about here and elsewhere, and which you, as '"computer pro" and "DAW seller" should have known or at least been able to easily confirm was otherwise (or maybe you lost your Intel distro's number?), and still refuse to so admit here).
At this point, another poster posted, quoting my earlier post (in which I had pointed out again that X79 had native PCI support and had jokingly suggested that maybe the reason PCI cards work on X79 is that X79 supports native PCI), and then (apparently missing my attempt at a joke) states that some PCI cards work on some bridged mobo's.
So I posted, responding to the other poster, pointing out how my statement which he quoted was meant as irony.
And then yet another poster posted, suggesting that my attempt at irony was actually sarcasm.
So then I posted, responding to the sarcasm/irony distinction suggester's post, by pointing out that I had thought it rather ironic when you (Scott ADK) had disputed my earlier statement that PCI can offer better performance (esp. lower latency) than FW/USB and had asserted your RME USB vs. "old low budget PCI" example, when I knew (from Chris' RME forum post) that even RME's PCI outperforms RME's FW/USB and noted that.
And, so, here we are now, with you accusing me of twisting your words and not being straight.
So where exactly was I not straight?
Was I wrong about the availability of native PCI support on certain 2011 and 1155 mobo's? Or were you perhaps the one who was wrong?
Was I wrong in stating that according to RME, their PCI offers better performance than their FW/USB?
All you've done so far in reply to the OP's inquiries is to give incorrect info about native PCI support availability on the platforms he asked about, dismiss PCI as dead, and tell him he needs a FW or USB interface. Not once did you even inquire what PCI interface he wanted to connect, so as to be in a position to advise on whether it could work (you know, based upon your vast knowledge and experience in such matters). Sorry Scott, but it all just smelled like a sales pitch to me. Margins on RME FW and USB gear must be pretty good..
jschild
and PCI is dead!
Yeah, so you keep saying. just like you kept saying that X79 doesn't support native PCI. Only that don't necessarily make it so. Does makes me wonder though...
Hmm, if PCI is dead, then why do Intel still offer PCI support on some of their latest greatest platforms?
Perhaps it's because there remains a sufficiently strong enough demand for native PCI support in the market that Intel and its mobo mfr chipset customers can turn a nice profit satisfying it?
Perhaps still enough demand for native PCI support in "industrial" and "telecom" and "data acquisition" and "scientific" and "medical" and "business", etc. market segments, where companies, institutions, hospitals, businesses etc. have invested heavily in PCI cards and software and systems running with them? The kinds of customers that buy mobo's with native PCI and long term support.
So, excuse me while I dispute that PCI is dead.
Maybe there aren't a great many PCI interfaces still available or supported (and sadly, maybe even fewer PCIe), but there are still some which can work in current mobo's (yeah, I know it depends upon the PCI card and the mobo) and which can offer comparable if not better performance to some USB and FW interfaces available (yeah, I know, it depends upon the interface). Point being, if someone still has a PCI interface and wants to upgrade to a more current system, they might be wise to find out if it will actually work in a newer mobo rather than to ditch it just because they see you keep announcing that PCI is dead. At least, that is, if they are concerned with audio latency and cpu load and system DPC latency.
But I don't benchmark these days. So I'll simply defer to someone who does:
http://www.dawbench.com/audio-int-lowlatency.htm and note that, in his Conclusion at the bottom of that page Vin stated:
Vin (TAFKAT)
Its quite clear from the testing results that for those that require ultra low latencies , the PCI/PCIe alternatives are always a good option .
Sadly there aren't a lot of professional PCI/PCie cards currently available on the market , developers opting more and more for FW/USB 2 as they can service a far wider market on both desktop and mobile, which is understandable in such a niche market such as ours. I can't help feeling that for some of the developers the focus has definitely shifted away from performance , which I find quite concerning.
LLP- Low Latency Performance is extremely important for users who require those lower latencies in their work flow, prime example being those that compose with virtual instruments and also guitar amp simulators where the round trip latency also comes into play , more so than those who are doing simple tracking of bands and mixing, for example.
However its increasingly common for audio interface units to be marketed , reviewed and sold on the bells and whistles , more so than the actual performance, which for some can lead to a rude awakening when the interface delivers poor performance at the preferred working latencies.
Finally, regarding WaveRT being a joke, I will just say that although there are only a very few PCI and PCIe audio interfaces with WaveRT driver support (for various reasons, including MS' screw-ups and reluctance by PCI/PCIe interface makers to develop WaveRT drivers when they had already developed and were supporting ASIO drivers) which do not actually reflect upon the technical merits of WaveRT, there are nonetheless a vast number of WaveRT driver supported audio devices deployed out there in user land: namely, almost every Windows "High Def Audio" (HDA) onboard soundchip in recent PCs.
So, anyone with a recent Windows PC with onboard HDA sound might want to check if it has WaveRT drivers, and if so, test their latency and cpu load and DPC latency under WaveRT for themselves. No, the audio quality might not compare very well with a higher quality pro/prosumer interface (t's just cheapo onboard sound after all), but that's not the point, which is the RTL, cpu load and DPC latency performance question.
'Cause you just never know, Scott just might be wrong (it's happened before, right here before our eyes) and your onboard HDA sound might perform decently enough that it could be used say, for non-critical playback or monitoring or running a digital feed out over.S/PDIF or whatever, and you might end up not needing to lug that expensive FW or USB interface around everywhere..
Selah....
post edited by Goddard - 2012/08/06 22:46:56