Scott Reams
Max Output Level: -56 dBFS
- Total Posts : 1918
- Joined: 2003/11/06 15:32:28
- Status: offline
RE: OT: Xeon vs. P4 vs. Itanium
2003/11/09 05:04:48
(permalink)
Hi Scott, the MSI uses a VIA chipset.. dont think i would suggest that for Audio.. Scott ADK I'm fully aware. It's not really an issue... considering that a) Via worked out any issues a while back for the most part and b) the Northbridge is now in the CPU... and Via is just handling Southbridge functionality. -S
|
Noel Borthwick [Cakewalk]
Cakewalk Staff
- Total Posts : 6475
- Joined: 2003/11/03 17:22:50
- Location: Boston, MA, USA
- Status: offline
RE: OT: Xeon vs. P4 vs. Itanium
2003/11/09 09:17:36
(permalink)
Except that its not dual CPU capable? ORIGINAL: Scott Reams The AthlonFX 51 may as well be called an Opteron 148. It's the exact same CPU.
|
Scott Reams
Max Output Level: -56 dBFS
- Total Posts : 1918
- Joined: 2003/11/06 15:32:28
- Status: offline
RE: OT: Xeon vs. P4 vs. Itanium
2003/11/09 14:35:04
(permalink)
ORIGINAL: Noel Borthwick [Cakewalk] Except that its not dual CPU capable? ORIGINAL: Scott Reams The AthlonFX 51 may as well be called an Opteron 148. It's the exact same CPU. Neither is the Opteron 148. 14x CPUs are single-CPU. 24x CPUs are dual. 84x can be setup in 8-way configurations. -S
|
jcschild
Max Output Level: -41 dBFS
- Total Posts : 3409
- Joined: 2003/11/08 00:20:10
- Location: Kentucky y'all
- Status: offline
RE: OT: Xeon vs. P4 vs. Itanium
2003/11/09 15:14:17
(permalink)
the Northbridge is now in the CPU... and Via is just handling Southbridge functionality Hmm good point. hadnt thought of that, dang it now i have another board to validate... Scott
|
GaryMedia
Max Output Level: -86 dBFS
- Total Posts : 217
- Joined: 2003/11/05 23:04:20
- Location: Cary, NC
- Status: offline
RE: OT: Xeon vs. P4 vs. Itanium
2003/11/09 22:34:43
(permalink)
This seems a good place to ask this quasi-hypothetical question: I have an opportunity to test Sonar 3 Producer on a 4-way Xeon server. It has 4 Intel Xeon 1.4ghz processors (Foster), and 5GB of RAM. These are the ones with the 512K L3 cache. So far it seems that I must at least use Win2000 Server for it to recognize 4 CPU's (can't guess whether the HyperThreading will work on this to give an 8-way appearance), nor is it a critical issue as long as the 4 real ones are used properly. My question is whether the I/O tasks and operating system overhead are sufficiently independent of the Audio Engine and the GUI to bring a benefit from having so many places to dispatch work on relatively weak 1.4ghz processors. (I use a 2.4 ghz P4 in WinXP-Home now as my DAW). It's fraught with potential compatibility issues with the device drivers for the Aardvark or MOTU cards, but I'm really excited about the possibility that I could get GigaStudio to work with Sonar in this environment. Tech support at Nemesys once warned me that dual processors weren't a stable environment for their product, so I'm not optimistic about this working on a 4-way. Anybody have some advice about what to try, and what NOT to try?
|
jcschild
Max Output Level: -41 dBFS
- Total Posts : 3409
- Joined: 2003/11/08 00:20:10
- Location: Kentucky y'all
- Status: offline
RE: OT: Xeon vs. P4 vs. Itanium
2003/11/10 09:54:39
(permalink)
Aardavark wont even work on a HT system, they claim no compatibility with dual therefore 4 CPUs no way.. however i have had an Aardvark working on duals. powercore wont work with HT either (unless they fixed it) nor will UAD-1. but we do have several running on dual systems, which i find odd. Giga seems fine on Dual. however with the Xeons with HT (shows 4 CPU) we have had a few issues and many have had to turn off HT. i would say its a major crop shoot. Scott ADK
|
Marquis42
Max Output Level: -88 dBFS
- Total Posts : 123
- Joined: 2003/11/07 13:34:53
- Status: offline
RE: OT: Xeon vs. P4 vs. Itanium
2003/11/10 18:58:05
(permalink)
With Giga and Sonar 3, I think you would find yourself I/O bound long before you found yourself CPU bound (presuming the quad Xeon). This would be particularly true of some of the very large .gigs that absolutely HAVE to be streamed from disk (think WHILE recording multiple tracks). It could possibly be done, but you'd have to invest a serious amount of money on good I/O. I would much rather have a second system dedicated to GigaStudio with just digital I/O, one relatively powerful CPU, and a decent MIDI interface (or interfaces). That is, my preference if this is what you want to do is just get two single-CPU boxes and dedicate one to GigaStudio. But then, that's me. I do precisely that with two GigaStudio-dedicated systems and one solely for Sonar 3 Producer.
|
GaryMedia
Max Output Level: -86 dBFS
- Total Posts : 217
- Joined: 2003/11/05 23:04:20
- Location: Cary, NC
- Status: offline
RE: OT: Xeon vs. P4 vs. Itanium
2003/11/11 10:03:24
(permalink)
Of course! If price was a factor in this, I wouldn't take this approach at all! I would never actually buy this 4-way monstrosity. It's just available for testing and I would like to explore its possibilities. In fact the fan noise alone from this beast would probably disqualify from a useful life as a DAW. I'll know when I actually get to see/touch/hear. For GigaStudio I think streaming will be a breeze, since the box comes with a native set of three 36GB SCSI 160 drives. Secondarily, with 5GB of RAM is attractive in that I might be able to fiddle with the registry to get GigaStudio to use more of the RAM than it usually does. One potential showstopper is the fact that there is no AGP slot in this thing, and the motherboard video runs on one of two or three (independent?) PCI busses. I haven't figured out how the PCI-X is implemented on this machine and I want to run dual monitors. I want to keep the monitor PCI bus isolated from the audio PCI bus and I'll be adding a PCI video card (remember those?). Right now I have a GigaStudio machine 1.9ghz P4 and separate Sonar DAW 2.4ghz P4. The 1.9 GS (512MB RAM) is a small form factor PC which has space for only one physical hard-drive. The native GigaStudio grand piano takes a long time (6 minutes?) to load from the logical D drive. Before I take the option to use external FireWire drives, or simply go to another full-sized machine (or XPC) I plan play with the 4-way Xeon solution.
|
Marquis42
Max Output Level: -88 dBFS
- Total Posts : 123
- Joined: 2003/11/07 13:34:53
- Status: offline
RE: OT: Xeon vs. P4 vs. Itanium
2003/11/11 11:16:03
(permalink)
If the box is yours to play with as you want (not in terms of ownership, but to monkey with), then that sounds great. I'd like to hear what you find out. I'd also try to make sure that the drives are setup with one physical boot disk and a RAID0 array with the other two. I would avoid using all three in a RAID5 particularly, although a three-disk JBOD array might work with one as boot, one as a GigaStudio source, and one for recording to. Might I ask what kind of machine it is (brand, model)? I could tell you what the PCI bus configuration is if you do.
|
GaryMedia
Max Output Level: -86 dBFS
- Total Posts : 217
- Joined: 2003/11/05 23:04:20
- Location: Cary, NC
- Status: offline
RE: OT: Xeon vs. P4 vs. Itanium
2003/11/11 16:11:58
(permalink)
The machine is an IBM xSeries 360. An 8686-1RX
|
Marquis42
Max Output Level: -88 dBFS
- Total Posts : 123
- Joined: 2003/11/07 13:34:53
- Status: offline
RE: OT: Xeon vs. P4 vs. Itanium
2003/11/12 08:42:21
(permalink)
Looking at the board, the two PCI slots closest to the center (not the Remote Supervisor Adapter slot which is THE closest to center) are Bus B and the other four are Bus A. While I can find no documentation about it, the existing integrated VGA would have to be on a separate PCI bus or else the PCI-X slots couldn't work in PCI-X mode (for compatibility, the busses downgrade to the lowest common denominator). I also could find no mention of what PCI bus the SCSI controller uses. Despite the vast wealth of information about deployment planning and such, IBM is surprisingly coy about technical documentation. Oh well... Hopefully that will at least help you.
|
GaryMedia
Max Output Level: -86 dBFS
- Total Posts : 217
- Joined: 2003/11/05 23:04:20
- Location: Cary, NC
- Status: offline
RE: OT: Xeon vs. P4 vs. Itanium
2003/11/12 22:57:09
(permalink)
Thanks for taking the time to look into this. It'll be at least six weeks before I get a chance to actually fire up this machine. Until that time I'll keep poking around for additional data. I'd like to develop a satisfactory remote consolte/dual screen solution since I saw the noise specification as 6.3 Bels !!! 63dB is one loud box! I just need my remote console solution to be CHEAP! Before I try any fancy KVM-over-LAN solutions, I think I'll simply try long cables and see how well my screen images hold together. Along with that I may have to go for a MOTU AudioWire solution since that'll allow more than 15 feet from the rack to the DAW. The other option is to minimize the noise in a DAW sound suppression box. I have a homemade version of the ISORAXX working today. It's adequate but not outstanding. I'll keep you posted. Thanks again.
|
Marquis42
Max Output Level: -88 dBFS
- Total Posts : 123
- Joined: 2003/11/07 13:34:53
- Status: offline
RE: OT: Xeon vs. P4 vs. Itanium
2003/11/13 07:07:20
(permalink)
I would try to build an iso box for the DAW, though ensuring adequate ventilation while still having fair isolation won't be easy. 63db is pretty freakin' loud, though not surprising considering it's a big quad-Xeon server.
|
GaryMedia
Max Output Level: -86 dBFS
- Total Posts : 217
- Joined: 2003/11/05 23:04:20
- Location: Cary, NC
- Status: offline
RE: OT: Xeon vs. P4 vs. Itanium
2004/03/29 13:55:21
(permalink)
I finally got this 4-way Xeon server and tested it under Win2003 server. I ran into the ultimate showstopper: None of my audio cards would go into the 3.3v PCI slots. As it turns out, this motherboard exclusively supports 3.3v PCI, (two bridges on the 32-bit slot), so no MOTU, Aardvark, ESI-WaveTerminal or Cport would fit! Game over! ...what a drag... It's really ok, though, this thing was so loud, it would have taken me two weeks to build a suitable box to muffle this beast while maintaining reasonable airflow. Furthermore, the 300+ watts it draws at idle would exacerbate my studio heat problem during the summer.
|
melloman
Max Output Level: -86 dBFS
- Total Posts : 228
- Joined: 2003/11/07 00:37:32
- Location: Portland OR, USA
- Status: offline
RE: OT: Xeon vs. P4 vs. Itanium
2004/03/29 14:13:48
(permalink)
That's the problem with server boards and consumer/PC cards. Anyone know of any PCI-X audio cards? Having built a 3 gig P4C, I am not sure how much power I need over that... If I needed to do more than what I could do with that machine, I would probably be doing stuff in another studio (I have a limited space, and limited outboard gear. I could probably upgrade to 16 tracks, maybe even 24 tracks without any problems recording, and then I could manage the processing from that without a lot of management issues. I would probably want to upgrade my memory to the four gigs my board would handle, if that were the case... But altogether that would be only about 500 for the memory, and 1600 for 2 more Q10s.
"There's no problem that the proper application of high explosives can't solve" Cpl Mallory www.mindlayer.com
|
Marquis42
Max Output Level: -88 dBFS
- Total Posts : 123
- Joined: 2003/11/07 13:34:53
- Status: offline
RE: OT: Xeon vs. P4 vs. Itanium
2004/03/29 15:35:35
(permalink)
Supposedly MOTU makes a variant of the PCI424 that works in Mac G5s which only has 3.3v slots (that's why they made it). My guess is that it's a 3.3v variant of the regular 5v-only PCI424 (that is, it's still only 32-bit/33MHz), though I haven't actually seen it so I'm not certain.
|