Greg Hess Posted April 24, 2002 Share Posted April 24, 2002 If your going dualy AMD's, look into asus motherboards. also, if you dont wanna pay through the nose for MP's look into the Duron processors... A quick note on this.... Dual XP's are no longer possible via the latest MPX core updates. Looks like you might be able to do dual XP's on the last revision of the Asus board, but as shown in Aces latest article, the board will detect the XP chip and disable the second CPU. Thats without getting into the whole instability issue of Dual XP's as seen on 2cpu.com and the support.discreet.com hardware forums. If you want stability without question, go Dual MP, if your feeling lucky, and wanna gamble, pick up some XP's. www.aceshardware.com (Workstation article, for screen caps proving the detection of the second XP procesor). Link to comment Share on other sites More sharing options...
cyclops Posted May 11, 2002 Share Posted May 11, 2002 Greg, I have a dual xp 1800+ and never had and instability problems. My board is Assus A7M266D. I looked on 2cpu.com but couldn't find any articles on dual xp instability. It sounds quite worrying anyway. I read somewhere that the only difference between xp and mp is just AMD's official support for dual operation. Any chance you could direct me to some relevant info/articles? Thanks. Link to comment Share on other sites More sharing options...
Greg Hess Posted May 12, 2002 Share Posted May 12, 2002 Hey Cyclops, If you got the Dual XP working, then your quite lucky. The original batches of MPX boards and MP/XP chips didn't have the current code in place to stop them from working. So if yours are working then you should probably be ok. I've only seen about a 5% failure on Dual XP's once their in place and working. Your not going to find alot of talk about failures on 2cpu.com, cause quite frankly, those people aren't stressing their systems. Now if you move over to the XSI, Maya, or Discreet forum's you'll see where some of the complaints are. So I suggest to check out those relevant cg sources of info first. My % numbers come from my own experience dealing with Dual XP's and personal customers/clients. As is currently, most motherboards out there will now detect the second processor and disable its function, if its an XP instead of an MP. I would be careful of flashing your bios in the future because of this. If you do, make sure to backup your old bios, in case the new bios has some detection code for the XP chips. As for differences between XP's and MP's... XP's are chips that are either not tested for SMP ability or chips that FAILED the MP tests. XP chips seem to have a slightly different L7 bridge. Though that could have just been a single batch. MP chips are processors which passed AMD's SMP tests and were verified to be stable under a variety of SMP tasks. Quite a few users have Dual XP's with zero problems, but I still firmly stand by my recommendations to pay the extra 50 dollars for a sound mind, knowing that your 72 hour render isn't going to die on the 71th hour and 56 minutes. Link to comment Share on other sites More sharing options...
Arnold Gallardo Posted May 12, 2002 Share Posted May 12, 2002 Ok I just became a member and just saw this post because of Greg Hess' post on the MAX forum. Well As for the L7.x being faster and the Athlons being slower I would definitely say that's because of the SSE 2 optimizations done by Newtek but realistically, LWers have reported that LW is only slightly faster (I recall it was around 10%) while doing raytracing and not when doing real world rendering+radiosity. You have to remeber that LW's radiosity is a sort of psuedo radiosity that's monte-carlo based so it cannot be compared to LVS (progressive refinement) or even maybe with VIZ4 (although that part is a guess since I have not used VIZ4 and could only extrapolate on the alogorithms used). LVS's core was probably written with x486 optimized compilers so the Athlon should be faster than the P4 as I would expect. Maybe we can do direct comparisons using one of the sample .ls files. I am interested in seeing the performance comparison results. I personally have been running a dual Athlon 1600+ and have been very happy with it performance with the several 3d applications that I use/work with. I used to buy Intel-based boxes before I built this first AMD box and will now only build Athlons (or Hammer) based boxes. Link to comment Share on other sites More sharing options...
cyclops Posted May 12, 2002 Share Posted May 12, 2002 Greg, thanks for your info again, it is all very useful to know. The bios update problem is one thing I had no ideea about. When I bought my xp's I thought it's just like buying mp's. The longest render I have done so far on this machine is 14 hours for a tight deadline. If something went wrong it would have been a nightmare, so I might consider changing them for mp. Good luck. Link to comment Share on other sites More sharing options...
Greg Hess Posted May 12, 2002 Share Posted May 12, 2002 Cyclops, Your system will most likely be fine. The warning is more for current users about to upgrade. Schlorby has a nice dual 1600+ XP and he's been using it for months now. He's had problems but they were due to something completely different. Just make sure to backup the old bios before updating Never know what those crazy people are going to try. Won't be long before AMD turns into Intel #2 . Arnold, thanks for the information about Lw 7, I'm sure everyone appreciates it. [ May 12, 2002, 07:14 PM: Message edited by: Greg Hess ] Link to comment Share on other sites More sharing options...
Arnold Gallardo Posted May 13, 2002 Share Posted May 13, 2002 Greg, Maybe we can include in the Athlon vs P4 benchmarks, LVS radiosity solution processing speed and if possible VIZ 4 as well? The sample Chamber.lp wil be good for this as well as the older Gallery.lp (~3.0 version). With LVS, we can separate the GUI performance with the pure radiosity processing by running the same .lp file under LSCAPE and LSRAD and then we can do the same for raytracing/rendering using LSRAY. We can even try to compare OpenGL based renders using LSRENDER. I think these will give a better perspective on the comparative performance of Athlons vs P4's. Link to comment Share on other sites More sharing options...
Greg Hess Posted May 13, 2002 Share Posted May 13, 2002 Originally posted by Arnold Gallardo: [QB]Greg, Hey Arnold, I'd love to include Viz 4 solutions in the render suite...the problem is time. I don't get paid or reimbursed for any articles and tests I write, and it currently it takes about 24 hours to benchmark just one machine. It looks like if I did all the tests that everyone wanted me to do, I would never leave my house and would just stare at my machines...blank faced waiting for some magical numbers to appear. So basically. Sure I'll run them. I just need step by step instructions, because I'm in no way shape or form a Viz or lightscape guru. (I'm not a max guru either, so don't get any ideas ) Link to comment Share on other sites More sharing options...
Administrators Jeff Mottle Posted May 13, 2002 Author Administrators Share Posted May 13, 2002 I'd be up for posting some benchmarks. I think I still have that old LS OpenGL tester somewhere and crunching LS files won't require much expertise. Just press go and record the time. Would be interesting to make up a quick "Hall of mirrors" to really kill the processors. Or maybe we can take one of the defualt scene as assign a glass material to almost everything. I know that those scenes used to be real pigs 5 years ago, but most will now process in a matter of minutes now. Let me know what you think and I'll post a scene and that benchmark and we can have an LS showdown. Link to comment Share on other sites More sharing options...
Arnold Gallardo Posted May 13, 2002 Share Posted May 13, 2002 Greg/Jeff, Well I too am busy nowadays this is why I suggested we use the sample .ls files with LVS..:)We just have to agree on what parameters to use. If I was not so busy I would be willing to specifically do a complex benchmark scene... It would be nice if we can have some LVS benchmarks to go by inlcuding the old OpenGl test. However doing a'hall of mirrors' will only test the raytracer. I was thinking of something that will be hard for the radiosity to compute well and these means occluders and curved surfaces that are near the light source (emitters) along with having many indirect illumination surfaces... Link to comment Share on other sites More sharing options...
Greg Hess Posted May 14, 2002 Share Posted May 14, 2002 Ya, all we really need is a radiosity solution of the titanium building in italy . Link to comment Share on other sites More sharing options...
Arnold Gallardo Posted May 16, 2002 Share Posted May 16, 2002 What titanium building in Italy? I am not familiar with that...You aren't thinking of Gehry's Guggenheim museum Bilbao are you? Actually I'd rather seen a Sant' Elia's unbuilt works in 3D or one of Kahn's if not FLW! Of course I'd settle for one of Michael Graves' too along with Isozaki! Link to comment Share on other sites More sharing options...
Greg Hess Posted May 18, 2002 Share Posted May 18, 2002 Arnold, Please forgive my crude mistake over the afformentioned structure. I was talking about the new Guggenheim museum in Spain. http://www.dnp.co.jp/museum/nmp/nmp_i/articles/world/bilbao/bilbao.html Link to comment Share on other sites More sharing options...
Recommended Posts
Create an account or sign in to comment
You need to be a member in order to leave a comment
Create an account
Sign up for a new account in our community. It's easy!
Register a new accountSign in
Already have an account? Sign in here.
Sign In Now