CompuServe test bed

CompuServe Thread

Autodesk Multimedia Forum · Hware/OpSys/video

#486 DX4 100 – Pentium 90

30 messages in this thread
#128607From: Smook van NiekerkOct 12, 1994 2:16 PM
More questions from a novice: Is the following statement true: "3D Studio does not (yet) fully exploit the architecture of the Pentium Chip. Accordingly, the rendering speed (for the same graphic) on a 486 DX4 100 will be better/equal (read shorter or similar) if compared to a Pentium 90" If if am setting out to purchase two rendering machines, it would make "cents" to purchase the cheaper 486 DX4 100 machine (compared to the Pentium 90 or even the Pentium 60) if I am getting the same (or slightly better) performance on the 486. I would appreciate any comment or suggestion.
#128625From: DAN BUENDIAOct 12, 1994 3:21 PM
Perhaps so. However, there is an untold future that may make use of the Pentium. If you don't mind the upgrade a few years down the road, it would make "cents" to me. You can always ad these machines to a network in the future. Just an opinion. Good luck!
#128630From: sherwood egbertOct 12, 1994 3:32 PM
i would like to know where you herd that comment<g>. the pentium 90 is world faster than the dx4 -100. 3dstudio runs just fine on the pentium 90, actualy alot faster than the dx4-100 or the pentium 60. i have all three of the above mentioned processors. sherwood
#128646From: Smook van NiekerkOct 12, 1994 5:08 PM
Hi Sherwood As a novice I need good info and do sometimes ask "stupid" questions. My source is my local dealer. Anyway, can you perhaps help by benchmarking the rendering time of the 57 chevy on the following platforms, respectively. PLEASE !! : a) 486 DX4 100 B) Pentium 60 c) Pentium 90 I assume these three machines will all have the same RAM and BUS. If not, please specify. Nevertheless, I would REALLY appreciate some help Regards Smook
#128674From: Pierre MenardOct 12, 1994 8:20 PM
Hi ! Ask Robert J. P. Smith #72640,1306 of the Montreal User Group. He have done many tests with these processors. He told me that his results will be soon published. Pierre Menard Montreal
#128855From: Smook van NiekerkOct 13, 1994 1:55 PM
Pierre Have mailed JP Smith for benchmarking info. Will inform you on the outcome Smook
#128690From: John TissavaryOct 12, 1994 8:45 PM
Lesson no. 1: Never believe what a dealer tells you. I have yet to find a computer hardware dealer that knows diddley squat about 3D Studio, except for Greg Pyros and a few forumites here (the minority, I assure you). Your dealer's probably got higher sales margins with the DX100. John Tissavary (LUNA cie, inc)
#128730From: Oct 12, 1994 10:58 PM
John: >> I have yet to find a computer hardware dealer that knows diddley squat >> about 3D Studio, except for Greg Pyros… Why, thanks, John! And that's ex-dealer, mind you! <bg> Did you ever get your Internet access going? What's your handle? Greg Pyros – gpyros@pyros.com
#128778From: John TissavaryOct 13, 1994 6:00 AM
my handle is "lunacie@hooked.net", where I do most of my e-mail these days. I'm definitely sold in internet. Hopefully I'll get a dial-up of some sort for my unix box one of these days, although I think it might be a tad pricey, still. John Tissavary (LUNA cie, inc)
#128835From: sherwood egbertOct 13, 1994 1:02 PM
hey no problem, will test them today and upload the results tommarro. and no question is stupid , just the ones not asked<g> sorry if i inplied that. sherwood
#128631From: MARTIN G FOSTEROct 12, 1994 3:47 PM
Smook, >> Is the following statement true: "3D Studio does not (yet) fully exploit the architecture of the Pentium Chip. Accordingly, the rendering speed (for the same graphic) on a 486 DX4 100 will be better/equal (read shorter or similar) if compared to a Pentium 90" << Yes to the first sentence and No to the second. Yes 3d Studio does not fully exploit the Pentium chip – it has not been compiled specifically for it. No, I'm almost certain a dx4/100 would not yield the same rendering speed as a Pentium. That's total rubbish. I would estimate that the Pentium is 2x faster than a 486 of the same clock speed. For example my Pentium 60 is 2x faster than my 486/66 in rendering the same scene. This is because even without specially compiled Pentium-specific software the Pentium's math performance is much better than the 486's. I hope others will confirm my assertions. Please don't waste your money on old technology. The price difference is not that much and the speed (now) is probably 2x times better. I'd guess this could improve to 4x or 5x faster with future Pentium specific 3ds versions. Martin
#128645From: Harley HugginsOct 12, 1994 5:02 PM
FWIW: I'm not much of a hardware person, but was told that while a DX4 100 is fast in most all situations, it is not necessarily faster than a DX2 66 in floating point calculations. I have a DX2/66 and a DX3/75, and the 75 is noticeably faster in everything except rendering. HH
#128709From: MARTIN G FOSTEROct 12, 1994 9:40 PM
Harley, >> it is not necessarily faster than a DX2 66 in floating point calculations. << well that's not the discussion here, fortunately. However, I would say in general that for rendering I've found 3d Studio can use almost every bit of extra clock speed that it can get. This might not be true for other applications where clock speed is only a small part of the equation, but it has been my experience with 3ds' intense floating point requirements.
#128729From: Kouros MoghaddamOct 12, 1994 10:56 PM
Martin, I'm getting part of my answer in this thread here. But if you can let me know about other componants such as VGA display, Hdrive, Monitor, etc… I would appreciate it
#128647From: Smook van NiekerkOct 12, 1994 5:08 PM
Martin, Thanks for well motivated answers. Some quantitative info would be greatly appreciated. Could you perhaps benchmark the rendering time of the 57 Chevy using the following machines : a) 486 dx 2 100 b) Pentium 60 c) Pentium 90 I assume that these machines will have the same RAM and Bus, etc. If not, please specify. The rendering time for each machine will greatly assist me to make the best "dollar per rendering minute" purchase decision. Thanks very much ! Smook
#128710From: MARTIN G FOSTEROct 12, 1994 9:48 PM
Smook, >> Thanks for well motivated answers. Some quantitative info would be greatly appreciated. << You are welcome. I was looking for a benchmark file in the libraries but the only one I could find was very old ('92) and didn't include Pentiums at all. >> Could you perhaps benchmark the rendering time of the 57 Chevy using the following machines : a) 486 dx 2 100 b) Pentium 60 c) Pentium 90 << I only have one of those (Pentium 60) – do you think I'm rich or something??? Just kidding, I will do the Pentium 60 and maybe someone else can do a) and c) >> I assume that these machines will have the same RAM and Bus << Bus is not too important when rendering since most of the time is spent in floating point, only a few seconds are taken for loading files typically. RAM is only important if you exceed the necessary amount for the animation. Then you build a swap file on the hard drive and your rendering times go to hell. >> The rendering time for each machine will greatly assist me to make the best "dollar per rendering minute" purchase decision. << What is the price difference (in approximate US dollars) between similarly configured systems but with the different motherboards? I would guess that it is only $100 or so here in the USA. Martin
#128822From: Steve OrmondeOct 13, 1994 11:46 AM
Martin, I believe Alan Iglesias has been keeping track of the Pentium benchmarks. – Steve Ormonde
#128845From: MARTIN G FOSTEROct 13, 1994 1:32 PM
Steve, >> I believe Alan Iglesias has been keeping track of the Pentium benchmarks. << thanks Steve! Alan are you there? Can you give Smook your benchmarks? TIA.
#128853From: Smook van NiekerkOct 13, 1994 1:54 PM
Martin, Me thought that all 3DS users are stinking rich ! (GRIN GRIN GRIN) Thanks for the offer to benchmark the Chevy on your Pentium 60. Looking forward to the result. Smook
#128689From: John TissavaryOct 12, 1994 8:45 PM
I would bet money that the DX100 is noticeably SLOWER than the P5.90, optimized or not. I _could_ be mistaken, but you should do some serious research before you go with the "cheaper" box. I've got a P5.90 and it renders typically 3 times faster than my AMI Enterprise III based 486/66's (fast machines in their own right). John Tissavary (LUNA cie, inc)
#128698From: Gus GrubbaOct 12, 1994 8:55 PM
Well, first of all, a DX4 runs at 25Mhz. It only runs at 75 internally. Thus, all data access is still performed at 25MHz. The Pentium 90 runs at, 90MHz. That's more than 3 times the speed. Second of all, the Pentium can perform most floating point instructions in half the amount of clock cycles of the 486 (at any given clock rate). That's at least twice the speed for a same clock speed and much greater when you take in consideration the 25 -> 90 MHz difference.
#128737From: jackOct 12, 1994 11:15 PM
<<from Chris Murray at 73300,2266>> Smook, I think the basic concensus is go for the p5/90. I agree with most of the comments, I have used all the above and the p5/90 is just plain the fastest. What you pay for now you will make-up in billable hours because you'll be able to do more work. If you're gonna get a 90, get a good box like a DEC lpx (?). Fast, expandable, and a 3 year warranty (at least in the US). Good luck. CM
#128854From: Smook van NiekerkOct 13, 1994 1:55 PM
Jack, Lucky beast (sitting in the USA with plenty of good p5/90 suppliers). Im in South Africa and have to turn every cent (twice). Must say I have been most pleasantly surprised with the comments. Agree that the p5/90 seems to be the best option to go. Thanks for you comments Smook
#128740From: John SchmidtOct 12, 1994 11:29 PM
Gus, >> The Pentium 90 runs at, 90MHz. According to my Intel M.B. manual, the P90 runs internally at 90mhz and externally at 60mhz. John
#128760From: Gus GrubbaOct 13, 1994 1:01 AM
You might be right. The original project was for a 90MHz (outside) chip. The 100Mhz one was the one with a double clock (50Mhz outside). The fact some people were having problems with the 90 led me to believe this was the case (remember the DX50?). If outside is 60MHz, where is the problem?
#128929From: John SchmidtOct 13, 1994 7:39 PM
Gus, It looks like they're doing the 1.5 clock rate for all the newer chips right now. My motherboard has jumper settings for 50, 60, and 66mhz, to go with the 75, 90, and 100mhz chips. Some people have reported on the Intel forum that they have been able to push their 90's to 100mhz with no problems, but mine got unstable when I tried it. 'Just didn't make the grade, I guess. <g> John
#129447From: Thomas FlanneryOct 17, 1994 3:02 PM
PMJI Gus, but no I don't remember the DX50. I do have one… Is it waiting to blow up on me or what? What was the issue that you refer to? Thanks Tom Flannery
#129463From: Gus GrubbaOct 17, 1994 4:16 PM
When they first showed up, none would work right because of being too fast. The problem was that most manufacturers used the same constructions technique they've been using for 33's and below and it didn't work. Signal leak, overheating, all kinds of problems. You most certainly have a newer model otherwise you would know for sure what these problems were. Note that this are true 50Mhz, not the 25MHz with doubled clock (you know that but some may not).
#129540From: will taitOct 17, 1994 11:35 PM
hi, just wanted to mention I run a "true" 50 EISA for a couple of years now. It is very fast and no problems. will tait
#129898From: Thomas FlanneryOct 19, 1994 12:05 PM
Gus, Thanks for the insight, and you are right my 50 is a Dell 450SE EISA machine that has worked like a champ for a couple of years now. I would like to get a new P590 but I dont want to part with my 450SE. It's a nice big tower, with cards that exit the top and it has 8 full slots which is non-existant in the current crop of Pentiums. I just don't have the room to keep all of my machines and still fit into my studio. Thanks, Tom