home *** CD-ROM | disk | FTP | other *** search
- Path: sparky!uunet!olivea!decwrl!csus.edu!netcom.com!jch
- From: jch@netcom.com (Jack Hwang)
- Newsgroups: comp.sys.ibm.pc.games
- Subject: Re: 3D-Bench Updated List (01/25/93)
- Message-ID: <1993Jan25.172410.10927@netcom.com>
- Date: 25 Jan 93 17:24:10 GMT
- References: <C1EE0L.77L@news.iastate.edu> <1993Jan25.160452.21703@newshost.lanl.gov>
- Organization: Netcom - Online Communication Services (408 241-9760 guest)
- Lines: 62
-
- In article <1993Jan25.160452.21703@newshost.lanl.gov> masten@beta.lanl.gov (David A. Masten) writes:
- >In article <C1EE0L.77L@news.iastate.edu> wity@iastate.edu (Senopati Pamungkas) writes:
- >>
- >>I have been receiving lots of 3dbench scores, so I decided to post the
- >>updated list twice a week. Here is the current list (01/25/93)
- >
- >(List deleted for brevity's sake)
- >
- >Thanks for volunteering to keep track of this. Maybe I won't have to
- >sift thru 100 old results to see someones new one now!
- >
- >Time for some analysis?
- >
- >First a question: The 3dBench graphics appear to be 320x200x256,
- >correct? If so they should be pretty relevant to games, but I have
- >doubts at the high end. Why?
- >
- >I've tested VLB and ISA cards in the same Gateway2000 486DX/33 and
- >see only about a 10-15% improvement with the VLB cards. When I say
- >test, I mean both 3DBench AND games that give framerate indications
- >(Falcon3, F15III, World Circuit). Note for the 33, the two methods
- >are in pretty good agreement.
- >
- >But looking at the 3DBench results for 50 and 66 mhz machines (DX2
- >or DX), we see HUGE benefits from VLB or LB. Why there and not in
- >33's? I submit that at the high framerates these machines are capable
- >of, the slow AT bus eventually becomes a bottleneck. However, it
- >appears to be the case only at fps>25 or so. But there, IT AIN'T
- >IMPORTANT, as no game will give that framerate (and if it does, it
- >really doesn't matter as 25fps is plenty smooth).
-
- It is apparent for high speed CPU, it can jam the bus more often than
- low speed ones. For 33MHz, the difference between VLB and ISA might
- not be too much.
-
- But I think this frame rate is only a benchmark. It will not be the
- real frame rate in the real games. Since the game companies are creating
- games with high details to reality, it will eventually meet the
- bottleneck on CPU power instead of grahic card speed.
-
- >I'll still get a VLB when I upgrade as it only costs a few $$ more.
- >But I don't expect it to perform 60% or so better than a non-VLB
- >for the same CPU.
- >
- >Can anyone test there DX2/66 (or 50) with both a VLB and ISA board
- >on an actual game and tell us what they find??
- >
- >Otherwise the results are fairly obvious: stick with your ET4000,
- >unless you plan on running Windoze a lot.
- >
- >Dave
- >masten@beta.lanl.gov
-
- Yes, get a VLB. I am glad to find my new PC(3 days) top at the 486/33
- ranks with 27.0 FPS. It was hard to decide what to get when I upgraded
- from the old AT. Now looks like I made a good decision.
-
- BTW, get an accelearte card anyway. Even a DOS-slow S3 card can perform
- good if it is a VLB card like the Diamond Stealth 24X. You would need
- to run Weedoow anyway.
-
- Jack
-