|
|
|
|
| Welcome, Guest | Home | Search | Login | Register | |
| Author | MiniBench Scores (Read 101323 times) | ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
Lixivial
2 MB  Posts: 3 
|
Reply #15 on: October 06, 2006, 01:18
I figured I'd try a bit of an intriguing test. The setup is this: Mac OS X 10.4.8, running SheepShaver as an emulated Power Macintosh 9500 with a 100MHz G4 -- yes, I know that such a thing does not exist, but... I ran a total of 15 tests and they average out to be approximately 136.5, with extensions on. The app would not load in SheepShaver without extensions on. I then ran a test with this setup: Mac OS X 10.4.8 running BasiliskII emulating a Quadra 900 with a 33MHz 68040 and 64 megabytes of RAM running System 7.5.3r2. Here I get -2.147484x10^8 I'm going to see if I can run this in Blue Box on Rhapsody -- I realize it does not work in Classic, but I'm curious if it works in Blue Box. I'm also going to see if it runs under A/UX -- I know that's a bit below the 7.5 requirement, because A/UX uses 7.0.1 to boot to it, but... I know these aren't pure scores and aren't really relevant for anything other than checking how well a program emulates the hardware. I just thought they'd be interesting until I can become motivated enough to plug my old Macs in. |
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
RacerX
|
32 MB Posts: 36
Reply #16 on: October 06, 2006, 04:03
|
Quote from: "Lixivial" I'm going to see if I can run this in Blue Box on Rhapsody -- I realize it does not work in Classic, but I'm curious if it works in Blue Box. Quote from: "RacerX" 79.7 I'm guessing you haven't read the other posts in the thread yet.
|
Lixivial
|
2 MB Posts: 3
Reply #17 on: October 06, 2006, 08:30
|
Indeed. I reread the thread after posting that message and noticed that you, as I should have guessed, mentioned Blue Box. In fact, I also noticed, too, that you had thrown out those times due to the performance hit. That makes a majority of my response irrelevant, as of course BasiliskII and SheepShaver would have performance hits as well. Yeah, so I was going to edit my post to accommodate that, but... 
|
dpaanlka
|
1024 MB Posts: 1646
Reply #18 on: October 06, 2006, 10:41
|
Quote from: "Lixivial" BasiliskII and SheepShaver would have performance hits as well. The emulator does not literally emulate the speed of a 33mhz Quadra - it just pretends to be a Quadra so that software of that vintage will run on it. It still runs it as fast as it possibly can on your machine. It's weird that you would get a massively negative number on the Quadra emulator, though. That isn't even close to being right... If you were running Basilisk on say a 1ghz Power Mac G4, I would expect the Quadra score to be around 150 or so... not -214748400.
|
Deer Steak
|
16 MB Posts: 25
Reply #19 on: October 06, 2006, 14:26
|
Quote from: "RacerX" 154.2 G4e takes an incredible hit on this, then: 176 System: iMac G4 Processor: G4e 800MHz, 256k L2 Memory: 512MB RAM Video: GeForce 4MX 32MB VRAM Operating System: MacOS 9.2.2 I'm thinking the seven-stage pipeline of the G4e vs. the 4-stage of the G3 and original G4 must have something to do with it. I'd bet the G3e (750cx and 750fx) with their four-stage pipeline is actually considerably faster than a G4e when not using AltiVec.
|
dpaanlka
|
1024 MB Posts: 1646
Reply #20 on: October 06, 2006, 18:55
|
Quote from: "Deer Steak" G4e takes an incredible hit on this, then: Wow that is incredible. Something has to be wrong here... was this machine booted into 9.2.2 directly or was this running under OS X's classic mode?
|
madmax_2069
|
16 MB Posts: 28
Reply #21 on: October 06, 2006, 19:40
|
that is weird. i can see that the altvec of the G4 cpu not helping any cause the program isn't designed to take advantage of it. but WOW you would figure a 800mhz cpu would at least score in the 200+ score if my G3 466mhz cpu in my Beige G3 got a 160.5. hummm maybe you should try making another program but try a different math method or scoring method. who knows but that G4 800mhz score is weird to say the least
|
Deer Steak
|
16 MB Posts: 25
Reply #22 on: October 06, 2006, 21:01
|
Quote from: "dpaanlka" Wow that is incredible. Something has to be wrong here... was this machine booted into 9.2.2 directly or was this running under OS X's classic mode? I tried both, with OS X being very similar in score. I still think it's the G4e's pipeline - remember, because the pipeline was extended, the G4e has a relationship to the original G4 that Prescott has to Northwood in the Pentium4 world - it's going to be slower per clock in order to get higher clock speeds. A 1.25GHz eMac running Tiger (in Classic - my wife's machine...it's X only, so no 9.2.2 install) gets 276, so over 100 points more, which again is in line - 55% increase in speed, 56.x% increase in score. It has double the L2 cache and it has a 67% faster memory bus (167 vs 100) and that kind of huge increase meant squat to the final score. To me the difference is within the margin of error; I ran it twice and took the higher score. That also makes me think that as long as your cache is at least 256k (which the 800 is), then the size of the cache isn't terribly important. Double the L2 cache didn't help the eMac; the score increased in a linear fashion. I have also observed something similar on a Digital Audio G4 tower with a 7447a CPU upgrade - the original dual 533MHz 7410 with 1MB of 2.5:1 L2 was not that much slower than a dual 1.6GHz 7447a when running SETI@Home. In terms of PPD, the 533 ran around 340ppd, my wife's eMac gets around 185 (single CPU, so dual would be around 370), and the 1.6GHz topped up around 400. I was flabbergasted, and expected the increase to be from somewhere around 340 to at least 600, which isn't even double...but to get a measley 20% was equally astounding.
|
Deer Steak
|
16 MB Posts: 25
Reply #23 on: October 06, 2006, 21:02
|
End of story: G4e is a dog. We all knew that though. 
|
Lixivial
|
2 MB Posts: 3
Reply #24 on: October 06, 2006, 23:24
|
Quote from: "dpaanlka" The emulator does not literally emulate the speed of a 33mhz Quadra - it just pretends to be a Quadra so that software of that vintage will run on it. It still runs it as fast as it possibly can on your machine. Indeed, fair enough, but through the act of emulation there are inherent performance hits over native hardware. The test, the way I see it, wasn't so much a test of the hardware as it was a test of BasiliskII/SheepShaver's ability to emulate CPU instructions, etc. This was running on a 2 GHz MacBook running the UB version of Basilisk II. I was also running the UB version of SheepShaver. I hadn't tested on native PowerPC OS X, though I'm interested to see what turns out. I'm also curious how the Windows version of Basilisk II would stack up. The test using BasiliskII takes a LONG time to complete, and I'm writing it off as the emulated hardware layer or BasiliskII's code translation. It doesn't handle CPU intensive tasks very well at all, as you'd might imagine. Anyroad, my tests are not really useful as they're not really testing the physical hardware of a given machine.
|
dpaanlka
|
1024 MB Posts: 1646
Reply #25 on: October 07, 2006, 04:20
|
I drove all the way back to Chicago to run this test on my 8600 which has a 400mhz G4 processor and I got 138.9 each time. So the question is, what would we expect all of these machines to score?
|
dpaanlka
|
1024 MB Posts: 1646
Reply #26 on: October 07, 2006, 04:49
|
Well I finally got around to running the test on an actual 68k computer, and WOW it took so long that five minutes later the first step of the progress bar wasn't even complete... So I think I'm going to go back to the drawing board for my benchmark aspirations...
|
Deer Steak
|
16 MB Posts: 25
Reply #27 on: October 07, 2006, 14:07
|
did the 68k have any cache?
|
Deer Steak
|
16 MB Posts: 25
Reply #28 on: October 07, 2006, 16:23
|
Beige G3 333 with 192MB RAM running OS 8.1 scored 113 Started it on my LC III and went and played about 3 sets in Guitar Hero, came back and it's about 60% of the way through. I'm going to let it finish...eventually.
|
dpaanlka
|
1024 MB Posts: 1646
Reply #29 on: October 07, 2006, 18:36
|
My PowerBook 540c scored a 0.6 Interesting...
|
|
Pages: 1 [2] 3 4
|
| |||||||||||||||
|
© 2021 System7Today.com. |
