50MHz SE/30 Benchmarks

[BadGoldEagle]

Check this out:

(unlisted video for now, I'll upload a new better one soon)
Enjoy the Carpenters
Audio recorded through the SE/30's audio jack, so mono only: powercache.mp3
There is definitely some background hiss when it goes quiet.

But still, I'm quite impressed with the results. 

Coming up, speedometer score...

 

[BadGoldEagle]

The results are in!

I'm happy to announce that

                      the winner of the  

                             50MHz SE/30 vs stock IIfx 

                                   CPU & Math CONTEST

                           IS....

Spoiler

THE SE/30

image courtesy of James D. Wages

 

Please ignore the disk score. I inadvertently selected the floppy drive... What an imbecile!

The IIfx still wins by a country mile the Graphics test, and some of the CPU tests but overall, the SE/30 wins.

BTW, I wonder what Nubus video card they used for the IIfx benchmark... since it does not have built in video... I don't think it's the 8•24GC though.

 

[Johnnya101]

:O

 

[Themk]

Very nice! Hopefully I can one day find one of these accelerators. Hopefully I can overcome the problem imposed by the graphics card and the accelerator, so I can have both 24-bit color, and 50MHz.

 

[BadGoldEagle]

Found it: The IIfx which was tested ran 6.0.7 and had the 8 bit Apple video card (standard for the Mac II, right?)

@Themk: Yeah I sure hope you find a solution to this. Maybe your card with the later ROM can be used with the Powercache? Only time will tell.

 

[Themk]

Rom version 1.3 (which I have) is fairly late in the game, coming in at 1991, maybe there was a bug that got patched? Let's hope so!

P.S. If someone else has a RasterOps Colorboard 264/SE30, and accelerator, I can give you a ROM dump for 1.3 that you could try out.

 

[Trash80toHP_Mini]

HEH! Hold your horses there, you're touting the only two benchmarks I'd have granted you before the race started.

As a matter of course your SE/30 comes out on top for CPU performance, at 50MHz that's a given. Same for the floating point operations, the FPU sits on the 50MHz internal bus of the PowerCache card.

You need to post memory intensive, I/O intensive and overall system performance benchmarks across a range of applications to have any kind of meaningful race with the IIfx.

The main question is whether cache operations of your accelerator can make up for the 16MHz system memory bottleneck of your config.

 

[BadGoldEagle]

Heeey Trash. Sure I can run any benchmark you like.  Gimme some names and I'll run some more benchmarks next week end after my finals.

But those I/O test results will depend on the HD, won't they? And mine's not that fast, scoring a mere 1.11 at the Speedometer Disk test, while a standard SE will score 1.70 at that test... I guess the last owner went for more bytes but less speed...

So what should we do about that? I have a couple of v5.0a SCSI2SDs... Let's load an image with all the stuff we'd like to test them with. What do you think?

The IIfx will smash the SE/30 at Graphics and I/O. I already know that. But it'd be interesting to know by how much it beats it.

Oh and if you plan on using your own IIfx for the test, please remove any add ons (I'm talking about your SCSI accelerator) .

The test will be more relevant if we compare the SE/30 Powercache to a stock IIfx...

And how much memory do you have? I have 8mb. My ROM is dirty as well, so we'll have to factor that as well. I do plan on upgrading the RAM and ROM in the future, but I have other priorities for now.

NO "Dash 30fx" allowed as well. That thing is the ultimate thing of things when it comes to 68030 machines.

Edit: I never said the SE/30 was the fastest, I just said it won the CPU and FPU tests ^_^

 

[Trash80toHP_Mini]

LOL! Okay, but I don't have any of those special IIfx double-dippy fast RAM Module thingamajiggers for an 8MB config, just the 32MB available on that 40MHz system bus that a IIfx owner would probably have had at a minimum. A $10,000 Mac II didn't use II or IIx era video cards. IIRC the 8•24GC was released along with the IIfx specifically for it. If you insist, I'll take the Fast/Wide SCSI2 HackHammer out and put in the Fast SCSI2 ATTO Express card to make it a bit more fair.

I think you can see what I'm getting at here. If you tricked the IIfx out with just a SCSI2 card worth about half(?) of what your PowerCache cost in the day the IIfx would blow the front bezel of that 50MHz injected SE/30 off in a heartbeat for overall system performance. Trying to get it down to a bone stock IIfx for comparison is fun and all, but not really reflective of the real world applications for each machine and how they'd be set up to do them.

I've never done any benchmarking, I look over the specs, check for the bottlenecks and then use the force. h34r:

 

[BadGoldEagle]

Actually, the Speedometer tests for the IIfx were done with a stock unit (and standard Apple 8 bit card, so definitely a test on the cheap side). So we don't really need to compare them again.

I see where you're getting at. My SE/30 has roughly the same performance as the stock IIfx, but you can expand the stock IIfx... a lot more. That makes the IIfx the clear winner. Actually, it'd be interesting to see how much faster your actual IIfx with its upgrades is compared to the maxxed out SE/30 and the stock IIfx.

Today some people use video rendering times as benchmarks, we could try to export a video using Adobe Premiere 3 or 4?

Perhaps we could use that mp3 encoding software MpegDec too? It has a file conversion feature... It took quite some time to convert the .mp3 to a .aiff. Could be fun.

Daystar had their own software to test the Powercache's performance. One of the few tests included some image unscrambler thing, maybe Photoshop has a similar feature? 

 

[eharmon]

Rom version 1.3 (which I have) is fairly late in the game, coming in at 1991, maybe there was a bug that got patched? Let's hope so!

P.S. If someone else has a RasterOps Colorboard 264/SE30, and accelerator, I can give you a ROM dump for 1.3 that you could try out.

Anyone else happen to have 1.3?