-
Updated 2023-07-12: Hello, Guest! Welcome back, and be sure to check out this follow-up post about our outage a week or so ago.
1400c 166 slow CF?
- Thread starter MacUp72
- Start date
Phipli
Well-known member
Just a sidenote while I'm running some benchmarks - remember that the disk cache set in the memory control panel heavily influences disk benchmarks. The Norton instructions encourage you to set it to 128k before running the benchmark to make it more of a comparison of disks than RAM.
Provisionally, my pismo gets 24MB/s peak read from my SD card adapter, which is about 4 to 5 times faster than a 9600 period SCSI disk.
I'll plot the data against some stock machines. I think I have ClarisWorks installed
Provisionally, my pismo gets 24MB/s peak read from my SD card adapter, which is about 4 to 5 times faster than a 9600 period SCSI disk.
I'll plot the data against some stock machines. I think I have ClarisWorks installed
Phipli
Well-known member
Here we go... as you can see, the SD card adapter kicks the proverbial bottom in the Pismo, even vs the stock disk in the faster Pismo.
All tests except the Pismo SD Card are built in scores in Norton System Info. My test (and Norton's) was done with the disk cache set to 128KB in the Memory control pannel. SD Card tests were all run as a block and took probably between a minute and two, I have know idea what throttling occurred if any.
All tests except the Pismo SD Card are built in scores in Norton System Info. My test (and Norton's) was done with the disk cache set to 128KB in the Memory control pannel. SD Card tests were all run as a block and took probably between a minute and two, I have know idea what throttling occurred if any.
MacUp72
Well-known member
thats a big difference between the 3400 and the 1400, maybe due to higher CPU clock 240 Mhz and more L2 cache of 256k vs 128k.
Apple dollars marketing..
is that Norton benchmarking tool on the Garden? looks good ..the pismo ahead of the pack..
yeah, that was to be expected anything G3..the 1400/117 on the last.
Phipli
Well-known member
I think you're looking at the 9600/200 - it picked similar colours sorry. The 5300, 1400 and 3400 are all basically identical.
Phipli
Well-known member
Yeah, its part of Norton Utilities. Different versions are baselined against different macs, SE, Q700, 6100 etc.
You have to know to turn on the more detail, and then select a couple of results and open the detail view to see the tabulated results... oh, and it defaults to a factor vs the baseline, so you have to select "absolute" to get the results in KB/s
System Info is a great tool, but it helps to know your way around it.Late OS 9 version (good for OS 9 and G3 / G4 era - 1999) :
Older (1994) :
Norton Utilities 3.1.x - Macintosh Garden
Norton Utilities 3.1.3. DL #1: 4 x Disk Copy 4.2, 1.4MB floppy disk images, in StuffIt v 4 ".sit" archive. DL #2: Second download is 3.1.
1995 : https://macintoshgarden.org/apps/norton-utilities-321
1998 : https://macintoshgarden.org/apps/norton-utilities-353
I use version 3.1 for my accelerated SEs and the like, and 3.2 for my beige PowerMacs I think.
Last edited:
MacUp72
Well-known member
no, I was quoting 3lectr1cPPCs benchmark, 3400/1400, yellow/blue ..
thanks for the links, this looks better than MacBench 4/5 I was using..
Phipli
Well-known member
Ah my bad - interesting because the provided Norton numbers tell us that stock, the 1400 and 3400 were similar, so you might expect a similar uplift by putting the SD card adapter in the 1400.
Worth noting w/ re 1400 vs. 2400/3400:
They are radically different architectures, with the 2400/3400 being one or two generations newer, on a much newer and better overall architecture. (this is even though they are basically contemporaries.)
The 1400, remember, is a mobile Performa/PMac 6200, whereas the 2400/3400 has a PCI architecture, closer to the Mac 6400.
MHz for MHz, the 1400 will always bench significantly worse than basically everything else. My own 1600/166 is handily outbenched by 120 (and probably 100, if I'm honest..) MHz PCI PowerMacs. (new thing to try on my 8500....) (this is all also true of the PowerBook 5300 and Duo2300). These are sometimes worse than the 6200/6300, because the 6200 at least also has 256k of L2 cache, and PowerBooks 5300/2300/1400 sometimes omitted it, or had only 128k of L2, depending. L1 is usually the 6300 "fixed" 32k of unified L1 vs. 8+8 split L1, so they at least don't have that part of the problem.)
One thing I didn't see addressed: What OS/software do you have on your 1400? Although these things can technically run 8.5 or 9.1, 7.6.1 or 8.1 might be a better compromise. You still get a lot more PPC code than 7.5 but they're significantly lighter-weight.
(though: disclaimer: my 1400's on a 30-gig IDE spinner, I haven't tried replacing it for SD or CF yet.)
One more thing: Getting a "good" SD card matters. These things "should" be able to turn in decent-for-the-time numbers. 6200 outbenches 6100 at disk... See if you can pick up a UHS-1 card or a card rated for sustained writes - that's where I had the best luck in my own SCSI2SD v6. (I know that's a different deployment scenario, but still.)
In terms of 1400s being weird with some storage: I've seen this cut both ways and I suspect it has, basically, to do with luck and all the exactly specific parts. Some people have reported no CF cards except for weird specific industrial ones work, and some people have reported varying luck with IDE<>SD adapters. Today, I'd say go to SD if you can since you can get bigger, faster cards and you'll be able to source them further into the future.
To both of those points... most CF cards, especially anything smaller than liek 4-8 or so gigs, wasn't meant in any way at all for computer use. These were meant for low speed, memory buffered, and/or read-oriented use cases. (although booting is read-oriented, booting a computer and, say, playing an MP3 are still different use cases, really.)
They are radically different architectures, with the 2400/3400 being one or two generations newer, on a much newer and better overall architecture. (this is even though they are basically contemporaries.)
The 1400, remember, is a mobile Performa/PMac 6200, whereas the 2400/3400 has a PCI architecture, closer to the Mac 6400.
MHz for MHz, the 1400 will always bench significantly worse than basically everything else. My own 1600/166 is handily outbenched by 120 (and probably 100, if I'm honest..) MHz PCI PowerMacs. (new thing to try on my 8500....) (this is all also true of the PowerBook 5300 and Duo2300). These are sometimes worse than the 6200/6300, because the 6200 at least also has 256k of L2 cache, and PowerBooks 5300/2300/1400 sometimes omitted it, or had only 128k of L2, depending. L1 is usually the 6300 "fixed" 32k of unified L1 vs. 8+8 split L1, so they at least don't have that part of the problem.)
One thing I didn't see addressed: What OS/software do you have on your 1400? Although these things can technically run 8.5 or 9.1, 7.6.1 or 8.1 might be a better compromise. You still get a lot more PPC code than 7.5 but they're significantly lighter-weight.
(though: disclaimer: my 1400's on a 30-gig IDE spinner, I haven't tried replacing it for SD or CF yet.)
One more thing: Getting a "good" SD card matters. These things "should" be able to turn in decent-for-the-time numbers. 6200 outbenches 6100 at disk... See if you can pick up a UHS-1 card or a card rated for sustained writes - that's where I had the best luck in my own SCSI2SD v6. (I know that's a different deployment scenario, but still.)
In terms of 1400s being weird with some storage: I've seen this cut both ways and I suspect it has, basically, to do with luck and all the exactly specific parts. Some people have reported no CF cards except for weird specific industrial ones work, and some people have reported varying luck with IDE<>SD adapters. Today, I'd say go to SD if you can since you can get bigger, faster cards and you'll be able to source them further into the future.
To both of those points... most CF cards, especially anything smaller than liek 4-8 or so gigs, wasn't meant in any way at all for computer use. These were meant for low speed, memory buffered, and/or read-oriented use cases. (although booting is read-oriented, booting a computer and, say, playing an MP3 are still different use cases, really.)
Similar threads
