bench test of Performa 6214CD internal IDE hard disk

[bigmessowires]

It appears in all likelihood that you have a 6300 board, my friend. I get the same <100MHz rating.

Thank you! Although that doesn't quite seem consistent with those MacBench scores from Reddit, you must be right. Meanwhile I also inspected my board under magnification, peeking through gaps in the heat sink, and it did not appear that the CPU had been reworked. So it must be a 6300 motherboard swap after all.

But this raises a new question. Both FWB HDT and Apple's System Profiler report a bus speed of 37.5 MHz and a CPU speed of about 94 MHz. Is that some kind of reporting bug, or is a 6300/100 really only 94 MHz and Apple just fudged the numbers? Short of sticking a scope on some pins, I'm not sure how we could really tell for sure.

I’m wondering if your Performa 6214 was involved in the Apple recall of certain Performa 62xx machines.

That's very interesting... could be! This computer also has a power supply with a 1999 sticker on it, which is too new for a 6214CD.

 

[NJRoadfan]

Reading the tech notes, the 6200 still uses the F108 for IDE like the 630 series did. Don't expect great performance from it. Apple claims PIO Mode 3, which is 11.1MB/sec... theoretical. Apple also oddly claims the drives need to support LBA, which wasn't a given when these machines were new!

 

[bigmessowires]

Yeah I'm more just curious about the drive throughput, but not expecting anything much. I actually have four different options to test: original 1.2GB 3600 RPM Conner IDE, 40GB 7200 RPM Seagate IDE, IDE to SATA adapter with a 750 GB SSD, IDE to SD card adapter with a class 10 SD card. I'll test most of these same options in my G4 Quicksilver too. Stay tuned!

But this raises a new question. Both FWB HDT and Apple's System Profiler report a bus speed of 37.5 MHz and a CPU speed of about 94 MHz. Is that some kind of reporting bug, or is a 6300/100 really only 94 MHz and Apple just fudged the numbers?

Researching this a little further, I think it's likely a reporting bug. All of the 62xx and 63xx models share the same Gestalt ID of 42, so it's possible that the OS actually has no way of determining the true bus speed and just uses a lookup table based on Gestalt ID. But some Macs with that ID will have a 37.5 MHz bus and others will have 40 MHz. If the OS reports 37.5 MHz for all of them, then it could explain the data we're seeing.

EDIT: Found another thread that says Mac Bench will attempt to actually measure the system bus speed, to an accuracy of 1-2 MHz.

 

[Phipli]

don't know what the -A and -B suffixes mean

A and B are the first and second production releases. Just board revisions. Usually this means the boards are effectively the same, but they made a tweak to the design.

 

[bigmessowires]

Gauge PRO reports 40.0 MHz bus speed, 100.1 MHz CPU speed. So it looks like the info from Apple System Profiler is just wrong, I have a normal 6300 motherboard in a 6214CD case, with 100 MHz 603e on a 40 MHz bus, and this was a lot of excitement over nothing. Oh well.

 

[Phipli]

Gauge PRO reports 40.0 MHz bus speed, 100.1 MHz CPU speed. So it looks like the info from Apple System Profiler is just wrong, I have a normal 6300 motherboard in a 6214CD case, with 100 MHz 603e on a 40 MHz bus, and this was a lot of excitement over nothing. Oh well.

This is quite a common thing. It was known to be unreliable at reporting speeds. It reports my G3 Upgraded 300MHz mac as a 75MHz machine, and some other upgraded machines with hundreds of MHz as things like "7MHz".

Its mostly right for stock computers though, but Apple... were a little prone to QC issues like this in the mid 90s.

 

[bigmessowires]

I'm running into a snag while trying to benchmark test other drives in this 6300. I removed the original IDE hard disk and swapped it for the 7200 RPM IDE drive from my G4 Quicksilver. It contains OSX 10.4.11 which won't run on this machine, but I'm not booting from it - I'm booting from an external disk with OS8.1. My hope was just to run some FWB HDT disk performance tests, but whenever this IDE disk is connected OS8.1 will crash part-way through booting.

My guess is that the IDE driver contained on the disk contains G3- or G4-specific code, and when OS8.1 tries to load and execute the driver, it crashes. Is there any way around this? It's kind of a catch-22 if I can't even boot the computer with this drive attached, so I can never get an opportunity to reformat it or install a different driver.

 

[bigmessowires]

Here's another try with a different disk: a cheap IDE-to-SD card adapter with a 16 GB class 10 SD card. FWB HDT 4.5.2 had some trouble with this. It detected the drive, but showed its capacity as only 7.87 GB. When I tried to format it, HDT complained that the driver needed to be updated. When I tried to update the driver, it complained that it couldn't be updated. The Apple standard Drive Setup utility from the OS 8.1 installer CD worked better. It detected the drive as 16 GB and had no trouble formatting it (14.9 GB formatted capacity).

I then went back to HDT, which still showed the drive as 7.87 GB but containing a 14.92 GB volume. Then I ran the disk I/O benchmarks. Here are the results, with comparison to the original drive:

Conner 1.2 GB 3600 rpm IDE: sustained read: 1789 KB/sec, sustained write: 2179 KB/sec, average access time 24.1 ms
IDE-to-SD with class 10 card: sustained read: 1720 KB/sec, sustained write: 1528 KB/sec, average access time 1.2 ms

The access time is much better, but throughput is actually worse than the original hard disk!

One of the product reviews from a PC user mentioned he was seeing around 25MB/s reads and 20MB/s writes with this adapter, which is like 10x better than I measured. So the IDE controller in this computer must truly be a dog. For comparison, I got sustained read speeds of 3682 KB/s on this computer with an external SCSI disk. But I'd really prefer to use the internal IDE. I still have a couple of other drive options to try, but this may be as fast as it's going to get.

 

[NJRoadfan]

You are lucky that Sintechi adapter worked. I'm 1 for 2 on buying them at the moment! They are artificially capped in speed, but that shouldn't matter on a machine with a slow PIO interface. The 7.87GB volume error is no surprise. That would be the "should be PC BIOS only" 8.4GB Int 13h addressing limit.

 

[bigmessowires]

The 7.87GB volume error is no surprise. That would be the "should be PC BIOS only" 8.4GB Int 13h addressing limit.

Would that mean Apple Drive Setup is communicating with the drive using some different method than HDT?

What kind of failure did you have when you tried this adapter? It just wasn't recognized, or something else?

I discovered something interesting - despite HDT's benchmarks showing this adapter was no faster than the original HD for reads and slower for writes, the startup time was reduced substantially. For a basic install of OS8.1, I was getting 1:20 for the original HD and this adapter starts up in 0:45.

But now I've completely broken something... I went back to HDT and tried updating the driver again, because I'd read that HDT's IDE driver was higher performance than Apple's. The update failed, and now if I attempt to boot the computer from the SD card I get an immediate chimes of death. Even if my SD card were corrupted, how could that cause chimes of death? I can get the computer to boot from another drive if I remove the SD card, and then insert it after booting. But when I do this both HDT and Drive Setup show the SD card as only having 32 MB capacity and Drive Setup fails when I try to re-initialize it. I'm not sure what I've broken, but it's not good.

 

[LaPorta]

A few things I can think of. One, that OS X disk may have GUID formatting, and not Apple Partition Map. That will definitely foul things up. I’d use disk utility on the G3/4 it came from and see what scheme it is. If you need to, reformat using Disk Utility under X as Apple Partition Map, and make sure to have OS 9 Disk Drivers checkbox selected.

As far as the SD adapter, I’d have to test mine in my 1400, but it feels pretty speedy. 1400 is pretty contemporary to the 6300.

 

[LaPorta]

If you can, reformat the card in an SD adapter in your G4 and try again.

 

[NJRoadfan]

But now I've completely broken something... I went back to HDT and tried updating the driver again, because I'd read that HDT's IDE driver was higher performance than Apple's. The update failed, and now if I attempt to boot the computer from the SD card I get an immediate chimes of death. Even if my SD card were corrupted, how could that cause chimes of death? I can get the computer to boot from another drive if I remove the SD card, and then insert it after booting. But when I do this both HDT and Drive Setup show the SD card as only having 32 MB capacity and Drive Setup fails when I try to re-initialize it. I'm not sure what I've broken, but it's not good.

This is the EXACT failure I had with the KOOBOOK branded adapter. It would intermittently see the drive correctly, but other times just show up as 32MB. No matter what I did (swapped IDE interfaces, SD cards, etc.), this would still happen. At least you were able to actually write something to the cards. I couldn't even get FDISK to partition a SD card.

My saga is here: https://www.vogons.org/viewtopic.php?f=46&t=96953

In my case I was using a 16GB SD card in a PC that could address drives up to 128GB (the 28bit LBA limit). I bought another adapter (GINTOOYUN is the brand if it matters) and it works fine.

 

[bigmessowires]

This is the EXACT failure I had with the KOOBOOK branded adapter.

Then lucky for you, I managed to salvage it! Maybe this method will also work for you, but it's Mac specific.

I used a USB SD card reader to mount the card in my MacBook Air (Ventura 13.4). The HFS+ OS8.1 partition was immediately mounted, which was pretty neat. That told me the problem was probably only a screwed-up driver partition on the SD card. I used Disk Utility to make a disk image of the OS8.1 partition, hoping this would save me from having to re-run the OS8.1 installer later. Then I reformatted the whole SD card as DOS MBR with FAT32 filesystem, intentionally choosing something that I hoped my 6300 would not recognize and try to mount.

Returning the SD card to my 6300, I no longer got chimes of death at startup. I re-ran Apple Disk Setup and it correctly detected the drive as 16 GB and formatted it with Apple File Partition. I then mounted the SD card on my MacBook Air a second time, as well as mounting the disk image I'd made earlier, and did a drag-and-drop copy of all the files from the image to the empty 16 GB partition.

Back on the 6300 again, this all worked as I'd hoped, except I had to drag the System file in and out of the System Folder to bless it.

This is super cool, I hadn't realized I could mount an old-school HFS+ parition from an SD card under the current version of OSX. That means it'll be very easy to get new software onto this 6300: just download it from the Garden, copy it to the OS8.1 partition on the SD card, and then stick the SD card back into the 6300.

It's actually easier than doing the comparable thing with a SCSI SD card emulator like Zulu SCSI or Blue SCSI, since as far as I know OSX and/or Disk Utility won't open the .hda files they use. You can still run them in a software emulator like Basilisk, but for simply moving files about that's not especially convenient.

As far as the SD adapter, I’d have to test mine in my 1400, but it feels pretty speedy. 1400 is pretty contemporary to the 6300.

Yes! I'm still getting 45 second boot times, which is a major improvement over the hard disk, even though HDT said the SD card was the same or worse performance as the hard disk. Maybe "sustained read speed" isn't the number I should be focusing on, and it's more about the access time? Maybe there isn't much performance difference between the two when reading a large file from contiguous sectors, but the SD card outperforms when doing lots of random reads with small transfer sizes? I'll need to play with the benchmark settings in HDT to see if I can find settings that are more representative of what happens during MacOS startup and typical Finder operations.

 

[NJRoadfan]

I suppose I could try that on say... an Amiga 4000. I don't have hope for this thing though. I suspect there is a different revision chip or ROM between the two I have. This thing failed miserably at working in a bog standard PC, even after blasting out partition tables and formatting.

 

[bigmessowires]

Running under OS 8.1, I'm now left with .fseventsd and .Spotlight-V100 folders on my disk. If I try to delete these it says "You cannot move .Spotlight-V100 to Trash, because you do not have the privilege to make changes."

EDIT: Oops, I get this same error if I try to modify ANY file on the disk. So it seems that all the owner/permissions info is screwed up. I may need to re-run the OS8.1 installer from scratch after all.

 

[Phipli]

Running under OS 8.1, I'm now left with .fseventsd and .Spotlight-V100 folders on my disk. If I try to delete these it says "You cannot move .Spotlight-V100 to Trash, because you do not have the privilege to make changes."

EDIT: Oops, I get this same error if I try to modify ANY file on the disk. So it seems that all the owner/permissions info is screwed up. I may need to re-run the OS8.1 installer from scratch after all.

Check with HDT if the drive is write protected.

 

[bigmessowires]

It doesn't appear to be locked, I think the owner on all the files is set incorrectly? I can just re-run the OS8.1 installer.

There are some practical problems with using this IDE to SD adapter in the 6300 though. It has the power and data cable connectors mounted vertically instead of using right-angle connectors like a real drive. With the extra vertical space needed for the cables, it barely fits in the bare hard drive bay but doesn't fit when mounted on the drive sled. So there's no good way to mount it except with something like velcro or tape.

The 6300 also expects the rear of the hard drive (with the power and data connectors) to be at the front of the computer. The cables are very short. When the adapter is mounted this way, the SD card ends up at the far end buried deep inside the computer. It's not accessible except by removing the floppy drive, unplugging the power and data cables, and sliding the whole adapter out of the computer. It's fine if you plan to install it and forget it, but if you want to use the SD card to transfer files like I'd discussed, it's impractical. A 12-inch IDE ribbon cable extension and 12-inch Molex power connector extension would probably solve this.

 

[Phipli]

I think the owner on all the files is set incorrectly?

But 8.1 doesn't have that type of file permissions? Only for network shares. The user can access all files, unless the disk or folder is locked.

But yes, a reformat is a good idea.

 

[LaPorta]

This is in my 1400, but you can imagine that space is at a premium. This adapter fits just perfectly, and a little foam tape helps secure it. You don't even need a power connector for it (will need an adapter to full size though, I suppose):

PowerBook 1400 Restoration: Redux

Earlier this year, I put a lot of time and effort into the recap and restoration of my 1400. Documenting it was part of the fun, for others to be able to learn from steps that I took and things that I found out. Well, the great crash wiped out my entire thread, and there appear to be no...

68kmla.org