CF AztecMonsters have landed

[Udo.Keller]

Perhaps it would help to try another benchmark utility too.

Absolutely. I took QuickBench, available as part of Intech HD SpeedTools, and did benchmark one of my CF cards w/ AztecMonster again:

FWB HDT driver

Image 6.tif (8.04 KiB) Viewed 1139 times

For reference, here is the FWB benchmark result, taken 4 weeks ago:
The results for both sustained read and sustained write are more or less the same. Random read and random write are comparable, too. Please note that the FWB diagram has a linear x-axis, whereas the QuickBench diagram has a logarithmic one.

But guess what? Following JDW's thought I just bought the Intech HD SpeedTools and re-initialized the CF card with HD SpeedTools, thus deleting the FWB driver with the HDST driver. Look at that:

Intech HDST driver

Image 5.tif (8.21 KiB) Viewed 1139 times

Lessons learned so far:
• Benchmarks made with FWB HDT and with Intech QuickBench show similar results. So presumably it's not the benchmark tool itself introducing the problem.
• CF card shows dramatically better read performance when formatted with Intech HDST. Regarding write performance, both FWB HDT and Intech HDST deliver the same, disappointing results.


For comparison, here are the QuickBench results of a Quantum Fireball 1280S real spinning platter hard disk.

Quantum Fireball 1280S

Image 7.tif (8.94 KiB) Viewed 1139 times

[ojfd]

Guys,
I still do not understand why you continiosly refuse to use ATTO for benchmarking and waste time trying out other sulutions
__insert picture of me banging head agaist the wall here__

Did you read what was said about ATTO tool at xlr8yourmac?

The ATTO Tools utility has a benchmarking function, which does not use the system disk cache (although it has an option to do so). Bypassing the system disk cache gives a better indication of the card and drive performance, eliminating the effect of a large cache setting in the memory control panel - which inflates scores in other benchmarks.

Please, use it and post the results. Only then, by excluding system disk cache from test, we can talk about drive's performance issues.

-----------
P.S. Udo, your tif images do not load on my machine. Any idea why?

[Udo.Keller]

I still do not understand why you continiosly refuse to use ATTO for benchmarking and waste time trying out other sulutions

Simple reason: The ATTO tools don't work on my SE/30 with System 7.1. According to their README file the require System 7.5.2 or later, and although they come as fat binaries, they crash on System 7.1.

Image 8.tif (9.45 KiB) Viewed 1131 times

[Udo.Keller]

The TIFF attachments seem to make problems for some comrades, sorry. From my side they looked O.K. Here comes a 2nd approach, this time with PNG attachments.

So we go again.

Perhaps it would help to try another benchmark utility too.

Absolutely. I took QuickBench, available as part of Intech HD SpeedTools, and did benchmark one of my CF cards w/ AztecMonster again:

FWB HDT driver

Image6.png (4.82 KiB) Viewed 1126 times

For reference, here is the FWB benchmark result, taken 4 weeks ago:
The results for both sustained read and sustained write are more or less the same. Random read and random write are comparable, too. Please note that the FWB diagram has a linear x-axis, whereas the QuickBench diagram has a logarithmic one.

But guess what? Following JDW's thought I just bought the Intech HD SpeedTools and re-initialized the CF card with HD SpeedTools, thus deleting the FWB driver with the HDST driver. Look at that:

Intech HDST driver

Image5.png (5.15 KiB) Viewed 1126 times

Lessons learned so far:
• Benchmarks made with FWB HDT and with Intech QuickBench show similar results. So presumably it's not the benchmark tool itself introducing the problem.
• CF card shows dramatically better read performance when formatted with Intech HDST. Regarding write performance, both FWB HDT and Intech HDST deliver the same, disappointing results.


For comparison, here are the QuickBench results of a Quantum Fireball 1280S real spinning platter hard disk.

Quantum Fireball 1280S

Image7.png (5.54 KiB) Viewed 1126 times

[ojfd]

Simple reason: The ATTO tools don't work on my SE/30 with System 7.1. According to their README file the require System 7.5.2 or later, and although they come as fat binaries, they crash on System 7

Oops!
7.5.2 , I see...(But that blurb about PCI machines in Read Me file is incorrect - it works on my Q605.)
I've just sent you an older ATTO test utility. Let's hope, that it works on your SE/30 and also bypasses system disk cache. Try it.

[Udo.Keller]

Let's hope, that it works on your SE/30 and also bypasses system disk cache.

QuickBench claims that it bypasses the system disk cache as well.

FWIW, here's the Kingston SSD I benchmarked last month. But this time, it has been initialized with HDST instead of HDT:

Kingston SSD, Acard ARS-2000SUP, Intech HDST

Image10.png (4.76 KiB) Viewed 1113 times

And here, for reference, is the FWB benchmark for the same SSD, taken 4 weeks ago:

Again, the same findings as above:
• HDST enhances the read performance to 1.8 MB/sec.
• Write performance is still limited to 407 KB/sec.

[JDW]

Since this thread is focused on flash drive solutions for our vintage Macs, I felt this would be appropriate.

Last night I commented on OWC's blog about their Black Friday 2011 specials. I noticed that they had specials on all their SATA SSDs but none on their "Legacy" IDE/ATA edition SSDs that are for vintage Macs. Having recently acquired some old PB G3 Wallstreets, I also commented about discounts on batteries too. Today, Grant at OWC kindly replied that he would take my words under consideration.

So as to not let this chance die (for truly, I am only one man asking about this), I wanted to post here to encourage those of you interested to reply in kind. Specifically, if you've ever pondered putting a true SSD, not just a CF card, in your vintage Mac, now is the time to seek out a discount and do it! I would therefore strongly encourage you to post a comment on OWC's blog if you would be encouraged to buy an OWC SSD if they offered a good enough discount on them.

Here's the Black Friday Blog:
http://blog.macsales.com/12693-holiday- ... ment-49219

And here are the 3 product pages I mentioned in my comment on the OWC blog...

40GB Legacy 2.5" SSD:
http://eshop.macsales.com/item/OWC/SSDMLP040/

PB G3 Wallstreet PRAM Battery:
http://eshop.macsales.com/item/NewerTech/PRAMPBG3WS/

PB G3 Wallstreet Main Battery:
http://eshop.macsales.com/item/NewerTech/BAPLLIWSRS/

And while pondering whether or not to post a comment there, keep in mind, "Ask and ye shall receive!"

Thanks!

[Udo.Keller]

The CF AztecMonster is built around an Acard SCSI-IDE bridge, the same one that is used in the AEC-7720U. So I took an AEC-7720U, a Toshiba MK3021GAS 2.5" HDD and my G3 Powermac running MacOS 9.2 and got the following results:

AEC-7720U, real HDD

Picture 4.png (11.3 KiB) Viewed 1172 times

And here's the AztecMonster with a CF card, attached to the same Powermac as above:

AztecMonster, CF card

Picture 3.png (11.38 KiB) Viewed 1172 times

Finally, a real SCSI HDD (Quantum Fireball 1280S) not needing the Acard SCSI-IDE bridge chip:

SCSI HDD

Picture 5.png (11.51 KiB) Viewed 1172 times

Powermac G3 Findings:
• In combination with an Acard SCSI-IDE bridge, a real HDD and a CF card show similar performance. Not the same, but similar.
• In both cases, the write performance reaches the same level as read performance. But it needs large transfer sizes to do so.
• Without the Acard SCSI-IDE bridge, a normal SCSI disk behaves more or less comparable to the CF card.

[MacJunky]

Just for fun, can you toss an IDE-CF adapter onto the AEC-7720U? (nothing with any fancy chips, just nice and passive)
It will likely not give better results, but just for the sake of curiosity and to deal with any questions about how the AztecMonster itself actually does stack against a normal SCSI-IDE adapter. With a passive CF-IDE dealy in the 7720U it *should* have the same results as the AztecMonster; since tests are going on atm if you have the stuff it might be neat.

[Trash80toHP_Mini]

< Bump > Neat stuff, I'd like to know if anything new has come up?
Are these adapters proving useful in real world applications?

[Udo.Keller]

Just for fun, can you toss an IDE-CF adapter onto the AEC-7720U? (nothing with any fancy chips, just nice and passive)
It will likely not give better results, but just for the sake of curiosity and to deal with any questions about how the AztecMonster itself actually does stack against a normal SCSI-IDE adapter. With a passive CF-IDE dealy in the 7720U it *should* have the same results as the AztecMonster; since tests are going on atm if you have the stuff it might be neat.

Did that. And since I was flashing the AEC-7720U to the Mac-specific firmware version 3.7m anyway, I took the AztecMonster, held it against the Acard firmware flashing tool and guess what? The firmware flashing tool identified the AztecMonster firmware as Acard version 3.7m. So from a firmware point of view, the AztecMonster is a 7720U running firmware version v3.7m.

Here are the measurements, done with my SE/30. IMHO, they are indistinguishable.

Left: AEC-7720U w/ passive IDE/CF adaptor; right: AztecMonster

[Udo.Keller]

Here's another comparison: AEC-7720U with CF card versus spinning platter HD (Toshiba MK3021GAS). Both the CF card and the HD have been initialized with Intech HD SpeedTools v3.6.

Left: AEC-7720U w/ passive IDE/CF adaptor; right: AEC-7720U w/ 2.5" IDE HD

The HD, hooked behind the AEC-7720U, does not perform better than a CF card. For me, the performance bottleneck seems to be the SCSI-IDE bridge. What do you think?

[ojfd]

For me, the performance bottleneck seems to be the SCSI-IDE
bridge. What do you think?

Udo.Keller,
You tested AEC-7720U vs. regular HD in SE/30 and in your G3 so far. How about repeating the same test in one of your Quadras? Q605 will do. I'd love to see how 53C96 performs under same conditions.

[JAG]

I had several AztecMonsters on order with Artmix and he never sent them, so I got a refund.

Looking at the board design, it looks like a pretty simple layout (not that I'm an EE by any means)

Also, the chip used is a standard ACARD ARC-760B, which if it could be sourced relatively cheaply, we (i.e. the retro Mac community) could probably come up with a design for it and have it fabbed over here, though who knows if it would be at a reasonable price.

I'd be interested in looking into it if there were any enthusiasm! The scarcity of drives for these old machines will only get worse as time goes on.

[JDW]

IMHO, they are indistinguishable.

As are those bars in your benchmark chart.

Once again, folks, please use something other than the bar or line graph on the SE/30. PLEASE! I like the bar and line graphics, but I like understanding what each bar and line mean even more. At the very least, use Photoshop to mark which is which.

Thanks!

[tt]

Was there a conclusion about whether the speed rating matters for the AztecMonster? I seem to remember someone showing it does not really matter, however I cannot find the discussion since there are multiple threads going on here about CF cards.

BTW, Manabu has more AztecMonsters in stock.

[mcdermd]

There has been AztecMonster discussion and testing going on at the NeXT Computer forum too.

[johnklos]

In my Acard ARS-2000SU with a CF to SATA adapter, I noticed that certain CF cards are much slower than others. Specifically, a 32 gig card will run at whatever speed the Quadra 605 it's in can run, but another card (a 64 gig SD card in an SD to CF adapter) works perfectly but much, much slower.

Some cards can only do very limited kinds of transfers, such as this 2 gig card in one of my Amigas:

wd0 at atabus0 drive 0: <CF500>
wd0: drive supports 1-sector PIO transfers, LBA addressing
wd0: 1959 MB, 3982 cyl, 16 head, 63 sec, 512 bytes/sect x 4013856 sectors

I don't care about the speed since it's only in there to boot the machine, but if I were using it as my primary disk I'd not want to use a 1-sector-at-a-time card.

As a separate issue, the Acard chip is definitely not a bottleneck. On other m68k machines I get 3 MB/sec write, 5 MB/sec read via an Acard and a 500 gig hardware mirrored SATA setup, and on a PowerMac 9600 I get 22.5 MB/sec write and 46 MB/sec read on two hardware mirrored 2 TB SATA drives, also via an Acard. They're quite fast.

[tt]

Some cards can only do very limited kinds of transfers, such as this 2 gig card in one of my Amigas:

wd0 at atabus0 drive 0: <CF500>
wd0: drive supports 1-sector PIO transfers, LBA addressing
wd0: 1959 MB, 3982 cyl, 16 head, 63 sec, 512 bytes/sect x 4013856 sectors

Is there a way to get this info from OS 7 on a AztecMonster or OS 10.6 on a USB cardreader?

[johnklos]

Only if you connect a CompactFlash adapter directly to an IDE bus. Adapting through a SCSI adapter or a USB adapter abstracts things so that this information isn't available.

[tt]

How about Linux? What sort of command would reveal the transfer type? I am going to do an Ubuntu install and I have an IDE to CF adapter I could use.

[LOOM]

Is it possible to use the CF AztecMonster inside a Apple HD-40 SC and be able to boot a Macintosh Plus like normal?

[tt]

Yes, the Apple 40SC is my current test bed for the AztecMonster used with my SE/30. I have booted from it and it should work just fine with a Plus as well.

[LOOM]

Great to know! Thanks

[uniserver]

ok ive seen 500mb SCSI drives on eaby for 20 bucks, how much are these? i saw one on ebay for 129.00 i said to my self that cant be right , no way. some one would pay 129.00 for something like this.

price has to be factored in, i'm to the point if a vintage mac does not come with a working HD, i will not bid,
hard drives are not easy to find and cost more then the machine in most cases…