Some (more) Interesting Times

[+Larry]

I decided to run some additional benchmark test on various HD and APE combinations.

 

These are pretty much "apples to apples" but there are some differences, namely some tests are run with hard drives and several with compact flash drives. Still, I think they are pretty illustrative and certainly interesting. The basic test is to copy 2000 double density sectors (no verification) using my Action! sector copier. Hias' HISIO.COM was used for the 3X SIO tests, except for the Black Box tests (which used the BB drivers). Timing is all internal to the program using the 18, 19 20 location timers. From earlier test, I've found that using CF drives are usually faster than a real hard drive, and I thin this would be especially true when reading/writing to the same drive. Incidently there is no buffer in the test program -- it reads/writes each sector -- a lot of head movement.

 

Baseline -- APE to APE

 

USB (28) or 1X SIO: 10:08

USB ( or 3X SIO: 3:00

USB (4) or ~4X SIO: 3:01 (I ran this twice and got the same results. Perhaps someone can speculate why it is no faster than 3X?)

 

MyIDE to MyIDE: 1:40 (compact flash drive)

MyIDE to APE: 2:42

 

Black Box to APE: 2:17 (Conner SCSI hard drive)

Black Box to Black Box: 1:50

Black Box to SCSI ZIP: 1:22

 

MIO to MIO using V1.1 firmware and ST-251 MFM drive: 3:35

MIO to APE: 2:48

 

KMK-JZ to APE (256-byte sectors): 2:13 (compact flash drive)

 

If possible, I'll add a couple more test of the KMK-JZ interface using 512-byte sectors.

 

Comments and (constructive) criticisms are very welcome.

 

-Larry

Edited January 10, 2009 by Larry

[+bob1200xl]

Larry:

 

Can we get a copy of your sector copier?

 

Bob

 

 

 

I decided to run some additional benchmark test on various HD and APE combinations.

 

These are pretty much "apples to apples" but there are some differences, namely some tests are run with hard drives and several with compact flash drives. Still, I think they are pretty illustrative and certainly interesting. The basic test is to copy 2000 double density sectors (no verification) using my Action! sector copier. Hias' HISIO.COM was used for the 3X SIO tests, except for the Black Box tests (which used the BB drivers). Timing is all internal to the program using the 18, 19 20 location timers. From earlier test, I've found that using CF drives are usually faster than a real hard drive, and I thin this would be especially true when reading/writing to the same drive. Incidently there is no buffer in the test program -- it reads/writes each sector -- a lot of head movement.

 

Baseline -- APE to APE

 

USB (28) or 1X SIO: 10:08

USB ( or 3X SIO: 3:00

USB (4) or ~4X SIO: 3:01 (I ran this twice and got the same results. Perhaps someone can speculate why it is no faster than 3X?)

 

MyIDE to MyIDE: 1:40 (compact flash drive)

MyIDE to APE: 2:42

 

Black Box to APE: 2:17 (Conner SCSI hard drive)

Black Box to Black Box: 1:50

Black Box to SCSI ZIP: 1:22

 

MIO to MIO using V1.1 firmware and ST-251 MFM drive: 3:35

MIO to APE: 2:48

 

KMK-JZ to APE (256-byte sectors): 2:13 (compact flash drive)

 

If possible, I'll add a couple more test of the KMK-JZ interface using 512-byte sectors.

 

Comments and (constructive) criticisms are very welcome.

 

-Larry

[+Larry]

Hi Bob-

 

Sure. I'll send it in a PM. There's nothing special about it, but if anyone else wants a copy, please PM me. It has one known bug and that is that if the jiffies are a large number, then the Action! math bug rears it's head, and the elapsed time calculation is very wrong (and easily noticeable). At least that's what I think is the cause. I wrote it as a HD backup program to back up my drives to APE images.

 

-Larry

 

 

Larry:

Can we get a copy of your sector copier?

Bob

Edited January 11, 2009 by Larry

[+Larry]

Here are some additional times.

 

Black Box - Black Box using a SCSI-IDE bridge with a Sandisk CF card: 1:07

Black Box to APE (with CF card drive): 2:17

 

So the writing to the CF card and writing without a lot of head-seeking made a really big difference in the first comparison.

OTOH, using the CF card made zero difference in the reading then writing to APE.

 

When I get all the updates done, I'll put them together in one post for easier comparisons.

 

-Larry

[+Larry]

KMK-JZ to KMK-JZ using Sandisk CF card: 0:53!

 

Last update until I figure out how to access the KMK-JZ in 512 bytes/sector "Native" mode.

 

-Larry