exxos's Atari, Amiga & retro forum

Supporting preinstalled CF card is against Amibay rules because WB is still under Licence..
maybe we just need to add empty CF cards in the order..🤔

terriblefire wrote: 19 Jul 2020 09:05 Have any of of you tried adding IDE buffers?

Is it 3 more SN74CB3T3245PWR that I need for the buffering? And remove the resistor arrays on the under side of that area?

All the cards that I use on the 330 work fine on the 536. But the ide code is identical to all these accelerators - nothing has really changed, so the behaviour is the same.

For the cards that state “no dos disk in DH0” you just have to offset the partition a bit away from the first and the last blocks and it will be fine.

I do offer CF cards when I actually have them in stock, which these days is not often considering how long it takes for delivery even though the amibay admins have been a big pain in the butt lately.

8 Bit Dreams wrote: 18 Jul 2020 19:33 As explained before - long booting is affected because of some things:
1. Because of setpatch feature (may be skipped by adding following string in SS:
2020-07-13 22.15.31.jpg
2.by changing block size to bigger one - this is critical to large disks & partitions!
Default block size is 512, just change it to 4096 and You'll be surprised how fast it becomes...
And use full CPU & MMULibs of course..
Cheers,

- 8 Bit Dreams

Thank you. I wouldn't figure this out myself.

Changing block size to 4096 did the trick on my 3.1.4 . I did load empty system in like 20sec now.

SysInfo speeds are still in 350kb, but I suspect that it is SysInfo thing, as speed to open drawers etc is the same like on 3.1
With MMULibs speeds are like in 3.1 now.

PS.
Not all HDTools have option to change block size.

I have a TF536 with gold top CPU. A500 with Rev 5 motherboad, 1mb chipmem mod. Kickstart 3.1, Workbench 3.1. 2GB SanDisk CF. 512 block size.

With CPU FASTROM, Sysinfo reports 4.5mb/sec

How big your CF Card / partition is? Probably small. Its not about weird Sysinfo numbers - these are useless, its about real booting time..You may change block size when you have large partition and it takes too long to boot.. :idea:
We can do it now since we have a lot of memory available, that's what it is for - to use it ☺

I noticed that there is some slight timing difference between the TF534 and the TF330.... Until now we were using the TF330 timing

terriblefire wrote: 20 Jul 2020 08:59 I noticed that there is some slight timing difference between the TF534 and the TF330.... Until now we were using the TF330 timing

tf536r2_2020_07_20_alpha.zip

I had a look and can't see any substantial difference really.

Now I have 2 Rev2 up an running, so I can compare it side by side. I will poke it a little bit more in the afternoon.

arkadiusz.makarenko wrote: 20 Jul 2020 10:39

terriblefire wrote: 20 Jul 2020 08:59 I noticed that there is some slight timing difference between the TF534 and the TF330.... Until now we were using the TF330 timing

tf536r2_2020_07_20_alpha.zip

I had a look and can't see any substantial difference really.

Now I have 2 Rev2 up an running, so I can compare it side by side. I will poke it a little bit more in the afternoon.

DTACK is asserted 1 clock sooner on IDE accesses.

Most people aren't even getting it to boot on A2000.
[/quote]


I've got a Rev 4.4 A2000 with 2 MB Chip RAM and WB 3.1.4.1, and the TF536 with buffered IDE works well in the Coprocessor slot. Not quite as fast as I'd like, about 1.5 MB/S but not a disaster.