S95Sedan wrote on 2023-10-04, 15:21:

Works on the ct1740.
Diagnoses like usual.

DMA clicking is something else, still present unfortunately.

Sad (( But it was epic quest ! ))

S95Sedan wrote on 2023-10-04, 15:21:

Works on the ct1740. Diagnoses like usual. […]
Show full quote

Works on the ct1740.
Diagnoses like usual.

DMA clicking is something else, still present unfortunately.
https://streamable.com/z0vg97

that doesn't sound like the single cycle DMA bug I'm familiar with...

what about a talkie game like DOTT or star trek judgment rites, it's very prevalent in those

maxtherabbit wrote on 2023-10-04, 15:31:

that doesn't sound like the single cycle DMA bug I'm familiar with...

what about a talkie game like DOTT or star trek judgment rites, it's very prevalent in those

Video should be changed but can have a look.
For DOTT its at the start isnt it?

Card is definately not liking that game, music doesnt play properly.

Noisy!
https://streamable.com/lyhsbd

Jill of the Jungle is a suitable test too, I mentioned that game several posts prior in this thread.

This is magnificent. Congratulation to Maelgrum!

my final notes after I dumped 4 x CT4170 cards already:

1. both "sb16dump_1.zip" and "sb16dump_2.zip" work good:
Re: The Soundblaster DSP project

2. all 4 dumps are :

1CRC32 (6k): 986E5CB9 (unknown)
2CRC32 (8k): 4A81FB32 (unknown)
3Saved     : 986E5CB9.416

(I guess all "integrated" version of V4.16 will be the same, maybe only with different padding)

3. note maybe relevant only to other DSP Version attack, but "sbcrack_st4.zip":

Re: The Soundblaster DSP project

"passed" from my side only 1 time from 3 runs on 2 different cards:

1Stage 4, addr 1227 ...failed
2Stage 4, addr 10D9 ...passed
3passed
4Stage 4, addr 1228 ...failed
5Stage 4, addr 10D8 ...passed
6passed

so, definitely the ST4 is the way to go, but it might be needed to be run few times.

sb16dump V0.12. Support dumping from V4.16

Maelgrum wrote on 2023-10-04, 15:42:

sb16dump V0.12. Support dumping from V4.16

I think something is buggy with the CRC32, because I get:

1DSP type  : unknown
2CRC32 (6k): 986E5CB9 (unknown)
3CRC32 (8k): 4A81FB32 (unknown)
4Saved     : 986E5CB9.416
5Success!

was expecting based on a glance in the source code to be recognized as "known".

S95Sedan wrote on 2023-10-04, 15:32:

Noisy!
https://streamable.com/lyhsbd

That noise is due to the CPU being too fast for DOTT when it attempts to initialize the FM synth, try slowing down the CPU.

On the flip side, I'm not hearing the single-cycle DMA clicking bug.

mattw wrote on 2023-10-04, 15:51:

Maelgrum wrote on 2023-10-04, 15:42:

sb16dump V0.12. Support dumping from V4.16

I think something is buggy with the CRC32

looking more:

1    case 13:
2    if (crc6k != 0x9E1B22C6l) return 0;
3    if (crc8k == 0xE22E9001l) return 1;
4    else return 2;
5    
6    case 16:
7    if (crc6k != 0x986E5CB9l) return 0;
8    if (crc8k == 0xB15514EFl) return 1;
9    return 0;

isn't the case for 16 be the same as the case for 13, i.e. to end with "else return 2;", indicating the "discrete" type?

Gmlb256 wrote on 2023-10-04, 15:52:

S95Sedan wrote on 2023-10-04, 15:32:

Noisy!
https://streamable.com/lyhsbd

That noise is due to the CPU being too fast for DOTT when it attempts to initialize the FM synth, try slowing down the CPU.

On the flip side, I'm not hearing the single-cycle DMA clicking bug.

I can't tell one way or the other over the noise tbh

Gmlb256 wrote on 2023-10-04, 15:52:

S95Sedan wrote on 2023-10-04, 15:32:

Noisy!
https://streamable.com/lyhsbd

That noise is due to the CPU being too fast for DOTT when it attempts to initialize the FM synth, try slowing down the CPU.

On the flip side, I'm not hearing the single-cycle DMA clicking bug.

Will have a look if i have a slower one.

This one is more audible though, especially at the end;
https://streamable.com/z0vg97

Everything else seems to run fine and sounds good, no issues whatsoever.

sb16dump V0.13. Support dumping from V4.16

Seems i am too tired ))

Maelgrum wrote on 2023-10-04, 16:02:

Seems i am too tired ))

take a deserved rest! frankly, I got a little tired from all those test i ran today as well.

Anyway nice job on the 4.16.
Hopefully it tells more what they were trying to achieve and changed compared to 4.13. Which might possibly lead to additional features or improvements.

maxtherabbit wrote on 2023-10-04, 15:59:

I can't tell one way or the other over the noise tbh

Yep, it is very hard to tell without adjusting the volume and careful listening.

I'm using headphones on my modern computer if that helps.

mattw wrote on 2023-10-04, 16:03:

Maelgrum wrote on 2023-10-04, 16:02:

Seems i am too tired ))

take a deserved rest! frankly, I got a little tired from all those test i ran today as well.

Agreed, Maelgrum deserves resting considering the huge activity on this thread.

S95Sedan wrote on 2023-10-04, 16:01:

Will have a look if i have a slower one. […]
Show full quote

Gmlb256 wrote on 2023-10-04, 15:52:

S95Sedan wrote on 2023-10-04, 15:32:

Noisy!
https://streamable.com/lyhsbd

That noise is due to the CPU being too fast for DOTT when it attempts to initialize the FM synth, try slowing down the CPU.

On the flip side, I'm not hearing the single-cycle DMA clicking bug.

Will have a look if i have a slower one.

This one is more audible though, especially at the end;
https://streamable.com/z0vg97

Everything else seems to run fine and sounds good, no issues whatsoever.

Run DOTT again with music disabled please

Maelgrum wrote on 2023-10-04, 16:01:

sb16dump V0.13. Support dumping from V4.16

Tested and Confirmed as OK:

1SB Reset: done
2DSP version: 4.16
3MPU-401 init: done
4Internal memory dump: done
5MPU-401 loopback check: done
6Dump complete !
7DSP type  : integrated (padding FC)
8CRC32 (6k): 986E5CB9 (known)
9Saved     : DSP_416I.bin
10Success!

My deepest thanks to all participants !
I couldn't do any of this without your support.