Topic: Baikal X10 ⚡OVERCLOCK⚡ Claim reward 0.2 BTC for TUTORIAL HOW TO :)⚡⚡⚡⚡ - page 11. (Read 22704 times)

I changed that value to 0x190 in sgminer, but it didn't have any impact on the hash rate. So probably you are right.

Without stupid hashrate limitter its running pretty fast. But only for 10sec. I asked Baikal If there is some hashrate limit - and they say yes.

Here is Giant B
[Baikal Logo]
BK1791R0
ND262

Cystal 12.5 Mhz

can we compare to a CPU?
can we overclock it?
on the electronic boards, are there the components allowing to vary the voltages and finally the frequency

or as mentioned above, quartz can't be modified ?

we somes god in electronics on this topic 

Then the question becomes, how do we make the miner, better?  What change would give it a boost?  It may be someone is trying to clone and improve it and may tap into the talent in this forum for future development .. never know

Its not a page, its a disassembly of the sgminer.
70 is not a value but offset to a variable..

I'm almost 99.9% sure that you cannot do anything to baikal without changing its crystal or charing STM firmware.
FInding the right crystal that will only give 10-20% boost would be hard..
Disassembling firmware even harder..

why dont you join in and leave the sidelines?

keep up the good work gents, still watching

How did you get this page up?

we are able to change fan speed, which here says 70 so why cannot it work for clock speed?

sgminer was hardcoded to report 300Mhz.
Its not picking up the value from STM board, what you see on sgminer interface is hardcoded string, nothing more.
I can make it report anything i want.

Moreover, Mhz setting to baikal_setoption is hardcoded as well, see below.


Changing supplied option to anything (over/undeclock) makes no difference in both calls to baikal_setoption.
I believe this setting is simply ignored by the STM firmware.

The X10 use 10Mhz crystal oscillator , B use 12MHz . Giant A2000 8MHz

exemple

the string lenght of quark if

[04:47:14] [THR1] Generated target 00000000000000000000000000000000000000000000000000a0a90a01000000
[04:47:14] [THR0] gen_stratum_work() - algorithm = quark
[04:47:14] [THR0] Generated stratum merkle a4cabd4cb6a2275c654cf7228c1f8fd427d0c520121a6921367d00ebeb5c5636
[04:47:14] [THR0] Generated stratum header 000000031237f750298c12396693229f8d92cc0e841efaac5f96e18d00000d1400000000a4cabd4 cb6a2275c654cf7228c1f8fd427d0c520121a6921367d00ebeb5c56365a81fd301a4a5bf3000000 000000000000000000000000000000000000
00000000000000000000000000000000000000000000000000000000000000
[04:47:14] [THR0] Work job_id ec nonce2 3 ntime 5a81fd30
[04:47:14] [THR0] Generated target 00000000000000000000000000000000000000000000000000a0a90a01000000
[04:47:14] Generated stratum work
[04:47:14] [THR0] Pushing work from to hash queue
[04:47:14] Pool 13 stratum share result lag time 0 seconds
[04:47:14] [THR0] PROOF OF WORK RESULT: true (yay!!!)

and the result with 400 mhz

[04:48:07] [THR2] Generated target 0000000000000000000000000000000000000000000000000000f6ff09000000
[04:48:07] [THR0] gen_stratum_work() - algorithm = quark
[04:48:07] [THR0] Generated stratum merkle f8e7f4689d8432d965b165b64a5be7b79b8d667e488ebc317f95f905173b7647
[04:48:07] [THR0] Generated stratum header 000000031fdbe60845f086328218ace1092ccd2821e136e6bbc73e87000023b900000000f8e7f46 89d8432d965b165b64a5be7b79b8d667e488ebc317f95f905173b76475a81fd711a489116000000 000000000000000000000000000000000000
00000000000000000000000000000000000000000000000000000000000000
[04:48:07] [THR0] Work job_id ee nonce2 551 ntime 5a81fd71
[04:48:07] [THR0] Generated target 00000000000000000000000000000000000000000000000000009cff63000000
[04:48:07] Generated stratum work


the baikal try to full up the 100mhz more with 000000000000000000000000000000000000000000
00000000000000000000000000000000000000000000000000000000000000


I know is not frequency put the data inside the baikal but the rehash logical power use the clock to calculed the end of the string return that it !

I think you need to provide more details on how you accomplished overclocking it, not keeping the 300MHz the same

I dont know if I win the 0.2 but the reason is simple, all giant a algo have a string lenght take and calculed by the baikal at the 300mhz clock so if you change the frequency the only think happen is a lot of invalid share that it !

There are thermal pads under the heatsink. I have replacements ordered, but I'm only running myr-gr, so not too worried about heat. You could probably replace the thermal pads with Arctic silver, but I'd be worried about it getting on the board.

There's a small heatsink on one side, and a large one on the other. The chips are all under the small one, so dont worry about taking them both off.

When you took off the heat sink, was there any thermal paste? And did you replace it?

Yes. If chips on Giant B are different only way is to modify firmware to comunicate with x10 chips. Or B to X10

I've tried flashing the Giant B firmware to an X10 and running the Giant B image on the Pi but no luck converting an X10 to GB or using that firmware image to overclock (so far).

Interesting that the controller boards are the same, but not surprising. I can't find the difference between those two STM model numbers anywhere online. Could the last couple lines just be batch numbers?

I DID take the heatsink off an X10 and the model number on the asic is this:

[Baikal Logo]
BK177280
ND33U

I'd upload a picture, but the chips are too small for my potato camera.

He only other thing I would try is to swap the STM board + OrangePi of Giant A and Swap with Giant B. Leave original code as is.

That would mean that we could order spare parts for each type and use hashboards

I EXTRACT THE FIRMWARE OF THE ARM WITH THE SCRIPT ON USR BIN AND I PUSH UP THE FIRMWARE ON THE GIANT B PLATE WITH THE SAME SCRIPT

fw-upgrade....