Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Feature: don't calculate asic_job_frequency but measure it. #514

Open
mutatrum opened this issue Nov 24, 2024 · 2 comments
Open

Feature: don't calculate asic_job_frequency but measure it. #514

mutatrum opened this issue Nov 24, 2024 · 2 comments
Labels
help wanted Extra attention is needed

Comments

@mutatrum
Copy link
Contributor

mutatrum commented Nov 24, 2024

Taken from Slack:

As a test, I extended the job_frequency to 5000ms, which is too long, and I see this:

I (18251) ASIC_task: Send next job: 20000000 Nonce 00000000 0000000000000000 diff 10000
I (18241) stratum_task: setup message accepted
I (19341) bm1368Module: Job ID: 18, Core: 45/1, Ver: 12AE2000
I (19341) asic_result: Ver: 32AE2000 Nonce DE4C035A diff 417.7 of 10000.
I (19651) bm1368Module: Job ID: 18, Core: 52/13, Ver: 1807A000
I (19651) asic_result: Ver: 3807A000 Nonce BC080368 diff 288.5 of 10000.
I (21201) bm1368Module: Job ID: 18, Core: 45/1, Ver: 12AE2000
I (21201) asic_result: Ver: 32AE2000 Nonce DE4C035A diff 417.7 of 10000.
I (21511) bm1368Module: Job ID: 18, Core: 52/13, Ver: 1807A000
I (21511) asic_result: Ver: 3807A000 Nonce BC080368 diff 288.5 of 10000.
I (23061) bm1368Module: Job ID: 18, Core: 45/1, Ver: 12AE2000
I (23061) asic_result: Ver: 32AE2000 Nonce DE4C035A diff 417.7 of 10000.

It starts looping, it returns the same 2 nonces of that range, every 1860ms.

This gives me an idea, instead of trying to calculate exactly what this period needs to be, we can just feed it a job for a few seconds, and time how long it took the chip to loop around. No need for difficult calculations, works on all chips and all nonce range masks and such.


The 1860ms is awfully close to the 1812ms from #420 Is the measurement wrong, or is there some time measurement overhead? It seems - in this small sample - to be consistent.

@WantClue WantClue added the help wanted Extra attention is needed label Dec 1, 2024
@mutatrum
Copy link
Contributor Author

Suprising measurement is that the time for a wrap-around is still 1860ms when the Supra is running at 600Mhz. Above was default at 525Mhz.

@adammwest
Copy link
Contributor

Suprising measurement is that the time for a wrap-around is still 1860ms when the Supra is running at 600Mhz. Above was default at 525Mhz.
yes it is surprising good find!

I have verified this at 490 and 420 mhz with the same value 1.86
All things equal this means
time = space/frequency
nonce_space ∝frequency
setting the frequency at least in some cases changes the nonce range.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
help wanted Extra attention is needed
Projects
None yet
Development

No branches or pull requests

3 participants