Question about staking (from a miner).

It’s well known that some Reddheads run nodes on their RasPi - or ReddPi. Since i’ve been running one for the last six months - with negligible balance variations - i’ve noticed that whenever it has to process a new block the CPU usually stays maxed out for some time.

So i was wondering, as i understand it the amount of coins you are staking is essentially your “hashing power” - in PoW terms. Under that assumption, as long as you have the - more or less - same amount of staking coins in your wallet, your hashing power remains the same. Think of it as if 1k RDD always equalled 1 kH/s of the total network hashing power at the time.

Now my question is if CPU usage is tied directly with the amount that you’re staking (i.e. the more coins you stake, the more CPU cycles it takes), with the current difficulty (i.e. since your “hashing power” is fixed, CPU load stays the same regardless of it), with block size (i.e. the number and value of TXs that the block contains), a combination of them all or whatever other factors i’m most likely missing.

I don’t own a crapload of coins and i don’t have a very good knowledge of the underlying algorithms at work, but i will also have to assume that as long as your CPU stays maxed out, you will potentially miss any chance of staking because you will be behind the network - basically catching up. I don’t know if the daemon runs in real-time priority under linux (which i doubt), but then again this is not exactly my field of expertise.

So, in a nutshell, i was wondering if the ReddPi will become obsolete once/if the RDD network grows substantially in size, or it’s still good as long as you’re not running it as a full node and staking some crazy millions. Thanks!

PS: Sorry if the question got a bit jumbled up but it’s holidays and i’ve done… things. @_@

Didi From what I have found a PI will run 100% or near that CPU usage at idle. Are you sure it is only when there is an incoming new block?

I stake very often with a old cpmputer and sometimes o have two stake rewards maturing2 at the same time which means second staking comes in less than 30 min of the one before

My wallet says im expected to earn rewards every 357777 days.
Do i need alot more coins? :open_mouth:

flyingman1 that figure will come down as it accumulates coinage, that said the more coins you have the more often you will stake

BrownSlaughter now its down to 34345 so i guess its getting closer

Didi Good question.

flyingman1 Which wallet are you using?

GrayPhoenix the official QT 5.2.0

I had to reboot the ReddPi so i took some snapshots. Here you’ll see the issue better:

Wallet loaded and idling:

img

Wallet staking:

img

Sudden spike one minute after:

img

Sometimes it gets maxed out for 10 or more seconds, and sometimes for only two or three.

kraus3742 Are you runing anything on the RasPi other than the wallet? O.o

GrayPhoenix Thanks… but that means you don’t have the answer. Lol. Ah well, it’s no biggie. The new dual core RasPi is far more powerful and still priced the same, so this should not really be an issue. Still curious though.

Didi I’m curious… Are you saying that larger blocks sometimes max out your CPU for longer times than smaller blocks, and you wonder if there is a certain block size threshold beyond which the CPU won’t be able to read the block in time before the next block? (Also, for clarification, you are referring to the client reading ~all~ new blocks, not just when you personally stake a new block, right?)

If so, you have an interesting thought, and one that I can’t argue for or against without thinking about some more. If it is the case, it could be that in such a scenario an upgrade might be warranted to the R Pi 2 as I think you mentioned.

Another thought though… have you tried running your node from the terminal instead of GUI? This would use far less computational power I would imagine.

Also, there is an additional factor should the network grow considerably (as we hope it will): full nodes will need to have much more storage capabilities. So, a good node setup down the road would likely look something like this: An R Pi 2 + 500 GB external HDD + headless w/ ssh support from a local network only + strong user passwords.

An alternative I am keeping my eye on is the newly announced Chromebit. If that is able to be flashed with a new OS like other Chromebooks and Chromeboxes, it would make for a pretty affordable and much higher powered solution that you could run on the back of your TV for easy monitoring.

1 Like

bigreddmachine said:

Didi I’m curious… Are you saying that larger blocks sometimes max out your CPU for longer times than smaller blocks, and you wonder if there is a certain block size threshold beyond which the CPU won’t be able to read the block in time before the next block? (Also, for clarification, you are referring to the client reading ~all~ new blocks, not just when you personally stake a new block, right?)

I think it pretty much boils down to that. It makes sense that the block size would be the main contributor to CPU usage, yet as you can see in the screenshots above, staking alone without processing any block (EDITED: any new block, regardless of who minted it) takes roughly 65% of the available CPU power (chip is o/c to 950Mhz). Here’s how it looks when a particularly big block (i assume) is being processed:

img

I suppose i could actually use reddsight and/or the console (console rather unfeasible bc it would alter usage reports) to make sure that’s the case. I thought that someone would know the answer straight away though.

If so, you have an interesting thought, and one that I can’t argue for or against without thinking about some more. If it is the case, it could be that in such a scenario an upgrade might be warranted to the R Pi 2 as I think you mentioned.

Another thought though… have you tried running your node from the terminal instead of GUI? This would use far less computational power I would imagine.

Nope… i have never tried to run the daemon alone. I’m aware that it must take less resources but as mentioned above, you can see that with the GUI and all CPU usage is negligible as long as you don’t unlock and start staking.

Also, there is an additional factor should the network grow considerably (as we hope it will): full nodes will need to have much more storage capabilities. So, a good node setup down the road would likely look something like this: An R Pi 2 + 500 GB external HDD + headless w/ ssh support from a local network only + strong user passwords.

Completely agree.

An alternative I am keeping my eye on is the newly announced Chromebit. If that is able to be flashed with a new OS like other Chromebooks and Chromeboxes, it would make for a pretty affordable and much higher powered solution that you could run on the back of your TV for easy monitoring.

…omg, that would be osom!!!