Block Size/Scratchpad Size - Why is it so difficult to find information about it

Hey there,

I am mining since 1 month with my small AMD Ryzen 3600 system, accompanied by an inexpensive AMD RX 560.

I usually focus on CPU mineable coins and every time I give another coin a try, especially exotic ones, I get a little bit more angry.

The so-called Block Size Scratchpad Size or Algorithm Size is one of the most important informations, imho!

Yet only a few pools/projects provide this information.

Because of that - and because of the differences this can make, depending on the Algorithm and CPU one has - I am pretty sure that a LOT of workers are running very inefficiently!

Just a few minutes ago, I tried a coin with Lyra2z330 algo. Usually, I run CPUMiner with 12 threads. At the same time, I googled for Algo Size for Lyra2z330 (which was quite easy this time, to be honest).

And quickly changed the CPUMiner config to 8 threads only. As the Block size is 4MB, multiplied by 8 Threads gives 32MB, which is the size of MY CPUs L3 cache!

I know there are people mining with older CPUs that usually had only 2 or 4MB of Level 2/Level 3 cache (exception are old AMD FX-CPUs which had up to 8MB). And I know that many of those people dont know or dont care or dont understand what's happening inside the miner or their system.

Back to my Lyra2z330 example:

12 Threads at 600H/s

8 Threads at 1000H/s

If it exceeds the cpu cache size, the performance decreases rapidly!

So my question again: Why is this very important information so hard to find?

Atm, I am searching for Argon2D16000 Block size. I can only guess that it's maybe 16 MB?

Who knows. So instead of providing this information on the pool website, I will try to find it out by running the CPU Miner with different threads - wasting precious mining time ^^ (PS: XMRig shows Algo Size when starting! *thumbsup* for XMRig)

Rant is over! Thanx for reading! ^^

EDIT: Yeah, Argon2d16000 Block size is 16MB. So for my Ryzen 3600 it doesnt make any difference if I use 2 threads or 12 threads, round about 300H/s

Well, there is a difference: When using 12 threads - 10 threads for nothing - it's a huge waste of electricity...