So I have an I5 3570k (Overclocked to 4.4 GHZ)
32gb 1400mhz ddr3 ram (Possibly only 1333, but it should all be matching modules)
and for whatever reason, I can only mine 80-100 h/s
For comparison, I also have an i5 2500k running at 3.4GHZ that is getting 180-200
16gb 1600mhz ddr3
It's not thermals (I'm at about 60C on hottest core)
and task manager says its getting about 80% of the CPU (roughly the same % on both computers)
Win 10 with winMiner
I was just curious if anyone might be able to narrow down the root cause of this slowdown. I'm not looking for a lecture on the unprofitability of CPU mining (I don't intend to mine with them, I just got winMiner and thought I would try it out for curiosity)
You'd use 3 threads (trying no affinity or affinity on cores 1,2,3 leaving 0), check in bios if ram interleaving is ON, clock ddr3 and timings as much as possible (your 2500k has 1600mhz ram
)
I Actually need to correct while the i5 2500k's ram is 1600mhz; its apparently only clocked at 1333 mhz. So it actually has slower ram than my i5 3570k... Also both Pc's have 4x sticks of matching ram, and perform 'as expected' when userbenchmark is ran.
Even If it was actually 1600; to me that should only account for 15-30% of the performance.
Try Different miners , play with the max core/thread you using for mining.
For me, sometimes the same cpu give me very different result in different mobos. In some case the same cpu, mining with 1 core is fast as the same cpu mining with 4 core in another motherboard .
There is my "
My ETN mining experience with CPU" topic
My Z77A-G41 has the i5 3570K and My Z68 Extreem4 Gen3 has the i5 2500k.
In your test, have you figured out any particular correlation for which boards behave poorly, and possibly why? I've always operated under the assumption that Motherboards themselves have a minimal impact aside from specific features like overclocking. and its usually best to get whatever mid-range board is on sale so-long as it meets your needs.
@all
I also tested SUMO and the results are similar (though the 2500k is only about 70% faster this time, instead of a full-fledged 100%)
Again I'm looking to solve the performance issue. (Or at least properly identify it) rather than grab a different miner and ignore it.
