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
Some problem on version 2.10 #2330
Comments
Yes, I will try 2.10.1, thanks you. |
I have tried 2.10.1, but the low_power_mode still not work when I mining monreo on i7-5930k. |
I am mining Monero and my CPU is i9-9900K. I use 4 cores only when using low power mode since CPU has 16M L3 cache. For 2.8 version: For 2.10.0 version: For 2.10.1 version: Is the new algo itself cannot provide as much efficiency as old algo by assigning more memory cache? |
CN-R has different actions in it than CN-V8, so it did shake up performances again but not as much as CN-V7 to V8 fork did. I have noticed it hurts older CPUs more (up to 50% is correct, but some of mine only lost like 30% depending on core type and compiler and cache size), and newer ones almost no reduction (maybe 7% avg). The hashrate also seems to vary a bit which is by design (self-morphing algo changes every so often on new-block, random actions added or removed change the overall efficiency for that round, etc). There could be more optimization possible for lesser CPUs on CN-R but I think there are other todos ahead of it. I have been considering switching all my old-CPU rigs to some other easier coin (turtle?) but the profit doesn't outweigh the triple hashrate so it's kind of pointless other than seeing a high hashrate onscreen again. I do not know about the low_power_mode I have never had positive gain from trying it 2.5 or 2.10 or otherwise, and I have a few large cache CPUs (xeon) where it should do something. |
Appears as though we do not have these enhancements yet but the increase looks about like what we're missing Test with xmrig and see how it does on the old CPUs, I may do the same |
Confirm switching to xmrig doubled or better all my old CPUs (same CN-R algo) |
have similar prblm |
I also confirmed my old CPU Xeon X5355 performed well by xmrig-2.14.1, as the xmr-stak-2.5.0. |
if this cpu has avx support this issue should be fixed with 2.10.1
|
The half of the issue about old CPU speeds on CN-R is not solved at all I have jumped ship to xmrig for my old CPUs until @SChernykh or someone applies their 2.6x speedup code here. Speeds are (generally) equal on AES capable, so yes that part works great now. |
Hi, problem with multihash is not solved for me on linux. Here are results with multi hash on two cpu threads: [2019-03-13 11:13:23] : Mining coin: cryptonight_r
|
I have update my miner from 2.5 to 2.10, and I found some problems.
The setting of "low_power_mode" in the cpu.txt seems not work. For example, one of my CPU is i7-5930k, that is 6 cores with 15M cache. In the old version, I set up the low_power_mode "true" will improve hashrate significantly. But that is no difference in the new version when I doubled the cache of one core.
For some very old CPU, such as Xeon X5355 or Pentium D, the "asm" setting in the cpu.txt needs be forced to off, or it will show "Illegal instruction (core dumped)". And set to off also reduce the hashrate about 50%. Is this related to AES?
The text was updated successfully, but these errors were encountered: