herrdrone (OP)
|
|
February 18, 2020, 06:30:31 PM |
|
I think Hashrate is okay, but if you possible for you please try mining monero with the fastest mining platform RagerX I think it can increase hashrate around 8% to 10% especially for Rayzen CPU's.
Which mining software did you use there? I'm mining monero using xmr-stak-rx hashrate is almost same as Xmrig but the advantage of this miner is it is free, I don't have a 1% fee as Xmrig.
Based on your post history. SHILL ON YOUR OWN THREAD
|
|
|
|
heavyarms1912
|
|
February 20, 2020, 12:15:44 AM |
|
anyone trying underclocks ?
i have a ryzen 9 3900x. set it to freq 3600. volts at 1.0
2 sticks of 8gb ram.
used 21 threads 🧵
using nicehash 2 ... 19
a evga 1660 doing phoinex eth.
rig pulls 217 watts. and earns 1.38 usd on nice hash.
217 watts is 5.1kwatts a day or 51 cents at ten cent power.
so 1.38-0.51= 87 cent daily profit.
i had a ryzen 7 1800x flashed my mobo to run the 3900x
i got the cpu for 430- 100 for the 1800x = 330 cost
roi will be a year. if it makes 87 cents a day
have set it to 1.0625v @ 3800 Mhz. 3600 CL14 ram kits (e-die overclocked). Gets about 13.5 Khs. Yeah I can get to 13,100 from my current 9,000 but the power used 2x so with cheaper ram and underclocked cpu using only 19 threads and my gear is doing much better at watts per hash then your rig. my cpu runs at 50c to 52c . I was interested if anyone has tired to underclock I am down to 3600 freq and 0.99 volt clocks 19 to 21 threads very stable and quiet rig. plus I can use it for other things if I want. I never mentioned power consumption so you can't make a comparison in efficiency. I experimented with 21 threads the power saved at wall is 10w for me while hashrate is 12khs down from 13.5khs. Efficiency no is 13500/198w ~ 68 and 12000/188w ~ 63. Note: I got a Radeon VII at fixed clocks also consuming power. HW-Info nos were roughly around 116w at 3.8-3.9 Ghz clocks and 111w at 3.6 Ghz clocks for 3900X. Also, simply undervolting doesn't translate to power savings. 1.0625v or 0.95v core voltage makes no difference to power consumption at wall. There should be more to it.
|
|
|
|
philipma1957
Legendary
Online
Activity: 4172
Merit: 8070
'The right to privacy matters'
|
|
February 20, 2020, 02:02:56 AM |
|
so with 20 threads and stock cooler i run at 59 temp.
clock set to 3500 volts to 0.095
two sticks of cheaper 8 gb ram.
savings are no water cooler needed.
unit burns 217 watts at kwatt meter. earns around 1.40 usd.
second unit burns 249 watts at kwatt meter earns around 1.28 usd.
using nicehash 3.05
1660 gpus set to 80% power and volts.
so i burn 466 watts or 11 kwatts a day.
my adjusted winter rates are 10 cents so i pay 1.10 in power a day and earn 2.68
gain 1.58.
stock coolers save 100 x 2 = 200 cheap ram save 100.
so i save 300 in parts maybe more like 330 in parts. my power is 1.10 so that is 300 days of mining saved.
yeah i do 8500 to 9000 hash vs 13000 hash.
i don’t claim this is smarter then pushing the gear hard.
i am asking if anyone else tries this.
|
|
|
|
heavyarms1912
|
|
February 20, 2020, 04:09:54 AM |
|
so with 20 threads and stock cooler i run at 59 temp.
clock set to 3500 volts to 0.095
two sticks of cheaper 8 gb ram.
savings are no water cooler needed.
unit burns 217 watts at kwatt meter. earns around 1.40 usd.
second unit burns 249 watts at kwatt meter earns around 1.28 usd.
using nicehash 3.05
1660 gpus set to 80% power and volts.
so i burn 466 watts or 11 kwatts a day.
my adjusted winter rates are 10 cents so i pay 1.10 in power a day and earn 2.68
gain 1.58.
stock coolers save 100 x 2 = 200 cheap ram save 100.
so i save 300 in parts maybe more like 330 in parts. my power is 1.10 so that is 300 days of mining saved.
yeah i do 8500 to 9000 hash vs 13000 hash.
i don’t claim this is smarter then pushing the gear hard.
i am asking if anyone else tries this.
And I guess what I am suggesting is that you're not saving much power reducing the threads. you might want to try out with 24 threads. When I experimented with raw undervolt+underclock it didn't save power in actual. Also, I am on stock cooler and micron e-die's are quite cheap. $65-$70 for 16gb kit. That's what I got. It's 414w at wall for the VII hashing at 85 Mhs + 3900X at 13.5 Khs. https://imgur.com/a/OcKi3S0
|
|
|
|
|
|
herrdrone (OP)
|
|
March 12, 2020, 09:29:58 PM Last edit: March 13, 2020, 04:40:25 AM by herrdrone |
|
thats a LOT! Is there any info on hashrate on the same setup on RagerX miner?
Coinluminary and Herry Toms only use their accounts to SHILL, you have done it before on this thread and you do on others particularly today as your post history shows. Yet, I will take your statement and point some facts: - RagerX is closed source with 3.5% fee. And was only created to mine on pool with same name (does not accept any other mining software), so miner/rig owner is not on control of it. - Those who have used it regret it, as they figure out they will never get paid. So they switch to XMRig and point rig to any other pool they choose. https://www.reddit.com/r/MoneroMining/comments/ff4bfh/ragerx/- You are breaking forum rules by spamming, and creating posts with no value to discussion. I encourage community to report spammers to moderators. Thanks.
|
|
|
|
herrdrone (OP)
|
|
May 11, 2020, 09:53:43 AM |
|
A new section has been added where CPU with 4+ records are listed with all details: #cores, #threads, default tdp, socket type, memory supported, cache size (L1, L2 & L3), lithography, hash per watts, average hashrate, and more. https://monerobenchmarks.info/list.php
|
|
|
|
Set Ready Go
Member
Offline
Activity: 273
Merit: 17
|
|
May 17, 2020, 01:18:09 PM |
|
If anyone has gotten their hands on a Ryzen 3 3100/3300x feel free to share the hashrate , settings etc...
Im thinking maybe swap a 1600 to the newer 3100..
|
|
|
|
|
Hueristic
Legendary
Offline
Activity: 3864
Merit: 5055
Doomed to see the future and unable to prevent it
|
|
September 08, 2020, 08:22:57 PM |
|
Anyone have a clue about Ampere vs Big Navi on RandomX?
|
“Bad men need nothing more to compass their ends, than that good men should look on and do nothing.”
|
|
|
sech1
|
|
September 09, 2020, 06:06:54 AM |
|
Anyone have a clue about Ampere vs Big Navi on RandomX?
Both will suck compared to CPUs. Big Navi will suck even more due to lack of native support for RDNA2 in xmrig.
|
|
|
|
sxemini
Member
Offline
Activity: 1558
Merit: 69
|
|
September 09, 2020, 11:08:09 AM |
|
Anyone have a clue about Ampere vs Big Navi on RandomX?
Both will suck compared to CPUs. Big Navi will suck even more due to lack of native support for RDNA2 in xmrig. XMRig have native support for Ampere? Or how do i understand you post?
|
|
|
|
sech1
|
|
September 10, 2020, 07:42:28 AM |
|
XMRig have native support for Ampere? Or how do i understand you post?
NVIDIA GPUs are efficient enough with straight-forward RandomX VM emulation, they even support directed rounding floating point operations in CUDA, so there's no need to add support for each new GPU. Generic CUDA code will just work. But even 2080 Ti does only 1600-1700 h/s. 3000 series will do a bit more, but still far from enough to compete with CPUs. AMD GPUs, on the other side, require native GCN/RDNA assembler code to run RandomX efficiently.
|
|
|
|
|
herrdrone (OP)
|
|
November 08, 2020, 04:50:24 PM |
|
|
|
|
|
herrdrone (OP)
|
|
November 08, 2020, 06:16:57 PM |
|
|
|
|
|
herrdrone (OP)
|
|
November 13, 2020, 10:17:44 PM |
|
|
|
|
|
herrdrone (OP)
|
|
December 04, 2020, 09:33:28 PM |
|
|
|
|
|
|
|