 Show Posts
|
|
Pages: [1] 2 3 »
|
In China, for false talk about coronavirus, The death penalty. And how is it in your country?
It's so ironically that the most "false talker" is the chinese government. But i doubt that they will use death penalty to themselves. Or under "false talks" you mean "talks that don't accept by chinese governmental propaganda"? The whole family (My neighbors) was shot only for false information in his video blog about Covid19 The truth is only for our government. Man has no freedom of speech. This is how we live. And I wonder. Is it everywhere like that? |
|
|
|
In the USA, Dr. Fauci, the CDC, and government allow... even authorize... false talk about Covid. But the question remains. In China, who decides if it is false talk, right? Any talk could be false talk if the government decides it that way, right?  yes bro |
|
|
|
|
In China, for false talk about coronavirus, The death penalty. And how is it in your country?
|
|
|
|
I'm bored and I want to play a little game, I will send someone $40 in BTC to see if they would send me $20 back from the $40 i send  the first person to comment I will start the game with them I want to play this game. |
|
|
|
im looking to setup to mine with ram and ssd on a laptop
This is a very bad idea. |
|
|
|
|
415Mh/s with 12x RX 580? with that power consumption, I highly doubt it
|
|
|
|
|
I had a similar issue before, did all the "known" solutions (new drivers, check risers, check power, check overclocking profiles and even tried stock); my rig would always crash after installing the 6th card.
Anyways, It seemed to work fine after changing my PCIs to GEN 2 instead of GEN 1, and disabling CSM Support from the boot settings in bios.
Whats still bothering me is that I dont understand how this solved my issue, it just does not make sense; but it worked on multiple rigs.
In your case it could be worth a try.
Good luck
|
|
|
|
|
NEVER Do that unless both PSUs ground are joined together. If you use the paper clip method, you will burn all your GPUs.
You can do it with ADD2PSU, or if you have an ASUS B250 mining motherboard, or any dual PSU adapters
|
|
|
|
|
These chinese PSU are built for 6 GPUs, and not 12.
It mentions having 12 PCIE cables for GPUs with 8+6 power connectors. But in reality, they are meant for 6 GPUs usage.
Do not connect the motherboard's SATA/Molex sockets that are near the PCIE slots if you are going to power the risers directly, and I do not recommend using powering the risers from the motherboard at all; with 12 PCIE slots, that means you can have upto 900W going through the motherboard, which is alot.
Always power the risers directly from the PSU using SATA or 6pin PCIE.
The risers can consume upto 75W of Power (though its rare you will reach that point), but powering 3 risers using 1 PCIE/SATA/Molex rail can melt your cables or blow the fuse or in the case of chinese PSUs, the whole PSU will not work (I use these chinese ones as well.. very cheap).
It happened several times with me; I had EVGA modular PSUs, 3 risers on one rail; some of the cables melted and I had to replace them, but that was during the dual mining era.. doubt it will ever happen if you are single mining, the whole GPU barely consumes 110W.
|
|
|
|
And on which timings will i copy the this strap? 777000000000000022CC1C00CEE55C46C0590E1532CD66090060070014051420FA8900A00300000 012123442C3353C19 1750 and 2000 ?  Thanks for the help! yes, 1750, and 2000 (just make sure when you paste it that there isnt any spaces. Set core to 1150, memory between 2100-2150 check this guide on how to overclock/undervolt using OverdriveNtool: https://mining.help/ |
|
|
|
|
I dont know who came up with the straps, but they are very popular for Samsung Memory and arguably provide the best hashrate for lowest clock (@2150 mhz)
If you google ubermix, you will find alot of resources and people recommending them.
Careful though, the strap you pasted has a space in the middle. (this is correct: 777000000000000022CC1C00CEE55C46C0590E1532CD66090060070014051420FA8900A00300000 012123442C3353C19), if you google the strap of ubermix, you will find a ton of info about them.
|
|
|
|
|
Your cards (According to the bios) are Samsung Memory, so use UBERMIX 3.2 Straps (777000000000000022CC1C00CEE55C46C0590E1532CD66090060070014051420FA8900A00300000 012123442C3353C19)
Set mem clock to somewhere between 2100-2150, core to 1150
Voltage between 820-880, start from 880 and go down each time to see what works well with you (usually 850mv works well).
I never tried MSI cards, but all my samsung memory cards work at the parameters above and consume 70-80W as per GPUZ (Around 100W from the wall) and hashing at 31.5 mhs
|
|
|
|
35000W power is 35000W worth of heat that needs to be cooled; getting an AC to cool that is going to be very expensive. As the people above have said, you will need to separate the cold air from the hot air and ventilate it outside. This is what I've done in my mining room https://ibb.co/mBEbQ8https://ibb.co/b1tECoI used 2mm plastic sheets to partition the room from both sides; so basically my room is now partitioned to 3 sections (Right side: hot air, left side: hot air, middle section: the miners and cold air). The outer partitions (right and left) has vents on the ceiling to suck the air outside. Its a bit costly to set up, especially the vents/ducts, I'm sure it can be done in a cheaper way.. but it saved me alot more on cooling and the power consumption that I would've otherwise spent on additional ACs. My setup is a bit inefficient, as there are some spaces open in the plastic sheets where cold air is being exhausted without going through the GPUs (such as the spaces behind the motherboard and empty sections of the rig frames). Edit: just to give you an example of how much it can save you on cooling spends. My rigs consume around 45000W; theoretically, without separating hot/cold air, I would need 45000W worth of cooling power (or 155000 BTU AC). My setup allows me to keep my gpus constantly below 58 degrees with only 72000 BTU ACs (21000W of cooling), the ACs consume around 8500W of power. Otherwise, I would've required twice that amount of cooling. Where im operating, the weather is at 42C right now; so this is kind of an extreme scenario; where in your country, you would require less than I do |
|
|
|
All firmly seated, I have 2 1000W gold rated psus with 3 gpu's each so power is not an issue. I believe the new windows 10 update might have messed everything up. I tried to roll back windows and now I get a kmode exception not handled bsod on boot.
I'm gonna try reinstall windows from scratch at this point.
This might be worth trying. I had an issue with 2 of my rigs, sudden crashes to a point that I thought the cards might be bricked. Anyways, turns out that I had CSM turned on in the motherboard's bios,, once disabled, everything worked like a charm |
|
|
|
|
The whole point of an ASIC is that it is built to do a very specific set of functions/calculations.
They are much simpler than other chips such as a GPU or a CPU, and you can't change how it works by using a modified program/tool.
Mining ASICs are basically just a very powerful and very specific calculators.
|
|
|
|
Given that the OP was able to replace stock fans with a more power efficient one, i'm thinking of doing the same for all my GPUs. Electricity is more expensive from where i mine so this is would boost profitability considerably while keeping my GPUs cooler. 2 birds with 1 stone.
Just wondering, can the fans be placed back and the GPU restored to its original state incase resale is needed?
Any other pros/cons of doing the same as OP?
It is possible, the fans are usually attached to the shroud, and depending on the card, the shroud is usually attached to the card through some screws (the gigabyte ones I had only needed 4 screws, but will require a magnetic screw driver as it isnt easy to reach screws spots). |
|
|
|
|
when you increase the number of rigs you have, at some point you would wish to have some sort of a server/hub machine that can access all your rigs.
I invested in buying a decent server machine (just a normal computer with a fast SSD; got Awesome Miner for my rigs and SRR.
Cost was a bit high at the beginning, but the more rigs that are running, the more the cost was offset among them.
SRR is able to shutdown/turnon/restart all the rigs.
Awesome Miner is able to manage all the miners running and monitor them.
And through the server machine, I'm able to access any of the rigs. I got a vpn running on the machine and I'm able to access it remotely.
|
|
|
|
Hi, thanks for being so many who will help.
I have now disabled Windows updates.
I have run the Windows Ethereum mining registry tweaks.
Unfortunately, it did not help.
What is interesting is that the reboot PC at AM 00.04 so AM 01.02 so AM 02.09. And like this every night until AM 06.04, than it work until noon. PM 05.30 and one more reboot PM 08.00
And so its like this every day. I do not think PSU can reboot it every night at the same time.
i check Changing active hours but i can change it to max 18 hours.
Do you have some more idea what can be the reason.
Thank you.
I notice some of your cards are running at memory clock of 2200+ mhz, check in HWMonitor if any of the cards produce memory errors; if you got a card producing alot of errors, then the card is most probably crashing after accumulating alot of errors. What about core clock speed? are you leaving it on default? cuz that would be too high for mining, and wont increase hashrate alot; alot of power draw, heat, and stress on PSU can cause the crashes. (Assuming those are AMD Cards, clock speed of 1150-1200 is good enough, and some will argue to lower it even more). Check windows 10 reliability monitor (assuming you are using windows 10), its a nice tool, but you will need some googling to understand how to work with it to understand system crashes. Try reading the last bits of claymore log files for when the crash happens, maybe you'll get a hint if some card acts funny. Maybe hardware logging (I think gpuz has a log feature to write to a file), and check for when the crash happened, what was ur hardware status.. if its a gpu crash, you will probably see that the clock and memory speeds went down to P0 state before the crash occurs (usually 300mhz). |
|
|
|
Great solution, thanks for sharing!
Surprising the temps are lower with those fans
I have some of the 1070 version of that card (and some of the 3-fan versions), ALL of which have had fans die or are starting to lose fans. I have a bunch of NMB 92mmx25mm fans around - server-case fan design for about 40 CFM rated .35 amp at 12VDC. I've been pulling the entire fan shroud assembly, twist-tie one NMB fan to each heatsink assembly, then wire them up to the *5 VDC* connection on a MOLEX - they toss enough air through the HS to do BETTER cooling than the original fans did at up to 80% or so, while using a LOT less power. The original Gigabyte fans are JUNK on their current Windforce models - dunno why they moved away from ball-bearing fans a few years back except on the Aorus models, but it has cost them any chance of selling ME any more cards. The AORUS models still have good ball bearing fans - but they tend to be high-end models with extra power connectors that make it harder to power a bunch of them for mining usage. EVGA and Sapphire are still "all ball bearing, all of the time" and are my go-to brands - I just wish Sapphire made Nvidia cards. Those 92mm fans have a similar CFM to the ones I got. My Question though, you have only 1 fan attached to the card? and does it do a good job? so far I'm tying 2 fans to each card, which does an insanely good job at cooling, but I think one fan could be enough, especially that one of these fans produces a higher CFM than both stock fans added together. |
|
|
|
Hello All, Just wanted to share my funny experience with these cards I purchased around a year ago. The cards Im talking about are the Gigabyte RX 570 Gaming 4G: https://www.gigabyte.com/Graphics-Card/GV-RX570GAMING-4GD#kfThey are pretty much the worst cards I got for mining, even though I bought them early on for 160$ each, and I got 30 of them. At first they were great, amazing value for money and ROI, elpida memory, they do 29.5mh/850mh eth/dcr; and 900 h/s on cryptonight. They worked great for 10 months, and then the fans started failing; these fans can not handle dust at all. They are very poor in cooling; and unless I set the fans at 80% or more, the gpu temps would easily reach 75C. The funny thing is that half of my cards would have the rear fan failing, either completely stop working, or work at a much slower rpm; to the point that I had to only mine cryptonight due to the lower power requirement for the algorithm compare to dual mining. So I got this idea.. A week ago, I purchased these fans: https://www.aliexpress.com/item/Brand-Original-Delta-AUB0812VH-pwm-DC12V-0-41A-80mm-computer-cpu-case-axial-cooling-fan/32541022969.html?spm=a2g0s.9042311.0.0.Fy7LDHAnd these splitters: https://www.aliexpress.com/item/SATA-Port-to-10-Way-12V-3-4Pin-Computer-CPU-Cooling-Fan-Splitter-Hub-with-PWM/32842778420.html?spm=a2g0s.9042311.0.0.Fy7LDHRemoved the shroud and fans from the gpu, and stuck the new delta fans I got on the heatsink (using some metal wires, and a bit of clamping). and this is the result https://image.ibb.co/hSBvY7/IMG_20180213_WA0006.jpgCurrently, out of all the cards I have; Nitro+ Limited Edition, XFX GTS Edition, Nitro+ 570 8GB, and Powercolor Red Dragon golden sample; these cheap elpida cards are my favorite xD Temps on the gpu are at around 39C under the same conditions, and they consume less power (compared to the original fans, 12v @0.5A, the new ones are 12v @0.41A). So I lose the warranty, but I've already recovered their cost and much more, and they still run perfectly, so I am very happy  Now im thinking to do the same for all my other cards, and probably turn one of the AC units I got off to save on electricity Thats great idea tbh, have you tough about upgrading your GPUs like getting new ones or ? There's no point to replace the gpus; as long as they are running, mining and generating profit, then I keep them, and if I got the budget to get more rigs, then I just add to what I got. |
|
|
|
|
