• Competitor rules

    Please remember that any mention of competitors, hinting at competitors or offering to provide details of competitors will result in an account suspension. The full rules can be found under the 'Terms and Rules' link in the bottom right corner of your screen. Just don't mention competitors in any way, shape or form and you'll be OK.

Nvidia Is About To Steal The Cryptocurrency Mining Crown From AMD

Except you get into the general problems with logistics as above, anyone that has space for 2 rigs, has space for 2 rigs, meaning 2 x 280x based computers will yield more profit and more significantly it will yield more hashpower in important times like at launches of new coins. If you have space for 5 rigs, 5 x 280 rigs will do a lot more.

Ultimately the investment cost is next to irrelevant, any given "hash" will pay itself back at the same rate, one computer or 10 pay back in the same time, you just make 10 times more profit from the day after they pay themselves back.

Who would chose having a room or warehouse or anything full of 3 times as many setups, which also needs 3 times the amount of work, building, monitoring, testing, switching, etc, etc. It's more work for no reason at all.

No serious miner would build more slower rigs rather than less faster rigs, there is zero point in doing so.

As for efficiency, my 7970 doesn't use 250W and undervolts while overclocking significantly. Almost everyone who uses a 7970/280x is both overclocking AND undervolting to get less than 250W power usage. Likewise your numbers seem to bump up the 750ti from 250 to 300, a 20% bump over what I've seen stated.

As for the high end Maxwell, NO higher end card will be more efficient than a low end one. In reality at the lowest end you can simply choose the best balance for power usage, you pick a memory bandwidth and scale shader cluster numbers to perfectly use it. At the high end, say with a 7970, you need more than a 256bit bus so you spend power on a 384bit bus but don't saturate it. That will never be as efficient but with a 256bit bus it would be woefully unbalanced. High end you have to deal with max die size vs yields vs power and compromise low end you don't have any hard limit so you just make the most effective card in that bracket possible.

The likely hood is that if the newest to date card, that don't forget was launched in Jan 2012 only JUST, and I really mean JUST manages to match the 7970 for mining efficiency..... where do you think AMD's next gen might be. In fact, given their mining capability it could be an area they've been thinking about architecturally in that past two years for all we know. The next gen might spank the current gen silly. Nvidia's almost certainly most power efficient card on their next architecture is only on par with AMD's two year old architecture, that paints a picture of an extremely low likely hood that they'll match AMD's next gen architecture.

i agree with "No serious miner would build more slower rigs rather than less faster rigs, there is zero point in doing so." upto a point then it doesn work or miners would only be buying the fastest amd card out there and 280x's wouldn't be changing in price or in demand
 
That hardware doesn't compete with GPU mining. Dedicated miners with huge hashrates are SHA-256 for bitcoin. The difficulty is already insane and only the latest miners can really keep up with the difficulty. Obtaining a new Asic soon is near impossible.

There have been attempts at Scrypt Asics but they don't have the value of most AMD cards when it comes to Scrypt mining after you consider the likely possibility of waiting months before receiving it. The 5gh/s miner by alphatech may cost $1,400 and seem amazing on paper but if you don't receive it for several months while difficulty is rising, then your profits go out the window.

Even if you do get it and your hashrate is still profitable, many other people will have them and the difficulty will be insane soon enough.

Then you take into account re-sale value. You will have no problem reselling any decent AMD GPU to gamers once it is not profitable mining with it. Once your dedicated miner is unprofitable, it essentially becomes a once expensive, electric powered brick.



5GH by Butterflly Labs cost $274 and then there is large Quadro Looking GPU's type that do 600GH but cost $2196

I found the link not sure if allowed here.
 
i agree with "No serious miner would build more slower rigs rather than less faster rigs, there is zero point in doing so." upto a point then it doesn work or miners would only be buying the fastest amd card out there and 280x's wouldn't be changing in price or in demand

It's a mix, just like GPUs for games. You buy what's more suited to your needs, resources and what's on the market (price and availability).
 
I remember when Bitcoins were given out freely but never really read in to it and since when did this AMD mining craze hit as its only recently I read about it (later 2013).

It is a craze like many before it and some who got in early enough will be lucky but it will end and personally I think the best times on it has now past.
 
They can have 1000Gh/s miners available for order today. Wont make a difference if these companies purposely dont ship out for months. Maybe though, we will see them ship a lot faster since it isnt just one company calling the shots and within a month or two the difficulty will be near enough BTC difficulty.

As for 750ti vs 280x for mining. I would imagine the 280x would have better resale value, which is just as important as the cost of the rigs for miners.

I stopped mining after i had maxed my hash rate and got bored because i had nothing to tweak. Must admit, trading cryptos was way more fun and profitable than mining them.
 
No serious miner would build more slower rigs rather than less faster rigs, there is zero point in doing so.

Equal or greater ratio of KH/s out to wattage in, plus significantly lower heat output, two very good reasons hence why serious miners DO build more slower rigs rather than less faster rigs. Once startup costs have been covered KH to watt is the only important factor. Also don't forget that the 6x750ti mining rig would take up less space than the 6x280X one.


Nvidia's almost certainly most power efficient card on their next architecture is only on par with AMD's two year old architecture, that paints a picture of an extremely low likely hood that they'll match AMD's next gen architecture.

300 KH/s at 66w is efficiency of 4.55 KH/s per watt, that's more than "on par" it's significantly better than anything AMD currently has barring the HD7990.


Nvidia is too late asics rolling out in march.

750ti 250 kh 60W

Asic 1000 kh 23w

ASIC's don't matter for GPU mining, the reason people mine Scrypt today is because SHA256 ASIC's made GPU mining SHA256 unviable. If ASICs make GPU mining Scrypt unviable then GPU miners will move onto new coins that Scrypt ASIC's cannot do like Vertcoin/Panda/etc.

ASICs hashing the hell out of Scrypt won't mean ASIC owners make megabucks it just means the difficulty on Scrypt coins will moon like SHA256 coins.
 
Because each thread (2 of them) are addressing different topics, latest GPU news topics from different sources.. Although related to the 750ti wouldn't be read if all in one place. The amount of posts per thread shows that.

So there are the last 4 threads then on the GTX750TI?? Since that is a GTX750TI vs console thread,a GTX750TI cryptocurrency thread,a GTX750TI review thread and a Maxwell thread with the GTX750TI in it.

That console one alone could have been put in the Maxwell thread or the GTX750TI review thread:

http://forums.overclockers.co.uk/showthread.php?t=18543967
http://forums.overclockers.co.uk/showthread.php?p=25871557

This cryptocurrency one could have been added to this one:

http://forums.overclockers.co.uk/showthread.php?t=18581542

That is a GTX750TI thread started a few days ago in the mining section.

Adding that information to existing threads means people who are interested can read about them. Otherwise a bazillion threads about the same thing just dilutes everything.

Remember the mahoosive GK104/Kepler thread?? That one included several other threads merged into it.

Lots of people have had their threads merged into a single one,I don't see why you should be treated differently.

Three of those threads are started by you. Hmm.
 
Last edited:
Equal or greater ratio of KH/s out to wattage in, plus significantly lower heat output, two very good reasons hence why serious miners DO build more slower rigs rather than less faster rigs. Once startup costs have been covered KH to watt is the only important factor. Also don't forget that the 6x750ti mining rig would take up less space than the 6x280X one.

300 KH/s at 66w is efficiency of 4.55 KH/s per watt, that's more than "on par" it's significantly better than anything AMD currently has barring the HD7990.

Again I'd like to know where 300kh/s comes from, it does 240kh/s at stock and 285kh/s overclocked, neither is 300kh/s and I've not seen power usage listed at 285kh/s neither provides 4.55kh/s and once again you've ignored thata 7970 at 750kh/s at 200W is pretty common.

Likewise I'd love to know how a mobo, case, psu and 6 gpu's takes up less space if the gpu's are 750ti's rather than 280x's. Apparently Nvidia are so good then transcend normal space and time, both achieving results I've not yet seen(300kh/s) and space that is not achievable in my reality.

241kh/s at 65W is 3.65kh/s, 750kh/s at 250W is 3kh/w, 750kh/s at 200W is 3.75kh/w. I'm not sure how much 750ti's undervolt, I would expect not much and due to the way voltage and power scale the saving on a 65W card would be significantly less than that of a 250W card.

They are in ballpark the same area, and once the cards are paid back efficiency is NOT the only factor, in fact it's the least important factor, profit is the only factor.

If I can have one card that costs £1 to run and provides £3 profit a day, or one card that cost £4 to run but brings in £10 profit a day, that means for every card I run I make more money off the faster, more powerfull, less efficient card.

Given the choice between a 250W card that does 750kh/s and a 100W card that does 600kh/s then the difference might be worth it. But If I have space/power to run X amount of computers, profit while mining is profitable is all that matters. Having three times the profitability per rig is a ridiculously obvious choice. If the 750ti gave me at least extremely similar amount of cash per rig, meaning after power, almost the same pound per computer/gpu/slot, fine, it doesn't come remotely close though. Complete waste of time, and this is Nvidia's two year near architecture, two years, not a couple months. This is also compared to my 7970 which I bought 2 months after launch with a reference(or crap as it's known) cooler. If AMD respun the 7970/280x and completely optimised it with every new trick and advance TSMC had made, and saved every little piece of power possible, it would be absolutely wiping the floor with the 750ti even without a new architecture. Expect new gen AMD cards to make the 7970 kh/w efficiency look like crap and the 750ti's mining credentials to become slightly more non existent than they are now.

Heat, irrelevant, power, almost irrelevant, profit per slot per computer.... relevant, 750ti..... absolutely and completely outclassed by gpu's that can triple your output for very little extra power per kh.
 
Last edited:
Just for you DM ;)

NVIDIA Coin Mining Performance Increases with Maxwell and GTX 750 Ti
Subject: General Tech, Graphics Cards | February 20, 2014 - 05:45 PM | Ken Addison
Tagged: nvidia, mining, maxwell, litecoin, gtx 750 ti, geforce, dogecoin, coin, bitcoin, altcoin

As we have talked about on several different occasions, Altcoin mining (anything that is NOT Bitcoin specifically) is a force on the current GPU market whether we like it or not. Traditionally, Miners have only bought AMD-based GPUs, due to the performance advantage when compared to their NVIDIA competition. However, with continued development of the cudaMiner application over the past few months, NVIDIA cards have been gaining performance in Scrypt mining.

The biggest performance change we've seen yet has come with a new version of cudaMiner released yesterday. This new version (2014-02-18) brings initial support for the Maxwell architecture, which was just released yesterday in the GTX 750 and 750 Ti. With support for Maxwell, mining starts to become a more compelling option with this new NVIDIA GPU.

With the new version of cudaMiner on the reference version of the GTX 750 Ti, we were able to achieve a hashrate of 263 KH/s, impressive when you compare it to the performance of the previous generation, Kepler-based GTX 650 Ti, which tops out at about 150KH/s or so.

View Full Size
As you may know from our full GTX 750 Ti Review, the GM107 overclocks very well. We were able to push our sample to the highest offset configurable of +135 MHz, with an additional 500 MHz added to the memory frequency, and 31 mV bump to the voltage offset. All of this combined to a ~1200 MHz clockspeed while mining, and an additional 40 KH/s or so of performance, bringing us to just under 300KH/s with the 750 Ti.

View Full Size
As we compare the performance of the 750 Ti to AMD GPUs and previous generation NVIDIA GPUs, we start to see how impressive the performance of this card stacks up considering the $150 MSRP. For less than half the price of the GTX 770, and roughly the same price as a R7 260X, you can achieve the same performance.

View Full Size
When we look at power consumption based on the TDP of each card, this comparison only becomes more impressive. At 60W, there is no card that comes close to the performance of the 750 Ti when mining. This means you will spend less to run a 750 Ti than a R7 260X or GTX 770 for roughly the same hash rate.

View Full Size
Taking a look at the performance per dollar ratings of these graphics cards, we see the two top performers are the AMD R7 260X and our overclocked GTX 750 Ti.

View Full Size
However, when looking at the performance per watt differences of the field, the GTX 750 Ti looks more impressive. While most miners may think they don't care about power draw, it can help your bottom line. By being able to buy a smaller, less efficient power supply the payoff date for the hardware is moved up. This also bodes well for future Maxwell based graphics cards that we will likely see released later in 2014.

Continue reading our look at Coin Mining performance with the GTX 750 Ti and Maxwell!!

To illustrate this example, we put together two builds of mining computers that should be capable of similar hashrates:

R9 270X Mining Rig GTX 750 Ti Mining Rig
Processor AMD Sempron 145 - $55 AMD Sempron 145 - $55
Motherboard GIGABYTE GA-990FXA-UD3 - $135 GIGABYTE GA-990FXA-UD3 - $135
System Memory Kingston Value RAM 4GB 1333MHz - $40 Kingston Value RAM 4GB 1333MHz - $40
PCIE Riser Cards 1 x 16X to 1X Converter - $10 3 x 16X to 1X Converter - $30
Power Supply Corsair CX750 Builder Series - $80 Corsair CX500 Builder Series - $50
Graphics Cards 4 x Radeon R9 270X - $1200 6 x MSI Graphics Cards N750Ti - $990
Price $1520 $1300
In these two builds, the core platform stays the same, with the AMD Sempron 145 Single Core processor. While this processor would be essentially useless for a lot of other tasks, Coin mining on a GPU is not a CPU intensive task, so we can get away with one of the cheapest CPUs on the market.

View Full Size
We chose the Gigabyte 990FXA-UD3, as it was the cheapest motherboard we could find for this platform with 6 PCI-Express ports.

As you have probably noticed not all of the PCI-E ports on this motherboard allow for a x16 card to be plugged in, and even the ones that are capable don't have proper spacing for 2 slot cards. To remedy this we have included the appropriate adapters.

Due to the fact that we are using PCIE risers, there is no case included. You would most likely be best served by building an open-air test bed for the system out of milk crates, shelving systems, wood, or some other building material. Just remember, it doesn't have to look pretty to be effective!

There is also no storage option included. For something like this you could either use any spare hard drive you have laying around, or even install a Linux distribution to a thumb drive. Due to this, we found storage to be a negligible option.

First, we have a more traditional build, using 4 x R9 270Xs, which we found available on Amazon right now for just above $300 each. With 4 of these cards running at about 450KH/s each, we should have a 1.8MH/s machine. With a power draw of 150W from each card, we get a total of 600W just for the GPUs alone. Throwing in another 75W for the 45W TDP processor and any additional overhead, we come to approximately 675W power draw for our entire mining rig.

At a total cost of around $1520, this machine would have a payoff period of about 113 days at the current Dogecoin rates, at 1.8MH/s

View Full Size
Our second build is based on the GTX 750 Ti. This time we instead opted for 6 x 750 Ti cards for a total of $900, which is still significantly lower than the $1200 for 4 270Xs. With 6 x 750 Ti cards, the estimated GPU power draw would only be 360W, just above half of the power draw of the 270X machine. Adding in the same 75W for additional system components the total estimated power draw works out to 435W, which allows us to purchase a cheaper power supply.

At a total cost of around $1300, this machine would have a payoff period of about 97 days at the current Dogecoin rates, at 1.8MH/s

View Full Size
As you can see, by cramming more of the lower end but impressive GTX 750 Ti's into a single machine you can create a similar performing machine for less money than the AMD alternative, which is contrary to all the advice given about coin mining up to this initial release of Maxwell. In addition, performance of the Maxwell-based machine should only improve as the Maxwell kernel for cudaMiner is developed further, whereas OpenCL performance for AMD mining has likely been as optimized as we will ever see it.

An additional factor you have to keep in mind is the fluctuating cryptocurrency market. Just because the payoff estimates today say you could be making a profit in 80 days, doesn't mean that will remain the same in the future. While the estimate could get better, it also could get a lot worse, leaving you with a lot of hardware to sell off in the future.

While no one is sure where the mining market will be as far as profitability is concerned when the high end Maxwell GPUs hit the market, NVIDIA could have a similar stock issues and an inability to deliver GPUs to gamers as we see AMD having today.

42d0a7ea1356950d44b79a341ec7bfda.jpg


b4233265cb978bf1a0b32ef6e8172214.jpg


1a3fe9ffdbd219da474574ef35f9fd9e.jpg


eb86ca99e679c85c81b28edd3e9f9d8f.jpg


It seems you are completely wrong DM.
 
if OCUK rape their customs on pricing due to demand then i wont buy from them again. I have spent thousands over the years but that will be it.

mining hurts gaming through inflation.
 
Based on what I’m seeing with the 750 Ti, Nvidia is poised to embarrass AMD in the performance-per-watt race

What's new? :p

AMD have been an embarrassment (compared to Intel/NVidia) for a few years when it comes to performance per watt.
 
this is being looked at in a weird way
if you went by this the 260x would be the #1 mining card right now
its not because you would need to build 3x the rigs

space has a cost too
if the performance/cost ratio of nvidias more powerful cards is the same then yeh they will be popular but it never works out that way, there's always a price premium
 
Again I'd like to know where 300kh/s comes from, it does 240kh/s at stock and 285kh/s overclocked, neither is 300kh/s and I've not seen power usage listed at 285kh/s neither provides 4.55kh/s

300 KH/s divided by 66 watts, is 4.55 Kh/s per watt.

265KH/s is "the norm" for a stock 750ti, 300KH/s is "the norm" for an over clocked 750ti while still staying within it's TDP. If you see reviews with lower numbers (I.E Tomshardware) it's because they are using older versions of Cudaminer (hence other Nvidia cards also scoring lower than normal).

Also 300KH/s wasn't the hardware limit of the reference card, it was the most that could be achieved with the MHz slider maxed out in PrecisionX. Third party cards apparently allow higher clocks with some people claiming 320-330KH/s. And lastly The author of Cudaminer has ordered some 750ti's so he can start on Maxwell optimisations and doesn't want to promise anything but has said he would be very surprised if the 750ti doesn't end up sitting north of 350 KH/s when all is said and done.
 

The Maxwell thread will incorporate all Maxwell designed cards as that is what it is for..

The other threads were about latest GPU news, that had GTX 750ti but were more about Nvidia recognizing mining as a another reason for gamers to move from AMD.

The console thread was about how a very cheap GPU could match / beat the 'next' gen consoles.

You already have a GTX 750ti review thread that you put up, those other news wouldn't be right in their..
 
this is being looked at in a weird way
if you went by this the 260x would be the #1 mining card right now
its not because you would need to build 3x the rigs

space has a cost too
if the performance/cost ratio of nvidias more powerful cards is the same then yeh they will be popular but it never works out that way, there's always a price premium
Yeah looking at my recent cost I spent about £400 odd for a 6 card, 1kW base unit, before the cost of GPUs is added.
 
some very rough numbers, but they need to take the average amount of cards a mining pc would have (3-4?) and divide the power usage without a gpu by that, so if it was 120W for the cpu alone (a guess) with average of 3 (best you can do without a special mining case) cards which would give 40W extra per gpu. Then add this to the highest and lowest power usage cards on that graph to get an idea on how inaccurate the gpu only figures are:

750 has 60W normally, 100W after including overheads which moves it from 4.38kh/W to 2.63kh/W
the 290x 300W, 340W after overheads which moves it from 2.79 to 2.46

In this case the 750 still wins out on electricity cost, but the difference is nowhere near as big. You would still have to buy 3 times as much overhead equipment for this reduced rate.
 
Back
Top Bottom