Fans stuck on my 7950
-
Those 750tis, someone on another Forum did a comparison of the PCB layout between the different manufacturers. From what I can remember, he suggested the EVGA might not be ideal for 24/7 mining, something about ironchokes and single phase??? Same guy praised the ASUS, but others later on in the thread claim they are less clockable.
EVGA do have a good warranty though.
-
Those 750tis, someone on another Forum did a comparison of the PCB layout between the different manufacturers. From what I can remember, he suggested the EVGA might not be ideal for 24/7 mining, something about ironchokes and single phase??? Same guy praised the ASUS, but others later on in the thread claim they are less clockable.
EVGA do have a good warranty though.
I’ve experienced the same with the ATI cards. The Asus 270x has a glass jaw where the Sapphire 270x goes hard…
-
I’m curious as to why the ASUS won’t clock though. It’s arguably the better built card.
-
-
I’m curious as to why the ASUS won’t clock though. It’s arguably the better built card.
It’s loads quieter at comparable fan speeds, I wish they would as well…
-
Asus at least for 280X seems to be set really high in voltage, and have good overclocking headroom, for someone trying to control the power handling however…
That being said, at least with Asus you dont get “reference” power phases, which kinda sucks (at least for AMD), i have melted 4x Gigabyte 7970 OC and i know someone that has melted 12 of em, same place that gets broken on all the cards.
I would however say that i actually pr today prefer the sapphire toxic versions.
As to 750ti i have no idea… :) Still contemplating about building an seccond park with nvidias…
Im thinking EVGA however, just got an feeling. with no hard proof to back it up ;D
-
Sven I’m so glad you are around to give advice, you may have just saved me and my investors a $50k mistake. So what’s the deal with the 750ti? My local IT supplier is about to get 50 of these in stock:
http://www.pccasegear.com/index.php?main_page=product_info&cPath=193_1617&products_id=27066
Same deal as ATI cards, can fit 6 on a mobo with risers just use CudaMiner instead?
I would however go for the single fan version… Why dual fan? dont overclock, get 250-260khash with low heat, allmost no noice, and allmost no power consumption.
Reviews have shown that for scrypt mining at least, your looking at quite the accelleration in regards to power usage 250khash vs 280khash, at 250khash the card use 38w!, with 280, your looking at 65watt! Thats not much
but think cooler operation, less stress on card, longer lifespan and so on.
Also out of experience my 7970 rigs back in the day one quad 7970 rig (gigabyte triple fan design thingy), the difference in poweruse total on the rig varied from 1350w to 1100w. depending on current fan speed (1350 at 85% speed).
Now if you add that to the mix and looking at its supposed to be LOOOOW consumption, adding an extra fan on it… Which isnt needed given that the reviews states around 56degrees running scrypt.
Just my 5 cent ^-^
Edit:
Just realised that my writing was a bit messy now, tired and sleepy… let me know if it sounds gibberish and ill rephrase.
-
Need to moderate myself btw.
With gridseed you can mine, scrypt, sha 256 or both btw.
But still, the gold carrot imo is the scrypt mining, cause look at the sha256 diff… soon enough it wont be worthwhile to mine sha256 with gridseed. Also the power usage goes through the roof on sha256 mining on em.
-
Need to moderate myself btw.
With gridseed you can mine, scrypt, sha 256 or both btw.
But still, the gold carrot imo is the scrypt mining, cause look at the sha256 diff… soon enough it wont be worthwhile to mine sha256 with gridseed. Also the power usage goes through the roof on sha256 mining on em.
I’ve got 40 gridseed incoming, as an experiment. I don’t like them because they look Chinky and I hear they crash every few days, which I can live with at 7w power consumption per unit. My only concern is, if some of the heavyweight scrypt currencies like Lite/Doge/Feather change their algorithm, I don’t want to be left with paperweights.
Which I doubt anyway, there are a lot of people that think ASICS are good for the eco system…
I will miss my GPU rigs if ASICS take off, there is something romantic about building a miner from a milk crate and a couple of planks of wood…
-
Remember, you only need a small modification on scrypt algo to make it resistant against scrypt-asic. adjust the ratio on thread/mem, make the algo not fit on an single thread etc. theres a lot of small tweak that will make on or the other unefficient.
Thats why i dont risk investment in scrypt-asic, or asic (basiclly cause its an ratrace on reaching ROI before next gen comes along…)
But only my personal opinion. Hopefully feathercoin will move away from scrypt algo soon, really dont like scrypt anymore. GPU killer number uno.
And an coin only really suited for specialized equipment makes it more a novelty item for the few.
-
Remember, you only need a small modification on scrypt algo to make it resistant against scrypt-asic. adjust the ratio on thread/mem, make the algo not fit on an single thread etc. theres a lot of small tweak that will make on or the other unefficient.
Thats why i dont risk investment in scrypt-asic, or asic (basiclly cause its an ratrace on reaching ROI before next gen comes along…)
But only my personal opinion. Hopefully feathercoin will move away from scrypt algo soon, really dont like scrypt anymore. GPU killer number uno.
And an coin only really suited for specialized equipment makes it more a novelty item for the few.
Could you stop scaring me please!
OK, so you think Litecoin is going to change it’s algorithm? Because I could always just point them at a Litecoin pool…
-
Remember, you only need a small modification on scrypt algo to make it resistant against scrypt-asic. adjust the ratio on thread/mem, make the algo not fit on an single thread etc. theres a lot of small tweak that will make on or the other unefficient.
Thats why i dont risk investment in scrypt-asic, or asic (basiclly cause its an ratrace on reaching ROI before next gen comes along…)
But only my personal opinion. Hopefully feathercoin will move away from scrypt algo soon, really dont like scrypt anymore. GPU killer number uno.
And an coin only really suited for specialized equipment makes it more a novelty item for the few.
Not exactly, if Litecoin were to change its algorithms often.
Companies building ASICs will just sell FGPA (Field programmable gate array) chips, those are faster than GPU and it’s underlying code can be changed.
You also have the option to customize the board like adding more RAMs, USB ports, etc.
-
Tbh i really dont care what litecoin does, they didnt invent scrypt anyways, and as of today they are like all the other alt coins that use scrypt (still this is based om my opinion). Their only advantage is that they are like mt gox, when crypto gets mentioned litecoin tags along with the bitcoin name.
Please do not get affended, this is just my perception on the matter ^-^
Yeah, FPGA sure was sweet back in the SHA256 times before the asic came along and ripped em an new one, i seem to recall people converting em into mining scrypt, but that it wasnt at all worth doing cause low hashrate/high power consumption.
They will still face the same issue with how the algo/decryption work, for example, put the size of the algo above the size of the internal memory cache on an CPU, and you make it kinda useless as CPU is an serial processing unit. Then again you can reformulate that the other way around making serialized process with fast connection to mem an better value, and making those 2048 stream processors (7970/280X) basiclly running idle or the work cant be devided properly among the “cores”.
Theres alot of different ways to make it perfect for one or the other.
As of now you have the CPU camp, the ASIC camp, The scrypt-Asic camp, FGPA camp (basiclly dead at this moment though), and the GPU camp.
Now my recent experience is that the GPU camp seems to be able to handle the obstacles on way more different algo than any of the other equipment. Just look at Nvidia and how it has progressed in regards to handling different algos, better performance etc. all without having to reinvest. My bet at least is still GPU.
And at the end of the day, if everything fails, at least you have equipment worth more than just beeing used as an paperweight.
I for example have ROI’d all my cards, so everything from now is just gravy. The question is can i afford to reinvest one more time?
If i do im sure what camp im gonna reinvest in 8)
-
Tbh i really dont care what litecoin does, they didnt invent scrypt anyways, and as of today they are like all the other alt coins that use scrypt (still this is based om my opinion). Their only advantage is that they are like mt gox, when crypto gets mentioned litecoin tags along with the bitcoin name.
Please do not get affended, this is just my perception on the matter ^-^
Yeah, FPGA sure was sweet back in the SHA256 times before the asic came along and ripped em an new one, i seem to recall people converting em into mining scrypt, but that it wasnt at all worth doing cause low hashrate/high power consumption.
They will still face the same issue with how the algo/decryption work, for example, put the size of the algo above the size of the internal memory cache on an CPU, and you make it kinda useless as CPU is an serial processing unit. Then again you can reformulate that the other way around making serialized process with fast connection to mem an better value, and making those 2048 stream processors (7970/280X) basiclly running idle or the work cant be devided properly among the “cores”.
Theres alot of different ways to make it perfect for one or the other.
As of now you have the CPU camp, the ASIC camp, The scrypt-Asic camp, FGPA camp (basiclly dead at this moment though), and the GPU camp.
Now my recent experience is that the GPU camp seems to be able to handle the obstacles on way more different algo than any of the other equipment. Just look at Nvidia and how it has progressed in regards to handling different algos, better performance etc. all without having to reinvest. My bet at least is still GPU.
And at the end of the day, if everything fails, at least you have equipment worth more than just beeing used as an paperweight.
I for example have ROI’d all my cards, so everything from now is just gravy. The question is can i afford to reinvest one more time?
If i do im sure what camp im gonna reinvest in 8)
So, I gotta ask you Svenn. If you had $50k to invest, how would you invest it? At the moment I’m spending a bit on Bitcoin ASIC mining, a bit on gridseed ASICs and a bit on more GPU miners…
-
i would leave 10k to hire me some extra programming skills and get some of my ideas of the ground.
the rest i would use to create an 750ti hashing farm.
-
i would leave 10k to hire me some extra programming skills and get some of my ideas of the ground.
the rest i would use to create an 750ti hashing farm.
glad im on the right boat with the 750ti’s ;)
-
But dont listen to me.
follow your own gut instinct:) -
But dont listen to me.
follow your own gut instinct:)I just discovered keccak… Feels like I’ve been wasting my time…
-
i would leave 10k to hire me some extra programming skills and get some of my ideas of the ground.
the rest i would use to create an 750ti hashing farm.Im going to take your advice… Lets form a technical partnership… Ive messaged you…
-
I just discovered keccak… Feels like I’ve been wasting my time…
What coins are using the keccak algo?