Topic: CCminer(SP-MOD) Modded NVIDIA Maxwell / Pascal kernels. - page 397. (Read 2347659 times)

sp_ lbry miner will always begin mining for sp_ then for you, on a cycle (2mn for him first, then 120mn for you, rince & repeat) so it will always begin by his fee, that's normal and to counter people that cut the miner out before fee is starting. (as saw on many occasions)

You own mining isn't locked to any pools. It's free to use on any pools you wish.

I like to make shit up to perpetuate my name too! Guess whatever works for you... On the same page as Chryo. I don't suppose you got this idea last year about a month after the release of 1070s when I was talking about all the different models of 1070s I was working with to find the best one and made posts on here?

Oh look, the date on your article in your profile is August 2, 2016... Weird... When were the 1070s released? June 10th... Hmmm...

But, you know, I'm sure no one notices what goes on around here or correlations between activity.

Here you go: https://bitcointalksearch.org/topic/m.15696535

Just to add flavor to the discussion, talking about the shit fans on the 1070 Gigabyte models. We can also talk about the shit Zotac models I pointed out with the absolutely terrible coolers that are now dying on people, the MSI cards that spontaneously exploded, or more recently the EVGA ones (which they've pretty much fixed).

1.66% fee.

2 minutes for me 120 minutes for you.

Hey @sp_ - I tried downloading your LBRY sp_mod #5  https://github.com/sp-hash/ccminer/releases/tag/LBRYSP5 and I tried running it with my regular .bat file for the suprnova pool but it just starts up mining to lbrmine.zpool.ca, with I assume your own wallet address.

So is this just a demo so you can sell an "unlocked" version for a "donation"? It only actually does a few percent better than the git-truvpot version I was already using anyway so meh.

What is the point of "releasing" something locked to a certain pool or whatever? Do I have to know how to compile from source to add my own pool URL, username, password, etc?

are there any LBRY (LBC) miners (private or otherwise) that perform better than epsylon's public one?

It's offtopic but:no; it is great that we can go to -50, already. And I don't know for any bios mods for it. What you want, work @ 10%?

Question :

I have gtx 1070 gigabyte 1 single 8 pin power connection . I use Afterburner for clock gpu and mem , and for set power limit to 50% for example .

Its possible set under 50% power limit ? With afterburner no .

yeah but the pool mine is not modded with sp mod(you get far less hash), and it doesn't even work last time i've tried

I see. I never had cards crashing doing SpreadX11 with my 'HSC' though. Nor they ever boosted to 2000+ Mhz with a basic OC so I'm not sure what you mean.

As much as I like SPR, it really is nothing extraordinary. Poolmining is possible for a very long time (https://spr.suprnova.cc/) so there's that.

New coin. Pool on suprnova, new algo - sha256t - triripple  
Can _sp make ccminer for this, plz...  

As always you've got the crystal ball. I've forgot it.

Yeah, I haven't got experience with vga and kernels. Tested more than 500 vgas since 2006 and in the scene, on a daily basis, since early 2014. Good continuation with your (useless) chats

@amph: sjs I suppose

it's not a g1 gaming, but a palit

Yeah and a 1060 costs more AND isn't supported. Software is part of the game dude. You can't just say 'oh if this where XYZ it'd be better', truth is it isn't. I started mining with Nvidia hardware because when I swapped AMD was very poorly supported and Nvidia had a ton of open options and community cooperation between developers. That's pretty much gone.

No, old information doesn't mean it's currently better. Prices change. I am one of the original people who argued about density close to two years ago in this very same thread. A 1060 has the same density as a 480 so that completely throws that argument out the window. A 1070 costs 2x as much for not anywhere close to 2x the density, density being hashrate.

You can't compare three different video cards to one and pretend they're the same thing. IF you're arguing for a 1060 vs 480, then you have to keep going with that. If you add up the bonus density for a 1070 and similiar cost of a 1060, it doesn't mean they're better then a 480. You can't make a mutated version between a 1060 and 1070, it doesn't even make sense.

A 1060 is a shitty version of a 480 right now. Only reason people are buying them is because they don't want to switch to AMD. What Nvidia used to do better then AMD, it doesn't anymore due to over saturation of Nvidia hashrate. There isn't a 'niche' for them anymore. 1070s and 1060s are currently inferrior to 480s, there isn't even a argument to be made here. You can essentially get 1.4x as much profit by buying a 480 compared to a 1060 and 1.5 compared to a 1070. That depends on your electricity, but that's at US average of $.1KWH.

That means a 480 is 50% better then a 1070 and 40% compared to a 1060. 1060s can't dual mine PASC and they don't have the same support as AMD hardware right now. Completely putting aside you can spend a lot of money and get really good bins out of Wolf0.


Algo switching can crash your GPU driver, that is entirely possible. Sometimes you need a full reboot to fix the issues. This usually happens at some bleeding edge clocks. I also have some cards that are factory OC'd to the point where 'light' algos will crash them when you either switch to them or off of them as the factory OC isn't entirely stable and the algo isn't drawing enough power to push down the clock through TDP. One of my Asus O8Gs does this.

Library is currently really good at crashing my drivers even though I can mine for days on them without any problems, it happens when switching.

Otherwise, most everything Nikolaj is talking about doesn't really make sense. The crashing part when switching algos does though.

I've got 9 types of gtx 1070's.. This is the basis of my statements in matter. You can't generalize with a single tipology of vga. I've got too your g1 model, but the others are very different. It's also worth to be considered that the real efficiency of gtx 1070 derives from the wingle 8 pin PCIE versions, because ot the capped TDP amount.

An sjs or amp extreme can lift higher than 1800/1850 MHz on average (from 8.4 to 8.7MH/s eventually) but with very bad efficiency (up to 220W in certain scenarios, opposed to 155W of a single 8 pin one).

These are the reasons why I am not interested in models that have got 2 power connectors.

From 8.4 to 9.4 we've got a theroretical performance increase of 11% on spr, with a giant loss in efficiency. The same happens with lbry, where it's even worst (less % gain in performance, similar efficiency loss)


Good for you, but these aren't day to day profit analysis . I would be cautious with lbry if I were you, btw. Holding this coin have brought a massive loss to many. It've been a lot better to sell on a daily basis


The algo switching it's internal to the SpreadX11 algorithm (hashing functions). It's not about closing ccminer . In a multigpu environment, with an high number of models and versions, it's hard to find the highest stable clock (let's call it HSC), and this leads to the need of a proper testing, which is a bit complex given the absence of a pool in this case (due the peculiar decentralized mining process, one of the key elements of this amazing project). Spr it's very undervalued, the technology behind it it's one of the best concepts in the markets.

No need to  dump your mined coins every day..
My farm have had more than 10% of the lbry hashrate for a while now. I might sell when the price goes up 300-1000%

Different models have different stable clocks though.

Algo switching doesn't crash cards, unless you use an OC that won't work with some algos. That is why I use the lowest stable OC.

Otherwise, algo switching doesn't push your card to 2035, there's no reason to do so. You might not close ccminer properly before an algo switch maybe?

it doesn't crash for me it run stable at 9500KH 2000-2100 core, but consumption is too high, i prefer 9MH at 18xx core(already few days running at this rate, perfectly stable), the onyl thing is the stupid noise...

You're talking about trading, not mining (daily revenue).

What you fail to realize is that in a multigpu environment, up to 6/7 vga in a farm, the clock frequency on average is far less than the one you always report with a single-gpu screen.

With the algo SpreadX11, if you go past 1800/1830 MHz on avg it crashes; the reason is the algo switching, that forces the chip to go past 2035 MHz, and it hangs.

Every algo is different, this one it's a bit complex to manage.

do it with a 1060, with dual mining eth+pascal, the earning would be the same as one 480 and they cost the same

i don't care about how much they cost in other country, here they cost the same around 200 euro

we all know that a 1070 is better because of density you need to pay for this bonus...in fact the 1060 already is better in roi than a 1070, you don't even need to look at amd to beat the 1070

You can't just copy what someone says and decide that it's also applicable against them. You were arguing that Nvidia was still very much relevant as far as hashrate goes... That's not true. Even if we had a Pasc+Eth miner (which we don't and also should be factored in that Claymore supports AMD first and foremost), the cards would still be 2x as expensive for not even close to double hashrate.