Topic: CCminer(SP-MOD) Modded NVIDIA Maxwell / Pascal kernels. - page 491. (Read 2347659 times)

214Sols/s  POSSIBLE--

The coder djeZo just posted, on another thread, a screencap of a GTX 1070 achieving 214Sols/s.  He says that more is possoble.       --scryptr

I mined ZEC at the start on multiple pools trying to gauge them and the payouts for the nicehash pool were terrible days after the launch.

That's probably the main reason miners left their pool (last block found was 20 hours ago) the second being that their pool is too basic for today's standards; there are no stats for miners to get an idea of their earnings and no transparency.

If they would setup an MPOS pool (like suprnova or coinmine) and allow their miner to work there as well as on nicehash I would actually be interested in their miner and I'm sure I'm not the only one. That is if I'd still mine ZEC/ZCL - which I'm not.

we are already sure that a 1070 can do 180sol, because nicehash has a work in progress miner that reach that speed, still unreleased, but it will be for their pool only, so i'm not really interested

but this mean that 200 is achievable

ASK NERDRALF--

The following is a quote from NerdRalph about AMD:

"While my initial analysis was focused on the external GDDR5 bandwidth limits, current ZEC GPU mining software seems to be limited by the memory controller/core bus.  On AMD GCN, each memory controller can xfer 64 bytes (1 cache line) per clock.  In SA5, the ht_store function, in addition to adding to row counters, does 4 separate memory writes for most rounds (3 writes for the last couple rounds).  All of these writes are either 4 or 8 bytes, so much less than 64 bytes per clock are being transferred to the L2 cache.  A single thread (1 SIMD element) can transfer at most 16 bytes (dwordX4) in a single instruction.  This means a modified ht_store thread could update a row slot in 2 clocks.  If the update operation is split between 2 (or 4 or more) threads, one slot can be updated in one clock, since 2 threads can simultaneously write to different parts of the same 64-byte block.  This would mean each row update operation could be done in 2 GPU core clock cycles; one for the counter update, and one for updating the row slot.

Even with those changes, my calculations indicate that a ZEC miner would be limited by the core clock, according to a ratio of approximately 5:6.  In other words, when a Rx 470 has a memory clock of 1750Mhz, the core would need to be clocked at 1750 * 5/6 = 1458Mhz in order to achieve maximum performance.

If the row counters can be kept in LDS or GDS, the core:memory ratio required would be 1:2, thereby allowing full use of the external memory bandwidth.  There is 64KB of LDS per CU, and the AMD GCN architecture docs indicate the LDS can be globally addressed; i.e. one CU can access the LDS of another CU.  However the syntax of OpenCL does not permit the local memory of one work-group to be accessed by a different work-group.  There is only 64KB of GDS shared by all CUs, and even if the row counters could be stored in such a small amount of memory, OpenCL does not have any concept of GDS.

This likely means writing a top performance ZEC miner for AMD is the domain of someone who codes in GCN assembler.  Canis lupus?"

NerdRalf is not an nVidia programmer, but I think someone donated a couple of nVidia cards to him just for Zcash programming.  If someone could explain the hashing speed limit on nVidia hardware in similar terms, we would have a real answer.

I don't think memory bus and memory bandwidth are the same.  Some AMD cards are specified at different values for bus, and for bandwidth.       --scryptr

'The pool had a bad luck in the start, so the interest of miners in using this pool was slowly dropping...'

bathrobehero, any thoughts on this?

I think  IMO  it can reach even more than 300 s/s because of the blake algo rounds embedded in it.

No, that's just their crappy pool. If they had a proper MPOS pool instead of their basic NOMP implementation then it wouldn't be crap. I'm not bashing NOMP but it basically has zero transparency.



AMD wouldn't be any different without Claymore's fee based miner.

factor=greedy community and developers. Even nicehash hardcoded their binary private protocol to nv miner lol. I can understand if it was 200 s/s miner, but it is 135 s/s . I thing there will not be any free nv miner... like no nv miner for monero.

IMHO

The only limit is the fantasy .

I think one gtx 1070 is easy get 200+ sol .

why so? memory bus and bandwith are the same, what are the limited factors here?

no.

so, see

https://bitcointalksearch.org/topic/m.16965755

a gtx 1070 can reach 200 easily, based on this speed for amd by clymore

Nicehash closed pool zcash :

https://www.nicehash.com/?p=news&id=115

I hope the new miner for nvidia its free and work in all pool .

EVGA Addresses GeForce GTX 1080 FTW PWM Temperature Problems (Update 2: All BIOS Updates Available)
http://www.tomshardware.com/news/evga-addresses-geforce-1080-temperature-problems,32967.html

It's all about shitty nvidia community -(

Sp is it possible that you can build that so i can use it on my vista without  msvs 2015?  I upgraded the hdd 3 yrs ago and its missing a file that updates anything even to win 7.  thx

yeah but it will be closed and tied to their pool so useless, unless you can reverse engineering it

The opensource NVIDIA developers have quit. But Nicehash will release a new miner that does 180SOl/s on the gtx 1070 soon. (zcash)

What makes a technology great?  Support for that technology.  So who is winning right now between AMD and Nvidia?  Obviously AMD.  And this thread is the holy grail for Nvidia support. 

It's funny.  It's ironic.  It's cruel.  It's disgusting.  It's a lot of four letter words and it can all be summed up by the best four-letter word of them all.  That's:

LIFE

only good thing is that there are chance of a pump after the block hit the maximum, i can see this coming, but currently zcash is producing too many btc per day

this is the reason why it can not be pumped higher enough