Claymore's ZCash/BTG AMD GPU Miner v12.6 (Windows/Linux)

[Kimax]

I have changed the wiki page, so its no longer restricted to the Claymore V8.0, but instead is a comparison table for ZCash in general.

I copied the data from the old table to the new and added a column for the miner version.

Hope you will enjoy this and please add your own results to the table.

https://en.wikibooks.org/wiki/ZCash_mining_GPU_Comparison

[agarwalajay]

With regards to the gpu hashing at 1000 h/s its a radeon pro duo
R9 390 hashes at 300 h/s with 40 compute units

Pro duo has 128 compute units hbm memory and tdp 300 watts

Dear All, I am new to Crypto Currency. Hence please do not mind my limited knowledge.

If Radeon Pro Duo gives 1000h/s at 300 watts then it is better than rx 480 rig.
Both initial cost(as compared to multiple rx480 for same 1000h/s) and electricity cost are lower.
also there will be less investment on PSU.
Right??

[Bitbit017]

Go mine on Dwarfpool!!! No stealing shares  like other pools..... I switched from flypool to dwarfpool, and I have 15% more shares!!!(my real shares)

@bitbit017 .. are you sure about it? what speed you have ?  idk, but something is wrong there today... and im not much impressed about shares...  my increase is just in the rejected ones, as i moved from 0,15%/0,25% to actual 0,76%/1,31% of rejected shares...

also my income is not impressive atm.. after 13h have just 0,072 ZEC with 2,7kh/s   lets see what will be the result after 24h.....  

does anyone else have some experience ??

Yes I'm right about that... My real power is about 11200Sol/s with v8!! I mined in  flypool from begining. I never saw a peak power greater than 10700sol/s in flypool. My average effective hashrate was about 10200sol/s in flypool. In dwarfpool my average hashrate it's about 11200sol/s or greater and it's exactly my power. The only thing that dwarfpool needs it's more users, to have more power to find more blocks.......
Flypool is PPLNS, dwarfpool is HBPPS

I used dwarfpool back in the ETH days for many reasons
But the fact is that dwarfpool lost the ZEC train out of complacency
Needs a lot more time and luck to be worth it to use dwarfpool for ZEC
Now it has 350 workers approx. and it mines 1-2 blocks per day

Let's see workers get to 5000 and then we can start thinking of using it again
Because it really sucks to mine all night with 2.000h/s and get nothing in the morning

Yes, you are right with that!
Dwarfpool needs more workers, more power. We can help with that....
Because to me it seems to be the fairest pool. On Eth was the same thing for me

[Assaro]

For me the Dwarfpool was one of the best ETH pools and pay close to calculator.  I have not tried ZEC pool but I think will also be one of the best.

not at the moment.... switched back to flypool, as i was disconnected by the dwarfpool and doent got jobs for more than 10 mins.... next i have to pay some bills, gave a shot for 24h and i got 0,16 zec which is approx 50% of normal earnings.

Also is there a high reject rate atm, had it after 24H from 0,94% - 1,09% which is little bit high...

they have to fix it and if it will be stable ill join the party gladly, as i was also happy with it on ETH... looking forward to get it fixed, GL....

[lexele]

yup even more impressive you can set the power limit down around -30 and still get around 260 h/s with i-4 and watage is down to around 120-130 watts
r9 nano and fury's in general are the undisputed kings of ZEC mining. Th00ber and Other fury experts tipped me off to this.

Let's see how long the ZEC price stays at this level.

Remeber too that eth price is depressed right now, it could shoot back up to 12 at any news then all the hash runs back to eth , either way miners win

Or It could continue to slide downhill... the curve is not really nice this times.

[KrokoTill]

Go mine on Dwarfpool!!! No stealing shares  like other pools..... I switched from flypool to dwarfpool, and I have 15% more shares!!!(my real shares)

@bitbit017 .. are you sure about it? what speed you have ?  idk, but something is wrong there today... and im not much impressed about shares...  my increase is just in the rejected ones, as i moved from 0,15%/0,25% to actual 0,76%/1,31% of rejected shares...

also my income is not impressive atm.. after 13h have just 0,072 ZEC with 2,7kh/s   lets see what will be the result after 24h.....  

does anyone else have some experience ??

Yes I'm right about that... My real power is about 11200Sol/s with v8!! I mined in  flypool from begining. I never saw a peak power greater than 10700sol/s in flypool. My average effective hashrate was about 10200sol/s in flypool. In dwarfpool my average hashrate it's about 11200sol/s or greater and it's exactly my power. The only thing that dwarfpool needs it's more users, to have more power to find more blocks.......
Flypool is PPLNS, dwarfpool is HBPPS

I can confirm it. Dwarfpool reports exactly the true hashrate and variable diff works also fine. It adjusts your worker diff to have ~200 shares per hour.

[Kimax]

With regards to the gpu hashing at 1000 h/s its a radeon pro duo
R9 390 hashes at 300 h/s with 40 compute units

Pro duo has 128 compute units hbm memory and tdp 300 watts

Please add your result to the GPU Comparison table.

https://en.wikibooks.org/wiki/ZCash_mining_GPU_Comparison

Thats is some insane speed.

[Altcoining]

Go mine on Dwarfpool!!! No stealing shares  like other pools..... I switched from flypool to dwarfpool, and I have 15% more shares!!!(my real shares)

@bitbit017 .. are you sure about it? what speed you have ?  idk, but something is wrong there today... and im not much impressed about shares...  my increase is just in the rejected ones, as i moved from 0,15%/0,25% to actual 0,76%/1,31% of rejected shares...

also my income is not impressive atm.. after 13h have just 0,072 ZEC with 2,7kh/s   lets see what will be the result after 24h.....  

does anyone else have some experience ??

Yes I'm right about that... My real power is about 11200Sol/s with v8!! I mined in  flypool from begining. I never saw a peak power greater than 10700sol/s in flypool. My average effective hashrate was about 10200sol/s in flypool. In dwarfpool my average hashrate it's about 11200sol/s or greater and it's exactly my power. The only thing that dwarfpool needs it's more users, to have more power to find more blocks.......
Flypool is PPLNS, dwarfpool is HBPPS

I used dwarfpool back in the ETH days for many reasons
But the fact is that dwarfpool lost the ZEC train out of complacency
Needs a lot more time and luck to be worth it to use dwarfpool for ZEC
Now it has 350 workers approx. and it mines 1-2 blocks per day

Let's see workers get to 5000 and then we can start thinking of using it again
Because it really sucks to mine all night with 2.000h/s and get nothing in the morning

Yes, you are right with that!
Dwarfpool needs more workers, more power. We can help with that....
Because to me it seems to be the fairest pool. On Eth was the same thing for me

I'm mining with Dwarf right now, the only thing you loose with Dwarf is payouts for the time being, their very slow because the number of found blocks is so low. Eth on there was the most fair pool I've ever used, and i think zec will be the same.

For contrast, Dwarfpool right now has 1.7% of the network, with 700k S/s.
Flypool has 35.68297116878886% of the network
Suprnova has >0.1%
Coinmine.pl has 15.96402616877132%
We need to distribute things more.

[Goruno]

How about the CPU mining software? . I have some servers to be free.
if possible , please give me the new link for that. Thanks

[Hotmetal]

yup even more impressive you can set the power limit down around -30 and still get around 260 h/s with i-4 and watage is down to around 120-130 watts
r9 nano and fury's in general are the undisputed kings of ZEC mining. Th00ber and Other fury experts tipped me off to this.

Let's see how long the ZEC price stays at this level.

Remeber too that eth price is depressed right now, it could shoot back up to 12 at any news then all the hash runs back to eth , either way miners win

Well, if ETH lands up being the coin to stay on, the Nano's aren't leading the RX cards by much.

[dArkjON]

This is what i get from v 8.0 with 6 x 470 RX 8 GB :

ZEC - Total Speed: 980.115 H/s, Total Shares: 43180, Rejected: 30, Time: 116:49
ZEC: GPU0 165.223 H/s, GPU1 163.291 H/s, GPU2 165.256 H/s, GPU3 161.303 H/s, GPU4 163.084 H/s, GPU5 161.958 H/s

[bmtrader]

This are 5x rx470 4G red devil + 1 r9 380 (GPU1)
All with Stock settings
win 10
h81 pro btc
4G ram

ZEC - Total Speed: 1027.998 H/s, Total Shares: 207, Rejected: 0, Time: 00:18
ZEC: GPU0 178.125 H/s, GPU1 138.477 H/s, GPU2 176.131 H/s, GPU3 180.199 H/s, GPU4 179.273 H/s, GPU5 175.810 H/s

[Trimegistus]

The ZEC dwarfpool is working flawlessly!

With 0% fee for the next days.

[Kimax]

This is what i get from v 8.0 with 6 x 470 RX 8 GB :

ZEC - Total Speed: 980.115 H/s, Total Shares: 43180, Rejected: 30, Time: 116:49
ZEC: GPU0 165.223 H/s, GPU1 163.291 H/s, GPU2 165.256 H/s, GPU3 161.303 H/s, GPU4 163.084 H/s, GPU5 161.958 H/s

Please add your results to the comparison table.

https://en.wikibooks.org/w/index.php?title=ZCash_mining_GPU_Comparison

[KrokoTill]

The ZEC dwarfpool is working flawlessly!

With 0% fee for the next days.

..was working. Now the vardiff port is down.

[Altcoining]

The ZEC dwarfpool is working flawlessly!

With 0% fee for the next days.

..was working. Now the vardiff port is down.

Working for me.

[KrokoTill]

The ZEC dwarfpool is working flawlessly!

With 0% fee for the next days.

..was working. Now the vardiff port is down.

Working for me.

zec-eu1..:3333 - no connection for me over an hour
..: 3335 works

and from pool statistics you can see that ~70 workers and 300k of hashrate is gone so I guess all of them were on eu vardiff

[Altcoining]

Hey, @Claymore I've just ran in to a problem?
My miner is spamming this hundreds of times per second.
GPU 1, GpuMiner kx(2) failed -36
GPU 0 is working fine...
This happens, so i restart the miner and it never starts until i reboot the machine.

[Altcoining]

The ZEC dwarfpool is working flawlessly!

With 0% fee for the next days.

..was working. Now the vardiff port is down.

Working for me.

zec-eu1..:3333 - no connection for me over an hour
..: 3335 works

and from pool statistics you can see that ~70 workers and 300k of hashrate is gone so I guess all of them were on eu vardiff

Ahh, I'm using the US server and it's fine, eu server is down here too.

[bardacuda]

On my 290 memory controller load is rarely over 60%. Big difference is that beside 290 memory bus being 2x wider, memory runs at 1250 MHz vs 2000 MHz on your 480. This means that you can do all possible tricks but no way can use that tight timings as on 290 at 1250 MHz or 390 at 1500 MHz. OK suppose that you reduce mem clock on 480 to 1500 or 1250 MHz to get the same timings but then you still do not get the speed that is possible with 2x wider bus.

This is not question of memory througput.
Reducing memclock almost twice affects only to 20% hashrate drop.
https://ip.bitcointalk.org/?u=http%3A%2F%2Fi.piccy.info%2Fi9%2Fce2e18589c91c75caab3b10a46a2c9f2%2F1480097808%2F34039%2F1051816%2Fmemdrop.png&t=571&c=zWfZWhuCJrGaPQ

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?

Core speed has more of an effect on 480s but they are still limited by memory bandwidth.

I'm very funny that you even still protect 290,390 here is about the memory of 4xx

According to AMD According to a CU efficiency increased by 15% compared to the Radeon R9 290. When processing tessellation in conjunction with heavy duty AA efficiency gains can be double or even triple. Supported data compression, thus improving memory bandwidth. In particular, supported by Delta Color compression algorithm that allows you to encode the color difference. On this technique we described in the description of the architecture NVIDIA Pascal. AMD has such compression is maintained including the Radeon Fury X, but the effectiveness of the algorithms at Polaris 10 above. With this increase in the efficiency of a data chip content bus word length of 256 bits. The Radeon RX 480 uses GDDR5 memory chips with an effective rate of communication 8 GHz.


And for that AMD has introduced new regulations of memory !!! such as FP16 and 16 Int.

Which I think Claymore's does not use, for this reason, the new data on the time of top 480 does not operate at full capacity. And at the same time using the old manual of memory with which he revived the old 7xxx to work at such speeds



And compared to the 7xxx, 290 and 390 may be given even greater speed, including 480 if you use the new instructions of memory that only support new models 290-390-480, though only suffer 7xxx model that greatly impact on mining in overall, since the data pattern immediately lose their significance in mining


http://i11.pixs.ru/storage/1/6/0/03amdradeo_6383058_24215160.png

I'm not "protecting" anything. I'm simply explaining why the Hawaii cards are faster and will remain faster than Ellesmere given that they have +50% higher memory bandwidth and this algo relies on memory bandwidth, and this has been reinforced by the devs that code these miners.

Memory bandwidth:

RX480: 256 GB/s
390X: 384 GB/s

ZEC uses a lot of memory operations. Do you still think that RX480 can work as fast as 390X?

I'm just letting you guys know so ppl stop whining in here that Claymore is somehow intentionally gimping the RX cards. Not only does he have no incentive to do that, he has a big incentive to do exactly the opposite. Otherwise, there's a large chance that people with RX cards are going to switch to Optiminer or ETH. In the case of Optiminer he gets no fees at all, and in the case of ETH he gets a much lower fee and that's only if they use his miner...which they probably won't because the no-fee miners like Genoil's are on par with his.

The fact that you quote AMD marketing stuff about color compression just shows that you are grasping at straws and wishing upon stars. Any speedup by these tweaks is only for changes in pixel colors in gaming scenarios, not compute tasks. You have to look at the "uncompressible data" part of the graph and you see that RX 480 gets no speedup due to delta color compression algos. Same thing goes for tesselation and AA optimizations in the compute units.

You also can't draw conclusions by underclocking the memory unless you use the same memory straps for every frequency...otherwise, every 125 or 250MHz drop the timings tighten, so a lot of the loss in bandwidth is made up in lower latency.