polaris is expecting to be released next month, so still in time, but first wave of gpu would be expensive as usual
i would look at the new 470 it should consume something like 40w and hash at 15MH at least
Topic: How will this change the world of mining?? GTX 1080 / 1070 - page 37. (Read 134091 times)
It was expected.
AMD wins in dagger-hashimoto because they have cards with wide bus - more mem controllers so faster random memory access/latency. I don't see how bigger page supported will help for "random" access
Obviously scaling on AMD is very very low after 1000-1200 shaders (7850/7870 265-270)
Its all because of bus width...
270 can make ~ 20MH/s with core:mem ratio of 2:3 (exp atm)
290 has 2x wider bus, 2x shaders... but afaik never was close to 40MH/s
Fury can make 35MH/s due to ver wide HBM memory.
So my prediction:
1070 - 20-24MH/s (the maximum for 256-bit bus)
Polaris10 - same lvl
AMD wins in dagger-hashimoto because they have cards with wide bus - more mem controllers so faster random memory access/latency. I don't see how bigger page supported will help for "random" access
Obviously scaling on AMD is very very low after 1000-1200 shaders (7850/7870 265-270)
Its all because of bus width...
270 can make ~ 20MH/s with core:mem ratio of 2:3 (exp atm)
290 has 2x wider bus, 2x shaders... but afaik never was close to 40MH/s
Fury can make 35MH/s due to ver wide HBM memory.
So my prediction:
1070 - 20-24MH/s (the maximum for 256-bit bus)
Polaris10 - same lvl
290 did get to 30MH though, so it was still scaling somewhat - just not 100% scaling vs. shader count.
Limit is definitely somewhere in the memory system though - and it's NOT just bus width, Fury/Nano have a 4096 bit wide memory bus (due to the structure of HBM) yet they're as fast as the 290 or about the SAME hash.
Genoil IIRC was speculating it was a limit in the TLB table hardware, but I dunno how far their research into that has gotten.
I don't see anything close to 20 MH out of my 7870s - the 270 has the SAME shader count, but it does have somewhat faster memory and IIRC 2x the memory bus width, but while it's certainly a good bit faster than the 7870 it's nowhere near 2X as fast, much less 2x PLUS the memory speed ratio.
TDP target on Polaris should be a LOT lower - the smaller node is a lot more efficient, even if you puch more transistors AND kick the clock rate up quite a bit.
On the other hand, it won't be out in time to do much Eth mining with, unless AMD pushes the release dates up a good bit.
I do hope AMD can release the Polaris earlier than November. Otherwise we will only have a few months to mine.
It was expected.
AMD wins in dagger-hashimoto because they have cards with wide bus - more mem controllers so faster random memory access/latency. I don't see how bigger page supported will help for "random" access
Obviously scaling on AMD is very very low after 1000-1200 shaders (7850/7870 265-270)
Its all because of bus width...
270 can make ~ 20MH/s with core:mem ratio of 2:3 (exp atm)
290 has 2x wider bus, 2x shaders... but afaik never was close to 40MH/s
Fury can make 35MH/s due to ver wide HBM memory.
So my prediction:
1070 - 20-24MH/s (the maximum for 256-bit bus)
Polaris10 - same lvl
AMD wins in dagger-hashimoto because they have cards with wide bus - more mem controllers so faster random memory access/latency. I don't see how bigger page supported will help for "random" access
Obviously scaling on AMD is very very low after 1000-1200 shaders (7850/7870 265-270)
Its all because of bus width...
270 can make ~ 20MH/s with core:mem ratio of 2:3 (exp atm)
290 has 2x wider bus, 2x shaders... but afaik never was close to 40MH/s
Fury can make 35MH/s due to ver wide HBM memory.
So my prediction:
1070 - 20-24MH/s (the maximum for 256-bit bus)
Polaris10 - same lvl
290 did get to 30MH though, so it was still scaling somewhat - just not 100% scaling vs. shader count.
Limit is definitely somewhere in the memory system though - and it's NOT just bus width, Fury/Nano have a 4096 bit wide memory bus (due to the structure of HBM) yet they're as fast as the 290 or about the SAME hash.
Genoil IIRC was speculating it was a limit in the TLB table hardware, but I dunno how far their research into that has gotten.
I don't see anything close to 20 MH out of my 7870s - the 270 has the SAME shader count, but it does have somewhat faster memory and IIRC 2x the memory bus width, but while it's certainly a good bit faster than the 7870 it's nowhere near 2X as fast, much less 2x PLUS the memory speed ratio.
TDP target on Polaris should be a LOT lower - the smaller node is a lot more efficient, even if you puch more transistors AND kick the clock rate up quite a bit.
On the other hand, it won't be out in time to do much Eth mining with, unless AMD pushes the release dates up a good bit.
Wide bus is very important. For the R9 390, I can run 1100/1200MHz, the memory controller usage is just 98%.
It was expected.
AMD wins in dagger-hashimoto because they have cards with wide bus - more mem controllers so faster random memory access/latency. I don't see how bigger page supported will help for "random" access
Obviously scaling on AMD is very very low after 1000-1200 shaders (7850/7870 265-270)
Its all because of bus width...
270 can make ~ 20MH/s with core:mem ratio of 2:3 (exp atm)
290 has 2x wider bus, 2x shaders... but afaik never was close to 40MH/s
Fury can make 35MH/s due to ver wide HBM memory.
So my prediction:
1070 - 20-24MH/s (the maximum for 256-bit bus)
Polaris10 - same lvl
AMD wins in dagger-hashimoto because they have cards with wide bus - more mem controllers so faster random memory access/latency. I don't see how bigger page supported will help for "random" access
Obviously scaling on AMD is very very low after 1000-1200 shaders (7850/7870 265-270)
Its all because of bus width...
270 can make ~ 20MH/s with core:mem ratio of 2:3 (exp atm)
290 has 2x wider bus, 2x shaders... but afaik never was close to 40MH/s
Fury can make 35MH/s due to ver wide HBM memory.
So my prediction:
1070 - 20-24MH/s (the maximum for 256-bit bus)
Polaris10 - same lvl
so it might be worth to wait again the 470, which should be the best and cheapest along with the 490 or the possible future polaris nano
according to some speculation tdp target seems a bit lower on polaris
It was expected.
AMD wins in dagger-hashimoto because they have cards with wide bus - more mem controllers so faster random memory access/latency. I don't see how bigger page supported will help for "random" access
Obviously scaling on AMD is very very low after 1000-1200 shaders (7850/7870 265-270)
Its all because of bus width...
270 can make ~ 20MH/s with core:mem ratio of 2:3 (exp atm)
290 has 2x wider bus, 2x shaders... but afaik never was close to 40MH/s
Fury can make 35MH/s due to ver wide HBM memory.
So my prediction:
1070 - 20-24MH/s (the maximum for 256-bit bus)
Polaris10 - same lvl
AMD wins in dagger-hashimoto because they have cards with wide bus - more mem controllers so faster random memory access/latency. I don't see how bigger page supported will help for "random" access
Obviously scaling on AMD is very very low after 1000-1200 shaders (7850/7870 265-270)
Its all because of bus width...
270 can make ~ 20MH/s with core:mem ratio of 2:3 (exp atm)
290 has 2x wider bus, 2x shaders... but afaik never was close to 40MH/s
Fury can make 35MH/s due to ver wide HBM memory.
So my prediction:
1070 - 20-24MH/s (the maximum for 256-bit bus)
Polaris10 - same lvl
Genoil is it really only 24Mhash for 1080? Cant be 
It is 24mhs at 140w that is for 1080 - the big question is 1070 that will be better choice for mining anyway (or even other cheaper Pascal based GPUs)well that's bad, because right now a 390 can do the same with the same wattage, now imagines what a 490 can do
so i think if the new gpu will not be optimized in some way they will be out of the cryptoscene altogether
It's not unthinkable that AMD is going the same way with their architecture. You already see that in their GCN1.2 line. Narrower buses, delta compression, all to save power. I wouldn't be surprised if 490 also sticks in the 20-25 range for ETH.
With 2MB page size support we may see performance up to full bandwidth utilization. But that's pure speculation.
mmh that's even worse, or maybe not so much, we would avoid upgrading to new gpu if they do not offer anything, minus a slightly efficency of some insignificant less wattage
Genoil is it really only 24Mhash for 1080? Cant be
It is 24mhs at 140w that is for 1080 - the big question is 1070 that will be better choice for mining anyway (or even other cheaper Pascal based GPUs)well that's bad, because right now a 390 can do the same with the same wattage, now imagines what a 490 can do
so i think if the new gpu will not be optimized in some way they will be out of the cryptoscene altogether
It's not unthinkable that AMD is going the same way with their architecture. You already see that in their GCN1.2 line. Narrower buses, delta compression, all to save power. I wouldn't be surprised if 490 also sticks in the 20-25 range for ETH.
With 2MB page size support we may see performance up to full bandwidth utilization. But that's pure speculation.
Do you have a chance to use the 1080/70 and optimise the Ethereum miner? I hope it will be faster.
No I have to wait like everybody. Optimization-wise there isn't much to do, as 10x0 appears to actually be very similar to 9x0 on the inside. Just a few more SM's die-shrink and very high clocks.
should ask how will amd top it and they will then nvidia will, as for mining i hope we get more coins like eth so i have a excuse to make a 9 card miner 
. Video card mining might just make a come back and stay, not with bitcoins but if we can do it with video cards again why not .I would very much welcome it .
. Video card mining might just make a come back and stay, not with bitcoins but if we can do it with video cards again why not .I would very much welcome it . Genoil is it really only 24Mhash for 1080? Cant be
It is 24mhs at 140w that is for 1080 - the big question is 1070 that will be better choice for mining anyway (or even other cheaper Pascal based GPUs)well that's bad, because right now a 390 can do the same with the same wattage, now imagines what a 490 can do
so i think if the new gpu will not be optimized in some way they will be out of the cryptoscene altogether
It's not unthinkable that AMD is going the same way with their architecture. You already see that in their GCN1.2 line. Narrower buses, delta compression, all to save power. I wouldn't be surprised if 490 also sticks in the 20-25 range for ETH.
With 2MB page size support we may see performance up to full bandwidth utilization. But that's pure speculation.
Do you have a chance to use the 1080/70 and optimise the Ethereum miner? I hope it will be faster.
Genoil is it really only 24Mhash for 1080? Cant be
It is 24mhs at 140w that is for 1080 - the big question is 1070 that will be better choice for mining anyway (or even other cheaper Pascal based GPUs)well that's bad, because right now a 390 can do the same with the same wattage, now imagines what a 490 can do
so i think if the new gpu will not be optimized in some way they will be out of the cryptoscene altogether
It's not unthinkable that AMD is going the same way with their architecture. You already see that in their GCN1.2 line. Narrower buses, delta compression, all to save power. I wouldn't be surprised if 490 also sticks in the 20-25 range for ETH.
With 2MB page size support we may see performance up to full bandwidth utilization. But that's pure speculation.
Genoil is it really only 24Mhash for 1080? Cant be
It is 24mhs at 140w that is for 1080 - the big question is 1070 that will be better choice for mining anyway (or even other cheaper Pascal based GPUs)well that's bad, because right now a 390 can do the same with the same wattage, now imagines what a 490 can do
so i think if the new gpu will not be optimized in some way they will be out of the cryptoscene altogether
Genoil is it really only 24Mhash for 1080? Cant be
It is 24mhs at 140w that is for 1080 - the big question is 1070 that will be better choice for mining anyway (or even other cheaper Pascal based GPUs)
Genoil is it really only 24Mhash for 1080? Cant be
I heard 6MH for ethereum, is that correct? :O
unimpressive so far...
290 stock
X11 12100
X13 4700
Keccak 300000
X15 3000
Qubit 12200
Quark 19000 vs 32000
Lyra2REv2 15000 vs 16000
All only with GPL miners, quite sure there are even faster versions in the wild, I don't have access to optimized decred miner, but probably will be around 3000
unimpressive so far...
290 stock
X11 12100
X13 4700
Keccak 300000
X15 3000
Qubit 12200
Quark 19000 vs 32000
Lyra2REv2 15000 vs 16000
All only with GPL miners, quite sure there are even faster versions in the wild, I don't have access to optimized decred miner, but probably will be around 3000
About 24MH/s on Linux, exactly like I predicted (can't find the proof tho
)Do you mean the 1080 mine the Ethereum at 24 MH/s?
What is the 290 stock above? Is that AMD R9 290?
Probably between 100 and 150
I heard 6MH for ethereum, is that correct? :O
unimpressive so far...
290 stock
X11 12100
X13 4700
Keccak 300000
X15 3000
Qubit 12200
Quark 19000 vs 32000
Lyra2REv2 15000 vs 16000
All only with GPL miners, quite sure there are even faster versions in the wild, I don't have access to optimized decred miner, but probably will be around 3000
unimpressive so far...
290 stock
X11 12100
X13 4700
Keccak 300000
X15 3000
Qubit 12200
Quark 19000 vs 32000
Lyra2REv2 15000 vs 16000
All only with GPL miners, quite sure there are even faster versions in the wild, I don't have access to optimized decred miner, but probably will be around 3000
About 24MH/s on Linux, exactly like I predicted (can't find the proof tho
)what is the wattage?
I heard 6MH for ethereum, is that correct? :O
unimpressive so far...
290 stock
X11 12100
X13 4700
Keccak 300000
X15 3000
Qubit 12200
Quark 19000 vs 32000
Lyra2REv2 15000 vs 16000
All only with GPL miners, quite sure there are even faster versions in the wild, I don't have access to optimized decred miner, but probably will be around 3000
unimpressive so far...
290 stock
X11 12100
X13 4700
Keccak 300000
X15 3000
Qubit 12200
Quark 19000 vs 32000
Lyra2REv2 15000 vs 16000
All only with GPL miners, quite sure there are even faster versions in the wild, I don't have access to optimized decred miner, but probably will be around 3000
About 24MH/s on Linux, exactly like I predicted (can't find the proof tho
)is that the 1080? very underwhelming if yes
I heard 6MH for ethereum, is that correct? :O
unimpressive so far...
290 stock
X11 12100
X13 4700
Keccak 300000
X15 3000
Qubit 12200
Quark 19000 vs 32000
Lyra2REv2 15000 vs 16000
All only with GPL miners, quite sure there are even faster versions in the wild, I don't have access to optimized decred miner, but probably will be around 3000
unimpressive so far...
290 stock
X11 12100
X13 4700
Keccak 300000
X15 3000
Qubit 12200
Quark 19000 vs 32000
Lyra2REv2 15000 vs 16000
All only with GPL miners, quite sure there are even faster versions in the wild, I don't have access to optimized decred miner, but probably will be around 3000
About 24MH/s on Linux, exactly like I predicted (can't find the proof tho
)
I heard 6MH for ethereum, is that correct? :O
unimpressive so far...
290 stock
X11 12100
X13 4700
Keccak 300000
X15 3000
Qubit 12200
Quark 19000 vs 32000
Lyra2REv2 15000 vs 16000
All only with GPL miners, quite sure there are even faster versions in the wild, I don't have access to optimized decred miner, but probably will be around 3000
unimpressive so far...
290 stock
X11 12100
X13 4700
Keccak 300000
X15 3000
Qubit 12200
Quark 19000 vs 32000
Lyra2REv2 15000 vs 16000
All only with GPL miners, quite sure there are even faster versions in the wild, I don't have access to optimized decred miner, but probably will be around 3000
Jump to: