Topic: Any word on amd vega hash rates? - page 157. (Read 202748 times)

Have you tried any other algos with it?  I've been thinking about picking a couple of these up but not if they're limited to only being potentially good with dagger.  Also, what kind of power draw are you seeing so far after tweaking?  The reviews that I've seen were saying that their cards were drawing >300w during their gaming and compute tests which doesn't sound so great unless they turn out to be mining monsters.

In an ideal world the coins I am holding go up a lot and I get 3 or 4 of either one.

I'd wait vega, i have a feeling nvidia might reduce 1080 ti's price a bit.

mostly zec  sometime skein   the density on the ti's is my main reason I use them.

you can do 700 sols at 180 watts.  and if you pick your shots the 1080 ti's are under 700 bucks.  the best price I ever paid for a new one was 577

What would you mine with TIs?

50 for eth. at 200 watts   would be nice
1000 for zec at 200 watts  would be nice


40 for eth would be a fail
as would 850 for zec.

for now  I will be waiting for some  info on zec rates

I sold off my  1070's   looking to go all 1080 ti or   part 1080 ti and  part  amd vega.

Ive been playing around with mine...nothing crazy to announce yet. Here are some tidbits for you guys

-Claymore's and a few other miners have incompatible kernels with the new GNC, so those don't even work yet
-Genoil's works, does about 35-40 with some basic overclocks
-Once i start tuning timings on this thing and over volt the memory should be able to hit 40+ on genoils generic kernel
-Im working with AMD for some driver optimizations for mining specific kernels (specifically on the linux side since currently its underperforming horrible on their new compute stack). AMD finally woke up and is addressing their mining customers (they could ignore us for so long when 90% of their profits are from us recently).
-Once these are done, and Claymore updates his miner for Vega NCU we should be seeing 50+

Verdict is still out on whether this can compete on price/hash and watts/hash on polaris, but if it can do 50+ it will still win out on density cost(i.e. achieve 160+% hash using the same hardware).

You trolls are asking the wrong question. The right question supposed to be --> optimized for high bandwidth memory, eth must not be such coin that benefits from hbm, i guess a new coin will be coming soon which will use high memory bandwidth.

 Keep in mind that video cards are aimed at GAMING as their primary market, not mining.

 The 5xx series moved to a slightly different process on the same node, that gave the ability to clock them higher at a minor cost in efficiency.


 I don't see the VEGA being a killer card for ETH/ETC and such mining, as it's still HBM and like the Fury line HBM has a lot of latency that hammers ETH hashrates.

 It SHOULD be a highly effective card on other mining though, like ZEC where the Fury line IS a good at mining.

 I can also see a high probability of it taking AMD 2 or 3 driver iterations to get a "good for mining" VEGA driver available.


 Don't compare the Vega Frontier Edition for $ - it's a workstation-targeted card (like the AMD WX line) and those are ALWAYS overpriced compared to consumer cards.
 (compare the WX 7100 to the RX 480 and RX 580 - slower clocks, single-slot, but it's higher priced than even the current GOUGE pricing on the RX equivalent cards).

 Also, AMD has already specified that there will be at least one "gaming" card with higher performance - and that will probably cost a fair bit less then the Frontier Edition.

 The consumer "Vega" cards are widely reported to be due for "release" as SIGGRAPH at the end of this month/early August - but no firm word YET on if it's going to be a "paper" release or if the cards will actually be available for sale at that time.
 I suspect it comes down to "has Hynix gotten the production issues with HBM2 fixed yet".

Two people two points of view .

I think the answer is in the middle.

I am guess this for sols at zec.

11gb 1080 ti can do 700sols at 180 watts

16/11 = 1.4545

so 1.4545 x 700 = 1018 sols at 180 watts.

which  makes the card a killer at zec

if a 1080 ti = 700 usd for 700 sols

and an amd vega = 1000 usd for 1000 sols

both are = at $ per sol.

but the amd  would be better since it is using 180 watts  for 1000 and the 1080 ti is using 180 watts for 700.

The 1080 ti  can be found at okay prices sometimes under 700 usd.

I would love to see what the amd does for zec. I am looking to buy a few but I am waiting for zec and eth hash rates.



stratum+tcp://skunk.usa.nicehash.com:3362

i dont think amd cares  about who buy their card... the more they sell the better it is

I wouldn't be surprised if AMD intentionally gimps their cards for mining. I mean look what happened with their 4xx/5xx cards. It's not good for either AMD or gamers.

It's too early to judge Vega, because it's only FE edition, with unoptimised drivers and high price tag. When ,,civil" Vega comes out with gaming drivers, then we can tell if its good or not.
@64dimmensions 5xx was quite improvement for gamers, I know it was just hard OC, but out of box was couple % faster than 1060 6gb, so casuals didn't have to bother with manual OC.

So far the AMD marketing direction is that vega is an improvement over previous AMD cards. I have never seen a concise explanation on what the 5XX series was improving

Found some reports from claymore's threads saying a vega was hashing at 30MH with ETH. So sad. Makes it pretty useless until someone develops drivers/miners. But they need to do it fast. ice age is coming.

There are no optimized miners or drivers for Vega. New update soon and my opinion Vega work 3x more hashrate from rx 580

Similar arch to the Fury line, so it shouldn't have the same issue as the Polaris cards.

Do you think that vega will have "polaris bug" or it will not be affected like nvidia gpus?.

I think we need to wait for the consumer cards to get into the wild so the programmers can play with them before we know their real potential.  By all accounts it seems it will be incremental instead of revolutionary anyway.

Guess we will need to wait for an update from Claymore before we see...

Anybody tried any other algos?