Author

Topic: Any word on amd vega hash rates? - page 152. (Read 202731 times)

legendary
Activity: 4256
Merit: 8551
'The right to privacy matters'
August 10, 2017, 11:28:20 PM
This topic has everything but vega hash rates.

They must suck.  If they really were good they would have leaked by now.

full member
Activity: 241
Merit: 100
To Hash or not to Hash, that's what the question
August 10, 2017, 10:50:42 PM
Quote
This topic has everything but vega hash rates.
- Monday the 14th, once someone get their hands on it and will upload video o youtube
YIz
hero member
Activity: 686
Merit: 502
August 10, 2017, 09:28:41 AM
Vega on ZEC, what's the hashrate? Impressive as I'd assume?
newbie
Activity: 26
Merit: 0
August 10, 2017, 09:27:37 AM
This topic has everything but vega hash rates.
legendary
Activity: 4256
Merit: 8551
'The right to privacy matters'
August 10, 2017, 08:06:18 AM
Quote
ALL FOUNDRIES THAT CAN MAKE CURRENT MODEL GPUs RUNNING AT 100%
- knowing the possibility of orders and not expanding manufacturing capabilities, is plain stupid. To my opinion, the expectations certainly were lower than reality due to high miners demand, but... no admin will wait with server upgrade until one of the resource will be loaded 100% all the time, same thing with any business, one need to understand capabilities and expand as needed, not wait until bottleneck becomes a serious problem. Business is loosing revenue. Another example if you get sick, you dont wait until you will be walking dead, you start treatment as soon as you know you are sick - right? Just my opinion, a better planning will help to remedy. I dont know details about silicon fabs loads, but i dont think they were prepped well - amd is not the only product they manufacture, so i bet they can allocate resources more efficiently, there musts be something else involved besides manufacturing capabilities... (money?) Time to sleep now...

Modern fabs always run at 100% capacity. They allocate production capacity years in advance and logistics is top-notch. Simply state of the art, no doubt about it.


 They might allocate some production to certain manufacturers for years in advance (AMD and Global Foundries long-term agreement, as modified usually once a year, is a good example of that) but a lot of it is only allocated months in advance.
 Still makes it hard to get additional capacity QUICKLY when you have a suddenly hot product with NO warning it's going to turn crazy-popular ahead of time.

 The big issue for the FOUNDRIES is the massive lead time TO expand their production - and the massive COST on recent production nodes - makes them hesitant to expand based on "speculation" when they don't have fairly firm contracts in hand AHEAD OF TIME to fill a newly-built production line.

 This then hits the manufacturers that USE the foundry if they do have a sudden demand spike.

 Keep in mind the TSMC was ALREADY full up, and NVidia had to go to Samsung to make their GTX 1050/1050ti models BECAUSE TSMC DIDN'T HAVE THE CAPACITY AVAILABLE - and that was BEFORE the big crypto-craze of 2017 hit.

 AMD has the advantage of their "locked in priority access" to a large chunk of the capacity of GF - but they also have Ryzen being made on the same node and selling VERY well, NOT just certain of their GPU models....


 Both AMD and NVidia have to be carefull about "after the crash" too, if they sell TOO MANY right now their sales DIE (probably early next year, possibly a bit longer) when the GPU mining CRASH hits and all of a sudden there are a ton of used cards on sale CHEAP killing their market for quite a while on NEW card sales.



The renewed rise in ETH price the last few days isn't going to help these calculations, though it certainly helps those that are MINING ETH already it will probably also fuel a renewed "new rig construction" boom just extending the "issue" from the point of view of the GPU chip makers, their allied card makers, and gamers (the LONG TERM and BIGGEST market for mid-to-high-end GPUs) and making the eventual "crash glut" even BIGGER.





look at market cap


https://coinmarketcap.com/


1   Bitcoin Bitcoin------------------------$55,762,056,812,,,,,,,,,,,,,,$3379.68   
2   Ethereum Ethereum-----------------$28,423,666,122,,,,,,,,,,,,,,,$302.69
   
3   Ripple Ripple--------------------------$6,985,059,332................$0.182169   
4   Bitcoin Cash Bitcoin Cash------------$4,862,084,192................$294.90
5   NEM NEM------------------------------$2,634,471,000...............$0.292719

6   Litecoin Litecoin-----------------------$2,507,514,730..................$47.86
7   IOTA IOTA------------------------------$1,521,325,869.................$0.547332
8   Dash Dash-----------------------------$1,507,227,839..................$201.41   
9   Ethereum Classic Ethereum Classic--$1,458,056,020..................$15.45   
10   NEO NEO------------------------------$1,377,680,000...................$27.55   


11   Qtum Qtum---------------------$820,412,700.........$13.91   
12   Monero Monero-----------------$741,660,765.........$49.73   
13   Stratis Stratis-------------------$725,984,964.........$7.37   
14   BitConnect BitConnect----------$567,828,072.........$88.64   
15   Waves Waves--------------------$531,835,000.........$5.32   


16   EOS EOS   $511,528,801   $1.87   
17   Zcash Zcash   $475,185,840   $246.18   
18   BitShares BitShares   $463,963,183   $0.178604   
19   OmiseGo OmiseGo   $446,450,631   $4.54   
20   Tether Tether   $318,860,283   $0.997994

   

Eth is at 28 billion  it has more value then alts 3 to 20 combined

This is a fundamental weakness  in gpu mining that may correct with a crash.

Or ETC could start to pick up slack.  I would prefer that  ETC has very strong growth for the next 3 months it should be fun to watch this unfold.

legendary
Activity: 1498
Merit: 1030
August 10, 2017, 04:22:05 AM
Quote
ALL FOUNDRIES THAT CAN MAKE CURRENT MODEL GPUs RUNNING AT 100%
- knowing the possibility of orders and not expanding manufacturing capabilities, is plain stupid. To my opinion, the expectations certainly were lower than reality due to high miners demand, but... no admin will wait with server upgrade until one of the resource will be loaded 100% all the time, same thing with any business, one need to understand capabilities and expand as needed, not wait until bottleneck becomes a serious problem. Business is loosing revenue. Another example if you get sick, you dont wait until you will be walking dead, you start treatment as soon as you know you are sick - right? Just my opinion, a better planning will help to remedy. I dont know details about silicon fabs loads, but i dont think they were prepped well - amd is not the only product they manufacture, so i bet they can allocate resources more efficiently, there musts be something else involved besides manufacturing capabilities... (money?) Time to sleep now...

Modern fabs always run at 100% capacity. They allocate production capacity years in advance and logistics is top-notch. Simply state of the art, no doubt about it.


 They might allocate some production to certain manufacturers for years in advance (AMD and Global Foundries long-term agreement, as modified usually once a year, is a good example of that) but a lot of it is only allocated months in advance.
 Still makes it hard to get additional capacity QUICKLY when you have a suddenly hot product with NO warning it's going to turn crazy-popular ahead of time.

 The big issue for the FOUNDRIES is the massive lead time TO expand their production - and the massive COST on recent production nodes - makes them hesitant to expand based on "speculation" when they don't have fairly firm contracts in hand AHEAD OF TIME to fill a newly-built production line.

 This then hits the manufacturers that USE the foundry if they do have a sudden demand spike.

 Keep in mind the TSMC was ALREADY full up, and NVidia had to go to Samsung to make their GTX 1050/1050ti models BECAUSE TSMC DIDN'T HAVE THE CAPACITY AVAILABLE - and that was BEFORE the big crypto-craze of 2017 hit.

 AMD has the advantage of their "locked in priority access" to a large chunk of the capacity of GF - but they also have Ryzen being made on the same node and selling VERY well, NOT just certain of their GPU models....


 Both AMD and NVidia have to be carefull about "after the crash" too, if they sell TOO MANY right now their sales DIE (probably early next year, possibly a bit longer) when the GPU mining CRASH hits and all of a sudden there are a ton of used cards on sale CHEAP killing their market for quite a while on NEW card sales.



 The renewed rise in ETH price the last few days isn't going to help these calculations, though it certainly helps those that are MINING ETH already it will probably also fuel a renewed "new rig construction" boom just extending the "issue" from the point of view of the GPU chip makers, their allied card makers, and gamers (the LONG TERM and BIGGEST market for mid-to-high-end GPUs) and making the eventual "crash glut" even BIGGER.



full member
Activity: 151
Merit: 100
August 10, 2017, 03:28:59 AM
Apparently there are works on dual vega gpu from ASUS (ARES III) /AMD something similar to R9 295X2 here some more info:

https://www.facebook.com/permalink.php?story_fbid=1155851951182552&id=348409301926825

 AMD already announced a "Nano" version of the VEGA - but I think it was going to be based on the Vega 56 not the Vega 64.
 Would not be hard to do a dual version of that - perhaps that's what ASUS meant in it's announcement?


If the power consumption is similar to R9 nano, I might buy one.
full member
Activity: 206
Merit: 100
August 10, 2017, 03:12:10 AM
Quote
ALL FOUNDRIES THAT CAN MAKE CURRENT MODEL GPUs RUNNING AT 100%
- knowing the possibility of orders and not expanding manufacturing capabilities, is plain stupid. To my opinion, the expectations certainly were lower than reality due to high miners demand, but... no admin will wait with server upgrade until one of the resource will be loaded 100% all the time, same thing with any business, one need to understand capabilities and expand as needed, not wait until bottleneck becomes a serious problem. Business is loosing revenue. Another example if you get sick, you dont wait until you will be walking dead, you start treatment as soon as you know you are sick - right? Just my opinion, a better planning will help to remedy. I dont know details about silicon fabs loads, but i dont think they were prepped well - amd is not the only product they manufacture, so i bet they can allocate resources more efficiently, there musts be something else involved besides manufacturing capabilities... (money?) Time to sleep now...

Modern fabs always run at 100% capacity. They allocate production capacity years in advance and logistics is top-notch. Simply state of the art, no doubt about it.

When they invest $10B-$20B in each new fab, they must have logistics figured out otherwise they lose billions. Imagine that one brand new fab costs like the whole annual NASA budget!

There is one problem though. There is usually no room for expansion within the fab because it's already full of equipment wall-to-wall. If you wish to "expand" you got to build another fab or scrap and rebuild an existing one. I don't think AMD or nVidia has much of a control over their expansion plans and even if they do, they have to wait for years for the new fab to get built and ramp up production.

By modern fabs I mean 10-20nm. You can find free capacity for older processes, but bleeding edge is always in high demand.

Google "semiconductor fab cost" for more info.

-SCSI
full member
Activity: 241
Merit: 100
To Hash or not to Hash, that's what the question
August 10, 2017, 01:42:09 AM
Quote
ALL FOUNDRIES THAT CAN MAKE CURRENT MODEL GPUs RUNNING AT 100%
- knowing the possibility of orders and not expanding manufacturing capabilities, is plain stupid. To my opinion, the expectations certainly were lower than reality due to high miners demand, but... no admin will wait with server upgrade until one of the resource will be loaded 100% all the time, same thing with any business, one need to understand capabilities and expand as needed, not wait until bottleneck becomes a serious problem. Business is loosing revenue. Another example if you get sick, you dont wait until you will be walking dead, you start treatment as soon as you know you are sick - right? Just my opinion, a better planning will help to remedy. I dont know details about silicon fabs loads, but i dont think they were prepped well - amd is not the only product they manufacture, so i bet they can allocate resources more efficiently, there musts be something else involved besides manufacturing capabilities... (money?) Time to sleep now...
hero member
Activity: 578
Merit: 508
August 09, 2017, 09:37:15 PM
FWIW, some data from the newegg vega frontier comments:

This review is from: AMD Radeon Vega Frontier Edition DirectX 12 100-506061 16GB 2048-Bit HBM2
Pros: Can run Ark Survival Evolved completely maxed out

Cons: No drivers or official overclocking software yet.

Other Thoughts: 1000 hashes per second of CryptoNight (Monero) mining. 11 mh/s of X11 Ghost.
legendary
Activity: 1498
Merit: 1030
August 09, 2017, 08:58:25 PM
AMD having had an experience of card shortages worldwide should of have been prepared for vega release

 AMD stated that one of the reasons they have delayed the release of the RX Vega was to make sure they HAD enough stock to get through the first few days without sellout issues.

 It's not their business model that is causing widespread shortages of the other RX series cards, it's "HUGH UNEXPECTED DEMAND OUT OF NOWHERE" due to the cryptocoin price explosion this spring coupled with ALL FOUNDRIES THAT CAN MAKE CURRENT MODEL GPUs RUNNING AT 100% CAPACITY on production of everything they are tasked to make.

 Keep in mind that it's NOT just AMD having shortages - NVidia has been getting hammered on the GTX 1070 and GTX 1060 as well (and some small indirect effect on the GTX 1080 probably more due to gamers having to "move up" from the unavailable 1070s than to miners buying a lot of THOSE cards).

 The irritating part - the gamers keep blaming "miners" for the entire shortage, even though GAMERS are still probably managing to buy more of the "shortage" card models than miners are.

full member
Activity: 241
Merit: 100
To Hash or not to Hash, that's what the question
August 09, 2017, 08:08:36 PM
AMD having had an experience of card shortages worldwide should of have been prepared for vega release, otherwise, their model of business needs serious update, or at least management part thats responsible for product supply/availability. But i bet first day will be sold out pretty quickly, but... if it turns out vega suck at mining, which we will know before the end of the monday the 14th, then it will all calm down, as all this hype is about mining and not a gaming. Cant wait to see first videos of VEGA MINING capabilities.
newbie
Activity: 53
Merit: 0
August 09, 2017, 06:56:09 PM
I would hope not, but to combat this issue there will be many more cards available in PC packages, probably paired with Ryzen parts.
newbie
Activity: 52
Merit: 0
August 09, 2017, 04:46:53 PM
Do you guys think the vega cards will sell out on release day? I want to build a rig, but don't want to have inflated prices and oos issues again.
legendary
Activity: 1498
Merit: 1030
August 09, 2017, 03:44:16 PM
Apparently there are works on dual vega gpu from ASUS (ARES III) /AMD something similar to R9 295X2 here some more info:

https://www.facebook.com/permalink.php?story_fbid=1155851951182552&id=348409301926825

 AMD already announced a "Nano" version of the VEGA - but I think it was going to be based on the Vega 56 not the Vega 64.
 Would not be hard to do a dual version of that - perhaps that's what ASUS meant in it's announcement?



legendary
Activity: 1498
Merit: 1030
August 09, 2017, 03:40:53 PM
There was also the dual Nano, the "Fury Pro" as I recall it being named, ballpark 500 watt TDP I think?
full member
Activity: 241
Merit: 100
To Hash or not to Hash, that's what the question
August 08, 2017, 11:08:22 PM
i got few 7970s diggin coins for years now... work like a charm, paid them selves back in B.C.  lol
besides r9 295x2 there is another hungry monster called DEVIL 13  one like this https://www.newegg.com/Product/Product.aspx?Item=N82E16814131680  one just got sold on ebay for like 600$ ... but i bet its more for a rarity/collectors item rather than gaming/mining as its not as powerful, but who knows... i had one broken coming in for repair like 2 months ago, took pictures of this beast. Its heavy as ph_ck - AIR cooled!

added: it has 4 CONENCTORS 8 PIN EACH! so 600W for sure!
full member
Activity: 584
Merit: 106
August 08, 2017, 04:38:44 AM
Can someone test Signatum speed with Vega FE?
newbie
Activity: 60
Merit: 0
August 08, 2017, 02:11:14 AM
Back in the BTC days those dual cards worked great, nice and efficient.

I still have the AMD HD 7990. They are fine on the ZEC.
hero member
Activity: 630
Merit: 502
August 07, 2017, 10:48:33 PM
Apparently there are works on dual vega gpu from ASUS (ARES III) /AMD something similar to R9 295X2 here some more info:

https://www.facebook.com/permalink.php?story_fbid=1155851951182552&id=348409301926825

I can't imagine this 600W+ monstrosity being a good choice for mining.
http://blockchurn.com/2017/07/04/is-the-r9-295x2-the-miners-holy-grail/
Jump to: