Welcome, Guest
Username: Password: Remember me

TOPIC: RX Vega Testing

RX Vega Testing 4 months 2 weeks ago #15812

  • Radioman
  • Radioman's Avatar
  • Offline
  • Lurker
  • Posts: 5
  • Karma: 1
Hi Steve

The pricing of RX Vega looks very attractive. It makes me wonder if AMD is attempting to Wall-Off its existing Polaris Customer Base from NVIDIA. With that is mind when you test RX Vega can you test it side by side with an RX 480/RX 580 ?

Radioman
The administrator has disabled public write access.

RX Vega Testing 4 months 2 weeks ago #15815

  • i7Baby
  • i7Baby's Avatar
  • Offline
  • Moderator
  • Posts: 663
  • Thank you received: 108
  • Karma: 24
R7-1700 @3900/1.375V, H110iGT, ASRock x370 Taichi, GSkill 2x8GB 3200/14 @3466/14 1.38V, Samsung 250gb EVO (W10), 2TB Barracuda, 2 x R9 Nano, EVGA G2 750, Enthoo Pro, LG 34UM88. Aizo kb, M6580, HyperX Cloud
Last Edit: 4 months 2 weeks ago by i7Baby.
The administrator has disabled public write access.
The following user(s) said Thank You: Radioman

RX Vega Testing 4 months 1 week ago #15871

  • Radioman
  • Radioman's Avatar
  • Offline
  • Lurker
  • Posts: 5
  • Karma: 1
Thanks for the reply I7Baby. Have you seen the new info about Vega that hit the street late last week ?

1) According to source trusted by Tweaktown the RX Vega 56 put up numbers in Benchmark Comparisons with the 1070 that thumped the 1070 in every one of 6 different games. That is far from complete but very encouraging nonetheless. Considering the RX Vega 56 is priced below the 1070 NVIDIA will either have to respond by adjusting the 1070 or accept its inevitable demise as a gaming card.

2) The second interesting tidbit comes OCUK (Over Clockers UK). In a forum post 'Gibbo' , an OCUK Staffer, claimed the Vega (did not specify which version) attained a hash rate of 75-100 Mhs. This compares 25 Mhs for the RX-580. There is still a great deal of question surrounding this claim. When pressed on the matter Gibbo says is was reported to them through a leak of an AIB Partner. If this number is indeed true it could well mean Vega could come under the same pressure from Crypto-Currency Miners that faced Polaris Cards. Again the lack of any information about which card the reported Hash rate does carry some degree of significance. The crucial parameter for Miners is the Hashrate per watt of power consumed. The cost of the card is a only partial determinant in how long it takes the Miner to make back his initially outlay of money. Hashrate per watt determines profit margin.

From what I have gathered at various sources it appears that Miners can make about 100 dollars/per card/per month using RX580s. A Ten Card Rig yields about 1000 dollars per month. Of course there are flucutations based upon the trading price of Ethereum and the Epoch Computation Formula. If the reports are correct and the Card tested was a Vega 64 which has a TDP 295 Watts a miner using RX-580s he could increase his income by 90% but upgrading his rig replacing the RX 580s with RX Vega 64s. If the card tested was an RX Vega 56 with a TDP of 210 Watts a Miner could increase his revenue over 150%. If this hashrate numbers are verified from my viewpoint it would seem the best strategy for AMD would be to place purchase restrictions on RX Vega 56s to preclude them from falling into the hands of Miners and then amping up RX Vega 64 production lines going Little Round Ones to the Wall in an effort to break the back of Miner demand. At the very least Polaris Cards will resurface on the second hand market once Miners begin to upgrade. That will set a very interesting dynamic in to motion. As the Polaris Cards hit the Used Market their prices will drop starting at 300 USD and head downward.Once word gets out to the Miner Community the value of Polaris Cards are dropping this compel Miners to upgrade to the more efficient Vega Card forcing up the demand for that product. The whole process feeds on itself and before you know it we will have massive numbers of RX-480s/580s whose price are in free fall whistling through 150 bucks. AMD makes a killing in this process selling Vega Cards to Miners at 499 USD who one year earlier were snapping up Polaris Cards for 229 USD. What it all comes down to does AMD have the extra manufacturing to produce the Vega 64 in the quantities to do this. Can AMD scare up the extra capacity to pull this off. My only guess where it could come from is trimming back Ryzen 5 and Ryzen 3 production. And I doubt anything will be done of this sort until after 4th Quarter Orders are fulfilled with a decent inventory built up.
The administrator has disabled public write access.

RX Vega Testing 4 months 1 week ago #15872

  • i7Baby
  • i7Baby's Avatar
  • Offline
  • Moderator
  • Posts: 663
  • Thank you received: 108
  • Karma: 24
I'm hoping GN can also add rendering performance to their testing. I'm after replacements for my R9 Nanos. They game like gtx980s, but render better than a 1080ti (I hope) - www.anandtech.com/bench/GPU15/1231
The Vegas COULD be a good replacement. Time will tell. Hopefully the mining boom will crash soon.
R7-1700 @3900/1.375V, H110iGT, ASRock x370 Taichi, GSkill 2x8GB 3200/14 @3466/14 1.38V, Samsung 250gb EVO (W10), 2TB Barracuda, 2 x R9 Nano, EVGA G2 750, Enthoo Pro, LG 34UM88. Aizo kb, M6580, HyperX Cloud
Last Edit: 4 months 1 week ago by i7Baby.
The administrator has disabled public write access.

RX Vega Testing - Power Profiles 3 months 2 weeks ago #15968

  • Slobodan
  • Slobodan's Avatar
  • Offline
  • Lurker
  • Posts: 1
  • Karma: 0
Hello all :-)

I see all people are complaining about RX Vega power consumption, and rightly so. But, I do not see many discussion on RX Vega power profiles.

The Tech Report site (and also AdoredTV) tested "power save", "balanced" and "turbo" profiles. The conclusion was that when you go with "power save" profile you lose about 10% or even less performance with massive power consumption reduction. So much that it comes close to Nvidia's cards.

That to me looks like an elephant in the room that no one wants to talk about. It looks like the power save is actually the balanced mode for RX Vega. It is like they so desperately wanted to match the Nvidia's cards that they just OC the brains out of RX Vega cards. Maybe they should just released them with power save setting as balanced (default) but with much better power consumption and thermals. And then work on drivers for performance boost (if it is possible).

My question is can you guys on GN do tests and analyze RX Vega performance under the power save profile? Both BIOS versions (I saw that two bioses on card have different power settings)? And, can you tweak the cards so their power consumption and performance is close to that of Nvidia's 1070 (for 56) and 1080 (for 64)? That would be very interesting to watch :-)

Cheers!
The administrator has disabled public write access.

RX Vega Testing - Power Profiles 3 months 2 weeks ago #15971

  • i7Baby
  • i7Baby's Avatar
  • Offline
  • Moderator
  • Posts: 663
  • Thank you received: 108
  • Karma: 24
I had hopes I could replace my R9 Nanos with Vega. BUT
1. The performance gain doesn't warrant the expenditure
2. The expenditure is too great when compared to nVidia
3. The power usage is too great which would require me to again replace my PSU. And the cost of electricity in Australia (40c/kWh) would again make the proposition financially unviable.
R7-1700 @3900/1.375V, H110iGT, ASRock x370 Taichi, GSkill 2x8GB 3200/14 @3466/14 1.38V, Samsung 250gb EVO (W10), 2TB Barracuda, 2 x R9 Nano, EVGA G2 750, Enthoo Pro, LG 34UM88. Aizo kb, M6580, HyperX Cloud
The administrator has disabled public write access.
Moderators: 2ndPlayer, i7Baby
Time to create page: 0.122 seconds
Powered by Kunena Forum

We moderate comments on a ~24~48 hour cycle. There will be some delay after submitting a comment.

Advertisement:

  VigLink badge