AMD Vega (FE and RX) Benchmarks [Updated Aug 10 - RX Vega 64 Unboxing]

[Veradun]

I was willing to deal with a little more power at 1080ti levels of performance because I would be able to cap FPS at 60 on some of the games I play (WoW, Hearthstone, HoTS etc)

Uhm. I get 70fps on HoTS at 1440p on an HD6870. I really doubt u need a 1080Ti to hit 60fps even at 16K

 

[Snarf Snarf]

Uhm. I get 70fps on HoTS at 1440p on an HD6870. I really doubt u need a 1080Ti to hit 60fps even at 16K

Lol no I don't the 1080ti is for the other things I play and for the eventual 1440p 144Hz monitor I will get. We have a used Acer Predator XB271huc sitting in front of me right now that my boss is willing to let go for $550 so I might actually bite on that if Vega doesn't wow me

 

[Elixer]

SK Hynix claims 1.2V for 4-Hi stacks only. Vega FE uses 8-Hi stacks... more stacks/layers obviously needs more voltage.

No, that is incorrect.
JEDEC clearly states what the voltages should be, as shown in this table.

Make no difference at all if it is 2/4/8 Hi stacks.
Same applies for GDDR5 / DDR4, the more RAM on a PCB/stick doesn't mean it needs more voltages.

They are pumping more voltages to either get more speeds than normal (OC), or, the yields are crap, and they have to compensate for that by pumping more voltage. Since JEDEC says 1.26 is MAX, can't say what running 1.35 constant does to the longevity of the HBM2.

 

[Malogeek]

What is the JEDEC max voltage for GDDR5 or DDR4?

 

[Elixer]

What is the JEDEC max voltage for GDDR5 or DDR4?

It isn't hard to find the info...go to the JEDEC site and...

 

[ElFenix]

I just updated the first post with some stuff I felt was worth updating. Any other solid reviews/benchmarks I should add?

Maybe a primer in how air conditioning works since people seem to have no idea about it


Jk, please don't

 

[Magic Hate Ball]

As I suspected, a user has had some luck undervolting from 1200mv max to 1075mv https://www.forum-3dcenter.org/vbulletin/showthread.php?p=11428407#post11428407

 

[nathanddrews]

Maybe a primer in how air conditioning works since people seem to have no idea about it

If my own AC unit worked 100%, I would certainly oblige.

 

[Azix]

As I suspected, a user has had some luck undervolting from 1200mv max to 1075mv https://www.forum-3dcenter.org/vbulletin/showthread.php?p=11428407#post11428407

the power consumption information is interesting on the page that the person linked.

http://www.pcgameshardware.de/Vega-...ase-AMD-Radeon-Frontier-Edition-1232684/2/#a3

looking forward to some tomshardware style power consumption measurements on rx vega.

could be promising with undervolt and if the performance improves markedly.

 

[Snarf Snarf]

If they can get perf to sit ~15% below 1080ti at those power numbers for RX Vega and sell it for $599-$650 for aftermarket or WC cooled variants I'd call that a huge success.

 

[96Firebird]

$600-$650 is too much for 15% lower performance, considering many recent aftermarket 1080 Ti deals are below $650...

 

[Cookie Monster]

If they can get perf to sit ~15% below 1080ti at those power numbers for RX Vega and sell it for $599-$650 for aftermarket or WC cooled variants I'd call that a huge success.

Its better than nothing but not great considering that Pascal based video cards are already a year old (maybe not so the Ti but the architecture). A hypothetical GV104 could come in to replace the GP102 based Ti and with lower power consumption/smaller die size.

 

[Snarf Snarf]

$600-$650 is too much for 15% lower performance, considering many recent aftermarket 1080 Ti deals are below $650...

Keep in mind those deals are not from Nvidia themselves, the AIB company gives rebates to consumer side distributors like Amazon and Newegg for short windows of time if the product isn't selling well enough at the time. The MSRP of 1080ti is still 699+ with most of the premium options being 750 (Aorus, Gaming X, FTW, and AMP!) leaving a nice window for a $599 Vega that (650 might be too high, you are correct) sits in between the 1080 and TI. Couple that with much cheaper FreeSync monitors and TCO is much lower for very comparable performance, which to me, would be a pretty big success after this FE debacle.

Edit: I'm pretty sure like 10 pages back I said if its 15% slower than 1080ti but is sold at $450 I will preorder one and wear a dress for a day.

 

[Tee9000]

PCGH undervolted their Vega FE to 1075mv at 1600 Mhz "rock stable" with 50% extra power limit. They said the card uses the same amount of power as stock @ 1200mv (their tests ran the card mostly at 1348-1440 Mhz range).

Vega FE Liquid Cooled is in stock and ships in 24 hours in the US.

 

[.vodka]

Aaaaaand AMD's typical factory overvoltage strikes again.

1.2v -> 1.075v for the same 1600MHz!?

This is ridiculous.

 

[Azix]

PCGH undervolted their Vega FE to 1075mv at 1600 Mhz "rock stable" with 50% extra power limit. They said the card uses the same amount of power as stock @ 1200mv (their tests ran the card mostly at 1348-1440 Mhz range).

Vega FE Liquid Cooled is in stock and ships in 24 hours in the US.

no power testing is reported. that was their guess tho since they were running lower clock before.

edit: they did say something about the total system power. should be the same as before if system power is similar. need to see what happens with lower power target.

 

[Tee9000]

The power consumption of the entire PC is at the same level as the factory setting

How can you tell they did not do any power testing?

How can you tell they were guessing?

 

[cytg111]

Updated July 12th but OP is not showing any. Am I blind?

 

[Magic Hate Ball]

How can you tell they did not do any power testing?

How can you tell they were guessing?

In the article the Google Translated sentence:
"The power consumption of the entire PC is at the level of the factory setting, although we have set the power limit to +50 percent."

(Note, that italicized "entire PC" is that way in the article, emphasis theirs)

Does this mean it stayed the same power usage with max clocks running? Or does it mean that it reduced power usage so much that the whole PC was now using the power that the video card alone was using before?

 

[Snarf Snarf]

Sounds like they mean that entire system power consumption remained the same as when they ran the card at stock settings, meaning power consumption didn't change.

 

[Magic Hate Ball]

Sounds like they mean that entire system power consumption remained the same as when they ran the card at stock settings, meaning power consumption didn't change.

Yeah, it's likely, but Google Translate makes stilted sentences sometimes.

Still better than Babelfish back in the early 2000's!

 

[Headfoot]

If they can get perf to sit ~15% below 1080ti at those power numbers for RX Vega and sell it for $599-$650 for aftermarket or WC cooled variants I'd call that a huge success.

"Great Success" yeah no. That would mean it's "Fury vs 980 Ti level flop" as opposed to a "2900 XT vs 8800 GTX level flop." Success means it wins in at least one category: price/perf, absolute perf. If it wins in neither, it is a flop.

 

[Snarf Snarf]

"Great Success" yeah no. That would mean it's "Fury vs 980 Ti level flop" as opposed to a "2900 XT vs 8800 GTX level flop." Success means it wins in at least one category: price/perf, absolute perf. If it wins in neither, it is a flop.

Remember that successful to you != successful to AMD RTG. If the card generates a return on investment costs, increases ASP of RX cards, gets market share back in the high end (which currently is at 0 for this gen), and gains market share in the data center it is a success to AMD. Also I said 2 posts down 650 is too high, it has to be 599 for the best of the best AIB cards. The success of Vega architecture is not going to be decided by the RX alone, it is going to have a full product stack with APU's included. Assuming they lose the top tier halo battle, there's still plenty of success that can happen below that point, some of you guys are way too prone to using hyperbole when talking about graphics cards

 

[tential]

Ya, I see the things you have to say in the abstract.
In the real world, nothing stated explains how this release will do better than Fiji did.
Nothing about Vega is impressive. Nothing. Absolutely nothing.

The apu talk is another just abstract theory with no concrete explanation as to how the product will be successful.

If people want to say wait for rx, then wait for the apu to come out before you claim its a relevant product.

Edit : also no one forced amd to mention Volta. They put that comparison put there themselves. All the back pedaling to "it just needs gtx 1080 performance at a 10% discount!".
The gtx 970 was torn apart for delivering no added performance in comparison to an old r9 290 yet this is OK?

The goalposts just get moved around and have no real meaning.

 

[thilanliyan]

To me if a stock Vega doesn't at least beat a 1080 by 10-15%, it is a failure IMO, regardless of price. This being their highest end card, they need to be closer to the top of the heap.