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

Page 94 - Seeking answers? Join the AnandTech community: where nearly half-a-million members share solutions and discuss the latest tech.
Status
Not open for further replies.

tential

Diamond Member
May 13, 2008
7,355
642
121
I really dont know why you guys so care what i am buying.Its my money not yours.You have some problem with it?

No one forced you to post your reasoning for why you were buying what you're buying on a DISCUSSION FORUM.
You're now surprised that people are talking about what you posted?

Given how AMD marketed this card, and other AMD actions taken, I'm surprised ANYONE is continuing your stance of AMD being a "good" company by any means. I must admit, from hearing everyone post this nonsense time and time again over the last 4 years it's sunk in a little to me too. I even bought into the Freeysnc = Gsync even though it's NOT the same even if freesync acts like Gsync, Gsync is a standard and Freesync isn't.

Sidegrading for your own personal beliefs though is nothing. So what if it costs you $50 or whatever to make the switch? That's pocket change if it will make you happier, so do it.
 

Head1985

Golden Member
Jul 8, 2014
1,866
699
136
30% performance is not sidegrade for me.
Lets say GTX1070 have 50fps and vega is 30% faster.Based on firestrike it is faster than aftermarket GTX1080 and the gap between GTX1070 and 1080 is 25-30% in new games.
50+30%=65fps
GTX970 to 1070 it was like how much 50%?And everyone called it biggest jump in years(wrong btw)
50+50%=75fps

Not to mention vega is new.I am 100% sure there will be big performance jumps with new drivers over time and vega will end up far more faster than GTX1070.Also fp16 can have some impact in future.HBCC will 100% have impact.
 
Last edited:

PeterScott

Platinum Member
Jul 7, 2017
2,605
1,540
136
30% performance is not sidegrade for me.
Lets say GTX1070 have 50fps and vega is 30% faster.Based on firestrike it is faster than aftermarket GTX1080 and the gap between GTX1070 and 1080 is 25-30% in new games.
50+30%=65fps
GTX970 to 1070 it was like how much 50%?And everyone called it biggest jump in years(wrong btw)
50+50%=75fps

Not to mention vega is new.I am 100% sure there will be big performance jumps with new drivers over time and vega will end up far more faster than GTX1070.Also fp16 can have some impact in future.HBCC will 100% have impact.

Everything that I have seen says Vega 64 ~= GTX 1080. Will you be able to get a Vega 64 for less money than GTX 1080? That remains to be seen, but you could have had this level of performance more than a year ago.

HBCC is unlikely to have any impact. It only matters if you have inadequate VRAM. To show a benefit AMD had to limit cards to 2GB. Compared to a NVidia card with 11/12GB, you won't seen any benefit of HBCC, ever.
 

Erithan13

Senior member
Oct 25, 2015
218
79
66


Alleged leaked gaming results, take with the usual grain of salt. Seen on r/amd.

64 is really going to struggle against the 1080 from the looks of things, we know it, AMD knows it and AMD knows we know it, hence the recent shift in focus toward the 56. At least that looks solid against the 1070, all comes down to price now.

One thing is for sure, the 1080ti has nothing to worry about, let alone Volta. I'm not even angry about this, just bitterly, bitterly disappointed. How pessimistic did you have to be this time last year to think this is where Vega would end up?
 
Reactions: ozzy702

Konan

Senior member
Jul 28, 2017
360
291
106
That "11 triangles per clock" figure is basically a lie from RTG's notoriously dishonest marketing department. It appeared only in the old architecture preview slides, and was part of a footnote claiming "over 2x peak throughput per clock" comparing to Fiji, which had no Primitive Discard Accelerator (introduced in Polaris). Note the careful use of the term throughput, not rendering. What they really mean is that it can render 4 triangles/clock and discard another 7 via the Primitive Discard Accelerator. There is NO WAY that Vega can actually render more than 4 triangles/clock no matter what the programmer does.

Note that this tendentious claim does NOT appear on any of the release day slide decks. Here is the release day architecture overview slides - please try to find anything backing up the 11 polygons/clock claim. It's not there, because that claim was always dishonest marketing BS. The emphasis in Vega slides is trying to reduce the amount of geometry the card actually has to render. When these tricks run out, Vega is bottlenecked. It WILL NEVER and CAN NEVER match GP102 in geometry-intensive games, because GP102 can actually render 6 triangles/clock and Vega can only actually render 4.

@JDG1980 thanks for articulating that for us.I was reading back about this and wondered what you thought about a couple of things -

AMD notes that the R9 Fury X offers four geometry engines and a peak of four polygons per clock. Comparatively, Vega offers the same number of engines, but they can handle up to 11 polygons per clock – a boost of 2.6x – which is a lot more pixel shifting power, and an impressive leap in performance to say the least.

My questions - (not looking for argument, just understanding)
  • If Fiji could render 4 triangles per clock with four engines wouldn't it be doubled with Vega to 8?
  • How do we know that 7 are discarded to the PDA? Does that require programming for? How does that work?
  • If Vega is "designed" for 11 per clock does this mean it is a "turned on" thing all the time? This implies that if you have 11 vs 6 there could be up to an 80% improvement for Vega with everything else being equal?

And while AMD's presentation and comments itself don't go into detail on how they achieved this increase in throughput, buried in the footnote for AMD's slide deck is this nugget: "Vega is designed to handle up to 11 polygons per clock with 4 geometry engines." So this clearly reinforces the idea that the overall geometry performance improvement in Vega comes from improving the throughput of the individual geometry engines, as opposed to simply adding more as the scalability improvements presumably allow. This is one area where Vega’s teaser paints a tantalizing view of future performance, but in the process raises further questions on just how AMD is doing it.

EDIT:

Source from AMD - Vega’s next-generation geometry pipeline enables the programmer to extract incredible efficiency in processing this complex geometry, while also delivering more than 200% of the throughput-per-clock over previous Radeon architectures.*

*Data based on AMD Engineering design of Vega. Radeon R9 Fury X has 4 geometry engines and a peak of 4 polygons per clock. Vega is designed to handle up to 11 polygons per clock with 4 geometry engines. This represents an increase of 2.6x. VG-3



Thanks
 
Last edited:

Mopetar

Diamond Member
Jan 31, 2011
8,005
6,451
136
I'm inclined to believe that it's a fake (Why would a Chinese site use English for text like Auto/Overclocking?). Even if it isn't, it's pretty useless since it contains no information about what resolution or settings were used.
 

Konan

Senior member
Jul 28, 2017
360
291
106
"AMD also asserts that RX Vega 64 will be able to run over 100 modern games at 4K 60FPS or better."

Just want to add I can't wait to see the reviews on this AMD statement!!
 

railven

Diamond Member
Mar 25, 2010
6,604
561
126
Just want to add I can't wait to see the reviews on this AMD statement!!

That's vague enough to be true. Tinker with enough settings, and you can get lots of games to do 4k@60. And "modern" doesn't exclude indie titles. Hell, "modern" can be argued as art syle versus tech involved.

Hoping for some early leaks, but tomorrow there gonna be a lot of internet traffic!
 
Reactions: Mopetar and Konan

UglyDuckling

Senior member
May 6, 2015
390
35
61
I shut Reddit up today with this video, thought i was bullshitting when i said my R9 Fury handles Fallout 4 on high settings at 4K VSR perfectly.


2 people with 1080Ti's even said... that's impressive.

AMD drivers have pushed Fiji way ahead than they were before.
 

UglyDuckling

Senior member
May 6, 2015
390
35
61
More Fallout 4 to come, put it under true load with some action going on.
A bottleneck appears which is completely weird but only small, video soon.
 

tviceman

Diamond Member
Mar 25, 2008
6,734
514
126
www.facebook.com


Alleged leaked gaming results, take with the usual grain of salt. Seen on r/amd.

64 is really going to struggle against the 1080 from the looks of things, we know it, AMD knows it and AMD knows we know it, hence the recent shift in focus toward the 56. At least that looks solid against the 1070, all comes down to price now.

One thing is for sure, the 1080ti has nothing to worry about, let alone Volta. I'm not even angry about this, just bitterly, bitterly disappointed. How pessimistic did you have to be this time last year to think this is where Vega would end up?

This is faker than fake, unless it's just a one game cherry-picked benchmark. 1080 TI performance is, on average, 35% faster at 4k and 29% faster at 1440p than a 1080.

I'll also attempt to eat a shoe if Vega 64 gets 15% real-world performance increase on air when OCing.
 

UglyDuckling

Senior member
May 6, 2015
390
35
61
Vega is easily going to gain 10-15% Fiji already does with it's non X Fury on air, that would be the Sapphire NITRO OC edition.
 

FatherMurphy

Senior member
Mar 27, 2014
229
18
81

UglyDuckling

Senior member
May 6, 2015
390
35
61
Notice something? optimization of FO brings performance drops from nowhere, but once something goes mental on screen, Fiji holds it's self together incredibly well.. barely drops it's shit.

 

UglyDuckling

Senior member
May 6, 2015
390
35
61
You're already on the fringe, claiming that a Vega 64 is going to get 15% OC gains. But if you're not trying to be edgy, at least try to be realistic. OC gains are going to be laughable, Vega FE already proved that.

Comparing what i would call an AIB, notice i mentioned Sapphire Nitro OC? to an incompetent (for overclocking) reference designed card that is not even aimed at gamers?

I'm not edgy i put forth realistic expectations, Fiji is a great example of how Vega may turn out because like Vega, Fiji is a high TDP GPU that is balls to the wall from the get go, one major difference being Fiji was the first try at HBM from AMD, and it sported from reference a power delivery system only capable of handling the voltage it ran at, unlike Sapphire's Nitro OC range which have beefed up power delivery, 2x 8 pin connector, an unlocked bios to 300 watts if you wish to use it at the touch of a button on the side of the GPU.

Sapphire's Nitro can be compared to the likes of EVGA's Classified edition GPU's, the cooler on the Fiji Nitro OC is amazing.
 

tviceman

Diamond Member
Mar 25, 2008
6,734
514
126
www.facebook.com
Comparing what i would call an AIB, notice i mentioned Sapphire Nitro OC? to an incompetent (for overclocking) reference designed card that is not even aimed at gamers?

I'm not edgy i put forth realistic expectations, Fiji is a great example of how Vega may turn out because like Vega, Fiji is a high TDP GPU that is balls to the wall from the get go, one major difference being Fiji was the first try at HBM from AMD, and it sported from reference a power delivery system only capable of handling the voltage it ran at, unlike Sapphire's Nitro OC range which have beefed up power delivery, 2x 8 pin connector, an unlocked bios to 300 watts if you wish to use it at the touch of a button on the side of the GPU.

Sapphire's Nitro can be compared to the likes of EVGA's Classified edition GPU's, the cooler on the Fiji Nitro OC is amazing.

You're not even talking about the same category card. You're trying to make full-fat Vega 64 OC predictions based on a cut-down Fiji card. Apples and oranges my friend, apples and oranges.
 

tviceman

Diamond Member
Mar 25, 2008
6,734
514
126
www.facebook.com
Please don't strawman me.

How am I strawmanning you? Your basing your Vega 64 OC predictions off your Fury Vanilla OC. Vega FE shows that full fat Vega is already at the brick wall. Reviews incoming in ~14 hours. Lets see what techpowerup, anandtech, hardocp, hardwarecanucks, techspot, and other reputable sites have to say.
 
Status
Not open for further replies.
sale-70-410-exam    | Exam-200-125-pdf    | we-sale-70-410-exam    | hot-sale-70-410-exam    | Latest-exam-700-603-Dumps    | Dumps-98-363-exams-date    | Certs-200-125-date    | Dumps-300-075-exams-date    | hot-sale-book-C8010-726-book    | Hot-Sale-200-310-Exam    | Exam-Description-200-310-dumps?    | hot-sale-book-200-125-book    | Latest-Updated-300-209-Exam    | Dumps-210-260-exams-date    | Download-200-125-Exam-PDF    | Exam-Description-300-101-dumps    | Certs-300-101-date    | Hot-Sale-300-075-Exam    | Latest-exam-200-125-Dumps    | Exam-Description-200-125-dumps    | Latest-Updated-300-075-Exam    | hot-sale-book-210-260-book    | Dumps-200-901-exams-date    | Certs-200-901-date    | Latest-exam-1Z0-062-Dumps    | Hot-Sale-1Z0-062-Exam    | Certs-CSSLP-date    | 100%-Pass-70-383-Exams    | Latest-JN0-360-real-exam-questions    | 100%-Pass-4A0-100-Real-Exam-Questions    | Dumps-300-135-exams-date    | Passed-200-105-Tech-Exams    | Latest-Updated-200-310-Exam    | Download-300-070-Exam-PDF    | Hot-Sale-JN0-360-Exam    | 100%-Pass-JN0-360-Exams    | 100%-Pass-JN0-360-Real-Exam-Questions    | Dumps-JN0-360-exams-date    | Exam-Description-1Z0-876-dumps    | Latest-exam-1Z0-876-Dumps    | Dumps-HPE0-Y53-exams-date    | 2017-Latest-HPE0-Y53-Exam    | 100%-Pass-HPE0-Y53-Real-Exam-Questions    | Pass-4A0-100-Exam    | Latest-4A0-100-Questions    | Dumps-98-365-exams-date    | 2017-Latest-98-365-Exam    | 100%-Pass-VCS-254-Exams    | 2017-Latest-VCS-273-Exam    | Dumps-200-355-exams-date    | 2017-Latest-300-320-Exam    | Pass-300-101-Exam    | 100%-Pass-300-115-Exams    |
http://www.portvapes.co.uk/    | http://www.portvapes.co.uk/    |