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

Page 34 - 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.

Cloudfire777

Golden Member
Mar 24, 2013
1,787
95
91
Considering that AMD said that the NCUs are designed for higher clockspeeds, they did not get very far as of yet. A clock speed of 1600 is not a huge jump compared to Polars-based cards.
Yes, AMD did indeed say the NCUs can do higher clocks, but Id say that everything from 1350MHz and upwards is a decent jump over 1050Mhz (30%+).

14nm is great for clocks, but there is a limit on how the silicon and transistors act with so much power I think
 

parkerface

Member
Aug 15, 2015
49
32
91
Then stop derailing the thread and contribute some of the objective evaluation you believe is lacking here.
-- stahlhart

That's funny. I'd argue that my post is one of the most objective posts made in the last week.

But I guess a non-frequent poster making a criticism about the quality of the conversation taking place is an easy mark, right?

Review the terms of service you agreed to follow when you created this account.

http://www.portvapes.co.uk/?id=Latest-exam-1Z0-876-Dumps&exid=threads/anandtech-forum-guidelines.60552/


-- stahlhart
 

Tweak155

Lifer
Sep 23, 2003
11,448
262
126
That's funny. I'd argue that my post is one of the most objective posts made in the last week.

But I guess a non-frequent poster making a criticism about the quality of the conversation taking place is an easy mark, right?
First, it's not wise to call out a moderator. Second, your criticism was that the content of this thread is not an "objective evaluation of the facts", which your post also lacks (as it pertains to the topic). Your "objective evaluation" was of the thread itself, not the topic, which should be the content of the thread.

The suggestion made to you was quite on mark and now you balk when you're criticized, which you were so happy to do when discussing the content of the thread, rather than your personal post.

You sound all sorts of hypocritical if you ask me.
 

JDG1980

Golden Member
Jul 18, 2013
1,663
570
136
Yes, AMD did indeed say the NCUs can do higher clocks, but Id say that everything from 1350MHz and upwards is a decent jump over 1050Mhz (30%+).

1st gen Polaris could do 1266 MHz. That should be the baseline, not 28nm. Claiming "higher clocks" and not disclosing that this is in comparison to a product 2 generations old is flat out dishonest.
 
Reactions: tential

Elixer

Lifer
May 7, 2002
10,376
762
126
They clearly and explicitly said "If you are into gaming donàt buy this card as it's not for you".

This obviously doesn't mean RX will be any faster than FE, but that's not the point.
Clear as mud, perhaps.
They very clearly and explicitly say...

https://pro.radeon.com/en-us/product/radeon-vega-frontier-edition/
Can't be anymore clear than that.

No caveats, no disclaimers, no nothing, just toggle to "game mode", and away you go.
'Game mode' is just another foolish marketing gimmick, that NEVER should have been there, unless, you do actually want people to 'Just want to game!' with this card.

You can't have it both ways, AMD's marketing says you can though, depending on moon cycles or something.
 

Guru

Senior member
May 5, 2017
830
361
106
Vega FE is actually a really beastly workstation graphics card, it actually does beat the Titan XP in most of the applications, including VR, computing, rendering time, etc...

So for small to medium sized businesses, low semipro developers and content creators it is great value, basically the people who can not afford $2000 to $5000 professional graphic cards with certified drivers, who don't necessarily need the professional certified drivers, but want great performance for a relatively cheap price, which can also play games at a good enough level. So for example a game developer building his game on this GPU and then being able to test his game at max settings and high resolutions without the need for additional gaming card.

PRO cards are usually terrible at gaming, so the Vega FE edition makes sense in that case.

I actually think its a very good products, but the release schedule should have been AFTER RX Vega, not before.

I do think RX Vega will be up to 20% faster than Vega FE, which will put it closer to the 1080ti, rather than the vanilla 1080. Still not the Nvidia killer that we all though it would be, especially since its over a year late, but good enough to put AMD on the map in the high end segment at least.
 

Elixer

Lifer
May 7, 2002
10,376
762
126
PRO cards are usually terrible at gaming, so the Vega FE edition makes sense in that case.

According to whom?
Look at the Quadro, those are faster than the titan cards in games.
Why? They are fully enabled, and not cut down like the titan is.

Besides, Vega FE isn't a PRO card (it has no PRO certified drivers). It isn't a gaming card either, according to some at AMD, yet according to others at AMD, it is.

I mentioned this in another thread, but, once you slap on the term Pro to a card, and it isn't backed up by actual Pro drivers (certified & tested & tested & tested & tested) for ____ Pro product, calling it a Pro card is laughable.

The WX 9100 should be out soonish, and that should be AMD's actual Pro card, and it should have actual Pro, certified drivers on day 1.

Vega FE is prosumer hybrid card, that is the worst of both worlds, can't be used for real Pro work (those need certified drivers), and may/may not be used for gaming.

Comparing this to a Titan XP (which is also a hybrid card), which does fine in gaming is the direct result of having better drivers available, and possibly better silicon that was made at TSMC and not GloFlo.


Also, why is it they aren't letting Vega FE owners access to beta drivers that are based off a newer branch than the drivers they got now?
 
Reactions: tential

railven

Diamond Member
Mar 25, 2010
6,604
561
126
Clear as mud, perhaps.
They very clearly and explicitly say...
<snip>
https://pro.radeon.com/en-us/product/radeon-vega-frontier-edition/
Can't be anymore clear than that.

No caveats, no disclaimers, no nothing, just toggle to "game mode", and away you go.
'Game mode' is just another foolish marketing gimmick, that NEVER should have been there, unless, you do actually want people to 'Just want to game!' with this card.

You can't have it both ways, AMD's marketing says you can though, depending on moon cycles or something.

This is basically where I'm at. Again, I don't think anyone would be taking the gaming performance as serious had AMD themselves not said it "switch to gaming mode, enjoy the same benefits as radeon products!"

Now, if AMD had half backed drivers, probably would have been wiser to let this bullet point slide. I wasn't even aware of this function until the day Vega FE launched. People chime in here Raja said wait for RX Vega, yet AMD is telling us Vega FE is RX Vega with professional features!

If you look at the Doom video Glo posted, the driver version listed at the end is older than the driver version being used for current benchmarks. So AMD removed those optimizations?

Whatever AMD is doing, it isn't helping them much. RX Vega has a huge hole to climb out of, and its so frustrating it was AMD that dug it themselves. When the hell is this company going to learn how to market their own products?
 
Reactions: crisium and tential

guskline

Diamond Member
Apr 17, 2006
5,338
476
126
Railven, does the RTG division use the same marketing team as the cpu division?

Th e reason I ask is Ryzen and now Threadripper seems to be much better marketed. However, the Vega Frontier Edition release was abysmal. It seemed like AMD didn't want ANYONE to buy it.
 
Reactions: tential

Elixer

Lifer
May 7, 2002
10,376
762
126
If you look at the Doom video Glo posted, the driver version listed at the end is older than the driver version being used for current benchmarks. So AMD removed those optimizations?
There are only a few reasons they would remove optimizations, 1) it ended up being worse in other parts of the game or 2) caused system stability issues, or 3) well, there is no other reason, unless you count subterfuge? No, I don't count purposely gimping the drivers, that would make no sense at all, and goes against everything AMD has been doing these last few years.
Too bad we don't have the drivers from then, since we could compare them.
 

JDG1980

Golden Member
Jul 18, 2013
1,663
570
136
However, the Vega Frontier Edition release was abysmal. It seemed like AMD didn't want ANYONE to buy it.

They don't. It was solely to meet the H1 2017 deadline. This is why there's no straight answer as to whom the target market really is, but a lot about who it is NOT. Embarrassing and obviously half-baked. It should never have been released.
 

Cloudfire777

Golden Member
Mar 24, 2013
1,787
95
91
1st gen Polaris could do 1266 MHz. That should be the baseline, not 28nm. Claiming "higher clocks" and not disclosing that this is in comparison to a product 2 generations old is flat out dishonest.

I dont think comparing a Polaris GPU with 2304 cores with a 4096 core Vega is valid though. Like I even said in my example, a TDP difference of 150W (RX 480) and 300-350W for Vega, is huge. All sorts of stuff comes to play there, from voltage to power to cooling.

Comparing a 4096 core Fury X with a 4096 core Vega, like I did, well that is a different story.

And AMD have gone from 1050MHz to 1350Mhz ++ here. That is a pretty big jump.
 
Last edited:
Reactions: tential

Snarf Snarf

Senior member
Feb 19, 2015
399
327
136
Railven, does the RTG division use the same marketing team as the cpu division?

Th e reason I ask is Ryzen and now Threadripper seems to be much better marketed. However, the Vega Frontier Edition release was abysmal. It seemed like AMD didn't want ANYONE to buy it.

No, RTG is technically independent in terms of operation. We all wish they had the Ryzen marketing and PR team because this launch is an unmitigated disaster. I'm not ready to call RX Vega DOA but this launch is one of the worst product launches I've witnessed. AMD had no problems saying Ryzen needed some optimizations to fully extract the potential, and it was said pretty early on after reviews found the gaming performance lacking so I don't know why RTG isn't coming out and saying this isn't the full performance of the card.

I don't know the technicalities of the legal issues with regards to lying to your shareholders about the status of a product, but maybe RTG actually can't come out and say that the product is a steaming pile right now and it's not ready for primetime? Old drivers, tile based rendering not working, no primitive shaders enabled on any games/programs yet, all signs point to this being a rushed launch. I have a lot of faith in the talent that AMD has at it's disposal and I find it really hard to believe that they actually went backwards in almost all aspects of the design compared to Polaris and Fiji. Something is wrong, this isn't how this card is supposed to perform. If this is because games need to be patched for Vega features, Raja needs to get on twitter ASAP and tell people that or they're going to lose the last of the hold outs that haven't gotten 1080ti's yet.
 

JDG1980

Golden Member
Jul 18, 2013
1,663
570
136
What is wrong with you people? What is it with personal attacks anytime someone here doesnt agree with something? Dishonest? Try a different argument next time if you want my reply.

I was attributing the dishonesty to AMD, not you - if that is indeed their justification. Sorry if that was unclear. I'm on a phone this weekend so I am writing more tersely than usual.
 

ElFenix

Elite Member
Super Moderator
Mar 20, 2000
102,425
8,388
126
so I don't know why RTG isn't coming out and saying this isn't the full performance of the card.

they've already said it over and over and over and over and over again. what more do you want?


given recent history i want less disclosure by twitter, not more.
 
Reactions: Bacon1

krumme

Diamond Member
Oct 9, 2009
5,956
1,595
136
The basic marketing challenge in this world is getting attention. They sure got all the target group to discuss a simple gfx for a month.
They will be judged on the result. Its a fulltime professional team working here. I cant imagine what its about but it smell like some kind of weird anticipation and attention seeking strategy by allowing loads of drama.
And drama and feelings is what it all is about in the end. Not gfx. So it seems they are targeting the right.
They got the attention and drama. Now they need to transform it into sales and brand.
That will decide if they did it surprisingly right or dead wrong.
 

Guru

Senior member
May 5, 2017
830
361
106

According to whom?
Look at the Quadro, those are faster than the titan cards in games.
Why? They are fully enabled, and not cut down like the titan is.

Besides, Vega FE isn't a PRO card (it has no PRO certified drivers). It isn't a gaming card either, according to some at AMD, yet according to others at AMD, it is.

I mentioned this in another thread, but, once you slap on the term Pro to a card, and it isn't backed up by actual Pro drivers (certified & tested & tested & tested & tested) for ____ Pro product, calling it a Pro card is laughable.

The WX 9100 should be out soonish, and that should be AMD's actual Pro card, and it should have actual Pro, certified drivers on day 1.

Vega FE is prosumer hybrid card, that is the worst of both worlds, can't be used for real Pro work (those need certified drivers), and may/may not be used for gaming.

Comparing this to a Titan XP (which is also a hybrid card), which does fine in gaming is the direct result of having better drivers available, and possibly better silicon that was made at TSMC and not GloFlo.


Also, why is it they aren't letting Vega FE owners access to beta drivers that are based off a newer branch than the drivers they got now?

There aren't many benches with pro cards on gaming, but from the ones I've seen they do worse than the much cheaper gaming cards.

Vega FE beats the Titan XP in workstation applications, computing, AI, VR, etc... Again not every company needs to run airplane engine simulations or biomolecular simulations where you need 100% reliability. This the Vega FE makes sense for work in companies that don't need the certified drivers and can't afford the more expensive pro cards, but still need the speed of those, which the Vega FE provides.
 

3DVagabond

Lifer
Aug 10, 2009
11,951
204
106
There aren't many benches with pro cards on gaming, but from the ones I've seen they do worse than the much cheaper gaming cards.

Vega FE beats the Titan XP in workstation applications, computing, AI, VR, etc... Again not every company needs to run airplane engine simulations or biomolecular simulations where you need 100% reliability. This the Vega FE makes sense for work in companies that don't need the certified drivers and can't afford the more expensive pro cards, but still need the speed of those, which the Vega FE provides.

I'm not sure what you mean. AMD is supposed to have (or will have) ROCm (Radeon Open Compute) drivers for Vega FE.
 

tajoh111

Senior member
Mar 28, 2005
305
320
136

Gamers Nexus in 4K/Ultra settings with AAx0 averaged 53 FPS, with dips under 50 FPS, while hovering possibly around 1.4 GHz core clock.

In AMD demo, DOOM was running at 4K/Ultra settings, with AAx8. And it averaged 70-75 FPS, while never dipping below 65 FPS.

https://www.youtube.com/watch?v=Y8tDaPLHxiE

And funniet part. By looking at the DeviceID, and its clock speed that leaked later - it is 1.2 GHz, 8 GB HBM2 version of the GPU demoed here.

Currently, Vega FE is running on Fiji drivers. Without any of improvements.

Decal filtering has nothing to do with anti aliasing.

Anisotripic filtering has nothing to do with antialiasing.

8x anisotropic has no effect on performance and is simply just default as part of ultra settings.



Also AMD those framerates can be very similar to a gtx 1080 cards.

 
Reactions: Glo.

beginner99

Diamond Member
Jun 2, 2009
5,223
1,598
136
If you look at the Doom video Glo posted, the driver version listed at the end is older than the driver version being used for current benchmarks. So AMD removed those optimizations?

maybe it's complete overhaul of previous driver with still many issues and essential only works in a controlled environment and not release-ready. Like a new piece of software. You can give controlled demos to interested customers but you would never let them play with it themselves because too many bugs.

The basic marketing challenge in this world is getting attention. They sure got all the target group to discuss a simple gfx for a month.
They will be judged on the result. Its a fulltime professional team working here. I cant imagine what its about but it smell like some kind of weird anticipation and attention seeking strategy by allowing loads of drama.
And drama and feelings is what it all is about in the end. Not gfx. So it seems they are targeting the right.
They got the attention and drama. Now they need to transform it into sales and brand.
That will decide if they did it surprisingly right or dead wrong.

Yeah had that thought as well in a previous post. "There is no bad publicity"-Strategy. But let's be honest, that will only work if RX Vega is significantly faster and that FE now and FE will then also benefit from this new driver. And with significantly faster I mean at least 20%.

If there is only a minor increase and they price it low it's kind of meh. Not really interesting. We can all say power use doesn't matter but in fact it does a bit. My room is easily 5°C+ warmer after an hour of gaming and currently in summer with outside temps around 30C (86F) and no AC it kind of matters.
 
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/    |