AMD Radeon VII Retested With Latest Drivers

[Fouquin]

It doesn't look like a lot, but that's a pretty decent chunk of performance left on the table for the review driver. Seems they could have delayed the launch by a week or two and shown better results.

Meanwhile Fury X can gain 3% from OC, bumps up the powerusage like crazy tho. OVERCLOCKERS DREAM!!!

Average core overclock for the Fury X is 1140MHz. That's 14%, not 3%. It still was not a great overclocker, but there's no need to fudge the numbers to push your narrative. Somewhat related, I tested an overclocked Fury X up against a Galax GTX 1070 EXOC and found the performance difference to be in the 5.5 - 11% range. Surprising to see the Fury X that close, to say the least.

 

[John Naylor]

Is the graph wrong? 5-6% is not multi-digit.

Whaddya talking about ? ... there's a 5 and a 6 . Two digits is more than one ! In the post 2016 era, nothing means what the dictionary says any more. But speaking of multi digit , have to wonder ... are the days of double digit performance increases after manual OC a memory from a bygone era ? Havn't seem much of it from AMD since the 2xx series but w/ nVidia we saw mid teens normally and even more than 31% on some cards. Now both camps seem to be aggressively overclocking the cards in the box, leaving is under 10% to grab on our own.

Stop harping on that. GTX1050Ti has MSRP of $139, RX570 has MSRP of $169. For most part of the lifetime relative prices have reflected that difference only lately moving to where the prices are now.

I still don't see as it matters ... the 1060 remains the better buy over both in "performance per dollar".

Well if you think about it, the maximum price tag for the highest of the high performance GPUs should be MAX $500. This is speaking about the Radeon VII, RTX 2080. Enthusiast price tag should be no more than $600. That would be a RTX 2080Ti and a Radeon VII+ with custom cooling
GPUs, especially the high end versions are all overpriced period.

If you do it accurately and account for inflation, since the year 2000, and excepting the current weirdness resulting from lack of competition at upper end, short supply and even tariffs, the average price of the top dog from nVidia hasn't strayed that much from $700 in 17 years

[/QUOTE]

Since you are fond of numbers that prove things scientifically, here is an example that proves that AMD GPUs aren't shown in their best form at their launch and this helps us customers to get an equal or better product in better price than will deserve in a few months only. And for a customer that keeps his hardware at least for 3 years, this is an opportunity..

There's two sides of that coin and I recently mentioned that in another thread. The 480 in particular improved significantly with the next driver release ... but you're forgetting just one thing. Like the oft heard comment "Well card B might be slower than Card A but when overclocked, card B is almost as fast as card A". The fact remains, card A can be overclocked too making the comment meaningless. Yes, the 480 git a nice bump afterwards and TOU dedicated an entire article to that subject. However, nVidias drivers provided improevements too. And when you look at the 1060 versus the 480, 580 and even the 590, when all 4 cards are manully overclocked, (based upon the data in TPUs test results on this site), the 1060 has still maintained the edge. The 480 OC'd about 6% .... 580 about 4.4% .... 590 did 3.9% .... the 1060 OCs over 18%. So any advantage we saw from those RX cards from aggressive clocking before putting them "in the box" was erased when users took them out of the box and did the manual OCs. Since 2xx series, AMD cards have been more aggressively clocked when taken "outta the box" typically will manually OC only in single digits, while nVidias cards have ranged from the mid teens most of the time (with rare single digit exceptions) to over 31% manual OC over reference. Even if we ignored driver improvements over time, that's a big hurdle to overcome.

Today, with 2xxx / Radeon VII we are seeing a watershed moment in that all of current cards are seeing manual OCs about 8% over reference which makes side by side comparisons easier when using those charts.

Getting back to the "new drivers" issue have to wonder how much of these improvements are real. Is the game actually performing better across the board ? Or has it just been tweaked a bit more to "look good" in that benchmark. That's what makes sites like TPU a "go to" source for me as Wiz doesn't use the IG demo. As was presented in the chart in Post #59, I have not seen any evidence that either side is doing better than the other in this respect .... that chart says otherwise.

https://www.techpowerup.com/forums/...ested-with-latest-drivers.252691/post-3997015

 

[Fluffmeister]

Ouch!!! Fact is, this turd of a card is getting too much attention, it was a token effort at best from AMD whilst they work with their Sony and Microsoft overlords to get Navi into check.... poor show all round AMD.

 

[efikkan]

The only purpose of this card is to have something until the delayed Navi is ready.
But the attention is not the bad part; the bad part is those who claim it's a decent buy, both in forums and among Youtube opinionators.

 

[John Naylor]

Average core overclock for the Fury X is 1140MHz. That's 14%, not 3%. It still was not a great overclocker, but there's no need to fudge the numbers to push your narrative. Somewhat related, I tested an overclocked Fury X up against a Galax GTX 1070 EXOC and found the performance difference to be in the 5.5 - 11% range. Surprising to see the Fury X that close, to say the least.

Core / memory Ocs don't scale to increases in FPS ... in most instances the best OCs on core and memory do NOT correspond at all to highest fps

As you can see above, the highest cores aren't delivering the most fps.

You'll note that "outta the box' the Fury X beat the "slower" 980 Ti (102.6) by 0.3 fps before overclocking .

However, when TPU OC'd the Fury, it brought just 5.1% to the table .... Now let's look at what happened when TPU OCd the 980 Tis

On the MSI Gaming X 980 Ti, they OC's it 27.2% to hit 130.5 fps, 20.7% faster than the Fury X
On the Giga G1 980 Ti, they OC's it 31.4% to hit 130.5 fps, 24.7% faster than the Fury X
On the Zotac AMP X 980 Ti, they OC's it 27.1% to hit 130.4 fps, 20.6% faster than the Fury X
On the Asus Strix X 980 Ti, they OC's it 28,4% to hit 131.7 fps, 21.8% faster than the Fury X

I'm not going to speak to the "narrative", but the math here is clear. I don't care about who wins, but I do care about facts. The overclocking room for the 980 Ti is 5 - 6 times that of the Fury X. If this is "competiing", it's the 1973 Belmont Stakes

 

[Prima.Vera]

This card is unfortunatelly 200$ more than it should have been. Nobody is going to pick this over an 1080Ti for example...

 

[moproblems99]

This card is unfortunatelly 200$ more than it should have been. Nobody is going to pick this over an 1080Ti for example...

and few years too late.

 

[sergionography]

This card is unfortunately 200$ more than it should have been. Nobody is going to pick this over an 1080Ti for example...

Unless they wanted 16GB of fast memory and weren't exclusively gamers. This card has a niche. I hear alot of complains about this card but its honestly not as bad as everyone makes it. Simply because it doesn't blow nvidia out of the water doesn't make it a bad card. rtx2070-2080 performance is not exactly "bad". Also this card is newer than the 1080ti and is likely to be supported for longer(for those who don't buy graphics cards every other year)

 

[EarthDog]

If you compute, it's a great card for the money. If you dont, then it performs worse, uses a lot more power, and is louder than its competition...at the same price point. Compute saves it if those other factors matter.

 

[B-Real]

This card is unfortunatelly 200$ more than it should have been. Nobody is going to pick this over an 1080Ti for example...

Radeon VII buyers get RE2, Division 2 and DMC5. 1080Ti buyers get ZERO games. Thats a $150-180 packet.

Ouch!!! Fact is, this turd of a card is getting too much attention, it was a token effort at best from AMD whilst they work with their Sony and Microsoft overlords to get Navi into check.... poor show all round AMD.

As poor as the whole RTX roundup is.

Yes, RX570 wipes the floor with GTX1050Ti. They are cards from different segments.
RX570 is the intended competitor of GTX1060 3GB.
GTX1050Ti is the intended competitor of RX560.

Pricing in the lowend and midrage is FUBAR.

False, 1050Ti doesn't have an AMD counterpart. GTX 1050 goes against (and trades blows) with the RX560, RX 470/570 against the 1060 3GB and the RX 570/580 against the 1060 6GB. Anyway, it's true that if there is someone who buys a 1050Ti for the same price as the RX 570 4GB, he is just literally stupis af.

 

[Xaled]

If you do it accurately and account for inflation, since the year 2000, and excepting the current weirdness resulting from lack of competition at upper end, short supply and even tariffs, the average price of the top dog from nVidia hasn't strayed that much from $700 in 17 years

Where is Titan in this table?

 

[Jism]

And in the meantime, the Vega VII scoring merely 300 points behind the 2080TI:

https://www.3dmark.com/hall-of-fame-2/timespy+3dmark+score+extreme+preset/version+1.0/1+gpu

It likes water > and many people archieved a clockspeed of over 2250Mhz. From a maximum boost of 1800Mhz that is not bad at all! It's a great OC'er actually.

 

[turbogear]

Radeon VII is not a bad card especially for 2K gaming with Freesync.
It may not match 2080 in every title but for people like me who come from Vega 64 and enjoy Freesync gaming it is a viable upgrade.
I know that Nvidia supports Freesync now but until now only officially on a handful of monitors.
I had a long thought if I should go for 2080 or Radeon VII, but finally decided to go with VII because of concerns with Freesync.
Nvidia does not support as far as I checked my Benq XL2730 freesync monitor and paying between 500€ to 700€ more to get a new good Gsync monitor would have been too much.
Only cooling performance of the refernce cooler on Radeon VII is not good but a good waterblock will fix that.

 

[bajs11]

What the hell are you smoking?

that guy is funny
he tried to turn amd vs nvidia to pc vs console
i wonder what a console peasant is doing here though
this forum is about pc hardware...
they should sit in their moms basement play Mario Kart

 

[metalfiber]

290X
7970
6970
5870
4870
x1900
x1950xtx
x850
9800 pro

Ah yes, I had a ATI 9800 Pro All In Wonder card...loved that card. Watched Tv, recorded tv and then edit that video. ATI was truly the undisputed king back then. I would not mind it too see that again one day.

 

[Rebe1]

But does the new drivers fix the Wattman? Can you do the OC / UV without any problems?

 

[cucker tarlson]

And in the meantime, the Vega VII scoring merely 300 points behind the 2080TI:

https://www.3dmark.com/hall-of-fame-2/timespy+3dmark+score+extreme+preset/version+1.0/1+gpu

It likes water > and many people archieved a clockspeed of over 2250Mhz. From a maximum boost of 1800Mhz that is not bad at all! It's a great OC'er actually.

it is,if you've cooling capacity to deal with a 500w gpu.

Average core overclock for the Fury X is 1140MHz. That's 14%, not 3%.

no,that's 8%

1140 is 1.08x of 1050

Somewhat related, I tested an overclocked Fury X up against a Galax GTX 1070 EXOC and found the performance difference to be in the 5.5 - 11% range. Surprising to see the Fury X that close, to say the least.

that's synthetics,amd always did great in 3dmarks

look at actual games

 

[cyneater]

Yes, RX570 wipes the floor with GTX1050Ti. They are cards from different segments.
RX570 is the intended competitor of GTX1060 3GB.
GTX1050Ti is the intended competitor of RX560.

Pricing in the lowend and midrage is FUBAR.

Pricing is fubur all over the board

 

[londiste]

And in the meantime, the Vega VII scoring merely 300 points behind the 2080TI:
https://www.3dmark.com/hall-of-fame-2/timespy+3dmark+score+extreme+preset/version+1.0/1+gpu

Not really.

Description: 2Way CF 2000MHz/1200MHz

 

[Deleted member 50521]

Not really.

Crossfire and still can’t match a single 2080Ti.

Welp

 

[INSTG8R]

Not really.

Looks like one card to me no?

 

[londiste]

Looks like one card to me no?

Click Detailed Result and read the description. Radeon VII is new and 3DMark's detection isn't the best even when tools are OK.

 

[trog100]

one of the first things i look for with any card are 3Dmark scores.. the radeon vii struck me as very low.. some reviewers include timespy in the testing results.. i think TPU should..

https://hexus.net/tech/reviews/graphics/126752-amd-radeon-vii/?page=4

trog

 

[EarthDog]

Looks like one card to me no?View attachment 116885

https://www.techpowerup.com/forums/threads/radeon-vii-gpu-z-support.252445/page-2#post-3997726

 

[Frutika007]

Remind me how 780Ti fared against 290x.
"Mythical" eh?


Yay, that feeling aspect of AMD products is really something special now, isn't it?

I mean, there should be reasons $139 1050Ti outsells 1.5-2 times faster 570 priced at $99, shouldn't there?

Kepler architecture's failure has nothing do to with AMD's 'finewine' myth. AMD didn't get better,Kepler got worse. Learn the difference. And 1050Ti 139$ and RX 570 99$?? Are you on drugs or something?? You wanna know the reason?? Here's your reason, GTX 1050Ti was launched in 2016 for 130$. RX 570 was launched in 2017 for 170$,also that's precisely when Mining scandal started and AMD gpu just dissapeared from the stock. 1 year gap with 40$ more money is not an appealing deal,not to mention the lack of availability due to mining.

Isn't it the case - the problem for Amd - that ALL games are tested and optimized for nVidia GPU'S?

No,there are many games that are tested and optimized for AMD.

hahaha....


but I hate their (NVIDIA) approach to reduce performance through a driver update to older GPU.
That's beyond my comprehension.

You are totally wrong. Nvidia doesn't reduce performance of older gpu through driver update. That's just a lie and misconception spread by AMD cultists.

Wow, already got -1 rating from that ignorant delusional braindead AMD cultist named medi01

From https://www.techpowerup.com/reviews/AMD/Radeon_VII/

Assassin's Creed Origins (NVIDIA Gameworks, 2017)

Battlefield V RTX (NVIDIA Gameworks, 2018)

Civilization VI (2016)

Darksiders 3 (NVIDIA Gameworks, 2018), old game remaster, where's Titan Fall 2.

Deus Ex: Mankind Divided (AMD, 2016)

Divinity Original Sin II (NVIDIA Gameworks, 2017)

Dragon Quest XI (Unreal 4 DX11, large NVIDIA bias, 2018)

F1 2018 (2018), Why? Microsoft's Forza franchise is larger than this Codemaster game.

Far Cry 5 (AMD, 2018)

Ghost Recon Wildlands (NVIDIA Gameworks, 2017), missing Tom Clancy's The Division

Grand Theft Auto V (2013)

Hellblade: Senuas Sacrif (Unreal 4 DX11, NVIDIA Gameworks)

Hitman 2

Monster Hunter World (NVIDIA Gameworks, 2018)

Middle-earth: Shadow of War (NVIDIA Gameworks, 2017)

Prey (DX11, NVIDIA Bias, 2017 )

Rainbow Six: Siege (NVIDIA Gameworks, 2015)

Shadows of Tomb Raider (NVIDIA Gameworks, 2018)

SpellForce 3 (NVIDIA Gameworks, 2017)

Strange Brigade (AMD, 2018),

The Witcher 3 (NVIDIA Gameworks, 2015)

Wolfenstein II (2017, NVIDIA Gameworks), Results different from https://www.hardwarecanucks.com/for...a-geforce-rtx-2080-ti-rtx-2080-review-17.html when certain Wolfenstein II map exceeded RTX 2080'

From top to the bottom,WRONG WRONG WRONG.

AC Origins gameworks?? It wasn't even nvidia sponsored,nobody sponsored that.
Battlefield 5 gameworks?? Are you having a giggle mate?? Do you even know what gameworks is?? Oh my god the ignorance and delusion in this comment is blowing my mind. Bf5 only has RTX and DLSS. It still uses Frostbite engine which vastly favours AMD.
Let me correct you.

"Assassin's Creed Origins (NVIDIA Gameworks, 2017)" - WRONG

"Battlefield V RTX (NVIDIA Gameworks, 2018)" - WORNG

"Darksiders 3 (NVIDIA Gameworks, 2018), old game remaster, where's Titan Fall 2." - WRONG

"Dragon Quest XI (Unreal 4 DX11, large NVIDIA bias, 2018)" - WRONG

"Ghost Recon Wildlands (NVIDIA Gameworks, 2017)" - WRONG

"Hellblade: Senuas Sacrif (Unreal 4 DX11, NVIDIA Gameworks)" -WRONG

Hitman 2 - it's an AMD title i think,not sure

"Monster Hunter World (NVIDIA Gameworks, 2018)" - not sure,but probably WRONG

"Middle-earth: Shadow of War (NVIDIA Gameworks, 2017)" - WRONG

"Prey (DX11, NVIDIA Bias, 2017 )" - WRONG

"Rainbow Six: Siege (NVIDIA Gameworks, 2015)" - WRONG

"Shadows of Tomb Raider (NVIDIA Gameworks, 2018)" - WRONG,it's only RTX but that hasn't even patched yet

"Wolfenstein II (2017, NVIDIA Gameworks)" - WRONG

Most of the games here aren't even sponsored by nvidia and doesn't have any gameworks in it. The only gameworks game in this list is Witcher 3. Most of the games in your nvidia list are even AMD sposnored title,like Wolfenstein 2 and Prey.

Radeon VII buyers get RE2, Division 2 and DMC5. 1080Ti buyers get ZERO games. Thats a $150-180 packet.


As poor as the whole RTX roundup is.


False, 1050Ti doesn't have an AMD counterpart. GTX 1050 goes against (and trades blows) with the RX560, RX 470/570 against the 1060 3GB and the RX 570/580 against the 1060 6GB. Anyway, it's true that if there is someone who buys a 1050Ti for the same price as the RX 570 4GB, he is just literally stupis af.

Who cares? RE2 already cracked,DMC5 will get cracked, Division 2 will suck as division 1 did. Not a single appealing game. I would rather take price reduction than taking those games. So ultimately 1080Ti is a much better choice. Not to mention 1080Ti came out 2 year ago for the same 699$ price tag. LOL. After 2 years with 7nm process and same 699$ price tag,Radeon 7 still can't beat 1080Ti. How pathetic! Here's the generational improvement i wonder? And RTX 2060 is a great gpu for the price. So the whole RTX roundup isn't poor.