Monday, February 18th 2019

AMD Radeon VII Retested With Latest Drivers

Just two weeks ago, AMD released their Radeon VII flagship graphics card. It is based on the new Vega 20 GPU, which is the world's first graphics processor built using a 7 nanometer production process. Priced at $699, the new card offers performance levels 20% higher than Radeon RX Vega 64, which should bring it much closer to NVIDIA's GeForce RTX 2080. In our testing we still saw a 14% performance deficit compared to RTX 2080. For the launch-day reviews AMD provided media outlets with a press driver dated January 22, 2019, which we used for our review.

Since the first reviews went up, people in online communities have been speculating that these were early drivers and that new drivers will significantly boost the performance of Radeon VII, to make up lost ground over RTX 2080. There's also the mythical "fine wine" phenomenon where performance of Radeon GPUs significantly improve over time, incrementally. We've put these theories to the test by retesting Radeon VII using AMD's latest Adrenalin 2019 19.2.2 drivers, using our full suite of graphics card benchmarks.
In the chart below, we show the performance deltas compared to our original review, for each title three resolutions are tested: 1920x1080, 2560x1440, 3840x2160 (in that order).



Please do note that these results include performance gained by the washer mod and thermal paste change that we had to do when reassembling of the card. These changes reduced hotspot temperatures by around 10°C, allowing the card to boost a little bit higher. To verify what performance improvements were due to the new driver, and what was due to the thermal changes, we first retested the card using the original press driver (with washer mod and TIM). The result was +0.2% improved performance.

Using the latest 19.2.2 drivers added +0.45% on top of that, for a total improvement of +0.653%. Taking a closer look at the results we can see that two specific titles have seen significant gains due to the new driver version. Assassin's Creed Odyssey, and Battlefield V both achieve several-percent improvements, looks like AMD has worked some magic in those games, to unlock extra performance. The remaining titles see small, but statistically significant gains, suggesting that there are some "global" tweaks that AMD can implement to improve performance across the board, but unsurprisingly, these gains are smaller than title-specific optimizations.

Looking further ahead, it seems plausible that AMD can increase performance of Radeon VII down the road, even though we have doubts that enough optimizations can be discovered to match RTX 2080, maybe if suddenly a lot of developers jump on the DirectX 12 bandwagon (which seems unlikely). It's also a question of resources, AMD can't waste time and money to micro-optimize every single title out there. Rather the company seems to be doing the right thing: invest into optimizations for big, popular titles, like Battlefield V and Assassin's Creed. Given how many new titles are coming out using Unreal Engine 4, and how much AMD is lagging behind in those titles, I'd focus on optimizations for UE4 next.
Add your own comment

182 Comments on AMD Radeon VII Retested With Latest Drivers

#1
ShurikN
So basically, aside from 2 titles, everything is within margin of error.
Posted on Reply
#2
Flyordie
Really, all they should do is focus on optimizing for the major game engines.

I mean over time, there will be more and more refinements as their driver teams discover more and more ways to optimize. Right now, this seems more like a bug fix and stability check update to the drivers more than anything.
Posted on Reply
#3
phill
I still like the card and performance it gives, it's a shame that it's not all it can be tho. Here's hoping some time it will be just where it needs to be :)
Posted on Reply
#4
bug
Since the first reviews went up, people in online communities have been speculating that these were early drivers and that new drivers will significantly boost the performance of Radeon VII, to make up lost ground over RTX 2080.
Can anyone remember the last AMD launch where disappointing performance wasn't expected to be mitigated in upcoming drivers?
There's also the mythical "fine wine" phenomenon where performance of Radeon GPUs significantly improve over time, incrementally.
Which won't go away, despite HardOCP looking at AMD's drivers over time and finding very modest increases, with a few (one) exceptions: https://www.hardocp.com/article/2017/01/30/amd_video_card_driver_performance_review_fine_wine/
Posted on Reply
#5
londiste
Taking a closer look at the results we can see that two specific titles have seen significant gains due to the new driver version. Assassin's Creed Odyssey, and Battlefield V both achieve multi-digit improvements, looks like AMD has worked some magic in those games, to unlock extra performance.
Is the graph wrong? 5-6% is not multi-digit.
Posted on Reply
#6
robert3892
I'd like to test this myself but sadly my Radeon 7 arrived as a Dead on Arrival unit and I had to send it back for a replacement. I suppose I'll get the replacement next week.
Posted on Reply
#7
medi01
W1zzard, post: 3996684, member: 1"
the mythical "fine wine"
Remind me how 780Ti fared against 290x.
"Mythical" eh?

bug, post: 3996827, member: 157434"
AMD launch where disappointing performance
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?
Posted on Reply
#8
W1zzard
londiste, post: 3996828, member: 169790"
Is the graph wrong? 5-6% is not multi-digit.
Whoops, I meant "multiple-percent", fixed now :)
Posted on Reply
#9
londiste
medi01, post: 3996838, member: 158537"
I mean, there should be reasons $139 1050Ti outsells 1.5-2 times faster 570 priced at $99, shouldn't there?
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.
Posted on Reply
#10
Zubasa
medi01, post: 3996838, member: 158537"
I mean, there should be reasons $139 1050Ti outsells 1.5-2 times faster 570 priced at $99, shouldn't there?
Yes, because you see nVidia logo pop up more often in Games, on ads etc.
The average user might not even have an FPS counter on, they just look at the "safe" option.
They see nVidia / Intel on system requirements they buy that.
Posted on Reply
#11
medi01
londiste, post: 3996847, member: 169790"
relative prices have reflected that differenc
BS. 570 wiped the floor with 1050Ti upfront.

There is the same story of 280/280x vs 960, which was a terrible thing to buy, if one had to buy from green, 970 was there.
But it seems to be hard to accept that hordes of consumers make ill informed GPU purchases, for some reason.
Posted on Reply
#12
londiste
medi01, post: 3996854, member: 158537"
BS. 570 wiped the floor with 1050Ti upfront.

There is the same story of 280/280x vs 960, which was a terrible thing to buy, if one had to buy from green, 970 was there.
But it seems to be hard to accept that hordes of consumers make ill informed GPU purchases, for some reason.
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.
Posted on Reply
#13
JB_Gamer
Isn't it the case - the problem for Amd - that ALL games are tested and optimized for nVidia GPU'S?
Posted on Reply
#14
xkm1948
FineWine hahahahahaha.

Nope. If I pay for full price on day 1 I expect full performance on day 1.
Posted on Reply
#15
bug
JB_Gamer, post: 3996866, member: 173137"
Isn't it the case - the problem for Amd - that ALL games are tested and optimized for nVidia GPU'S?
Not really. Games are calling DX API, they don't know what's going on below that level. Optimizations are more on the level of "don't try to use that feature extensively, because it will bring the hardware to its knees" or "while you do X, you can also do Y, because the hardware can process that in parallel".
It's on;y when you take those optimizations and try to apply them to all hardware that you can get in trouble. But besides a couple of titles that went overboard with hairworks, I don't recall developers being that dumb.
Posted on Reply
#16
INSTG8R
My Custom Title
JB_Gamer, post: 3996866, member: 173137"
Isn't it the case - the problem for Amd - that ALL games are tested and optimized for nVidia GPU'S?
Many on that list are AMD games. FC5,AC, and a couple more I can’t be arsed to install and confirm.
Posted on Reply
#17
las
medi01, post: 3996838, member: 158537"
Remind me how 780Ti fared against 290x.
"Mythical" eh?
Remind he how 980 Ti fared against Fury X.
7970 was the last good AMD GPU. It's been 7 years.
AMDs GPU business is in a worse state than ever now. I miss ATi.
Fury X, Vega 64 and now Radeon VII - All cards have been a complete joke.
Posted on Reply
#18
xkm1948
las, post: 3996885, member: 111974"
Remind he how 980 Ti fared against Fury X.
Yeah man dont wanna even think about that overclocker’s dream...
Posted on Reply
#19
Countryside
Let try to keep this thread clean of red is bad and green is good or vice versa
Posted on Reply
#20
las
Countryside, post: 3996889, member: 131094"
Let try to keep this thread clean of red is bad and green is good or vice versa
I'm simply stating facts. I would wish that AMD GPU's were able to compete.. All their "high-end" solutions have been terrible recently. Leaving this segment for Nvidia.
Posted on Reply
#22
medi01
xkm1948, post: 3996868, member: 50521"
Nope. If I pay for full price on day 1 I expect full performance on day 1.
You mean, when you buy 780Ti, you expect it to be beaten by 960 later on, right?

las, post: 3996885, member: 111974"
Remind he how 980 Ti fared against Fury X.
Why the hell not!
You might discover interesting things, if you check how results were changing over time in TP charts.
Fury used to beat 980Ti only at 4k at launch (and even then, barely) at 1440p it was about 10% behind:
https://www.techpowerup.com/reviews/AMD/R9_Fury_X/31.html


And were are we now:
https://www.techpowerup.com/reviews/AMD/Radeon_RX_Vega_64/31.html


From 10% behind to several % ahead. Not bad, is it?

And one more point: FineWine in general refers to graceful aging of AMD cards, especially in contrast with that nVidia does with its customers.
Not about some magic dust coming into play later on or AMD purposefully crippling it at launch (the way certain other company does)
Posted on Reply
#23
Countryside
las, post: 3996890, member: 111974"
I'm simply stating facts. I would wish that AMD GPU's were able to compete.. All their "high-end" solutions have been terrible recently. Leaving this segment for Nvidia.
Everybody has a right to their own opinion but with hardware opinion matters not and facts can be bent to serve once narrative.
Posted on Reply
#24
bug
Imsochobo, post: 3996891, member: 66457"
290X
7970
6970
5870
4870
x1900
x1950xtx
x850
9800 pro
So, over 5 years ago. Or 5 generations ago. Gotcha.
Posted on Reply
#25
notb
medi01, post: 3996897, member: 158537"
And one more point: FineWine in general refers to graceful aging of AMD cards, especially in contrast with that nVidia does with its customers.
Not about some magic dust coming into play later on or AMD purposefully crippling it at launch (the way certain other company does)
Man, it's you again on your vendetta against Nvidia. I hoped you would give up after R VII launch.

FineWine doesn't refer to "graceful aging of AMD cards". It refers to AMD not being able to provide proper drivers at the time of launch. So with Nvidia you get that extra 10% the first day, and with AMD you have to wait.

Basically, you just said that instead of getting $1000 today you'd rather get by monthly installments over a year, because then you'd have the sense of earning money.

Also, I would love to learn a way to revoke your rating rights, because you're just running around giving a -1 to anyone who doesn't share your love for Radeon chips. It undermines the already little sense that ranking system has.
Posted on Reply
Add your own comment