AMD 390x Improvement over 290X is because of driver

[dn7309]

Last week when I saw the new AMD driver release notes, they say it only affect new 390x, 200 series are not supported. I decide to dig around and see what gives. I came across people posting image of the two cards and the specs are identical (except for 8gb). Google send me to an online forum where some one posted a dutch review site who claimed that all of the improvement that came with the 390x is a because of the driver.

I haven't seen anyone post it here yet.

http://www.pcgameshardware.de/AMD-Radeon-Grafikkarte-255597/Specials/Radeon-R9-390X-Test-1162303/#a5

And to be honest, I'm a little pissed that AMD did this But I guess business is business.

To be fair the site did speculate that future "Omega" driver may improve all GCN cards... but I'll hold on to my opinion until I see that I happen.

Just for anyone who may care I have an XFX DD 290x

 

[YBS1]

To be fair the site did speculate that future "Omega" driver may improve all GCN cards... but I'll hold on to my opinion until I see that I happen.

Yeah, I wouldn't be too quick to jump on AMD about this issue just yet for that very reason. However if this goes bad, there are a lot team red fans here that will need to eat some crow over the Kepler/Maxwell hubbub.

Read the rules for the AMD sub forums. This sub forum is not for AMD vs Nvidia
Moderator Subyman

 

[antihelten]

http://www.pcper.com/news/Graphics-...rformance-Check-Validating-AMD-R9-390-Testing

Apparently the new drivers generally doesn't have a great influence on actual game performance.

 

[96Firebird]

If I'm reading that right, AMD improved tessellation performance of Hawaii with 15.15. I'd still like to see a flashed 290X OC'd to 390X levels compared.

 

[gamervivek]

No tessellation improvements on 290X using the 300 driver on Heaven benchmark here,

http://www.reddit.com/r/hardware/co...sellation_is_not_improved_performance/csby37k

But there are some improvements,

AMD'd 15.5

API overhead test

DX11 multi 874 308
DX11 Single 931 772
Mantle 9 777 904

Modded 15.15

DX11 multi 1 120 471
DX11 single 1 165 430
Mantle 9 808 061

http://forums.overclockers.co.uk/showpost.php?p=28195575&postcount=6122

 

[3DVagabond]

If I'm reading that right, AMD improved tessellation performance of Hawaii with 15.15. I'd still like to see a flashed 290X OC'd to 390X levels compared.

This^^^ Rather than downclocking the 390X I'd like to see tests of the 290X at the 390X's clocks.

One reason I haven't seen hypothesized for Grenada's improvement is if the VRAM is better quality. Possibly offering better timings at 6GHz than what was used with the 290X? Just a thought.

 

[96Firebird]

Ryan Smith pointed to better timings of the VRAM in his 300 series preview, or whatever it was. It's weird though, looking at the part number and specs for the GDDR5 chips given by W1zzard over at TPU in the MSI 390X review, they are actually spec'd for 5GHz, not 6. I thought that was strange, but I didn't see any further explanation...

I would link, but I'm on my phone and it's a pain.

 

[gamervivek]

http://www.pcper.com/news/Graphics-...rformance-Check-Validating-AMD-R9-390-Testing

Apparently the new drivers generally doesn't have a great influence on actual game performance.

He should test at lower resolutions if that improved draw calls thing is right. And witcher with hairworks on.

 

[happy medium]

future "Omega" driver may improve all GCN cards...

But for now we are just making you think the 390 cards are better in all the reviews you have read..

Love the word "may"

 

[3DVagabond]

But for now we are just making you think the 390 cards are better in all the reviews you have read..

Love the word "may"

Sure, they could have improved the performance of the 290/X months ago to be more competitive with the 980/970 but AMD decided to just tank sales against nVidia and sell their cards below market value so they could sell some model in the future.

Read the rules for the AMD sub forum. This is not for creating nvidia vs AMD arguments.
Moderator Subyman

 

[krumme]

But for now we are just making you think the 390 cards are better in all the reviews you have read..

Love the word "may"

Facts:
20% higher membandwith
Better timings for memory
5% faster core
Better board control eg silent fan control
As fast as 980
Silent and cool

Guessing:
Probably 5% faster than 980 in a year.

Ofcource the rest gets the same benefits of driver ones the reviews is written. Those minor tessalation benefits is of no importance vs the above improvements.

Amd improving driver and nv using their energy on crapworks i have no problem with amd for once focusing on first impression last.
Hell what have nv done the last 2 years. Nothing but crapworks and first impression last focus? Spin spin spin all the way.
While amd invented a new driver generation nv gave us crapworks and 180w cards in our desktop. Nothing like the old nv i knew.
The old nv gdx i have in the house is crap because of that strategy.
Never more nv unless they prove they care just a little about the consumer.

Keep it about AMD.
Moderator Subyman

 

[cusideabelincoln]

Yeah, I wouldn't be too quick to jump on AMD about this issue just yet for that very reason. However if this goes bad, there are a lot team red fans here that will need to eat some crow over the Kepler/Maxwell hubbub.

No one should have to eat crow because it's bad no matter who does it.

Curse words are not allowed in the technical forums
Moderator Subyman

 

[railven]

Now we see AMD doing something similar, <in regards to holding back optimizations used on the new line from the old line>

<We shouldn't allow either company to cause us to pick one brand over the other for our gaming needs> I hope AMD responds the exact same way Nvidia does "oh there was a bug, we found it, it will be fixed."

AMD is slowly starting to copy Nvidia page for page. <If AMD does follow suit and Gaming Evolved games reach Greenwork levels of optimization, we could be facing situations were the games we play might as well be considered "exclusive">

"Some people say Nvidia" in the AMD sub forum and they get an infraction. Do not respond to those posts, report them.
Moderator Subyman

USER EDIT: Cleaning up my mess, apologies to anyone who misunderstood my post.

 

[3DVagabond]

Some of us say Nvidia tanked Kepler performance, some people respond "they already hit max optimization on Kepler, move on."

Now we see AMD doing something similar, some people say "LOL, so much for improved hardware."

The flag waving around here is amazing. No better than consolers at times. I hope AMD responds the exact same way Nvidia does "oh there was a bug, we found it, it will be fixed."

AMD is slowly starting to copy Nvidia page for page. If Gaming Evolved games reach Nvidia stages of "optimizations" oh man will I die of popcorn consumption.

Since they were running the 290X on prior drivers how exactly did they tank it's performance? Voodoo?

 

[railven]

Since they were running the 290X on prior drivers how exactly did they tank it's performance? Voodoo?

What are you talking about? I didn't say they tanked 290X performance.

EDIT: OHHHH, you inferred me saying some of us claimed Kepler got tanked as 290X got tanked, no no. More so, the 390X "optimizations" weren't given to the 290X, yet, thus an artificial performance improvement (of course this is all based on current "guys, I read this on the internet so it must be true" witch hunt claims).

IF, big if, this is true and AMD DID do as implicated, I said I hope they respond the same way Nvidia did - thus effectively puncturing the complainers' rebuttal. Then I went into hyper "us vs them" mode with my finisher. (Because I really don't want to have to swap my GPU out to play different games on the PC.)

 

[Udgnim]

compare [H]'s #s vs pclabs with HairWorks on

http://www.hardocp.com/article/2015/06/18/msi_r9_390x_gaming_8g_video_card_review/3#.VYa5i_lVhV8

http://pclab.pl/art64437-33.html

pclabs labs shows the R9 39X with a performance lead but no where near as large as [H]

other thing I noticed was that after market R9 390 kept performing similar to after market R9 290X in their Witcher 3 benchmarks

 

[gamervivek]

Did they benchmark using the older driver or just added a few percent more performance to the 290X?

 

[Abwx]

From what Ryan Smith published it s clear that the silicon is slightly better, 20% faster memory controler is a hint that internal latencies could have been reduced as well, hence the reason why they grabbed some perf/Mhz over the previous iterations.

For this reason i wouldnt advise flashing 290X with 390X bios as the former isnt automaticaly up to the low level new settings, those who say that it s possible should buy a 290X and do the experiment at their own expenses eventualy...

 

[Subyman]

Stop making this a Nvidia vs AMD thread. Check the forum we are in right now. AMD. Keep it about AMD. If you want to discuss Nvidia vs AMD, do it in general. Anyone that mentions Nvidia drivers after this will get an infraction. If someone tries to make it about Nvidia vs AMD, report them and move on.

Moderator Subyman

 

[Hitman928]

From pcper link.

So what's the take away? While in Grand Theft Auto V there is some performance delta between Catalyst 15.5 and Catalyst 15.15 with the R9 290, that is not the case in Battlefield 4 or Metro: Last Light. (Note: the results at 4K testing, which I did run, showed identical behavior to the 2560x1440 testing shown above.) It makes sense that GTA V would see some improvement in gaming experience with a newer driver as it is still a new title getting updates and fixes from both AMD and NVIDIA. BF4 and Metro: LL are quite a bit longer in the tooth and thus the lack of change is expected.
But, there does not appear to be any kind of smoking gun to point to that would indicate AMD was purposefully attempting to improve its stance through driver manipulation. And that's all I wanted to make sure of with this testing today and this story. I obviously didn't test every game that users are playing today, but all indications are that AMD is in the clear.

I'd like to see it tested some more with more games, but initially at least it doesn't seem to matter in games.

 

[monstercameron]

From pcper link.







I'd like to see it tested some more with more games, but initially at least it doesn't seem to matter in games.

Just wow, another gpu brand handicaps their previous gen gpu and we don't hear mum about it from the tech press but the rumor gets spread that amd is doing the same thing and they get interrogated. I don't see how amd can ever be prosperous with the press doing all it can to slander and paint amd in such a light.
/rant

 

[Innokentij]

I saw this test where they clocked the 290X and 390X the same and if u look from the benchmark it's the same card.

http://www.hardocp.com/article/2015/06/18/msi_r9_390x_gaming_8g_video_card_review/9

 

[railven]

From pcper link.



I'd like to see it tested some more with more games, but initially at least it doesn't seem to matter in games.

That's good to read, and at least the press covered it promptly this time. If I had to guess, they didn't want to end up with egg on their face like last time when it was users who discovered issues in other situations, before they did.

 

[railven]

I saw this test where they clocked the 290X and 390X the same and if u look from the benchmark it's the same card.

http://www.hardocp.com/article/2015/06/18/msi_r9_390x_gaming_8g_video_card_review/9

Wish that would include power consumption. If AMD put some efficiency tweaks, it would show up there.

 

[IEC]

I saw this test where they clocked the 290X and 390X the same and if u look from the benchmark it's the same card.

http://www.hardocp.com/article/2015/06/18/msi_r9_390x_gaming_8g_video_card_review/9

That test is flawed, because they underclocked the 390X to match the 290X speeds. So any gains you would obtain from the better memory speed and/or timings at the 390X memory speed are lost.