Regarding the RX 480 AOTS benchmark at Computex

[zMeul]

14 minutes ago, shdowhunt60 said:

Someone's mad.

gee .. you think?

 

Robert Hallock, Roy Taylor and Richard Huddy in particular have done more harm to AMD than anyone else before them

 

---

 

here's some other stuff, maybe @Fulgrim should add to the OP:

good question: how the fuck did AMD get their hands on a pre-release driver?

if AMD would've bought the GTX1080, like everyone else, they would've used the WHQL driver - this means that someone broke the NDA and gave AMD access

[MrDynamicMan]

3 hours ago, zMeul said:

gee .. you think?

 

Robert Hallock, Roy Taylor and Richard Huddy in particular have done more harm to AMD than anyone else before them

 

---

this means that someone broke the NDA and gave AMD access

Lol. Think you might be the only one who cares.

[alamox]

4 hours ago, zMeul said:

gee .. you think?

 

Robert Hallock, Roy Taylor and Richard Huddy in particular have done more harm to AMD than anyone else before them

 

---

 

here's some other stuff, maybe @Fulgrim should add to the OP:

good question: how the fuck did AMD get their hands on a pre-release driver?

if AMD would've bought the GTX1080, like everyone else, they would've used the WHQL driver - this means that someone broke the NDA and gave AMD access

 

really thats where your outrage stops ? why the hek does Nvidia have press only drivers with broken renderer that affect benchmark, and never released to the public ?

i think the last one is way more concerning.

from now on i will not take into consideration any benchmark using press only driver, reviews have to use public ones, and stop this shady crap.

[zMeul]

32 minutes ago, alamox said:

really thats where your outrage stops ? why the hek does Nvidia have press only drivers with broken renderer that affect benchmark, and never released to the public ?

i think the last one is way more concerning.

really?! that's your concern?

 

it affected the benchmark how? except visually, since that was already established - and can you explain to me how no one in AMD's camp never looked at the videos side by side and saw the differences? only the people watching the stream ?!

AMD got their hands on a driver that they weren't supposed to have

• they didn't bother to link to the tests that showed the exact numbers showed on stage
• they didn't bother to re-do the tests

as for your concern nVidia knew about AMD's RX480 comparison, somehow , and cheated - there are a lot of runs on Ashes bench page with 1440p crazy preset with results higher than AMD's GTX1080 run - those were done with the publicly available WHQL driver

 

AMD fucked up and people responsible for that PR bullcrap should burn in hell for their lies

nVidia didn't put that driver in AMD's hands - AMD is solely responsible

 

----

 

do you know what's funny? none of the GTX1080 runs on RadeonDemo's profile are earlier than the 27th: http://www.ashesofthesingularity.com/metaverse#/personas/b0db0294-8cab-4399-8815-f956a670b68f/match-details/cfe81b3f-f8d8-4c98-9e03-41978350fa02

[79wjd]

19 minutes ago, alamox said:

 

really thats where your outrage stops ? why the hek does Nvidia have press only drivers with broken renderer that affect benchmark, and never released to the public ?

i think the last one is way more concerning.

from now on i will not take into consideration any benchmark using press only driver, reviews have to use public ones, and stop this shady crap.

Wait wait wait....its concerning to you that Nvidia is still working on drivers and basically only has beta drivers available for reviewers? How dare they not have the official release drivers ready weeks in advance. 

[Trixanity]

6 minutes ago, zMeul said:

really?! that's your concern?

 

it affected the benchmark how? except visually, since that was already established - and can you explain to me how no one in AMD's camp never looked at the videos side by side and saw the differences? only the people watching the stream ?!

AMD got their hands on a driver that they weren't supposed to have

• they didn't bother to link to the tests that showed the exact numbers showed on stage
• they didn't bother to re-do the tests

as for your concern nVidia knew about AMD's RX480 comparison, somehow , and cheated - there are a lot of runs on Ashes bench page with 1440p crazy preset with results higher than AMD's GTX1080 run - those were done with the publicly available WHQL driver

 

AMD fucked up and people responsible for that PR bullcrap should burn in hell for their lies

nVidia didn't put that driver in AMD's hands - AMD is solely responsable

There is a large probability that the less visually demanding run would result in less demand on the GPU resulting in higher scores (albeit probably negligible values).

Once again, they worked with what they have. The release driver was available after the deadline. Don't know how many times you have to be told that. 

 

I'll pretty much guarantee you that Nvidia will have access to the RX480 and pre-release drivers before it's released to the public. That's how the industry operates. They all have their sources. I'm certain Intel will also get slipped some Zen samples in Q3 when they will be sampled around to OEMs.

 

AMD didn't create the debacle. Children raging, like you do, did. It was a non-issue until people were saying that the 1080 was visually much better looking, so someone had to point out that it was in fact not because it didn't run the game properly. 

 

It's just a matter of bad timing. The release of the card itself and corresponding drivers coincided with Computex. If it happened a week or two later, none of it would have happened. The only other alternative would have been to test against Maxwell, which they probably should have since people could have extrapolated performance from that and they could have pointed out that it was also faster than the 1080. However, that would probably have resulted in you saying "But 1080 is just released. It's so much better. Why didn't they use the 1080??!??!?!?!".

 

This is why we can't have nice things.

[Kobz360]

14 hours ago, zMeul said:

what are you talking about?

 

AMD used pre-release drivers instead of the publicly available 368.25 WHQL, since May 26th

This entire driver fiasco was highly unfortunate for everyone and caused avoidable conflicts of interest. However, all this does is reaffirm the fact that the 480s were running at the same settings as the 1080.. after all, that was the entire point of this thread and you, along with others were saying earlier that the 1080 was running at higher settings. 

 

In that aspect, you were wrong.

 

 

[Kobz360]

40 minutes ago, zMeul said:

do you know what's funny? none of the GTX1080 runs on RadeonDemo's profile are earlier than the 27th: http://www.ashesofthesingularity.com/metaverse#/personas/b0db0294-8cab-4399-8815-f956a670b68f/match-details/cfe81b3f-f8d8-4c98-9e03-41978350fa02

Uh, isn't that because the 1080 launched on the 27th? lol

 

Edit: I see the context of your post now, very interesting. But once again the purpose of this thread has been reached as I've discussed briefly above

[zMeul]

12 minutes ago, Kobz360 said:

480s were running at the same settings as the 1080.. after all

I still see some image discrepancies in the areas not covered with snow, in the mountain ridges - my guess is that will discuss it again one the RX480 after release when people get their hands on two of them

 

and there's the matter of the missing 2 x RX480 runs from AMD's Ashes profile - the only existing result doesn't even come close to the numbers AMD showed: http://www.ashesofthesingularity.com/metaverse#/personas/b0db0294-8cab-4399-8815-f956a670b68f/match-details/ac88258f-4541-408e-8234-f9e96febe303

at most, it puts the 2x RX480 on same numbers as a single GTX1080

 

Quote

Uh, isn't that because the 1080 launched on the 27th? lol

368.25 was available since the 26th: http://www.geforce.com/drivers/results/103610

how did AMD get their hands on 368.19 ?!?! and how did they tested on the 27th with pre-release drivers instead of the release one!?!?

[Kobz360]

1 minute ago, zMeul said:

I still see some image discrepancies in the areas not covered with snow, in the mountain ridges

 

and there's the matter of the missing 2 x RX480 runs from AMD's Ashes profile - the only existing result doesn't even come close to the numbers AMD showed: http://www.ashesofthesingularity.com/metaverse#/personas/b0db0294-8cab-4399-8815-f956a670b68f/match-details/ac88258f-4541-408e-8234-f9e96febe303

at most, it puts the 2x RX480 on same numbers as a single GTX1080

Bruh are we seriously still disputing this? An AotS developer himself literally said it was the driver glitch. Let it go mate. 

 

And yes, the 1080p runs (of both the 1080 and the 480s) shown at Computex aren't on that account. Perhaps their on another one, perhaps their somewhere else, who knows. Regardless, looking at the 1440p results of the 480s vs. the 1080 runs, the story is pretty similar. The 1440p runs show the 480s sometimes edging out the 1080s, and sometimes losing to the 1080s. There's also obviously 1 (crazy preset) 1440p (dual 480) run to compare the multiple 1080 runs with though which is problematic. 

 

Anyway, we'll see when the card launches. 

[xAcid9]

52 minutes ago, Trixanity said:

AMD didn't create the debacle. Children raging, like you do, did. It was a non-issue until people were saying that the 1080 was visually much better looking, so someone had to point out that it was in fact not because it didn't run the game properly. 

+1 

[alamox]

3 hours ago, zMeul said:

really?! that's your concern?

 

it affected the benchmark how? except visually, since that was already established - and can you explain to me how no one in AMD's camp never looked at the videos side by side and saw the differences? only the people watching the stream ?!

AMD got their hands on a driver that they weren't supposed to have

• they didn't bother to link to the tests that showed the exact numbers showed on stage
• they didn't bother to re-do the tests

as for your concern nVidia knew about AMD's RX480 comparison, somehow , and cheated - there are a lot of runs on Ashes bench page with 1440p crazy preset with results higher than AMD's GTX1080 run - those were done with the publicly available WHQL driver

 

AMD fucked up and people responsible for that PR bullcrap should burn in hell for their lies

nVidia didn't put that driver in AMD's hands - AMD is solely responsible

 

----

 

do you know what's funny? none of the GTX1080 runs on RadeonDemo's profile are earlier than the 27th: http://www.ashesofthesingularity.com/metaverse#/personas/b0db0294-8cab-4399-8815-f956a670b68f/match-details/cfe81b3f-f8d8-4c98-9e03-41978350fa02

do you even understand what's going on ? just so that you know any information released to  reviewers or even AIBs/OEMs end up at both vendors, thats just how it works.

the issue again that a GPU vendor send specific drivers to reviewers ONLY, these reviewers do benchs with those drivers to show case the performance of that GPU to the public, if those special drivers contain more of these renderer discrepancies they could have a significant impact on the reviewed performance and the real world performance of a GPU.

that is why this is leagues more concerning than AMD getting hold of that very special driver, and if you dont see it then you have a serious problem.

this could very well be an isolated issue, or could be much worse, it's not from me to investigate, but you seem to completly dismiss it to try and hang to something really trivial in comparaison, and wide spread in the industry, corporate spying isn't new, and this hardly qualifies, if you release driver to hundreds of ppl and expect it not to leak.

[zMeul]

1 hour ago, alamox said:

, these reviewers do benchs with those drivers to show case the performance of that GPU to the public, if those special drivers contain more of these renderer discrepancies they could have a significant impact on the reviewed performance and the real world performance of a GPU.

oh yeah I understand - and it's the reviewers job to verify the release driver against what nVidia has provided them; nVidia gave them all the tools, all they needed was to download them

and you know the interesting part? most reviewers used the 368.10 or 368.13 and not 368.19

 

Quote

if you release driver to hundreds of ppl and expect it not to leak.

I expect that AMD's people to behave like professional - they didn't; and it's not the 1st time they blame anyone but themselves for the mess they created

[alamox]

36 minutes ago, zMeul said:

oh yeah I understand - and it's the reviewers job to verify the release driver against what nVidia has provided them; nVidia gave them all the tools, all they needed was to download them

and you know the interesting part? most reviewers used the 368.10 or 368.13 and not 368.19

 

I expect that AMD's people to behave like professional - they didn't; and it's not the 1st time they blame anyone but themselves for the mess they created

easier yet, why not make them public, reviewers in the first place shouldn't accept to test hardware with drivers that are never released to the public, if the driver is not ready then the hardware is not ready, and often these type of things are spoted by chance when millions use it, so it's not the reviewer's job to find that out.

do you really believe Nvidia doesn't spy on AMD ? with this behave like professionals ? if a corp doesnt spy on their competitor then there must be a problem, and the release of latest top end GPUs peformance and time frame should be clear enough on how good nvidia is at spying on AMD.

[Bouzoo]

48 minutes ago, zMeul said:

and it's not the 1st time they blame anyone but themselves for the mess they created

I'm sorry, I must be missing something. AMD blames who for what? 

[MrDynamicMan]

13 hours ago, Bouzoo said:

I'm sorry, I must be missing something. AMD blames who for what? 

Just leave it. I seriously doubt anyone outside of two or three people on the entire planet care.