Honestly, when I first received this news, I thought it was a mistaken re-announcement of the contest from a few months ago. The original Order of 10 challenge was made up of a series of puzzles, and the first handful of people to solve it, received a GTX 10-Series graphics card. Turns out, NVIDIA is doing it again.
For four weeks, starting on July 21st, NVIDIA will add four new challenges and, more importantly, 100 new “chances to win”. They did not announce what those prizes will be or whether all of them will be distributed to the first 25 complete entries of each challenge, though. Some high-profile YouTube personalities, such as some of the members of Rooster Teeth, were streaming their attempts the last time around, so there might be some of that again this time, too.
Is the winner going to be the
Is the winner going to be the first to solve the High DPC Latency and stuttering
https://forums.geforce.com/default/topic/941579/geforce-1000-series/gtx-1080-high-dpc-latency-and-stuttering/
Nvidia cheating in
Nvidia cheating in Doom?
https://www.youtube.com/watch?v=FVb25eomcrI
3DMark is NOT using async?
http://www.overclock.net/t/1605674/computerbase-de-doom-vulkan-benchmarked/220#post_25351958
Can we hope for an article in PCPer about those two? Of course NOT.
How about you email us when
How about you email us when you see an interesting story? I didn't see either of these before now (albeit it's quite late, so I may or may not get to one or both of them tonight).
I fear you will direct my
I fear you will direct my mails to the spam folder, or directly to the trash can. :p
In any case people who send stuff to newsposters/reviewers expect – sometimes demand – to see an article about that in the next hour or day at worst. So sending a mail wouldn’t change something. A mail would have been an option if it was a story no one else was talking about.
Anyway, both look interesting and both could be nothing.
The Doom example could be a GTX 970 specific thing, trying to compensate for the 3.5GB limit. Or maybe it is something more serious, like Nvidia using a technique, something like HiAlgo BOOST to get higher framerates in some games, hoping that no one will notice. (HiAlgo is the company AMD bought a few weeks ago, so if they try to fake higher framerates without giving a clear “enable/disable” option to the user, you also should come out and point the finger at them)
3DMark on the other hand is the only application where Nvidia’s Pascal GPUs gain something from Time Spy’s – whatever type of – Async is using, if it is using any. At the same time the benchmark seems to not use AMD’s hardware as it should. Really convenient if you want to assure people that Pascal cards are future proof and $700 for a 1080 is not a bad long term investment.
One more link about this that could help anyone wanting to have a look is this
http://www.overclock.net/t/1606224/various-futuremarks-time-spy-directx-12-benchmark-compromised-less-compute-parallelism-than-doom-aots-also
Took a while, but my take on
Took a while, but my take on the Asynchronous Compute post is up.
Yes the real argument is
Yes the real argument is about or NOT having fully in the GPU’s hardware that hardware form of Asynchronous Compute. And not trying to mix up the software form of preemption and context switching of software threads with the hardware form of preemption and context switching with the processor threads.
And still not much focus on AMD’s instruction pre-fetch and how that may help even more on AMD’s GCN/Polaris based SKUs.
Now on the CPU side the online review sites where all over Intel’s Version of SMT, HyperThreading(TM), and that is a good example of hardware asynchronous compute on a CPU’s core. A GPU’s cores are no different with respect to benefiting from fully in the GPU core’s hardware the same sort of ability to manage the GPUs many more processor threads on many GPU cores.
But those analogies my draw a certain GPU makers ire, and these things can not be done to show that JHH emperor that he is wearing no asynchronous hardware metallic clothing. Man the emperor’s marketing department would not like any form of end user education of the unwashed masses who shall remain ignorant of the truth so as to be the most exploitable.
Thanks for your
Thanks for your effort.
Things like this should come out in public before companies think that they can do whatever they want, absolutely sure that no one will care to see how things are done. You, with FCAT, proved that just an fps counter is NOT enough to warranty a good gaming experience. In other cases people and tech sites should question what hardware companies are saying and what numbers, benchmarks that have the power to sell cards, are producing.
Futuremark came out with a statement and I think it is a win for PCPer to have an article published just before that statement.
Now it is AMD that have to take the initiative. If they do nothing for any reasons, then either there is no problem with Time Spy, either they are just stupid or afraid of something. Anyway, it is their problem now. Nvidia didn’t had a problem to come out and start shooting at AoTS when they saw that the benchmark was favoring AMD cards. AMD should start learning from Nvidia how marketing is done, because Nvidia is years ahead in that area. Like comparing GTX 1080 with an HD 3870.
What I ask from the press is to publish and question as much as possible. Nothing more, nothing less.
Thanks again.
How about the pile of email
How about the pile of email I’ve sent you about NASA faking the manned moon landing!?!?
I’m waiting by more than 40 years now had have yet to see an article on PCPer…
Looking for an ant mound to
Looking for an ant mound to turn in to mount Everest i see.
With doom how offen is a person standing around looking at a rock face? Its a face paced FPS so its not something anyone will notice less they are one those just looking for something to complain about.
If you want to go there we could have a talk about AMD and tessellation where they cheat on that?
Everyone by now knows that
Everyone by now knows that you will rush to do damage control for nvidia. And while you can point to my avatar, the thing is that I had in the past point my finger at AMD. You on the other hand will prefer to cut your whole arm than pointing a finger at Nvidia.
The Doom case looks like HiAlgo boost. If you don’t give the user a clear option in their settings, then you are cheating.
Any links about AMD’s cheating on tessellation are welcomed.
He trolls several forums
He trolls several forums doing the same thing. I think he is part of the Nvidia defence team.
Probably.
Everyone has a
Probably.
Everyone has a preference, we can’t say that we are all objective. But there are some people who post like they have a specific agenda and their posting looks to follow specific rules. You can show to them 50 different photos of a cube and they will always insist that they see a sphere.
“AMD and tessellation where
“AMD and tessellation where they cheat on that”
Cite your source.
I watched the first video
I watched the first video —
He *may* be onto something, but it could also be that playing Doom at 4K means there’s a lot more shifting around going on due to 3.5GB of VRAM usable vs. 8 GB on the rx 480 tested. This may not be the case with the GTX 1060 @ 6 GB or the 1070/1080 @ 8 GB..
The 970 is effectively obsoleted out of the market now so this really needs to be done on a 1060..
It looks like its forcing a
It looks like its forcing a different LOD which is a cheat if it is true.