Nvidia Ampere Discussion [2020-05-14]

[SimBy]

3070 launch delayed from 15th to 29th: https://videocardz.com/newz/nvidia-pushes-geforce-rtx-3070-launch-back-to-october-29th

Jebaited! -Nvidia

 

[Kyyla]

Keep in mind these undervolting quick tests tell you nothing about stability in long term use.

 

[trinibwoy]

3070 launch delayed from 15th to 29th: https://videocardz.com/newz/nvidia-pushes-geforce-rtx-3070-launch-back-to-october-29th

Was the review date delayed too? The upside for Nvidia is that AMD won't be able to use the 3070 in their marketing materials on the 28th. They will be forced to compare to Turing or Big Ampere and won't have the opportunity to crap all over the 3070.

 

[Malo]

These stupid marketing ploys are pathetic. What if AMD delay by 2 days, will Nvidia delay again and they both play corporate leapfrog?

 

[Pinstripe]

Was the review date delayed too? The upside for Nvidia is that AMD won't be able to use the 3070 in their marketing materials on the 28th. They will be forced to compare to Turing or Big Ampere and won't have the opportunity to crap all over the 3070.

Or you know, they're just gonna compare it to their own product stack.

 

[Scott_Arm]

Keep in mind these undervolting quick tests tell you nothing about stability in long term use.

Tons of people undervolt their cards and if you can get a long multi-hour run you're usually good. I've had my card running undervolted for a long time now. I always use msi kombuster with the artifact scanner turned on to check and make sure there's nothing going wrong. Might be harder to test stability with the tensor and RT cores.

 

[trinibwoy]

Or you know, they're just gonna compare it to their own product stack.

Of course but where's the fun in that.

 

[trinibwoy]

These stupid marketing ploys are pathetic. What if AMD delay by 2 days, will Nvidia delay again and they both play corporate leapfrog?

Maybe it's not a marketing ploy and they really want to ramp up inventory. Though I don't get how waiting 2 weeks will help. It's still going to be the same total number of cards that make it to market whether it's a drip or a drop.

 

[DegustatoR]

Was the review date delayed too?

I suspect that the delay was mostly because the reviewers wouldn't be able to get the cards to publish the reviews on time, again.
But maybe it has something to do with AMD, I dunno. Doubtful though.

 

[Jawed]

Maybe it's not a marketing ploy and they really want to ramp up inventory. Though I don't get how waiting 2 weeks will help. It's still going to be the same total number of cards that make it to market whether it's a drip or a drop.

Less people will want to buy 3070 once AMD shows their wares?

 

[DegustatoR]

https://www.nvidia.com/en-us/geforce/news/geforce-rtx-3070-available-october-29/

 

[Jawed]

Getting desperate showing stuff that isn't games.

 

[Voxilla]

Yes, quite a curious selection. Looks good compared to 2070.
In 4K the 2080Ti will still have a good edge, due to significant more memory bandwidth and texturing power.

 

[Malo]

Well they did state that 3070 beats the 2080ti so they want to show examples of where it does in earnest. Their original comparisons showed it on par in games which is still very good imo.

 

[Deleted member 2197]

Getting desperate showing stuff that isn't games.

Surprisingly quite a few professional review sites include these models in their reviews.

 

[Benetanegia]

It's the same games and benchmarks that they showed 3080 and 3090 with. I don't think there's much to interpret from that. Based on actual reviews many of those games aren't even the ones where Ampere shines the most, with the exception of Doom Eternal maybe.

 

[Jawed]

Reviewers should have Ryzen 5000 series by the time this launches, shouldn't they? I'm not as hyped as some for the "single thread IPC gains" it should bring, but I suppose there's a decent chance that 3070 will benefit at 1440p from these new CPUs.

 

[Cyan]

tldr: There's nothing wrong with the capacitor configurations. The new driver has the same performance, but eliminates crashes. The clock frequency vs voltage curve is nearly identical and power consumption is nearly identical. There were minor tweaks, probably in the boosting algorithm so it wouldn't change quite as rapidly. The linux driver was always stable. It was the windows driver that had crashing problems.

this guy speaks the truth, and unlike the "tech youtuber I am not gonna mention by name" -his words- he never jumped on the back of that theory. Igor the guy who came up with that theory also said that it was a possible cause. He could be wrong, but at least he offered a possible explanation 'cos he knows his stuff and it was interesting, but when a "I am cool know-it-all" like jayztwocents appeared for giving his unwanted, not knowledgeable opinion, the issue got a bit more complicated and drama ensued.

I blocked jayztwocents videos on youtube --took the bait once, not anymore. He believes he is something he is not. This is a feeling I have with some youtubers, even the guy with the long hair that start believing that they speak higher words than those who are smarter than them and work for a company like nVidia.

You won me a subscription to his channel, at least he is more humble, and smarter.

 

[Rootax]

Yeah I like Hardware Unboxed channel. The other dude on it is doing a good job too. For a weird reason I hate their intro music

Anyway, I hope that with the news drivers, this drama is behind us now.

 

[Entropy]

Surprisingly quite a few professional review sites include these models in their reviews.

Who pays their bills though? There are barely any truly independent review sites. But there are differences in integrity.